article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
populations of self - sustained oscillators can exhibit various synchronization phenomena @xcite .
for example , it is well known that a limit - cycle oscillator can exhibit phase locking to a periodic external forcing ; this phenomenon is called the forced synchronization @xcite .
recently , it was also found that uncoupled identical limit - cycle oscillators subject to weak common noise can exhibit in - phase synchronization ; this remarkable phenomenon is called the common - noise - induced synchronization @xcite . in general , each oscillatory dynamics is described by a stable limit - cycle solution to an ordinary differential equation , and the phase description method for ordinary limit - cycle oscillators has played an essential role in the theoretical analysis of the synchronization phenomena @xcite . on the basis of the phase description ,
optimization methods for the dynamical properties of limit - cycle oscillators have also been developed for forced synchronization @xcite and common - noise - induced synchronization @xcite .
synchronization phenomena of spatiotemporal rhythms described by partial differential equations , such as reaction - diffusion equations and fluid equations , have also attracted considerable attention @xcite ( see also refs .
@xcite for the spatiotemporal pattern formation ) .
examples of earlier studies include the following . in reaction - diffusion systems ,
synchronization between two locally coupled domains of excitable media exhibiting spiral waves has been experimentally investigated using the photosensitive belousov - zhabotinsky reaction @xcite .
in fluid systems , synchronization in both periodic and chaotic regimes has been experimentally investigated using a periodically forced rotating fluid annulus @xcite and a pair of thermally coupled rotating fluid annuli @xcite .
of particular interest in this paper is the experimental study on generalized synchronization of spatiotemporal chaos in a liquid crystal spatial light modulator @xcite ; this experimental synchronization can be considered as common - noise - induced synchronization of spatiotemporal chaos .
however , detailed theoretical analysis of these synchronization phenomena has not been performed even for the case in which the spatiotemporal rhythms are described by stable limit - cycle solutions to partial differential equations , because a phase description method for partial differential equations has not been fully developed yet . in this paper , we theoretically analyze common - noise - induced phase synchronization between uncoupled identical hele - shaw cells exhibiting oscillatory convection ; the oscillatory convection is described by a stable limit - cycle solution to a partial differential equation .
a hele - shaw cell is a rectangular cavity in which the gap between two vertical walls is much smaller than the other two spatial dimensions , and the fluid in the cavity exhibits oscillatory convection under appropriate parameter conditions ( see refs .
@xcite and also references therein ) . in ref .
@xcite , we recently formulated a theory for the phase description of oscillatory convection in the hele - shaw cell and analyzed the mutual synchronization between a pair of coupled systems of oscillatory hele - shaw convection ; the theory can be considered as an extension of our phase description method for stable limit - cycle solutions to nonlinear fokker - planck equations @xcite ( see also ref .
@xcite for the phase description of spatiotemporal rhythms in reaction - diffusion equations ) . using the phase description method for oscillatory convection
, we here demonstrate that uncoupled systems of oscillatory hele - shaw convection can be in - phase synchronized by applying weak common noise .
furthermore , we develop a method for obtaining the optimal spatial pattern of the common noise to achieve synchronization .
the theoretical results are validated by direct numerical simulations of the oscillatory hele - shaw convection .
this paper is organized as follows . in sec .
[ sec:2 ] , we briefly review our phase description method for oscillatory convection in the hele - shaw cell . in sec .
[ sec:3 ] , we theoretically analyze common - noise - induced phase synchronization of the oscillatory convection . in sec .
[ sec:4 ] , we confirm our theoretical results by numerical analysis of the oscillatory convection .
concluding remarks are given in sec .
[ sec:5 ] .
in this section , for the sake of readability and being self - contained , we review governing equations for oscillatory convection in the hele - shaw cell and our phase description method for the oscillatory convection with consideration of its application to common - noise - induced synchronization . more details and other applications of the phase description method are given in ref .
@xcite .
the dynamics of the temperature field @xmath0 in the hele - shaw cell is described by the following dimensionless form ( see ref .
@xcite and also references therein ) : @xmath1 the laplacian and jacobian are respectively given by @xmath2 the stream function @xmath3 is determined from the temperature field @xmath0 as @xmath4 where the rayleigh number is denoted by @xmath5 .
the system is defined in the unit square : @xmath6 $ ] and @xmath7 $ ] .
the boundary conditions for the temperature field @xmath0 are given by @xmath8 where the temperature at the bottom ( @xmath9 ) is higher than that at the top ( @xmath10 ) .
the stream function @xmath3 satisfies the dirichlet zero boundary condition on both @xmath11 and @xmath12 , i.e. , @xmath13 to simplify the boundary conditions in eq .
( [ eq : bcty ] ) , we consider the convective component @xmath14 of the temperature field @xmath0 as follows : @xmath15 inserting eq .
( [ eq : t_x ] ) into eqs .
( [ eq : t])([eq : p_t ] ) , we derive the following equation for the convective component @xmath14 : @xmath16 where the stream function @xmath3 is determined by @xmath17 applying eq .
( [ eq : t_x ] ) to eqs .
( [ eq : bctx])([eq : bcty ] ) , we obtain the following boundary conditions for the convective component @xmath14 : @xmath18 that is , the convective component @xmath14 satisfies the neumann zero boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath12 .
it should be noted that this system does not possess translational or rotational symmetry owing to the boundary conditions given by eqs .
( [ eq : bcpx])([eq : bcpy])([eq : bcxx])([eq : bcxy ] ) . the dependence of the hele - shaw convection on the rayleigh number @xmath5 is well known , and the existence of stable limit - cycle solutions to eq .
( [ eq : x ] ) is also well established ( see ref .
@xcite and also references therein ) . in general , a stable limit - cycle solution to eq .
( [ eq : x ] ) , which represents oscillatory convection in the hele - shaw cell , can be described by @xmath19 the phase and natural frequency are denoted by @xmath20 and @xmath21 , respectively .
the limit - cycle solution @xmath22 possesses the following @xmath23-periodicity in @xmath20 : @xmath24 . inserting eq .
( [ eq : x_x0 ] ) into eqs .
( [ eq : x])([eq : p_x ] ) , we find that the limit - cycle solution @xmath22 satisfies @xmath25 where the stream function @xmath26 is determined by @xmath27 from eq .
( [ eq : t_x ] ) , the corresponding temperature field @xmath28 is given by ( e.g. , see fig .
[ fig:2 ] in sec .
[ sec:4 ] ) @xmath29 let @xmath30 represent a small disturbance added to the limit - cycle solution @xmath22 , and consider a slightly perturbed solution @xmath31 equation ( [ eq : x ] ) is then linearized with respect to @xmath30 as follows : @xmath32 as in the limit - cycle solution @xmath22 , the function @xmath33 satisfies the neumann zero boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath12 .
note that @xmath34 is time - periodic through @xmath20 .
therefore , eq . ( [ eq : linear ] ) is a floquet - type system with a periodic linear operator . defining the inner product of two functions as @xmath35\!\!\bigr ] } } = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dx \int_0 ^ 1 dy \ , u^\ast(x , y , \theta ) u(x , y , \theta ) , \label{eq : inner}\end{aligned}\ ] ] we introduce the adjoint operator of the linear operator @xmath34 by @xmath36\!\!\bigr ] } } = { \ensuremath{\bigl[\!\!\bigl [ { \cal l}^\ast(x , y , \theta ) u^\ast(x , y , \theta ) , \ ,
u(x , y , \theta ) \bigr]\!\!\bigr]}}.
\label{eq : operator}\end{aligned}\ ] ] as in @xmath33 , the function @xmath37 also satisfies the neumann zero boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath12 .
details of the derivation of the adjoint operator @xmath38 are given in ref .
@xcite . in the following subsection
, we utilize the floquet eigenfunctions associated with the zero eigenvalue , i.e. , @xmath39 we note that the right zero eigenfunction @xmath40 can be chosen as @xmath41 which is confirmed by differentiating eq .
( [ eq : x0 ] ) with respect to @xmath20 . using the inner product of eq .
( [ eq : inner ] ) with the right zero eigenfunction of eq .
( [ eq : u0 ] ) , the left zero eigenfunction @xmath42 is normalized as @xmath43\!\!\bigr ] } } = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dx \int_0 ^ 1 dy \ , u_0^\ast(x , y , \theta ) u_0(x , y , \theta ) = 1.\end{aligned}\ ] ] here , we can show that the following equation holds ( see also refs .
@xcite ) : @xmath44 = 0.\end{aligned}\ ] ] therefore , the following normalization condition is satisfied independently for each @xmath20 as follows : @xmath45 we now consider oscillatory hele - shaw convection with a weak perturbation applied to the temperature field @xmath0 described by the following equation : @xmath46 the weak perturbation is denoted by @xmath47 . inserting eq .
( [ eq : t_x ] ) into eq .
( [ eq : t_p ] ) , we obtain the following equation for the convective component @xmath14 : @xmath48 using the idea of the phase reduction @xcite , we can derive a phase equation from the perturbed equation ( [ eq : x_p ] ) .
namely , we project the dynamics of the perturbed equation ( [ eq : x_p ] ) onto the unperturbed solution as @xmath49 \nonumber \\ & = \int_0 ^ 1 dx \int_0 ^ 1 dy \ , u_0^\ast(x , y , \theta ) \left [ \nabla^2 x + j(\psi , x ) - \frac{\partial \psi}{\partial x } + \epsilon p(x , y , t ) \right ] \nonumber \\ & \simeq \int_0 ^ 1 dx \int_0 ^ 1 dy \ , u_0^\ast(x , y , \theta ) \left [ \nabla^2 x_0 + j(\psi_0 , x_0 ) - \frac{\partial \psi_0}{\partial x } + \epsilon p(x , y , t ) \right ] \nonumber \\ & = \int_0 ^ 1 dx \int_0 ^ 1 dy \ , u_0^\ast(x , y , \theta ) \left [ \omega \frac{\partial}{\partial \theta } x_0(x , y , \theta ) + \epsilon p(x , y , t ) \right ] \nonumber \\ & = \int_0 ^ 1 dx \int_0 ^ 1 dy \ , u_0^\ast(x , y , \theta ) \ , \biggl [ \omega \ , u_0(x , y , \theta ) + \epsilon p(x , y , t ) \biggr ] \nonumber \\ & = \omega + \epsilon \int_0 ^ 1 dx \int_0 ^ 1 dy \ , u_0^\ast(x , y , \theta ) p(x , y , t),\end{aligned}\ ] ] where we approximated @xmath14 by the unperturbed limit - cycle solution @xmath22 .
therefore , the phase equation describing the oscillatory hele - shaw convection with a weak perturbation is approximately obtained in the following form : @xmath50 where the _ phase sensitivity function _ is defined as ( e.g. , see fig .
[ fig:2 ] in sec .
[ sec:4 ] ) @xmath51 here , we note that the phase sensitivity function @xmath52 satisfies the neumann zero boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath12 , i.e. , @xmath53 as mentioned in ref .
@xcite , eq . ( [ eq : theta_p ] ) is a generalization of the phase equation for a perturbed limit - cycle oscillator described by a finite - dimensional dynamical system ( see refs .
however , reflecting the aspects of an infinite - dimensional dynamical system , the phase sensitivity function @xmath52 of the oscillatory hele - shaw convection possesses infinitely many components that are continuously parameterized by the two variables , @xmath11 and @xmath12 . in this paper , we further consider the case that the perturbation is described by a product of two functions as follows : @xmath54 that is , the space - dependence and time - dependence of the perturbation are separated . in this case , the phase equation ( [ eq : theta_p ] ) can be written in the following form : @xmath55 where the _ effective phase sensitivity function _ is given by ( e.g. , see fig .
[ fig:5 ] in sec .
[ sec:4 ] ) @xmath56 we note that the form of eq .
( [ eq : theta_q ] ) is essentially the same as that of the phase equation for a perturbed limit - cycle oscillator described by a finite - dimensional dynamical system ( see refs .
we also note that the effective phase sensitivity function @xmath57 can also be considered as the collective phase sensitivity function in the context of the collective phase description of coupled individual dynamical elements exhibiting macroscopic rhythms @xcite .
in this section , using the phase description method in sec .
[ sec:2 ] , we analytically investigate common - noise - induced synchronization between uncoupled systems of oscillatory hele - shaw convection .
in particular , we theoretically determine the optimal spatial pattern of the common noise for achieving the noise - induced synchronization .
we consider @xmath58 uncoupled systems of oscillatory hele - shaw convection subject to weak common noise described by the following equation for @xmath59 : @xmath60 where the weak common noise is denoted by @xmath61 .
inserting eq .
( [ eq : t_x ] ) into eq .
( [ eq : t_xi ] ) for each @xmath62 , we obtain the following equation for the convective component @xmath63 : @xmath64 as in eq .
( [ eq : p_x ] ) , the stream function of each system is determined by @xmath65 the common noise @xmath66 is assumed to be white gaussian noise @xcite , the statistics of which are given by @xmath67 here , we assume that the unperturbed oscillatory hele - shaw convection is a stable limit cycle and that the noise intensity @xmath68 is sufficiently weak . then , as in eq .
( [ eq : theta_q ] ) , we can derive a phase equation from eq .
( [ eq : x_xi ] ) as follows ) can be slightly different from the natural frequency given in eq .
( [ eq : x_x0 ] ) ; however , this point is not essential in this paper because eq .
( [ eq : lambda ] ) is independent of the value of the frequency .
the theory of stochastic phase reduction for ordinary limit - cycle oscillators has been intensively investigated in refs .
@xcite , but extensions to partial differential equations have not been developed yet . ] : @xmath69 where the effective phase sensitivity function @xmath57 is given by eq .
( [ eq : zeta ] ) .
once the phase equation ( [ eq : theta_xi ] ) is obtained , the lyapunov exponent characterizing the common - noise - induced synchronization can be derived using the argument by teramae and tanaka @xcite . from eqs .
( [ eq : xi])([eq : theta_xi ] ) , the lyapunov exponent , which quantifies the exponential growth rate of small phase differences between the two systems , can be written in the following form : @xmath70 ^ 2 \leq 0 .
\label{eq : lambda}\end{aligned}\ ] ] here , we used the following abbreviation : @xmath71 .
equation ( [ eq : lambda ] ) represents that uncoupled systems of oscillatory hele - shaw convection can be in - phase synchronized when driven by the weak common noise , as long as the phase reduction approximation is valid . in the following two subsections ,
we develop a method for obtaining the optimal spatial pattern of the common noise to achieve the noise - induced synchronization of the oscillatory convection .
considering the boundary conditions of @xmath52 , eqs .
( [ eq : bczx])([eq : bczy ] ) , we introduce the following spectral transformation : @xmath72 for @xmath73 and @xmath74 . the corresponding spectral decomposition of @xmath52 is given by @xmath75 by inserting eq .
( [ eq : z_cossin ] ) into eq .
( [ eq : zeta ] ) , the effective phase sensitivity function @xmath57 can be written in the following form : @xmath76 where the spectral transformation of @xmath77 is defined as @xmath78 the corresponding spectral decomposition of @xmath77 is given by @xmath79 for the sake of convenience in the calculation below , we rewrite the double sum in eq .
( [ eq : zeta_double ] ) by the following single series : @xmath80 in eq .
( [ eq : zeta_single ] ) , we introduced one - dimensional representations , @xmath81 and @xmath82 , where the mapping between @xmath83 and @xmath84 is bijective .
accordingly , we obtain the following quantity : @xmath85 ^ 2 = \sum_{n=0}^\infty \sum_{m=0}^\infty s_n s_m q_n'(\theta ) q_m'(\theta ) , \label{eq : dzeta_squared}\end{aligned}\ ] ] where @xmath86 . from eqs .
( [ eq : lambda])([eq : dzeta_squared ] ) , the lyapunov exponent normalized by the noise intensity , @xmath87 , can be written in the following form : @xmath88 ^ 2 = \sum_{n=0}^\infty \sum_{m=0}^\infty k_{nm } s_n s_m , \label{eq : lambda_k}\end{aligned}\ ] ] where each element of the symmetric matrix @xmath89 is given by @xmath90 by defining an infinite - dimensional column vector @xmath91 , eq .
( [ eq : lambda_k ] ) can also be written as @xmath92 which is a quadratic form . using the spectral representation of the normalized lyapunov exponent , eq .
( [ eq : lyapunov ] ) , we seek the optimal spatial pattern of the common noise for the synchronization . as a constraint
, we introduce the following condition : @xmath93 that is , the total power of the spatial pattern is fixed at unity . under this constraint condition
, we consider the maximization of eq .
( [ eq : lyapunov ] ) . for this purpose
, we define the lagrangian @xmath94 as @xmath95 where the lagrange multiplier is denoted by @xmath96 . setting the derivative of the lagrangian @xmath94 to be zero ,
we can obtain the following equations : @xmath97 \frac{\partial f}{\partial \lambda } & = - \left ( \sum_{n=0}^\infty s_n^2 - 1 \right ) = 0,\end{aligned}\ ] ] which are equivalent to the eigenvalue problem described by @xmath98 these eigenvectors @xmath99 and the corresponding eigenvalues @xmath100
satisfy @xmath101 because the matrix @xmath89 , which is defined in eq .
( [ eq : k ] ) , is symmetric , the eigenvalues @xmath100 are real numbers .
consequently , under the constraint condition given by eq .
( [ eq : unity ] ) , the optimal vector that maximizes eq .
( [ eq : lambda ] ) coincides with the eigenvector associated with the largest eigenvalue , i.e. , @xmath102 therefore , the optimal spatial pattern @xmath103 can be written in the following form : @xmath104 where the coefficients @xmath105 in the double series correspond to the elements of the optimal vector @xmath106 associated with @xmath107 . from eq .
( [ eq : lyapunov ] ) , the lyapunov exponent is then given by @xmath108 finally , we note that this optimization method can also be considered as the principal component analysis @xcite of the phase - derivative of the phase sensitivity function , @xmath109 .
in this section , to illustrate the theory developed in sec .
[ sec:3 ] , we numerically investigate common - noise - induced synchronization between uncoupled hele - shaw cells exhibiting oscillatory convection . the numerical simulation method is summarized in ref .
modes for the dirichlet zero boundary condition and a cosine expansion with @xmath110 modes for the neumann zero boundary condition .
the fourth - order runge - kutta method with integrating factor using a time step @xmath111 ( mainly , @xmath112 ) and the heun method with integrating factor using a time step @xmath113 were applied for the deterministic and stochastic ( langevin - type ) equations , respectively . ] .
considering the boundary conditions of the convective component @xmath114 , eqs .
( [ eq : bcxx])([eq : bcxy ] ) , we introduce the following spectral transformation : @xmath115 for @xmath73 and @xmath74 . the corresponding spectral decomposition of the convective component @xmath114 is given by @xmath116 in visualizing the limit - cycle orbit in the infinite - dimensional state space , we project the limit - cycle solution @xmath22 onto the @xmath117-@xmath118 plane as @xmath119 the initial values were prepared so that the system exhibits single cellular oscillatory convection .
the rayleigh number was fixed at @xmath120 , which gives the natural frequency @xmath121 , i.e. , the oscillation period @xmath122 .
figure [ fig:1 ] shows the limit - cycle orbit of the oscillatory convection projected onto the @xmath117-@xmath118 plane , obtained from direct numerical simulations of the dynamical equation ( [ eq : x ] ) .
snapshots of the limit - cycle solution @xmath22 and other associated functions , @xmath28 and @xmath52 , are shown in fig .
[ fig:2 ] , where the phase variable @xmath20 is discretized using @xmath123 grid points .
we note that fig .
[ fig:1 ] and fig .
[ fig:2 ] are essentially reproductions of our previous results given in ref .
details of the numerical method for obtaining the phase sensitivity function @xmath52 are given in refs .
@xcite ( see also refs .
@xcite ) . as seen in fig .
[ fig:2 ] , the phase sensitivity function @xmath52 is spatially localized .
namely , the absolute values of the phase sensitivity function @xmath52 in the top - right and bottom - left corner regions of the system are much larger than those in the other regions ; this fact reflects the dynamics of the spatial pattern of the convective component @xmath22 . as mentioned in ref .
@xcite , the phase sensitivity function @xmath52 in this case possesses the following symmetry . for each @xmath20 , the limit - cycle solution @xmath22 and the phase sensitivity function @xmath52 , shown in fig .
[ fig:2 ] , are anti - symmetric with respect to the center of the system , i.e. , @xmath124 where @xmath125 and @xmath126 .
therefore , for a spatial pattern @xmath127 that is symmetric with respect to the center of the system , @xmath128 the corresponding effective phase sensitivity function @xmath57 becomes zero , i.e. , @xmath129 that is , such symmetric perturbations do not affect the phase of the oscillatory convection .
the optimal spatial pattern is obtained as the best combination of single - mode spatial patterns , i.e. , eq .
( [ eq : aopt ] ) .
thus , we first consider the following single - mode spatial pattern : @xmath130 then , the effective phase sensitivity function is given by the following single spectral component : @xmath131 from eq .
( [ eq : lambda ] ) , the lyapunov exponent for the single - mode spatial pattern can be written in the following form : @xmath132 ^ 2,\end{aligned}\ ] ] where @xmath133 .
figure [ fig:3](a ) shows the normalized lyapunov exponent for single - mode spatial patterns , i.e. , @xmath134 .
owing to the anti - symmetry of the phase sensitivity function , given in eq .
( [ eq : anti - symmetric ] ) , the normalized lyapunov exponent @xmath134 exhibits a checkerboard pattern , namely , @xmath135 when the sum of @xmath136 and @xmath137 , i.e. , @xmath138 , is an odd number .
the maximum of @xmath134 is located at @xmath139 ; under the condition of @xmath140 , the maximum of @xmath134 is located at @xmath141 .
the single - mode spatial patterns , @xmath142 , @xmath143 , and @xmath144 , are shown in figs .
[ fig:4](b)(c)(d ) , respectively .
we note that @xmath142 and @xmath143 are anti - symmetric with respect to the center of the system , whereas @xmath144 is symmetric .
these spatial patterns are used in the numerical simulations performed below .
we now consider the optimal spatial pattern .
figure [ fig:3](b ) shows the spectral components of the optimal spatial pattern , i.e. , @xmath105 , obtained by the optimization method developed in sec .
[ subsec:3c ] ; figure [ fig:4](a ) shows the corresponding optimal spatial pattern , i.e. , @xmath103 , given by eq .
( [ eq : aopt ] ) . as seen in fig .
[ fig:3 ] , when the normalized lyapunov exponent for a single - mode spatial pattern , @xmath134 , is large , the absolute value of the optimal spectral components , @xmath145 , is also large . as seen in fig .
[ fig:4](a ) , the optimal spatial pattern @xmath103 is similar to the snapshots of the phase sensitivity function @xmath52 shown in fig .
[ fig:2 ] .
in fact , as mentioned in sec .
[ subsec:3c ] , the optimal spatial pattern @xmath103 corresponds to the first principal component of @xmath109 . reflecting the anti - symmetry of the phase sensitivity function , eq .
( [ eq : anti - symmetric ] ) , the optimal spatial pattern @xmath103 is also anti - symmetric with respect to the center of the system .
figure [ fig:5 ] shows the effective phase sensitivity functions @xmath57 for the spatial patterns shown in fig .
[ fig:4 ] . when the normalized lyapunov exponent @xmath134 is large , the amplitude of the corresponding effective phase sensitivity function @xmath57 is also large .
for the spatial pattern @xmath144 , which is symmetric with respect to the center of the system , the effective phase sensitivity function becomes zero , @xmath146 , as shown in eq .
( [ eq : zeta_s ] ) . to confirm the theoretical results shown in fig .
[ fig:5 ] , we obtain the effective phase sensitivity function @xmath57 by direct numerical simulations of eq .
( [ eq : x_p ] ) with eq .
( [ eq : p_aq ] ) as follows : we measure the phase response of the oscillatory convection by applying a weak impulsive perturbation with the spatial pattern @xmath77 to the limit - cycle solution @xmath22 with the phase @xmath20 ; then , normalizing the phase response curve by the weak impulse intensity @xmath147 , we obtain the effective phase sensitivity function @xmath57 .
the effective phase sensitivity function @xmath57 obtained by direct numerical simulations with impulse intensity @xmath147 are compared with the theoretical curves in fig .
[ fig:6 ] .
the simulation results agree quantitatively with the theory .
therefore , the phase response curve normalized by the impulse intensity @xmath147 converges to the effective phase sensitivity function @xmath57 as @xmath147 decreases . as shown in fig .
[ fig:6](d ) , when the impulsive perturbation is not weak , the dependence of the phase response curve on the impulse intensity @xmath147 becomes nonlinear . in general , when the impulsive perturbation is not weak , the phase response curve is not equal to zero , even though the effective phase sensitivity function is equal to zero , @xmath146 .
we also note that the linear dependence region of the phase response curve on the impulse is generally dependent on the spatial pattern @xmath77 of the impulse . ] . in this subsection , we demonstrate the common - noise - induced synchronization between uncoupled hele - shaw cells exhibiting oscillatory convection by direct numerical simulations of the stochastic ( langevin - type ) partial differential equation ( [ eq : x_xi ] ) .
theoretical values of both the lyapunov exponents @xmath148 for several spatial patterns @xmath77 with the common noise intensity @xmath149 and the corresponding relaxation time @xmath150 toward the synchronized state are summarized in table [ table:1 ] .
figure [ fig:7 ] shows the time evolution of the phase differences @xmath151 when the common noise intensity is @xmath149 .
the initial phase values are @xmath152 for @xmath153 .
figure [ fig:8 ] shows the time evolution of @xmath154 , which corresponds to fig .
[ fig:7 ] .
the relaxation times estimated from the simulation results agree reasonably well with the theory ( d ) should be constant because the effective phase sensitivity function is equal to zero , @xmath146 , for this case . as shown in fig .
[ fig:6](d ) , when the perturbation is not sufficiently weak , the phase response curve is not equal to zero ; this higher order effect causes the slight variations shown in fig .
[ fig:7](d ) . ] . as seen in fig .
[ fig:7 ] and fig .
[ fig:8 ] , the relaxation time for the optimal spatial pattern @xmath103 is actually much smaller than those for the single - mode spatial patterns . for the cases of single - mode patterns ,
the relaxation time for the single - mode spatial pattern @xmath142 is also smaller than those for the other single - mode spatial patterns , @xmath143 and @xmath144 .
we also note that the time evolution of both @xmath151 and @xmath154 for @xmath142 is significantly different from that for @xmath144 in spite of the similarity between the two spatial patterns of the neighboring modes ; this difference results from the difference of symmetry with respect to the center , as shown in eq .
( [ eq : zeta_s ] ) .
figure [ fig:9 ] shows a quantitative comparison of the lyapunov exponents between direct numerical simulations and the theory for the case of the optimal spatial pattern @xmath103 .
the initial phase values are @xmath155 for @xmath156 , i.e. , the initial phase difference is @xmath157 .
the results of direct numerical simulations are averaged over @xmath158 samples for different noise realizations .
the simulation results quantitatively agree with the theory .
figure [ fig:10 ] shows the global stability of the common - noise - induced synchronization of oscillatory convection for the case of the optimal spatial pattern @xmath103 ; namely , it shows that the synchronization is eventually achieved from arbitrary initial phase differences , i.e. , @xmath159 $ ] .
although the lyapunov exponent @xmath148 based on the linearization of eq .
( [ eq : theta_xi ] ) quantifies only the local stability of a small phase difference , as long as the phase reduction approximation is valid , this global stability holds true for any spatial pattern @xmath77 with a non - zero lyapunov exponent , namely , the lyapunov exponent is negative , @xmath160 , as found from eq .
( [ eq : lambda ] ) .
the global stability can be proved by the theory developed in ref .
@xcite , i.e. , by analyzing the fokker - planck equation equivalent to the langevin - type phase equation ( [ eq : theta_xi ] ) ; in addition , the effect of the independent noise can also be included .
our investigations in this paper are summarized as follows . in sec .
[ sec:2 ] , we briefly reviewed our phase description method for oscillatory convection in the hele - shaw cell with consideration of its application to common - noise - induced synchronization . in sec .
[ sec:3 ] , we analytically investigated common - noise - induced synchronization of oscillatory convection using the phase description method .
in particular , we theoretically determined the optimal spatial pattern of the common noise for the oscillatory hele - shaw convection . in sec .
[ sec:4 ] , we numerically investigated common - noise - induced synchronization of oscillatory convection ; the direct numerical simulation successfully confirmed the theoretical predictions .
the key quantity of the theory developed in this paper is the phase sensitivity function @xmath52 .
thus , we describe an experimental procedure to obtain the phase sensitivity function @xmath52 . as in eq .
( [ eq : z_cossin ] ) , the phase sensitivity function @xmath52 can be decomposed into the spectral components @xmath161 , which are the effective phase sensitivity functions for the single - mode spatial patterns @xmath162 as shown in eq . ( [ eq : zjk ] ) . in a manner similar to the direct numerical simulations yielding fig
. [ fig:6 ] , the effective phase sensitivity function @xmath161 for each single - mode spatial pattern @xmath162 can also be experimentally measured .
therefore , in general , the phase sensitivity function @xmath52 can be constructed from a sufficiently large set of such @xmath161 .
once the phase sensitivity function @xmath52 is obtained , the optimization method for common - noise - induced synchronization can also be applied in experiments .
finally , we remark that not only the phase description method for spatiotemporal rhythms but also the optimization method for common - noise - induced synchronization have broad applicability ; these methods are not restricted to the oscillatory hele - shaw convection analyzed in this paper .
for example , the combination of these methods can be applied to common - noise - induced phase synchronization of spatiotemporal rhythms in reaction - diffusion systems of excitable and/or heterogeneous media .
furthermore , as mentioned above , also in experimental systems , such as the photosensitive belousov - zhabotinsky reaction @xcite and the liquid crystal spatial light modulator @xcite , the optimization method for common - noise - induced synchronization could be applied .
y.k . is grateful to members of both the earth evolution modeling research team and the nonlinear dynamics and its application research team at ifree / jamstec for fruitful comments .
is also grateful for financial support by jsps kakenhi grant number 25800222 .
h.n . is grateful for financial support by jsps kakenhi grant numbers 25540108 and 22684020 , crest kokubu project of jst , and first aihara project of jsps .
e. m. izhikevich , _ dynamical systems in neuroscience : the geometry of excitability and bursting _
( mit press , cambridge , ma , 2007 ) . g. b. ermentrout and d. h. terman , _ mathematical foundations of neuroscience _ ( springer , new york , 2010 ) .
h. nakao , t. yanagita , and y. kawamura , procedia iutam * 5 * , 227 ( 2012 ) .
y. kawamura , h. nakao , k. arai , h. kori , and y. kuramoto , phys .
* 101 * , 024101 ( 2008 ) . [ arxiv:0807.1285 ] + y. kawamura , physica d * 270 * , 20 ( 2014 ) . [ arxiv:1312.7054 ] | we investigate common - noise - induced phase synchronization between uncoupled identical hele - shaw cells exhibiting oscillatory convection . using the phase description method for oscillatory convection
, we demonstrate that the uncoupled systems of oscillatory hele - shaw convection can exhibit in - phase synchronization when driven by weak common noise .
we derive the lyapunov exponent determining the relaxation time for the synchronization , and develop a method for obtaining the optimal spatial pattern of the common noise to achieve synchronization .
the theoretical results are confirmed by direct numerical simulations . |
road traffic prediction plays an important role in intelligent transport systems by providing the required real - time information for traffic management and congestion control , as well as the long - term traffic trend for transport infrastructure planning @xcite .
road traffic predictions can be broadly classified into short - term traffic predictions and long - term traffic forecasts@xcite .
short - term prediction is essential for the development of efficient traffic management and control systems , while long - term prediction is mainly useful for road design and transport infrastructure planning .
there are two major categories of techniques for road traffic prediction : those based on non - parametric models and those based on parametric models .
non - parametric model based techniques , such as k - nearest neighbors ( knn ) model @xcite and artificial neural networks ( ann ) @xcite , are inherently robust and valid under very weak assumptions @xcite , while parametric model based techniques , such as auto - regressive integrated moving average ( arima ) model @xcite and its variants @xcite@xcite , allows to integrate knowledge of the underlying traffic process in the form of traffic models that can then be used for traffic prediction .
both categories of techniques have been widely used and in this paper , we consider parametric model based techniques , particularly starima ( space - time autoregressive integrated moving average)-based techniques . as for the estimation of parameters and coefficients in starima model , overfitting easily
occurs which makes the predictive performance poor as it overreacts to minor fluctuations in the training data @xcite .
furthermore , the same model and hence the same correlation structure is used for traffic prediction at different time of the day , which is counter - intuitive and may not be accurate . to elaborate , consider an artificial example of two traffic stations @xmath0 and @xmath1 on a highway , where traffic station @xmath1 is at the down stream direction of @xmath0 .
intuitively , the correlation between the traffic observed at @xmath0 and the traffic observed at @xmath1 will peak at a time lag corresponding to the time required to travel from @xmath0 to @xmath1 because at that time lag , the ( approximately ) same set of vehicles that have passed @xmath0 now have reached @xmath1 . obviously , the time required to travel from @xmath0 to @xmath1 depends on the traffic speed , which varies with the time of the day , e.g. peak hours and off - peak hours .
accordingly , the time lag corresponding to the peak correlation between the traffic at @xmath0 and the traffic at @xmath1 should also vary with time of the day and , to be more specific , should approximately equal to the distance between @xmath0 and @xmath1 divided by the mean speed of vehicles between @xmath0 and @xmath1 .
therefore , in designing the starima model for traffic prediction , the aforementioned time - varying lags should be taken into account for accurate traffic prediction . ] ] to validate the aforementioned intuition , we analyze the cross - correlation function ( ccf ) of traffic flow data at two traffic stations ( stations 6 and 3 ) , denoted as @xmath2 , from i-80 highway ( more details of data are discussed in section [ sub : data_collection ] ) with the formulation : @xmath3}{\sigma_{uu}\sigma{}_{yy}}\label{eq : corr_u_y}\ ] ] where @xmath4 and @xmath5 are the traffic flow data collected in @xmath6 time slots from the two traffic stations , @xmath7 is the temporal order in the range of @xmath8\subset\mathbf{n}$ ] , @xmath9 and @xmath10 are respectively the standard deviation of @xmath4 and @xmath5 .
a higher value of ccf indicates a stronger correlation of the traffic at both stations . as shown in fig.[fig : flow_3_6 ] , the correlation between traffic at stations 6 and 3 peaks at different time lags depending on the time of the day . during on - peak period ( approximately from 6:30am - 8:30am ) , the correlation peaks at a time lag of @xmath11 ( one time lag corresponds to @xmath12 ) while during off - peak period ( approximately from 19pm - 24pm ) , the correlation peaks at a time lag of @xmath13 , where one time lag corresponds to a time of 30@xmath14 .
we observe that at peak hours , the time lag corresponding to the maximum correlation is larger than that for off - peak hours . in the latter section , we will further show that this time lag approximately equals to the distance between the two traffic stations divided by the average speed .
therefore , our intuition explained in the previous paragraph is valid .
the above observation motivates us to design a starima - based traffic prediction with time - varying lags which better matches the time - varying correlation structure between traffic of different stations and hence can potentially deliver more accurate traffic prediction .
more specifically , the contributions of the paper are : * we analyze the ccf between the speed and traffic flow data between different detector stations and establish the relationship between the changes in the temporal lag ( corresponding to the aforementioned maximum correlation ) and the speed variations . *
an unsupervised classification algorithm based on isodata algorithm is designed to classify different time periods of the day according to the variation of the speed .
the classification helps to determine the appropriate time lag to use in the starima model . * a starima - based model with time - varying lags
is developed for short - term traffic prediction .
experimental results using real traffic data show that the developed starima - based model with time - varying lags has superior accuracy compared with its counterpart developed using the traditional cross - correlation function and without employing time - varying lags .
the the rest of the paper is organized as follows . in section [ sec : related_work ] , we briefly discuss related work . section [ sec : data_method ] introduces the starima model and the isodata algorithm in section [ sec : starima_time_lag_variation ] , we present the details the proposed algorithm .
the experimental results are presented in section [ sec : results ] .
finally , section [ sec : conclusion ] concludes the paper .
there is previous work , which predicts traffic flow using a modified arima models @xcite . in @xcite and @xcite , a multivariate arima based model , arimax ,
was applied for better traffic flow prediction .
the difference is that the former paper considered the varibility of the speed from upstream to downstream , the other one considered different model specifications during different time periods of the day .
similarly , the authors in @xcite also different configurations of temporal lags in arima model . more concretely , they firstly applied a hidden markov model ( hmm ) model along with an expectation - maximization ( em ) algorithm to evaluate the traffic state ( one of \{major accident , minor incident , instability , normal driving } ) in next time slot .
after that , the arima models with different configurations of temporal lags were used to predict the state of the traffic flow .
all these models have improved the accuracy of the forecasting results compared with arima model . however , the spatial information was less considered in these models . in this way , the starima based models @xcite have aroused more and more concern . ] in @xcite , the authors proposed a dynamic starima model by combining the dynamic turn ratio prediction ( dtrp ) model and the starima model . in this paper ,
a dynamic space weigh matrix is used to capture different impact of traffic at upstream locations on traffic at downstream locations .
similarly , the research in @xcite also applied the starima model with the consideration of the dynamic space weight matrix .
our work distinguish from theirs in that in our work , the space weight matrices vary with on - peak and off - peak periods to capture the time - varying correlations of road traffic at different locations . from the above related work , we can find that the " `` dynamic' of a arima or starima model in existing research is often used to indicate the dynamic of the space weight matrix , the traffic state during different time periods @xcite . however , sometimes it is not enough to only consider these aspects .
for example , fig.[fig : speed_flow_data ] presents the normalized average speed and normalized flow data collected at traffic station 6 every 30 minutes in one day .
theoretically , the weight matrix in time slot 3 ( or 4 , the left hollow rectangle ) and slot 15 ( the right hollow rectangle ) should be different since they are respectively in the off - peak period and peak period . however , their average speed are the same .
this is caused by an inaccurate evaluation of the time range of peak or off - peak period .
furthermore , few research considered the relationship between speed and the parameters ( temporal or spatial lag ) in starima model .
specially , a great majority of research use pacf to evaluate temporal lag which easily causes overfitting problem .
motivated by the above observations , in this paper we investigate a more efficient method to evaluate these parameters in starima model with the consideration of spatial information and the variation of average speed during different time periods .
] [ sub : data_collection ] in urban environment , the road structure is often complex .
also , the sensors ( such as loop detectors or cameras ) @xcite are not deployed at every road .
therefore , it is difficult to obtain comprehensive data . for simplicity ,
we only consider highway in this paper where there are only on - ramp / off - ramps so the traffic condition is comparatively simpler than that in the urban area .
we use data collected from a segment of interstate 80 ( i-80 ) freeway located in emeryville , california @xcite .
available data are collected every 30 seconds from six traffic stations numbered by 1 , 3 , 4 , 5 , 6 and 7 within 10 days .
there are two traffic stations for upstream and downstream traffic respectively .
the road topology is shown in fig.[fig : real_scenario_and_topology ] .
note that there is no data at station 2 and there are too many interference caused by on or off - ramps at station 1 and 7 . for simplicity
, we only use the data collected at stations 3 , 4 , 5 , and 6 .
the travel direction is from station 3 to 6 .
in this section , we first present a simple way to evaluate the temporal lag in relation to speed variation .
then we propose a classification algorithm based on isodata by which we can respectively obtain a set of speed clusters and a set of time period clusters .
finally , we describe a modified starima model with the varying temporal lag . from fig.[fig :
ccf_3_6_different_time_period ] , the temporal lags with maximal ccf between stations 6 and 3 are different during peak and off - peak periods .
this is attributable to the variation of speed . assuming the distance between two detector stations @xmath0 and @xmath1 is @xmath15 , and the vehicles keep a stable average speed @xmath16 , then approximately @xmath17 is needed for vehicles to travel from @xmath1 to @xmath0 . in other words , the traffic flow collected at station @xmath0
is strongly correlated with that at @xmath1 @xmath18 time ago .
thus the temporal lag with the maximum correlation should be @xmath19 $ ] where @xmath20 is the length of one temporal lag .
note that @xmath15 depend on the spatial order @xmath21 in the starima model and @xmath16 can often be measured by loop detectors @xcite .
furthermore , the advance in telecommunication and electronic technology also brings a number of new techniques that allows us to estimate the travel time , e.g. via smartphones .
indeed , the observation discussed in the introduction section suggests another novel way to estimate travel time : we can infer travel time from the correlation of the observed traffic . in order to validate the above discussion
, we further analyze the results in fig.[fig : ccf_3_6_different_time_period ] by using the f average speed information at stations 3 and 6 , which is collected in the same day as the traffic flow data used in fig.[fig : ccf_3_6_different_time_period ] .
specifically , the average speed from station 3 to station 6 between 6:30 am and 8:30 am is 44.45 feet / second .
the average speed is 67.05 feet / second between 19 pm and 24 pm .
the maximal temporal lag during these two time periods is respectively 3 and 2 with 30 seconds in each temporal lag . as @xmath22 ,
given @xmath23and @xmath24 during two time periods along with the corresponding best temporal lag @xmath25 and @xmath26 , we are able to obtain the following equation according to the theoretical analysis above : @xmath27 substituting the data into formulation , it is easy to find @xmath28 .
this result agrees with our theoretical analysis and verifies our speculation that temporal @xmath29 is a function of the variation of average speed @xmath16 in section [ sec : introduction ] .
an easy way to classify speed data is by dividing into peak time and off - peak periods .
after that , the temporal lag can be calculated using @xmath30 , where @xmath31 is the average speed in time period @xmath32\{peak , off - peak}. however an empirical classification is often prone to error and inaccuracy .
recall the analysis in section [ sec : related_work ] , the evaluation of the average speed is sensitive to the time range selected for peak or off - peak period .
it is obvious that the speed is not always fast even during off - peak period from fig .
[ fig : speed_flow_data ] .
therefore , in this paper an isodata based speed data classification algorithm is developed to deliver an accurate classification . using this algorithm , we firstly classify the speed data collected in each time slot into different clusters .
after that , the time period clusters are confirmed based on the time slots contained in different speed clusters .
* input : * @xmath33 , @xmath34,@xmath35,@xmath36,@xmath37,@xmath38,@xmath39 * return : * @xmath40,@xmath41 @xmath40**@xmath42isodata**(@xmath33 , @xmath34,@xmath35,@xmath36,@xmath43,@xmath38 ) [ algo - isodata ] @xmath44 [ algo - for - start-1 ] @xmath45 , @xmath46 @xmath47 @xmath48 @xmath49 @xmath50 and @xmath51 @xmath52 @xmath49 @xmath53 @xmath54 [ algo - if-1 ] @xmath55[algo - min - start ] @xmath56[algo - min - end ] @xmath57 [ algo - endif-1 ] @xmath58 [ algo - for - end-1 ] assuming there is a set of speed data @xmath59 in which @xmath60 is the speed in time slot @xmath61 .
the purpose here is to confirm a set of speed clusters , denoted as @xmath62 , where @xmath63 with cluster center @xmath64 and @xmath65 .
based on @xmath66 , we can obtain another set of time period clusters , denoted as @xmath67 , in which @xmath68 .
in addition , let @xmath69 be a set of continuous time slots , termed as a time range and defined as follows : @xmath70 where @xmath71 is the number of time slots contained in @xmath72 , @xmath39 is a threshold defined as the minimal number of time slots included in a time period .
the speed data classification algorithm is given in algorithm [ alg : speed - data - classification ] .
in line [ algo - isodata ] , the isodata algorithm is implemented to get speed clusters .
the time period clusters are obtained from line [ algo - for - start-1 ] to [ algo - for - end-1 ] .
it is worth mentioning that a decision is made to decide whether @xmath73 belongs to @xmath74 by comparing its capacity and threshold @xmath71 ( from line [ algo - if-1 ] to [ algo - for - end-1 ] ) .
if @xmath73 does not belong to @xmath74 , line [ algo - min - start ] and [ algo - min - end ] are implemented to allocate each @xmath75 to other @xmath76 by the operation @xmath77 .
@xmath78 is defined as the absolute difference between speed recorded in time slot @xmath79 and the average speed calculated during time period @xmath80 , which is presented in .
@xmath81 according to the speed and time period clusters obtained from section [ sub : classification_speed ] , we propose a modified starima model , denoted as @xmath82 .
the definitions of parameters @xmath83 , @xmath84 and @xmath85 in this model are the same as the original starima model , except that the temporal lag @xmath29 will vary with the spatial order @xmath21 and the average speed in different time periods .
more precisely , given a time period @xmath86 , @xmath82 is defined as follows : @xmath87 in [ eq : starima_tv_model ] , @xmath88 is a @xmath89 vector in which each element @xmath90 represents the temporal lag between two station @xmath91 and @xmath92 with the spatial order @xmath21 .
@xmath93 is calculated by @xmath94 where @xmath95 is the distance between these two stations and @xmath96 is the average speed in time period @xmath73 which is equal to @xmath97 .
note that when @xmath98 , the " `` 0th order neighbor' of a station is itself such that the temporal lag is evaluated with the pacf used in arima .
based on the data collection introduced in section [ sec : data_method ] , we utilize the speed and traffic flow data at stations 3 , 4 , 5 and 6 . at each station , there are 2880 data recorded in one day and the length of one time slot is 30 seconds . in order to eliminate noise in the data , we make a " `` smooth' operation by calculating the mean traffic flow every @xmath99 data points and regarding it as one data point . the experimental results are divided into two parts . in the first part , we provide the speed and time period clusters classified by our proposed algorithm . for the speed data @xmath100
, we choose @xmath101 . in the second part
, we present the forecast results of traffic flow in different time periods and stations using @xmath82 model .
we choose @xmath102 to calculate the mean traffic flow using original traffic flow data within 2 minutes .
firstly , the configuration of input parameters of algorithm is given in table [ tab : the - configuration ] in which the speed data @xmath100 is the results after the smooth operation on the speed data collected from four stations . with this setting ,
the smallest length of time range @xmath86 is 120 minutes .
the speed and time period clusters classified by algorithm [ algo - isodata ] is presented in table [ tab : speed - time - cluster ] .
.the input parameters in algorithm [ algo - isodata][tab : the - configuration ] [ cols="^,^,^,^",options="header " , ] in table [ tab : mse_mape_diff_algo_one_day ] , we compare the mape / mse of one day using @xmath103 , chaos and @xmath104 at four stations . except station 3 , all the mape / mse achieved by our model are better than those of the other two models . furthermore , in table [
tab : the - mape / mse - of - two - periods - s6 ] , we present the mape / mse of the forecast results of station 6 using these three models , in which the forecast time ranges are 9 - 10am and 23 - 24pm . it can be seen that the performance of our model is almost coincident with the true data .
and chaos comes to the second in the prediction during 23 - 24pm .
motivated by the observation that the correlation between traffic at different traffic stations is time - varying and the time lag corresponding to the maximum correlation approximately equals to the distance between two traffic stations divided by the speed of vehicles between them , in this paper , we developed a modified starima model with time - varying lags for short - term traffic flow prediction .
experimental results using real traffic data collected on a highway showed that the developed starima - based model with time - varying lags has superior accuracy compared with its counterpart developed using the traditional cross - correlation function and without employing time - varying lags . in an urban environment
, the correlation between traffic tends to be much more intricate .
it is part of our future work plan to develop prediction technique for urban roads that incorporates the knowledge of the underlying road topology .
p. dellacqua , f. bellotti , r. berta , and a. de gloria , " time - aware multivariate nearest neighbor regression methods for traffic flow prediction , _ ieee trans .
intelligent transportation systems _ , vol .
16 , no . 6 , pp .
33933402 , 2015 .
kim , j .- s .
et al . _ , " urban traffic flow prediction system using a multifactor pattern recognition model , _ ieee trans .
intelligent transportation systems _ , vol . 16 , no . 5 , pp . 27442755 , 2015 .
e. i. vlahogianni , m. g. karlaftis , and j. c. golias , " short - term traffic forecasting : where we are and where we re going ,
_ transportation research part c : emerging technologies _ , vol .
43 , pp . 319 , 2014 .
m. lippi , m. bertini , and p. frasconi , " short - term traffic flow forecasting : an experimental comparison of time - series analysis and supervised learning , _ ieee trans .
intelligent transportation systems _ , vol . 14 , no . 2 , pp . 871882 , 2013 .
l. song , " improved intelligent method for traffic flow prediction based on artificial neural networks and ant colony optimization .
_ journal of convergence information technology _ , vol . 7 , no . 8 , 2012 . b. l. smith , b. m. williams , and r. k. oswald , " comparison of parametric and nonparametric models for traffic flow forecasting , _ transportation research part c : emerging technologies _ ,
10 , no . 4 , pp . 303321 , 2002 .
x. min , j. hu , q. chen , t. zhang , and y. zhang , " short - term traffic flow forecasting of urban network based on dynamic starima model , in _ 2009 ieee int . conf .
intelligent transportation systems _ , pp . 16 .
a. stathopoulos and m. g. karlaftis , " a multivariate state space approach for urban traffic flow modeling and prediction ,
_ transportation research part c : emerging technologies _ , vol .
11 , no . 2 , pp . 121135 , 2003 .
g. mao and b. d. anderson , " graph theoretic models and tools for the analysis of dynamic wireless multihop networks , in _ 2009 ieee wireless communications and networking conference_.1em plus 0.5em minus 0.4emieee , 2009 , pp . 16 .
a. a. kannan , b. fidan , and g. mao , " robust distributed sensor network localization based on analysis of flip ambiguities , in _ ieee globecom 2008 - 2008 ieee global telecommunications conference_.1em plus 0.5em minus 0.4emieee , 2008 , pp . 16 . | based on the observation that the correlation between observed traffic at two measurement points or traffic stations may be time - varying , attributable to the time - varying speed which subsequently causes variations in the time required to travel between the two points , in this paper , we develop a modified space - time autoregressive integrated moving average ( starima ) model with time - varying lags for short - term traffic flow prediction .
particularly , the temporal lags in the modified starima change with the time - varying speed at different time of the day or equivalently change with the ( time - varying ) time required to travel between two measurement points .
firstly , a technique is developed to evaluate the temporal lag in the starima model , where the temporal lag is formulated as a function of the spatial lag ( spatial distance ) and the average speed .
secondly , an unsupervised classification algorithm based on isodata algorithm is designed to classify different time periods of the day according to the variation of the speed .
the classification helps to determine the appropriate time lag to use in the starima model .
finally , a starima - based model with time - varying lags is developed for short - term traffic prediction . experimental results using real traffic data show that the developed starima - based model with time - varying lags has superior accuracy compared with its counterpart developed using the traditional cross - correlation function and without employing time - varying lags . |
the hadronic particle - antiparticle correlation was already pointed out in the beginning of the nineties . however , the final formulation of these hadronic squeezed or back - to - back correlations was proposed only at the end of that decade@xcite , predicting that such correlations were expected if the masses of the mesons were modified in the hot and dense medium formed in high energy nucleus - nucleus collisions . soon after that , it was shown that analogous correlations would exist in the case of baryons as well .
an interesting theoretical finding was that both the fermionic ( fbbc ) and the bosonic ( bbbc ) back - to - back correlations were very similar , both being positive and treated by analogous formalisms . in
what follows , we will focus our discussion to the bosonic case , illustrating the effect by considering @xmath0 and @xmath1 pairs , considered to be produced at rhic energies@xcite .
let us discuss the case of @xmath3-mesons first , which are their own antiparticles , and suppose that their masses are modified in hot and dense medium .
naturally , they recover their asymptotic masses after the system freezes - out .
therefore , the joint probability for observing two such particles , i.e. , the two - particle distribution , @xmath4 , can be factorized as @xmath5 $ ] , after applying a generalization of wick s theorem for locally equilibrated systems@xcite .
the first term corresponds to the product of the spectra of the two @xmath3 s , @xmath6 , being @xmath7 and @xmath8 the free - particle creation and annihilation operators of scalar quanta , and @xmath9 means thermal averages .
the second term contains the identical particle contribution and is represented by the square modulus of the chaotic amplitude , @xmath10 .
together with the first term , it gives rise to the femtoscopic or hanbury - brown & twiss ( hbt ) effect .
the third term , the square modulus of the squeezed amplitude , @xmath11 , is identically zero in the absence of in - medium mass - shift .
however , if the particle s mass is modified , together with the first term it leads to the squeezing correlation function . the annihilation ( creation ) operator of the asymptotic , observed bosons with momentum @xmath12 , @xmath13 ( @xmath14 ) , is related to the in - medium annihilation ( creation ) operator @xmath15 ( @xmath16 ) , corresponding to thermalized quasi - particles , by the bogoliubov - valatin transformation , @xmath17 , where @xmath18 , @xmath19 .
the argument , @xmath20 $ ] , is the _ squeezing parameter_. in terms of the above amplitudes , the complete @xmath0 correlation function can be written as c_2(k_1,k_2 ) = 1 + + , [ fullcorr ] where the first two terms correspond to the identical particle ( hbt ) correlation , whereas the first and the last terms represent the correlation function between the particle and its antiparticle , i.e. , the squeezed part .
the in - medium modified mass , @xmath21 , is related to the asymptotic mass , @xmath22 , by @xmath23 , here assumed to be a constant mass - shift .
the formulation for both bosons and fermions was initially derived for a static , infinite medium @xcite .
more recently , it was shown@xcite in the bosonic case that , for finite - size systems expanding with moderate flow , the squeezed correlations may survive with sizable strength to be observed experimentally .
similar behavior is expected in the fermionic case . in that analysis , a non - relativistic treatment with flow - independent squeezing parameter
was adopted for the sake of simplicity , allowing to obtain analytical results .
the detailed discussion is in ref .
@xcite , where the maximum value of @xmath24 , was studied as a function of the modified mass , @xmath21 , considering pairs with exact back - to - back momentum , @xmath25 ( in the identical particle case , this procedure would be analogous to study the behavior of the intercept of the hbt correlation function ) . although illustrating many points of theoretical interest , this study in terms of the unobserved shifted mass and exactly back - to - back momenta was not helpful for motivating the experimental search of the bbc s
. a more realistic analysis would involve combinations of the momenta of the individual particles , @xmath26 , into the average momentum of the pair , @xmath27 .
since the maximum of the bbc effect is reached when @xmath28 , this would correspond to investigate the squeezed correlation function , @xmath29 , close to @xmath30 . for a hydrodynamical ensemble , both the chaotic and the squeezed amplitudes , @xmath31 and @xmath32 , respectively ,
can be written in a special form derived in @xcite and developed in @xcite . therefore , within a non - relativistic treatment with flow - independent squeezing parameter , the squeezed amplitude is written as in @xcite , i.e. , @xmath33 + 2 n^*_0 r_*^3 \exp\bigl[-\frac{(\mathbf{k}_1-\mathbf{k}_2)^2}{8 m _ * t}\bigr ] \exp \bigl[-\frac{im\langle u\rangle r(\mathbf{k_1 } + \mathbf{k_2})^2}{2 m _ * t_*}\bigr ] \exp\bigl[- \bigl ( \frac{1}{8 m _ * t _ * } + \frac{r_*^2}{2 } \bigr ) ( \mathbf{k_1 } + \mathbf{k_2})^2\bigr ] \bigl\ } $ ] , and the spectrum , as @xmath34 , where @xmath35 and @xmath36 @xcite .
we adopt here @xmath37 . inserting these expressions into eq .
( [ fullcorr ] ) and considering the region where the hbt correlation is not relevant , we obtain the results shown in figure 1 .
part ( a ) shows the squeezed correlation as a function of @xmath38 , for several values of @xmath39 .
the top plot shows results expected in the case of a instant emission of the @xmath0 correlated pair .
if , however , the emission happens in a finite interval , the second term in eq .
( [ fullcorr ] ) is multiplied by a reduction factor , in this case expressed by a lorentzian ( @xmath40 ^ -1 $ ] ) ,
i.e. , the fourier transform of an exponential emission .
the result is shown in the plot in the middle of figure 1(a ) .
we see that this represents a dramatic reduction in the signal , even though its strength is sizable for being observed experimentally . if the system expands with radial flow ( @xmath41 ) , the result is shown in the plot at the bottom of figure 1(a ) , again considering that the @xmath3 s are emitted during a finite period of time , @xmath42 fm / c .
we see that , in the absence of flow , the squeezed correlation signal grows faster for higher values @xmath43 than the corresponding case in the presence of flow .
however , this last one is stronger in all the investigated @xmath43 region , showing that the presence of radial flow enhances the signal .
the sensitivity of the squeezed - pair correlation to the size of the region where the mass - shift occurs is shown in figure1(b ) for two values of radii , @xmath44 fm and @xmath45 fm , keeping @xmath46 gev / c fixed . the differences are reflected in the inverse width of the curves , plotted as a function of @xmath47 . in case of no in - medium mass modification , the squeezed correlation functions would be unity for all values of @xmath48 in both plots .
fm ( top ) and @xmath45 fm ( bottom).,title="fig : " ] fm ( top ) and @xmath45 fm ( bottom).,title="fig : " ] in the case of the squeezed correlations of @xmath1 pairs , we show in figure 2(a ) results for the generated momenta of the pairs within the narrow interval @xmath49 mev / c , by plotting the squeezed correlation , @xmath50 versus @xmath21 and @xmath51 .
for the kaons , we can fixe the value of the shifted mass to be @xmath52 mev , corresponding to one of the maxima in figure 2(a ) , and then procedure similarly to what was done in the @xmath0 case .
the result is shown in figure 3 of ref.@xcite .
also in this case the intensity of the squeezed correlation would be large enough to be searched for experimentally .
next , we investigate how the behavior of the identical particle correlations could be affected in case of in - medium mass modification , since the femtoscopic correlation function also depends on the squeezing factor , @xmath53 . the hbt correlation function is obtained by inserting the chaotic amplitude , @xmath54 + n^*_0 r_*^3 ( |c_{_0}|^2+|s_{_0}|^2 ) \exp\bigl[-\frac{(\mathbf{k_1}+\mathbf{k}_2)^2}{8 m _ * t_*}\bigr ] \exp\bigl[-\bigl(\frac{im\langle u\rangle r}{2 m _ * t_*}\bigr)(\mathbf{k}_1 ^ 2-\mathbf{k}_2 ^ 2)\bigr ] \exp\bigl[-\bigl ( \frac{1}{8 m _ * t } + \frac{r^2_*}{2}\bigr)(\mathbf{k_1}-\mathbf{k_2})^2\bigr ] \bigr\ } , $ ] together with the expression for the spectrum , into eq.([fullcorr ] ) .
we use the case of identical @xmath55 pairs as illustration , as seen in figure 2(b ) .
the investigation is extended to both the cases of instant emission ( @xmath56 ) and finite emission ( @xmath57=2fm / c ) . in this figure
, we can see the well - known result corresponding to the narrowing of the femtoscopic correlation function with increasing emission times , as well as the broadening the curve with flow in the absence of squeezing , as expected .
however , if the squeezing originated in the mass - shift is present , its effects tend to oppose to those of flow ( for large @xmath58 , it practically cancels the broadening of the correlation function due to flow ) , another striking indication of mass - modification , even in hbt !
in the present work we suggest an effective way to search for the back - to - back squeezed correlations in heavy ion collisions at rhic , and later at lhc energies , by investigating the squeezed correlation function , @xmath29 , in terms of @xmath59 , for different values of @xmath60 .
we showed that , in the presence of flow , the signal is stronger over the momentum regions analyzed in the plots , suggesting that flow may help to effectively discover the bbc signal experimentally .
another important point that we find , within this simplified model and in the non - relativistic limit considered here , is that the squeezing would distort significantly the hbt correlation function as well , tending to oppose to the flow effects on those curves , practically neutralizing it for large values of @xmath58 .
the analysis in terms of the variable @xmath61 would not be suited for a genuine relativistic treatment . in this case , however , a momentum variable could be constructed , as @xmath62 . in fact , it would be preferable to redefine this variable as @xmath63 , whose non - relativistic limit is @xmath64 , as discussed in ref.@xcite . finally , it is important to emphasize that all the effects and signals discussed here would exist only if the particles analyzed had their masses modified by interactions in the hot and dense medium .
m. asakawa m , t. csrg and m. gyulassy _ phys .
lett . _ * 83 * , 4013 ( 1999 ) .
p. k. panda , t. csrg , y. hama , g. krein and sandra s. padula , _ phys
b * 512 * , 49 ( 2001 ) .
sandra s , padula , y. hama , g. krein , p. k. panda and t. csrg , _ phys .
_ c * 73 * , 044906 ( 2006 ) .
m. gyulassy , s. k. kaufmann , and l. w. wilson , phys .
c * 20 * , 2267 ( 1979 ) .
a. makhlin and yu .
sinyukov , _ sov .
phys . _ * 46 * , 354 ( 1987 ) ; yu .
sinyukov , _ nucl .
phys . _ a * 566 * , 589c ( 1994 ) .
sandra s. padula , o. socolowski jr .
, t. csrg , m. i. nagy , proc .
quark matter 2008 , _ j. phys .
g : nucl . part .
phys . _ * 35 * , 104141 ( 2008 ) .
sandra s. padula , o. socolowski jr . and danuce m. dudek , arxiv:0812.1784v1[nucl - th ] , to be published in the proc . of the xxxviii international symposium on multiparticle dynamics ( ismd 2008 ) | a novel type of correlation involving particle - antiparticle pairs was found out in the 1990 s .
currently known as squeezed or back - to - back correlations ( bbc ) , they should be present if the hadronic masses are modified in the hot and dense medium formed in high energy heavy ion collisions .
although well - established theoretically , such hadronic correlations have not yet been observed experimentally . in this phenomenological study
we suggest a promising way to search for the bbc signal , by looking into the squeezed correlation function of @xmath0 and @xmath1 pairs at rhic energies , as function of the pair average momentum , @xmath2 .
the effects of in - medium mass - shift on the identical particle correlations ( hanbury - brown & twiss effect ) are also discussed . |
in the unification scheme of agn the difference between type 1 and type 2 agn is explained by angle - dependent circumnuclear obscuration of the accretion disk and broad - line region @xcite . this obscuring dusty medium
commonly referred to as `` dust torus '' is optically and geometrically thick and probably extends from sub - parsec scales outward to several 10s of parsecs , or beyond for high luminosity objects .
the dust in the torus absorbs the incident uv / optical radiation an re - emits the received energy in the infrared
. observations have shown that type 1 agn show significantly more emission in the near - ir than type 2 agn for the same given intrinsic luminosity ( e.g. * ? ? ?
this is consistent with the picture where the face - on view onto the torus in type 1 agn exposes the innermost hot dust to the observer . on the other hand in type 2 agn the torus
is seen edge - on so that internal obscuration blocks the line - of - sight to the hot dust . owing to this effect
, it is expected that for a given agn luminosity the infrared emission of type 1 agn is generally stronger than from type 2s . in the light of attempts at forming isotropic agn samples based on ir fluxes it seems important to know exactly how strong of a bias towards type 1s over type 2s
may occur when invoking flux limits .
moreover , probing the wavelength dependence of this anisotropy in the infrared has some constraining power on our understanding of how the torus obscures the agn . it may be possible to distinguish torus models where the dust is smoothly distributed from those where the dust is arranged in clouds ( e.g. * ? ? ?
* ) : if the dust is smoothly distributed within the torus , a large degree of anisotropy is expected .
if , however , the dust is arranged in clouds the anisotropy is expected to be smaller .
a problem commonly encountered when studying agn samples in the local universe is a significant contribution of the host galaxy to the ir .
this is related to the typical lower luminosity seyfert galaxies which dominate the nearby agn population .
one way around this problem is the use of high - spatial resolution observations , as possible with the largest ground - based telescopes or interferometers , which are able to resolve out the host and isolate the agn emission ( for details see * ? ? ?
* ) . however , it is difficult to set up representative samples owing to the observational limitations .
another possibility is the use of high luminosity objects typically at higher redshift where the agn outshines the host galaxy by a large factor in the optical and near - ir .
if pah features are absent , the agn most likely dominates the mid - ir wavelength region as well ( in our sample : host @xmath4 for wavelengths @xmath5 17@xmath2 ) . in this paper
we aim at quantifying the wavelength dependence of the anisotropy of the agn emission in the infrared from @xmath1 . for that we use a ( nearly ) isotropically selected and complete sample of quasars and radio galaxies with hidden quasars at @xmath0 as recently presented in @xcite and described in sect .
[ sec : sample ] .
here we improve the analysis by using host - galaxy subtraction for the radio galaxies and use clumpy torus models for interpretation . in sect .
[ sec : results ] we show the average seds of each of the subsamples which are representing obscured ( type 2 ) and unobscured ( type 1 ) agn .
we further analyze the origin of the anisotropy by fitting extinction models and clumpy torus models to the observations in sect .
[ sec : analysis ] . in sect .
[ sec : obsaniso ] we discuss our results by comparing them to previous ir anisotropy estimates in literature .
the results are summarized in sect .
[ sec : summary ] .
the object sample for this paper comprises all 3crr @xcite radio galaxies and quasars with @xmath6 this lobe dominated sample presents a well matched set of radio galaxies and quasars in terms of their intrinsic luminosity ( @xmath7erg / s for the quasars and @xmath8erg / s for the radio galaxies ; errors indicate standard deviation of the sample ) .
the data used here have been presented previously in @xcite and @xcite and we refer the reader to these papers for further details on the source selection , data reduction , and building of the average seds . to summarize briefly , we obtained mid - ir photometry in all six filters from 3.6@xmath2 to 24@xmath2 and spectroscopy from 19@xmath2 to 38@xmath2 utilizing all three instruments onboard the _ spitzer _
space telescope @xcite .
after the data reduction in a standard manner , the individual source seds were interpolated onto a common rest frame wavelength grid .
the quasars and radio galaxies were then averaged into a mean sed for each class of objects .
the individual seds ( including observed and interpolated photometry ) as well as the the average seds are presented in ( * ? ? ?
* their figs.1 & 2 ) . for this paper , additional corrections have been applied to the radio galaxy data before the averaging process outlined above : at the shortest wavelengths considered here ( @xmath9@xmath10 m , rest frame ) the radio galaxy seds show contributions from the host galaxy .
since we want to isolate the emission coming from the active nucleus , we have to correct for the stellar emission in these cases .
this correction was performed by fitting the observed irac photometry with a combination of a moderately old ( 5 - 10gyr ) elliptical galaxy sed to represent the stellar emission ( taken from the grasil webpage ; @xcite ) and a hot black body for the agn dust whose temperature we allowed to vary ( see e.g. * ? ? ?
* ; * ? ? ?
the resulting black body temperatures in the radio galaxies range from 600 to 970k , with a median value of 860k .
the fraction of host galaxy light contributing to the flux measured at the observed frame wavelengths 3.6 , 4.5 , 5.8 , and 8.0@xmath10 m was found to be 0.9 , 0.6 , 0.3 , and 0.1 , respectively . despite a negligible contribution from the host galaxy
, we performed similar fits to the quasar sample as well to obtain characteristics of the hot dust emission and compare it to the radio galaxies .
for the quasars we find notably hotter temperatures in the range from 8801250k ( median 1020k ) .
this is consistent with the idea that the hottest dust is obscured in radio galaxies , while directly seen in quasars . in the radio galaxies we then subtracted the estimated host galaxy contribution from the observed irac photometry using the scaled template sed .
for the irs and mips measurements the host galaxy contributions were considered negligible at restframe wavelength @xmath11 and no corrections have been applied .
we refrain from further corrections related to possible starformation .
as pointed out in @xcite neither the individual seds nor the averages showed any pah features ( see also fig .
[ fig : aver_sed ] ) , which indicates that any contribution from starformation to the mid - ir is probably negligible .
the radio galaxy 3c368 has a galactic m - star superimposed close to the position of the radio galaxy nucleus ( e.g. * ? ? ?
* ; * ? ? ?
both sources are partly blended even at the shortest irac wavelengths which makes the correction for the host galaxy emission in this source quite uncertain .
consequently , we removed this source from the sample considered here .
this leaves us with 11 quasars and 8 radio galaxies from which the average seds have been calculated .
in fig . [ fig : aver_sed ] we show the average sed of @xmath13 quasars ( red ) and radio galaxies ( blue ) respectively .
the error bars reflect the mean absolute deviation while the shaded areas show the range of the respective subsample at each wavelength point of the interpolated data ( see sect .
[ sec : sample ] ) . for each of the object types we calculated a spline fit through the mean data points in order to guide the eye .
it is obvious that the radio galaxies are systematically lower in infrared emission than the quasars .
the discrepancy is largest in the near - ir and flattens out towards longer wavelengths .
@xcite showed similar average seds .
here we used additional host galaxy subtraction , which isolates the agn light much better ( see sect . [
sec : sample ] ) .
this is most obvious in the near - ir part of the radio galaxies shortward of 5@xmath2 .
the sed keeps on falling toward shorter wavelengths consistent with the wien tail from hot dust emission , instead of making an upward turn ( see * ? ? ?
since both types have the same radio luminosities due to our selection , this difference between quasars (= type 1 agn ) and radio galaxies (= type 2 agn ) is a generic property of the sample .
it either reflects a difference in line - of - sight extinction ( e.g. by cold dust in the host galaxy ) , or traces the anisotropy in re - emission of the agn - heated dust .
@xcite tested the former possibility and found a surprisingly good match of the difference between radio galaxies and quasars in the mid - ir by a single extinction law .
this , however , breaks down in the near - ir . in this paper
we will test if , instead , a single absorber and emitter (= the dust torus ) may be responsible for the radio galaxy / quasar anisotropy ( see sect .
[ sec : modcomp ] ) . to quantify the wavelength dependence of the anisotropy we plot the quasar / radio galaxy ratio in fig .
[ fig : ratio ] . also shown is the mean absolute deviation of the ratio calculated by propagating the standard deviations of each subsample . from 2 to 8@xmath2 the emission ratio gradually decreases from 20 to 2 . in the silicate feature
this ratio increases again up to about 3 and flattens out towards 15@xmath2 at a value of @xmath14 .
our sample comprises a range of sed shapes , meaning that a range of ratios is observed .
most of this sample range comes from the radio galaxies which show much less uniformity than the quasars ( see fig .
[ sec : sample ] ) .
we illustrated the range of ratios covered by the radio galaxies in fig .
[ fig : range ] where we plot the wavelength dependence of the ratio of each radio galaxy using the average quasar sed and normalize it to 15@xmath2 . at around 10@xmath2
the radio galaxies show the silicate feature in absorption while quasars display a weak silicate emission feature .
we used the spline fits described above ( see fig .
[ fig : aver_sed ] ) to locate the centers of the silicate features following the method outlined by @xcite .
the silicate absorption feature center is found to be at @xmath15 while the peak of the silicate emission feature was measured at @xmath16 .
this difference in central wavelength is similar to the ones observed in local galaxies .
as recent high spatial resolution studies of seyferts suggest , the `` shift '' in wavelength is not a pure radiative transfer effect due to the location or distribution of the dust , but implies a change of dust chemistry within the torus @xcite .
as demonstrated by @xcite the silicate emission feature may be located at around 10.5@xmath2 if a fraction of the hot silicate dust consists of porous silicate grains .
we note that the transition from quasars to radio galaxies is not as smooth as one may expect . in spite of some overlap in the range of seds of both types in the near - ir part in fig .
[ fig : aver_sed ] , quasars generally show infrared emission characteristics expected for a type 1 agn ( blue ir color ; silicate emission feature ) and radio galaxies have ir seds with type 2 emission properties ( redder sed ; silicate absorption feature ) .
in this section we analyze the scenarios that can lead to the observed anisotropy in radio galaxies and quasars .
there are two possibilities : ( 1 ) extinction by a cold dust screen and ( 2 ) absorption and emission in a warm dusty medium .
the former possibility may be associated with cold dust in the host galaxy while the latter one is equivalent to the dust torus in the agn unification scheme ( which we may call `` intrinsic anisotropy '' ) .
we will first discuss the plausibility and consequences of a cold dust screen on the seds . if the ir anisotropy is dominated by cold host galaxy dust
, then we have a situation where the ir emission originates from the torus while the absorption is coming from a different component ( i.e. dust in the host ) .
this means that we require an additional component outside the agn to model the data ( e.g. as used for a significant minority of high-@xmath17 type 2 qsos in @xcite ) .
however , the objects suffering extinction ( here : radio galaxies ) would be offset from quasars by only a single extinction law .
this has been tested and ruled out by @xcite for the same set of objects as presented here . in fig .
[ fig : ratio ] we show the anisotropy ratio between quasars and radio galaxies ( black circles with error bars ) .
we overplot a standard ism extinction curve , resembling cold screen extinction , scaled to the observed 15@xmath2 anisotropy ( dark - blue dotted line ; using a mixture of 53% silicates and 47% graphite , based on updated dust opacity cuves by @xcite , and a grain size distribution according to @xcite ) .
the extinction curve significantly overpredicts the anistropy in the silicate feature with better agreement in the near - ir . for reference we also plot the extinction curve based on (
* light - blue dotted line ; pixie dust ) , as used in @xcite , which results in better agreement within the silicate feature but significant offsets in the near - ir .
in fact , @xcite pointed out that a good correspondence of quasars and radio galaxies can be achieved only if the quasar sed is attenuated by at least _ two _ instead of one extinction components which is reminiscent of radiative transfer (= absorption _ and _ emission ) within the torus rather than a cold screen
. moreover , we found a mean anisotropy of about 1.4 at 15@xmath2 ( see sect . [ sec : results ] ) .
dust opacity curves typically have opacity ratios @xmath18 . in order to obtain the observed anisotropy
, the dust screen would have to have @xmath19 .
while such optical depth values are in reach for galactic dust lanes ( e.g. extinction towards our own galactic center ) , it requires very edge - on views onto disk galaxies since the scale height of galactic disks is small . on statistical grounds
this possibility may be viable for a minority of all radio galaxies , but it is unlikely that the whole population is dominated by host extinction .
we note that the same line of argument can be made using the near - ir anisotropy leading to even higher @xmath20 and illustrating the need for more then just one cold extinction screen in this scenario .
we use our clumpy torus models _ cat3d _
@xcite to test if the observed anisotropy ratio can be explained by the intrinsic anisotropy as predicted in the unification scheme .
generally smooth dust torus models predict stronger anisotropy than clumpy models @xcite .
@xcite argue that the observed small anisotropy in the mid - ir / x - ray correlation ( see sect .
[ sec : obsaniso ] ) is qualitatively in agreement with a clumpy torus .
here we aim at being more quantitative and show consistency of the observed anisotropy with torus orientation . in fig .
[ fig : ratio ] we overplot the observed ratio with predicted ratios of a clumpy torus model ( red dashed line ) . as mean torus inclination in quasars
we assumed 39@xmath21 while the mean radio galaxy inclination is set to 75@xmath21 .
this corresponds to a mean opening angle of the torus of 60@xmath21 or a type 1/type 2 ratio of 1:1 , which is consistent with the number statistics of our complete and isotropic sample @xcite .
[ fig : ratio ] shows that the model is following the continuum anisotropy curve quite well .
there is , however , a slight deviation within the silicate feature where in the center of the feature the model curve predicts slightly lower anisotropy than shown by the observed curve .
this may be indicative of additional , off - torus obscuration in some objects ( e.g. from the host galaxy ) , which effectively deepens silicate absorption features in radio galaxies ( the denominator of the plot in fig .
[ fig : ratio ] ) and , to a lesser degree , changes the spectral slope .
if this happens in individual objects , the average curve and scatter will tend to be slightly more anisotropic .
some local examples of host obscuration are centaurus a , ngc 5506 , or the nucleus of the circinus galaxy where host - galactic dust lanes are projected onto the nucleus producing deep silicate features .
in fact the silicate absorption feature in the average radio galaxy sed seems to be deeper than in typical local seyfert 2 galaxies without host obscuration @xcite .
note that this requires a statistical alignment of cold host dust with that of the torus ( see sect .
[ sec : extinct ] ) .
the model used for reproducing the ratio uses a dust cloud distribution with radial power law @xmath22 and 5 clouds along the line - of - sight in equatorial direction ( for details see * ? ? ?
* ) . in comparison
to models for seds and mid - ir interferometry of local seyfert galaxies , these parameters suggest only a slightly more centrally condensed and transparent torus @xcite , while the half - opening angle may be wider ( 60@xmath21 instead of 45@xmath21 ) .
in fact a range of torus model parameters satisfies the observed type 1/type 2 anisotropy spectrum within the error bars of the sample ( e.g. various steeper and shallower dust distributions ) . from bayesian inference analysis we found that the torus model parameters are generally poorly constrained ( broad posterior distributions for individual parameters ) .
a weakness is certainly that modeling the flux ratio is not very constraining for model parameters since it does only take relative fluxes into account , while absolute fluxes ( e.g. actual silicate strength of type 1s or type 2s ) are not included . on the other hand , what the modeling shows is that the observed small ratios at long wavelengths and the change of anisotropy from the near- to the mid - ir are in agreement with expectations from clumpy torus models without fine - tuning parameters . in summary , the torus model seems to reproduce the observed anisotropy reasonably well over most of the wavelength range , while single extinction laws result in much worse fits .
the model parameters used in the clumpy torus model fit are reasonable in comparison to fits to local agn .
it demonstrates plausibility of the scenario that the anisotropy is related to torus orientation .
this strongly suggests that the observed anisotropy is a measure for the intrinsic anisotropy of luminous type 1 and type 2 agn at @xmath12 .
in sect . [ sec : analysis ] we argued that the observed ir anisotropy is probably reflecting the `` intrinsic anisotropy '' as caused by the dust torus .
the isotropic and complete selection of the sample helps to minimize any biasing effects on the anisotropy . on the other hand , optically- and
x - ray - selected samples often suffer from missing some of the most obscured objects .
moreover , since the ratio is @xmath23 at 15@xmath2 , flux - limited mid - ir selected samples are potentially biased towards type 1 agn as well .
thus , even using 12@xmath2-selected agn samples will slightly suffer from the assumption of isotropy .
one popular way of comparing type 1s and type 2s is the correlation between x - ray and mid - ir luminosity . in case
the x - luminosity is emitted isotropically ( and traces the dust - heating emission ) , and the mid - ir emission is radiated very anisotropically , this correlation is expected to be different for type 1 and type 2 agn . however , using high spatial resolution observations @xcite showed that local seyfert 1 and seyfert 2 galaxies , up to column densities of few @xmath24 @xmath25 , essentially follow the same correlation .
a conservative estimate suggests that at 12@xmath2 the difference between both types is smaller than a factor of 3 . despite including some `` mildly '' compton - thick objects in this study
, the most obscured objects are still missing due to the lack of intrinsic x - ray data .
this essentially makes anisotropy estimates from the mid - ir / x - ray - correlation a lower limit on the `` true '' anisotropy .
nevertheless this result is fully consistent with our finding for powerful @xmath12 agn . at 12@xmath2
we find a type 1/type 2 ratio of @xmath26 . assuming that x - ray selection misses the highest - inclination objects our result would predict that the anisotropy in the mid - ir / x - ray correlation is @xmath27 , at least for powerful agn as presented here .
@xcite used a sample of local seyfert galaxies and compared the average 535@xmath2 sed of type 1s and type 2s , scaled to their respective 8.4ghz emission .
however since the sources have been selected according to a flux limit at 12@xmath2 , the sample can not be considered isotropic .
@xcite report generally higher fluxes for type seyfert 1 agn as compared to seyfert 2s .
the anisotropy decreases from a factor of about 8 to @xmath32.5 from 5 to 8@xmath2 .
this is significantly larger than what we find for our isotropically selected radio - loud sample . at longer wavelengths the seyfert 1/seyfert 2 ratio of @xcite flattens to about a factor of 23 ( with no convergence to unity as expected at long wavelengths ) which is , again , larger than what we found .
the discrepancy between our results and @xcite may be either due to ( 1 ) the different selection criteria chosen , ( 2 ) a difference between low and high luminosity agn , or ( 3 ) a difference between radio - quiet and radio - loud agn . while the data in @xcite has not been corrected for host galaxy , the contribution by starformation should not be significant given the lack of pah features in the average spectrum .
if a host - correction were applied , it would predominantly affect type 2 agn , thus making the anisotropy even larger .
furthermore , if the mid - ir selection had any effect , then the sample would miss out agn at highest obscuration , so that the real type 1/type 2 anisotropy would again be larger . in conclusion
, possible selection effects and host contamination in the @xcite study would tend to result in an underestimated anisotropy , making the difference to our findings even stronger .
it is well possible that the difference in anisotropy between our high - luminosity radio - loud sample and the low - luminosity radio - quiet sample is real .
this would imply that either luminosity or radio power are the drivers for the observed characteristics .
radio jets are highly collimated so that any influence of the jet can be expected perpendicular to the torus plane and just in a very small solid angle .
it is , therefore , more reasonable to assume that the higher agn luminosity would cause the lower anisotropy than the jet .
the classical receding torus picture changes the opening angle of the torus for higher luminosity ( e.g. * ? ? ?
* ; * ? ? ?
this mainly affects the relative number of type 1s and type 2s in a sample .
the relatively high fraction of about 50% unobscured agn in our sample would support this scenario .
however , to change the anisotropy between both types , it would be necessary to also change the obscuration properties ( i.e. type 2s must on average look more like type 1s ) .
such a scenario of `` radiation - limited obscuration '' has been proposed by @xcite and supported by observations of @xcite . in this case about the same effect is expected for radio - quiet and radio - loud agn .
we use the sample of quasars and radio galaxies at @xmath12 recently presented in @xcite . since the sample was selected isotropically
, it should cover all torus inclination angles ( weighted by solid angle ) .
for these objects an infrared 117@xmath2 restframe sed has been constructed .
average seds were calculated for the quasar (= type 1 agn ) and radio galaxy (= type 2 agn ) samples , respectively , in order to study the intrinsic anisotropy of the ir emission of the dust torus .
it is shown that the ratio between type 1 and type 2 agn in our parameter space of very luminous radio galaxies , the value gradually decreases from 20 to 2 at wavelengths 2 to 8@xmath2 . within the 10@xmath2 silicate feature the ratio raises slightly . at longer wavelength
the mid - ir emission becomes more isotropic .
the intrinsic ratio between our type 1 and type 2 agn is @xmath14 at 15@xmath2 . when using ir - selected flux - limited samples this anisotropy has to be taken into account .
by analyzing the silicate feature in the sample averages we find the well - established `` shift '' of the central peak of the silicate emission feature with respect to the center of the absorption feature .
the resulting central wavelengths at @xmath15 for the absorption and @xmath16 for the emission feature are in agreement with previous reports ( e.g. * ? ? ?
* ; * ? ? ?
we discussed our results in the frame of other anisotropy estimators .
our findings are consistent with upper limits derived from the x - ray / mid - ir correlation of local seyfert galaxies @xcite .
some discrepancy exists with respect to a similar study of @xcite for nearby seyferts .
if real it would imply that nearby , radio - quiet lower - luminosity agn show a higher degree of anisotropy in the ir than higher luminosity , radio - loud sources .
this may be explained by a receding torus model with luminosity - dependent obscuration .
we also show that the overall relatively small degree of anisotropy is consistent with the torus being clumpy rather than smooth .
our clumpy torus model reproduces the observed type 1/type 2 ratio reasonably well .
we would like to thank our referee prof .
andy lawrence for helpful and constructive comments which significantly improved the paper , as well as poshak gandhi who also commented on this manuscript .
the paper was made possible by deutsche forschungsgemeinschaft ( dfg ) in the framework of a research fellowship ( `` auslandsstipendium '' ) for sh .
mh is supported by the nordrhein - westflische akademie der wissenschaften und der knste .
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 .
antonucci , r. 1993 , ara&a , 31 , 473 best , p. n. , longair , m. s. , & roettgering , j. h. a. 1997 , mnras , 292 , 758 buchanan , c. l. , gallimore , j. f. , odea , c. p. , baum , s. a. , axon , d. j. , et al .
2006 , aj , 132 , 401 chiar , j. e. , & tielens , a. g. g. m. 2006 , apj , 637 , 774 de breuck , c. , seymour , n. , stern , d. , willner , s. p. , eisenhardt , p. r. m. , et al .
2010 , apj , 725 , 36 draine , b. t. 2003 , apj , 598 , 1026 gandhi , p. , horst , h. , smette , a. , hnig , s. , comastri , a. , et al .
2009 , a&a , 502 , 457 haas , m. , willner , s. p. , heymann , f. , ashby , m. l. n. , fazio , g. , et al .
2008 , apj , 688 , 122 hammer , f. , proust , d. , & le fevre , o. 1991 , apj , 374 , 91 hnig , s. f. , & beckert , t. 2007 , mnras , 380 , 1172 hnig , s. f. , kishimoto , m. , gandhi , p. , smette , a. , asmus , d. , duschl , w. , polletta , m. , weigelt , g. 2010 , a&a , 515 , 23 hnig , s. f. , & kishimoto , m. 2010 , a&a , 523 , 27 laing , r. a. , riley , j. m. , & longair , m. s. 1983 , mnras , 204 , 151 lawrence , a. 1991 , mnras , 252 , 586 leipski , c. , haas , m. , meusinger , h. , siebenmorgen , r. , chini , r. , et al .
2005 , a&a , 440 , l5 leipski , c. , haas , m. , siebenmorgen , r. , meusinger , h. , albrecht , m. , et al .
2007 , 473 , 121 leipski , c. , haas , m. , willner , s. p. , ashby , m. l. n. , wilkes , b. j. , et al .
2010 , apj , 717 , 766 levenson , n. a. , radomski , j. t. , packham , c. , mason , r. e. , schaefer , j. j. , telesco , c. m. 2009 , apjl , 703 , 390 mathis , j. s. , rumpl , w. , & nordsieck , k. h. 1977 , apj , 217 , 425 polletta , m. , weedman , d. w. , hnig s. f. , lonsdale , c. j. , smith , h. e. , houck , j. r. 2008 , apj , 675 , 960 schartmann , m. , meisenheimer , k. , camenzind , m. , wolf , s. , tristram , k. r. w. , & henning , t. 2008 , a&a , 482 , 67 seymour , n. , stern , d. , de breuck , c. , vernet , j. , rettura , a. , et al .
2007 , apjs , 171 , 363 shi , y. , rieke , g. h. , hines , d. c. , gorjian , v. , werner , m. w. , et al . 2006 , apj , 653 , 127 silva , l. , granato , g.l . , bressan , a. , danese , l. , 1998 , apj , 509 , 103 simpson , c. 2005 , mnras , 360 , 565 sirocky , m. m. , levenson , n. a. , elitzur , m. , spoon , h. w. w. , armus , l. 2008 , apj , 678 , 729 smith , h. a. , li , a. , li , m. p. , khler , m. , ashby , m. l. n. , et al .
2010 , apj , 716 , 490 sturm , e. , schweitzer , m. , lutz , d. , contursi , a. , genzel , r. , et al .
2005 , apj , 629 , l21 treister , e. , virani , s. , gawiser , e. , urry , c. m. , lira , p. , et al .
2009 , apj , 693 , 1713 urry , c. m. , & padovani , p. 1995 , pasp , 107 , 803 werner , m. w. , et al .
2004 , apjs , 154 , 1 | we use restframe near- and mid - ir data of an isotropically selected sample of quasars and radio galaxies at @xmath0 , which have been published previously , to study the wavelength - dependent anisotropy of the ir emission . for
that we build average seds of the quasar subsample (= type 1 agn ) and radio galaxies (= type 2 agn ) from @xmath1 and plot the ratio of both average samples . from 2 to 8@xmath2 restframe wavelength
the ratio gradually decreases from 20 to 2 with values around 3 in the 10@xmath2 silicate feature .
longward of 12@xmath2 the ratio decreases further and shows some high degree of isotropy at 15@xmath2 ( ratio @xmath31.4 ) .
the results are consistent with upper limits derived from the x - ray / mid - ir correlation of local seyfert galaxies .
we find that the anisotropy in our high - luminosity radio - loud sample is smaller than in radio - quiet lower - luminosity agn which may be interpreted in the framework of a receding torus model with luminosity - dependent obscuration properties .
it is also shown that the relatively small degree of anisotropy is consistent with clumpy torus models . |
in this letter we address dynamical processes in highly ordered complex plasmas associated with _ spontaneous symmetry breaking_. spontaneous symmetry breaking ( ssb ) plays a crucial role in elementary particle physics but is also very common in classical physics @xcite .
it happens whenever the system goes from a state which has a certain symmetry , e.g. rotational symmetry , into an ordered state , which does not have this symmetry anymore . in general
, this state not necessarily has to be the ground ( vacuum ) state and the transition to the new state may or may not be associated with a phase transition .
for example , in the case of magnetization the spins point all in one direction ( ordered state ) whereas above the curie temperature there is no preferred direction .
another example from a mechanical system without phase transition is a vertical stick which bends under a sufficiently high force from above to one side breaking the rotational symmetry of the system without the force .
different symmetries coexisting in the same phase , and symmetry transformations escorting phase transitions are widely spread in nature . for instance , the mechanisms of symmetry breaking are thought to be inherent in the molecular basis of life @xcite .
ssb is also an important feature of elementary particle physics @xcite .
the universe itself is believed to have experienced a cascade of symmetry - breaking phase transitions which broke the symmetry of the originally unified interaction giving rise to all known fundamental forces @xcite .
symmetry effects are crucial either in 3d and 2d systems .
chiral ( mirror - isomeric ) clusters @xcite , magic clusters of a new symmetry frozen - in by a solid surface @xcite , or dynamical symmetry breaking by the surface stress anisotropy of a two - phase monolayer on an elastic substrate @xcite are examples of the importance of 2d or quasi-2d systems in many applications .
low pressure , low temperature plasmas are called _ complex plasmas _ if they contain microparticles as an additional thermodynamically active component . in the size domain of 1 - 10@xmath0 m ( normally used in experiments with complex plasmas ) these particles
can be visualized individually , providing hence an atomistic ( kinetic ) level of investigations @xcite .
the interparticle spacing can be of the order of 0.1 - 1 mm and characteristic time - scales are of the order of 0.01 - 0.1 s. these unique characteristics allow to investigate the microscopic mechanism of ssb and phase transitions at the kinetic level .
common wisdom dictates that symmetry breaking is an inherent attribute of systems in an active state .
hence these effects are naturally important in complex plasmas where the _ particle cloud - plasma _ feedback mechanisms underlying many dynamical processes are easy to vitalize .
also in complex plasmas where different kind of phase transitions exist , e.g. in the elelectrorheological plasmas @xcite , one can find examples for classical ssb . another option , interesting in many applications , is the clustering of a new phase which is dissymmetric with regard to a background symmetry ( as an example of fluid phase separation in binary complex plasmas see @xcite ) .
it is important to mention that the microparticles , collecting electrons and ions from the plasma background , become charged ( most often negatively @xcite ) and hence should be confined by external electric fields .
the configuration of the confining forces might deeply affect the geometry and actual structure of the microparticle cloud . in rf discharge complex plasmas
the particles are self - trapped inside the plasma because of a favorable configuration of the electric fields @xcite .
one of the interesting things is the possibility to levitate a monolayer of particles under gravity conditions . in this case
the particle suspension has a flat practically two dimensional structure .
this is , of course , a very attractive simplification ( from a theoretical point of view ) , significantly lowering the description difficulties .
below we concentrate mostly on 2d complex plasmas . depending on the discharge conditions , the monolayer can have crystalline or liquid order .
2d configurations of dust particles either in crystalline or liquid state were successfully used to study phase transitions , dynamics of waves and many transport phenomena in complex plasmas @xcite .
a symmetry disordering escorting a crystalline - liquid phase transition has been investigated experimentally in @xcite .
dislocation nucleation ( a shear instability ) has been reported in @xcite , albeit the importance of ssb for this phenomenon has not been explained .
the results of these recent experimental observations can not be properly addressed without a deep understanding of this important issue .
we would like to highlight this in the paper and report on the physics of spontaneous disordering of a cold plasma crystal , simulated melting and crystallization process , including associated defect clusters nucleation , dissociation , and symmetry alternation .
these options are realizable in experimental complex plasmas , and can be mimicked in simulations , as we demonstrate below .
it is well known that two broken symmetries distinguish the crystalline state from the liquid : the broken translational order and the broken orientational order . in two dimensions for ordinary crystals
it is also well known that even at low temperatures the translational order is broken by spontaneous disordering mediated by thermal fluctuations @xcite . as a result ,
the fluctuation deflections ( disordering ) grow with distance and translational correlations decay ( algebraically , see @xcite ) .
2d plasma crystals also obey this common rule .
the character of disordering may be deeply affected by the confinement forces , though .
usually such an in - plane confinement is due to the bowl - shaped potential well self - maintained inside the discharge chamber , which to first order is approximately parabolic ( see , e.g. @xcite ) , that is @xmath1 , where @xmath2 is the distance , @xmath3 is the particle mass , and @xmath4 is the _ confinement parameter _ @xcite .
( the out - of - plane confining forces , controlling the position of the entire lattice , are normally much stronger ; below we consider the pure 2d - case , assuming , hence , an absolutely stiff out - of - plane confinement . ) the fluctuation spectra can be calculated in the following manner .
the long - range phonon contribution to the free energy of a 2d system of particles interacting via the yukawa potential and confined by a shallow isotropic parabolic well can be conveniently represented as @xcite : @xmath5 @xmath6 where @xmath7 are the transverse ( shear wave ) and the longitudinal ( compressional wave ) sound speed , and @xmath8 are the fourier components of the vorticity @xmath9 and the divergency @xmath10 of the particle displacements @xmath11 , and @xmath12 is the wave vector ( @xmath13 ) .
the unperturbed crystal is supposed to be hexagonal .
the relationship ( [ eq.a ] ) provides ( see , e.g. @xcite ) the probability of the fluctuation @xmath14 .
next , using it , we can calculate the averaged fluctuation spectral intensity per unit mass as @xmath15 where @xmath16 is the particle thermal velocity .
it is known ( see @xcite ) that a lattice layer of a finite size @xmath17 is stably confined if roughly : @xmath18 here the parameter @xmath19 stands for an effective number of the nearest neighbors of any edge particle .
note that according to ( [ eq.4 ] ) formally @xmath20 at @xmath21 . without confinement ( @xmath22 ) the fluctuation spectrum
( [ eq.3 ] ) apparently diverges @xmath23 at @xmath24 , and , as a consequence , in agreement with @xcite , the crystal ordering decays algebraically with the distance @xmath2 , i.e. the density - density correlation behaves as @xmath25 here @xmath26 , @xmath27 is the vector of the reciprocal lattice , and @xmath28 .
it is assumed that @xmath29 is large compared to the interparticle separation @xmath30 . in the experiments @xmath31 is always finite ( though noticeably small , one or two orders of magnitude less than the frequency of the local caged oscillations of the individual particles @xcite ) . from ( [ eq.3 ] ) at non - vanishing @xmath4 it immediately follows that the fluctuations remain finite even at @xmath32 .
this absence of a singularity alters the character of disordering from algebraic ( [ eq.5 ] ) to exponential at a scale depending on the confinement parameter : @xmath33 it is essential that both asymptotes algebraic and exponential must be treated as near - field ( @xmath34 ) and far - field ( @xmath35 ) approximations .
hence it would be logical to assume that the ordering decay alternates with distance from algebraic to exponential .
this is indeed in qualitative agreement with observations @xcite .
remarkably ( [ eq.a])-([eq.3 ] ) are formally similar to the equations describing director fluctuations in nematic crystals in the presence of a magnetic field @xcite .
the action of the magnetic field is known as suppressing the large - scale director fluctuations in liquid crystals .
the length scale @xmath36 seems to be of a fundamental importance .
the particles , experiencing a horizontal confinement , are distributed non - uniformly .
the steady - state displacements of the particles @xmath37 in the plasma crystal from their ideal locations in a uniform 2d crystal represent a growing function with distance @xmath38 @xcite .
the lattice breaks up when @xmath39 where @xmath40 is the lindemann parameter . since @xmath40=0.16 - 0.18 ( see , e.g. @xcite ) , it follows that the first row of defects most probably appears at @xmath41 . making use of ( [ eq.4 ] ) , ( [ eq.7 ] )
one can estimate the size of the domains ( or equivalent correlation length ) as : @xmath42 the correlation length ( [ eq.9 ] ) does not depend explicitly on the temperature .
in other words , for purely topological reasons the big crystal spontaneously splits , assembling an array of sub - domains , even at zero temperature .
the estimated values of @xmath43 agree well with those obtained in the simulation
see fig .
[ fig : cluster ] , and in experiments . for instance
, it has been observed in @xcite that the crystal orientational order had a power law decay at distances @xmath44 in fairly good agreement with @xmath45 following from ( [ eq.9 ] ) .
the one - plus correlation length ( [ eq.9 ] ) , unavoidably introducing a network of sub - domains to a lattice layer , is of crucial importance , e.g. , for observations of the so called _ hexatic state _ in the plasma crystals that is still an outstanding and controversial issue in complex plasma studies @xcite .
one of the possible scenarios for melting ( recrystallization ) in a 2d complex plasma is a precipitous increase ( decrease ) in the density of the dislocations and the dislocation aggregates ( such as defect clusters , grain boundaries etc . )
@xcite . to realize this scenario in simulations , it is desirable to avoid any aforementioned complications associated with the lattice layer sectioning from the very beginning. a promising tool in that sense
, allowing to create a defect - free initial lattice layer , is a hexagonal confinement cell proposed in @xcite .
we performed a series of simulations that revealed several peculiarities in symmetry that are worth to mention .
first , the order parameter of the paired defects dislocations , was systematically lower for 7-fold cells .
this is not surprising actually from a purely geometric point of view because the 5-fold cell in a pair is more compact .
second , simulations manifested that not only isolated pairs dislocations ( @xmath46@xmath47 ) , but also compact triplets like ( @xmath46@xmath47@xmath46 ) , quadruplets ( @xmath46@xmath47@xmath48@xmath49 ) etc . , or even elongated defect chains were quite frequent .
actually they dominantly defined the symmetry of the entire particle suspension .
it would certainly be promising to connect the cluster formation in ordered complex plasmas @xcite with the general percolation process known in many similar applications ( see , e.g. , @xcite ) .
third , in such melted clusters , in agreement with recent experimental observations @xcite , the defect density permanently decreased upon cooling . at higher temperatures in the beginning of the recrystallization process , while the mutual interparticle collisions were still frequent , the defect density dropped exponentially .
then , at lower temperatures , the decay rate significantly slowed down ( see fig.[fig : defect ] ) .
a sharp drop in the defect numbers followed by a quasi - saturation resembles the well - known situation @xcite in which both thermal activation and tunneling events occur .
hence , by analogy , the fact that in our case the system of defects behaves in a similar way could be naturally explained by an annihilation scenario which is presumably of the _ dissipative tunneling _
type @xcite at lower mean kinetic energies .
nucleation of dislocations is another important example of spontaneous symmetry breaking on a scale of elementary cells . whatever the melting scenario would be true ,
still there would remain a question what mechanism explains nucleation of the primary dislocation clusters . recently
this issue has been studied experimentally : spontaneous nucleation of the edge - dislocation pairs ( followed by their dissociation ) has been successfully observed at the kinetic level in the experiments with plasma crystals @xcite .
since the burgers vector of the entire lattice is kept constant ( e.g. zero ) spontaneously created dislocations must be paired forming defect quadruplets of the type ( @xmath46@xmath47@xmath48@xmath49 ) .
( the burgers vector characterizes the magnitude and direction of the crystalline lattice distortion by a dislocation @xcite . )
these dislocation clusters were created in the lattice locations where the internal shear stress exceeded a threshold .
it has also been shown that even an elementary act of nucleation is in fact a multi - scale process consisting of the latent pre - phase , prompt nucleation of a defect cluster , and dissociation of the cluster followed by the escape of free dislocations @xcite . in the experiments @xcite
it was suggested that the stress that finally caused nucleation was affected by the differential crystal rotation .
the exact reason of nucleation , however , was difficult to determine consistently . in simulations the nucleation conditions are certainly easier to identify .
to demonstrate nucleation in simulations a deformable hexagonal cell is used ( fig.[fig : nucleation ] ) .
it confines a 2d cloud of equally charged particles interacting pairwise via the yukawa ( the screened coulomb ) force : @xmath50 where @xmath51 is the relative coordinate and @xmath52 is the distance between the particles @xmath53 with the coordinates @xmath54 ; @xmath55 is the particle charge and @xmath56 is the screening length .
the cell design is similar to that applied in @xcite to simulate melting and recrystallization process of the plasma crystal .
the hexagonal simulation cell has the evident advantage of flexible shape , compared to , e.g. , a parabolic cell confinement . deforming the boundary of the cell , it is simple to manipulate the particles in a tractable way .
an additional option of variable geometry enables an opportunity to separate or consolidate pure shear and simple shear deformation @xcite if desirable .
the strain rate is controllable during deformation as well .
[ fig : nucleation ] shows a simple - sheared particle lattice layer . at a properly chosen loading rate deformation affects the _ shear instability _ that ends up with nucleation of defect clusters in the bulk of the lattice layer . after a while , when deformation becomes stronger , the components of the clusters decoupled and the newly born free dislocations glided away in a similar manner as the dislocations observed in experiments .
symmetry alternation is of primary importance for understanding nucleation of dislocation clusters .
the compact cluster design is _ magic _ in the sense that the _ hexagonal _ symmetry of the particle system neatly turns into a nearly _ tetratic _ symmetry of the cluster core ( like lead turns into gold when touched by the philosophers stone ) , see fig .
[ fig : topology ] . despite an apparent simplicity of the cluster interior
only four nearest neighbor particles ( marked abcd in fig .
[ fig : topology ] ) , the centers of the 5- and 7-fold cells , are in the core , to discover the cluster topology was certainly a challenge @xcite . in our case
the interparticle interaction potential is of the screened yukawa type , hence more compact in contrast to the @xmath57 interaction potential in case of magnetically interacting super - paramagnetic colloid particles considered in @xcite .
thus there is a unique opportunity to verify whether the core topology is universal .
let us start with a simple model treating the cluster as constituted of two point - like dislocations , which are set apart at a distance @xmath2 and allowed to glide only along two fixed crystallographic planes separated by one lattice period @xmath30 , so that @xmath58 , where @xmath59 is the angle of mutual orientation of the cluster components with respect to the gliding plane .
the interaction energy of the point dislocations having the counter - directed burgers vectors is @xcite : @xmath60+const.\ ] ] it has a minimum , a _
stable ground state _
, at @xmath61 .
it corresponds to @xmath62 , hence tetragonal symmetry of the cluster core might be considered as preferred .
this prediction agrees noticeably well with the results of simulations of finite clusters : on average in fig . [
fig : topology ] the edge - to - diagonal angle in the cluster core is @xmath63 .
it is worth noting that the cluster core is nearly cyclic .
a measure of it immediately follows from the famous ptolemy s inequality valid for any quadrilateral : @xmath64 where @xmath65 denote the ( ordered ) sides , and @xmath66 are the diagonals of the quadrilateral . over 80 % of the recognized clusters
have @xmath67 for the simulation results shown in fig .
[ fig : topology ] . for comparison
a hexagonal four - side cell corresponds to @xmath68 .
note also that a stable defect cluster could not be obtained only by shifting positions of four central particles from a hexagonal configuration to tetragonal one .
such deformation would be reversible , hence unstable .
a weakly deformed environment , impeding relaxation of the core particles back to the stable hexagonal configuration , is indeed a necessary lock making the deformation plastic , i.e. irreversible . at sufficiently strong external stress even a stable cluster dissociates .
whatever is the orientation of the cluster as a whole , escaping dislocations can glide only along two crystallographic directions ( along burgers vectors , see fig . [
fig : topology ] ( a , b ) ) .
this naturally explains the asymmetry of the escape directions and chirality of the defect configurations revealed by the newly nucleated dislocations in experiments @xcite .
spontaneous symmetry breaking is a common and inherent feature of many systems in physics as well as other fields , it plays an important role , for example , from classical one - component plasmas to modern string representations @xcite , from the evolution of the early universe @xcite to the dynamics of a wide variety of small - scale systems @xcite .
therefore it is not surprising that ssb is present also in the physics of plasma crystals .
as example we have considered dislocations in plasma crystals which exhibit a spontaneous disordering , involved in the process of melting , and form clusters that are caused by a shear instability and show an interesting topological symmetry .
the authors appreciate valuable discussions with dr .
ivlev and dr . nosenko .
k. r. stterlin , a. wysocki , a. v. ivlev , c. rth , h. m. thomas , m. rubin - zuzic , w. j. goedheer , v. e. fortov , a. m. lipaev , v. i. molotkov , o. f. petrov , g. e. morfill , and h. lwen , phys .
lett . * 102 * , 085003 ( 2009 ) . | spontaneous symmetry breaking is an essential feature of modern science .
we demonstrate that it also plays an important role in the physics of complex plasmas .
complex plasmas can serve as a powerful tool for observing and studying discrete types of symmetry and disordering at the kinetic level that numerous many - body systems exhibit . |
in the present almost frenetic rate of advance of cosmology it is useful to be reminded that the big news this year is the establishment of evidence , by two groups ( @xcite , @xcite ) , of detection of the relativistic curvature of the redshift - magnitude relation .
the measurement was proposed in the early 1930s .
compare this to the change in the issues in particle physics since 1930 .
the slow evolution of cosmology has allowed ample time for us to lose sight of which elements are reasonably well established and which have been adopted by default , for lack of more reasonable - looking alternatives .
thus i think it is appropriate to devote a good part of my assigned space to a discussion of what might be included in the standard model for cosmology .
i then comment on additions that may come out of work in progress .
main elements of the model are easily listed : in the large - scale average the universe is close to homogeneous , and has expanded in a near homogeneous way from a denser hotter state when the 3 k cosmic background radiation was thermalized . the standard cosmology assumes conventional physics , including general relativity theory .
this yields a successful account of the origin of the light elements , at expansion factor @xmath0 .
light element formation tests the relativistic relation between expansion rate and mass density , but this is not a very searching probe . the cosmological tests discussed in 3 could considerably improve the tests of general relativity .
the model for the light elements seems to require that the mass density in baryons is less than that needed to account for the peculiar motions of the galaxies .
it is usually assumed that the remainder is nonbaryonic ( or acts that way ) .
our reliance on hypothetical dark matter is an embarrassment ; a laboratory detection would be exceedingly welcome . in the past decade
many discussions assumed the einstein - de sitter case , in which there are negligibly small values for the curvature of sections of constant world time and einstein s cosmological constant @xmath1 ( or a term in the stress - energy tensor that acts like one ) .
this is what most of would have chosen if we were ordering .
but the evidence from the relative velocities of the galaxies has long been that the mass density is less than the einstein - de sitter value @xcite , and other more recent observations , notably the curvature of the redshift - magnitude relation ( @xcite , @xcite ) , point in the same direction .
now there is increasing interest in the idea that we live in a universe in which the dominant term in the stress - energy tensor acts like a decaying cosmological constant ( @xcite - @xcite ) .
this is not part of the standard model , of course , but as discussed in 3 the observations seem to be getting close to useful constraints on space curvature and @xmath1 .
we have good reason to think structure formation on the scale of galaxies and larger was a result of the gravitational growth of small primeval departures from homogeneity , as described by general relativity in linear perturbation theory .
the adiabatic cold dark matter ( acdm ) model gives a fairly definite and strikingly successful prescription for the initial conditions for this gravitational instability picture , and the acdm model accordingly is widely used in analyses of structure formation .
but we can not count it as part of the standard model because there is at least one viable alternative , the isocurvature model mentioned in 3.3 .
observations in progress likely will eliminate at least one , perhaps establish the other as a good approximation to how the galaxies formed , or perhaps lead us to something better .
the observational basis for this stripped - down standard model is reviewed in references @xcite and @xcite . here
i comment on some issues now under discussion .
pietronero @xcite argues that the evidence from redshift catalogs and deep galaxy counts is that the galaxy distribution is best described as a scale - invariant fractal with dimension @xmath2 .
others disagree ( @xcite , @xcite ) .
i am heavily influenced by another line of argument : it is difficult to reconcile a fractal universe with the isotropy observed in deep surveys ( examples of which are illustrated in figs . 3.7 to 3.11 in @xcite and are discussed in connection with the fractal universe in pp . 209 - 224 in @xcite ) .
-1.0truecm fig .
1 shows angular positions of particles in three ranges of distance from a particle in a fractal realization with dimension @xmath3 in three dimensions . at @xmath3
the expected number of neighbors scales with distance @xmath4 as @xmath5 , and i have scaled the fraction of particles plotted as @xmath6 to get about the same number in each plot .
the fractal is constructed by placing a stick of length @xmath7 , placing on either end the centers of sticks of length @xmath8 , where @xmath9 , with random orientation , and iterating to smaller and larger scales .
the particles are placed on the ends of the shortest sticks in the clustering hierarchy .
this construction with @xmath10 ( and some adjustments to fit the galaxy three- and four - point correlation functions ) gives a good description of the small - scale galaxy clustering @xcite .
the fractal in fig . 1
, with @xmath3 , the dimension pietronero proposes , does not look at all like deep sky maps of galaxy distributions , which show an approach to isotropy with increasing depth .
this can not happen in a scale - invariant fractal : it has no characteristic length .
a characteristic clustering length for galaxies may be expressed in terms of the dimensionless two - point correlation function defined by the joint probability of finding galaxies centered in the volume elements @xmath11 and @xmath12 at separation @xmath13 , dp = n^2[1+_gg(r)]dv_1dv_2 .
[ eq : xigg ] the galaxy two - point function is quite close to a power law , = ( r_o / r)^,= 1.77 , 10hr10 , [ eq : xiggparameters ] where the clustering length is r_o=4.50.5 , and the hubble parameter is h_o=100h^-1 ^ -1 .
the rms fluctuation in galaxy counts in a randomly placed sphere is @xmath14 at sphere radius @xmath15 mpc , to be compared to the hubble distance ( at which the recession velocity approaches the velocity of light ) , @xmath16 mpc .
the isotropy observed in deep sky maps is consistent with a universe that is inhomogeneous but spherically symmetric about our position .
there are tests , as discussed by paczyski and piran @xcite .
for example , we have a successful theory for the origin of the light elements as remnants of the expansion and cooling of the universe through @xmath17 mev @xcite .
if there were a strong radial matter density gradient out to the hubble length we could be using the wrong local entropy per baryon , based on conditions at the hubble length where the cbr came from , yet the theory seems to be successful .
but to most people the compelling argument is that distant galaxies look like equally good homes for observers like us : it would be startling if we lived in one of the very few close to the center of symmetry .
mandelbrot @xcite points out that other fractal constructions could do better than the one in fig . 1 .
his example does have more particles in the voids defined by the strongest concentrations in the sky , but it seems to me to share the distinctly clumpy character of fig .
1 . it would be interesting to see a statistical test .
a common one expands the angular distribution in a given range of distances in spherical harmonics , a_l^m = d()y_l^m ( ) , [ eq : alm ] where @xmath18 is the surface mass density as a function of direction @xmath19 in the sky .
the integral becomes a sum if the fractal is represented as a set of particles .
a measure of the angular fluctuations is e_l = l_-l < m < l |a_l^m|^2/(a_0 ^ 0)^2 , where ^2/^2 - 1 = _ l 1 e_l / l . [ eq : variance ] in the approximation of the sum as an integral @xmath20 is the contribution to the variance of the angular distribution per logarithmic interval of @xmath21 . it will be recalled that the zeros of the real and imaginary parts of @xmath22 are at separation @xmath23 in the shorter direction , except where the zeros crowd together near the poles and @xmath22 is close to zero .
thus @xmath20 is the variance of the fractional fluctuation in density across the sky on the angular scale @xmath24 and in the chosen range of distances from the observer .
i can think of two ways to define the dimension of a fractal that produces a close to isotropic sky .
first , each octant of a full sky sample has half the diameter of the full sample , so one might define @xmath25 by the fractional departure of the mean density within each octant from the mean in the full sample , ( e_2)^1/2~2 ^ 3-d - 1.[eq : e_2 ] thus in fig .
1 , with @xmath3 , the quadrupole anisotropy @xmath26 is on the order of unity .
second , one can use the idea that the mean particle density varies with distance @xmath13 from a particle as @xmath27 .
then the small angle ( large @xmath21 ) limber approximation to the angular correlation function @xmath28 is @xcite 1+w(=/l)~^l e_l dl / l _ 0 ^ 1du [ u^2 + ( /l)^2]^-(3-d)/2 . [
eq : e_l ] to find @xmath20 differentiate with respect to @xmath21 . at @xmath3
this gives @xmath29 : the surface density fluctuations are independent of scale . at @xmath30 , @xmath31 .
the x - ray background fluctuates by about @xmath32 at @xmath33 , or @xmath34 .
this is equivalent to @xmath35 in the fractal model in eq .
( [ eq : e_l ] ) .
the universe is not exactly homogeneous , but it seems to be remarkably close to it on the scale of the hubble length
. it would be interesting to know whether there is a fractal construction that allows a significantly larger value of @xmath36 for given @xmath20 than in this calculation .
expansion that preserves homogeneity requires that the mean rate of change of separation of pairs of galaxies with separation @xmath4 varies as the hubble law , v = hr.[eq : hl ] the redshift - distance relation for type ia supernovae gives an elegant demonstration of this relation ( @xcite , @xcite ) .
arp ( @xcite , @xcite ) points out that such precision tests do not directly apply to the quasars , and he finds fascinating evidence in sky maps for associations of quasars with galaxies at distinctly lower redshifts .
but there is a counterargument , along lines pioneered by bergeron @xcite , as follows .
a quasar spectrum may contain absorption lines characteristic of a cloud of neutral atomic hydrogen at surface density @xmath37 atoms @xmath38 .
if this absorption system is at redshift @xmath39 a galaxy at the same redshift is close enough that there is a reasonable chance observing it , and with high probability an optical image does show a galaxy close to the quasar and at the redshift of the absorption lines ( @xcite , @xcite ) .
also , when a galaxy image appears in the sky close to a quasar at higher redshift then with high probability the quasar spectrum has absorption lines at the redshift of the galaxy .
we have good evidence the galaxy is at the distance indicated by its redshift .
we can be sure the quasar is behind the galaxy : the quasar light had to have passed through the galaxy to have produced the absorption lines .
if quasars were not at their cosmological distances we ought to have examples of a quasar appearing close to the line of sight to a lower redshift galaxy and without the characteristic absorption lines produced by the gas in and around the galaxy .
arp s approach to this issue is important , but i am influenced by what seems to be this direct and clear interpretation of the bergeron effect , that indicates redshift is a good measure of distance for quasars as well as galaxies . in the relativistic friedmann - lematre cosmological model the wavelength of a freely propagating photon
is stretched in proportion to the expansion factor from the epoch of emission to detection : 1+z=_obs_em = a_obsa_em .
[ eq : redshift ] the first expression defines the redshift @xmath40 in terms of the ratio of observed wavelength to wavelength at emission . the cosmological expansion parameter @xmath41 is proportional to the mean distance between conserved particles . the most direct evidence that the redshift is a result of expansion is the thermal spectrum of the cbr @xcite . in a tired light model in a static universe
the photons suffer a redshift that is proportional to the distance travelled , but in the absence of absorption or emission the photon number density remains constant . in this case
a significant redshift makes an initially thermal spectrum distinctly not thermal and inconsistent with the measured cbr spectrum .
one could avoid this by assuming the mean free path for absorption and emission of cbr photons is much shorter than the hubble length , so relaxation to thermal equilibrium is much faster than the rate of distortion of the spectrum by the redshift .
but this opaque universe is quite inconsistent with the observation of radio galaxies at redshifts @xmath42 at cbr wavelengths .
that is , the universe can not have an optical depth large enough to preserve a thermal cbr spectrum in a tired light model . in the standard world model
the expansion has two effects : it redshifts the photons , as @xmath43 , and it dilutes the photon number density , as @xmath44 .
the result is to cool the cbr while keeping its spectrum thermal .
thus the expanding universe allows a self - consistent picture : the cbr was thermalized in the past , at a time when when the universe was denser , hotter , and optically thick .
i have not encountered any serious objection to this argument ; the issue is the expansion factor . in the relativistic friedmann - lematre model the expansion of the universe traces back at least as far as redshift @xmath0 , when the light elements formed in observationally reasonable amounts @xcite . in the model of arp _ et al .
_ @xcite the expansion and cooling traces back to a redshift only moderately greater than the largest observed values , @xmath45 , when there would have been a burst of creation of matter and radiation followed by rapid clearing of the dust that thermalized the radiation . the arp _
_ picture for the origin of the light elements has not been widely debated .
if it were agreed that it is viable then a choice between this and the friedmann - lematre model would depend on other tests , such as the angular fluctuations in the cbr , as discussed next . -1.0truecm
the tests in table 1 are organized in four categories : spacetime geometry , galaxy peculiar velocities , structure formation , and early universe physics .
i offer grades for three sets of parameter choices .
as the tests improve we may learn that one narrowly constrained set of values of the cosmological parameters receives consistent passing grades , or else that we have to cast our theoretical net more broadly . in the relativistic friedmann - lematre cosmological model the mean spacetime geometry ( ignoring curvature fluctuations produced by local mass concentrations in galaxies and systems of galaxies )
may be represented by the line element ds^2 = dt^2 - a(t)^2 , [ eq : lineelement ] where the expansion rate satisfies the equation h^2 = ( aa ) ^2 = 83 g + 3 , which might be approximated as h^2= h_o^2[(1+z)^3 + ( 1+z)^2 + ] .
[ eq : cos_pars ] the last equation defines the fractional contributions to the square of the present hubble parameter @xmath46 by matter , space curvature , and the cosmological constant ( or a term in the stress - energy tensor that acts like one ) . the time - dependence assumes pressureless matter and constant @xmath1 .
other notations are in the literature ; a common practice in the particle physics community to add the matter and @xmath1 terms in a new density parameter , @xmath47 .
i prefer keeping them separate , because the observational signatures of @xmath19 and @xmath48 can be quite different . by 1930
people understood how one would test the space - time geometry in these equations , and as i mentioned there is at last direct evidence for the detection of one of the effects , the curvature of the relation between redshift and apparent magnitude ( @xcite , @xcite ) .
as indicated in line 1b , the measured curvature is inconsistent with the einstein - de sitter model in which @xmath49 and @xmath50 .
the measurements also disagree with a low density model with @xmath51 , though the size of the discrepancy approaches the size of the error flags , so i assign a weaker failing grade for this case .
the measurements are magnificent .
the issue yet to be thoroughly debated is whether the type ia supernovae observed at redshifts @xmath52 are drawn from essentially the same population as the nearer ones . in a previous volume in this series
krauss @xcite discusses the time - scale issue .
stellar evolution ages and radioactive decay ages do not rule out the einstein - de sitter model , within the still considerable uncertainties in the measurements , but the longer expansion time scales of the low @xmath19 models certainly relieve the problem of interpretation of the measurements .
thus i enter a tentative negative grade for the einstein - de sitter model in line 1a . in the analysis by falco _
_ @xcite of the rate of lensing of quasars by foreground galaxies ( line 1d ) for a combined sample of lensing events detected in the optical and radio , the @xmath53 bound on the density parameter in a cosmologically flat ( @xmath54 ) universe is @xmath55 .
the sneia redshift - magnitude relation seems best fit by @xmath56 , @xmath57 , a possibly significant discrepancy .
a serious uncertainty in the analysis of the lensing rate is the number density of early - type galaxies in the high surface density branch of the fundamental plane at luminosities @xmath58 , the luminosity of the milky way .
if further tests confirm an inconsistency of the lensing rate and the redshift - magnitude relation the lesson may be that @xmath48 is dynamical , rolling to zero , as ratra & quillen @xcite point out .
the relation between the mass density parameter @xmath19 and the gravitational motions of the galaxies is an issue rich enough for a separate category in table 1 .
it has been known for the past decade that if galaxies were fair tracer of mass then the small - scale relative velocities of the galaxies would imply that @xmath19 is well below unity @xcite .
if the mass distribution were smoother than that of the galaxies , the smaller mass fluctuations would require a larger mean mass density to gravitationally produce the observed galaxy velocities .
davis , efstathiou , frenk & white @xcite were the first to show that such a biased distribution of galaxies relative to mass readily follows in numerical n - body simulations of the growth of structure , and the demonstration has been repeated in considerable detail ( @xcite , @xcite , and references therein ) .
this is a serious argument for the biasing effect . but here are three arguments for the proposition that galaxies are fair tracers of mass for the purpose of estimating @xmath19 .
first , in many numerical simulations dwarf galaxies are less strongly clustered than giants .
this is reasonable , for if much of the mass were in the voids defined by the giant galaxies , as required if @xmath59 , then surely there would be remnants of the suppressed galaxy formation in the voids , irregular galaxies that bear the stigmata of a hostile early environment .
the first systematic redshift survey showed that the distributions of low and high luminosity galaxies are strikingly similar @xcite .
no survey since , in 21-cm , infrared , ultraviolet , or low surface brightness optical , has revealed a void population . there is a straightforward interpretation : the voids are nearly empty because they contain little mass .
second , one can use the galaxy two - point correlation function in eq .
( [ eq : xigg ] ) and the mass autocorrelation function @xmath60 from a numerical simulation of structure formation to define the bias function b(r , t ) = ^1/2 .
[ eq : bias ] in numerical simulations @xmath61 typically varies quite significantly with separation and redshift @xcite .
that is , the galaxies give a biased representation of the statistical character of the mass distribution in a typical numerical simulation .
the issue is whether the galaxies , or the models , or both , are biased representations of the statistical character of the real mass distribution .
what particularly strikes me is the observation that the low order galaxy correlation functions have some simple properties .
the galaxy two - point function is close to a power law over some three orders of magnitude in separation ( eq . [ eq : xiggparameters ] ) .
the value of the power law index @xmath62 changes little back to redshift @xmath63 . within the clustering length
@xmath64 the higher order correlation functions are consistent with a power law fractal .
a reasonable presumption is that the regularity exhibited by the galaxies reflects a like regularity in the mass , because galaxies trace mass .
i am impressed by the power of the numerical simulations , and believe they reflect important aspects of reality , but do not think we should be surprised if they do not fully represent other aspects , such as relatively fine details of the mass distribution .
-1.0truecm the third argument deals with the idea that blast waves or radiation from the formation of a galaxy may have affected the formation of nearby galaxies , producing scale - dependent bias . in this case
the apparent value of the density parameter derived from gravitational motions within systems of galaxies on the assumption galaxies trace mass would be expected to vary with increasing scale , approaching the true value when derived from relative motions on scales larger than the range of influence of a forming galaxy .
2 shows a test .
the abscissa at the entry for clusters of galaxies is the comoving radius of a sphere that contains the mass within the abell radius .
the estimates at larger scales are plotted at approximate values of the radius of the sample . if it were not for the last two points at the right - hand side of fig . 2
, one might conclude that the apparent density parameter is increasing to the true value @xmath65 at @xmath66 mpc . but considering the last two points , and the sizes of the error flags , it is difficult to see any evidence for scale - dependent bias .
i assign a strongly negative grade for the einstein - de sitter model in line 2a in table 1 , based on galaxy motions on relatively small scales , because biasing certainly is required if @xmath49 and i have argued there is no evidence for it . the more tentative grade in line 2b
is based on fig . 2 :
the apparent value of the density parameter does not seem to scale with depth .
the friedmann - lematre model is unstable to the gravitational growth of departures from a homogeneous mass distribution .
the present large - scale homogeneity could have grown out of primeval chaos , but the initial conditions would be absurdly special .
that is , the friedmann - lematre model requires that the present structure the clustering of mass in galaxies and systems of galaxies grew out of small primeval departures from homogeneity .
the consistency test for an acceptable set of cosmological parameters is that one has to be able to assign a physically sensible initial condition that evolves into the present structure of the universe .
the constraint from this consideration in line 3c is discussed by white _
@xcite , and in line 3b by bahcall _
( @xcite , @xcite ) . here
i explain the cautious ratings in line 3a .
as has been widely discussed , it may be possible to read the values of @xmath19 and other cosmological parameters from the spectrum of angular fluctuations of the cbr ( @xcite and references therein ) .
this assumes nature has kept the evolution of the early universe simple , however , and we have hit on the right picture for its evolution .
we may know in the next few years .
if the precision measurements of the cbr anisotropy from the map and planck satellites match in all detail the prediction of one of the structure formation models now under discussion it will compel acceptance .
but meanwhile we should bear in mind the possibility that nature was not kind enough to have presented us with a simple problem .
-1.0truecm an example of the possible ambiguity in the interpretation of the present anisotropy measurements is shown in fig .
the two models assume the same dynamical actors cold dark matter ( cdm ) , baryons , three families of massless neutrinos , and the cbr but different initial conditions . in the adiabatic model
the primeval entropy per conserved particle number is homogeneous , the space distribution of the primeval mass density fluctuations is a stationary random process with the scale - invariant spectrum @xmath67 , and the cosmological parameters are @xmath68 , @xmath69 , and @xmath70 ( following @xcite ) .
the isocurvature initial condition in the other model is that the primeval mass distribution is homogeneous
there are no curvature fluctuations and structure formation is seeded by an inhomogeneous composition . in the model shown here
the primeval entropy per baryon is homogeneous , to agree with the standard model for light element production , and the primeval distribution of the cdm has fluctuation spectrum p(k)k^m , m = -1.8 .
[ eq : piso ] the cosmological parameters are @xmath71 , @xmath72 , and @xmath73 .
the lower density parameter produces a more reasonable - looking cluster mass function for the isocurvature initial condition @xcite . in both models
the density parameter in baryons is @xmath74 , the rest of @xmath19 is in cdm , and space sections are flat ( @xmath75 ) .
both models are normalized to the large - scale galaxy distribution .
the adiabatic initial condition follows naturally from inflation , as a remnant of the squeezed field that drove the rapid expansion .
a model for the isocurvature condition assumes the cdm is ( or is the remnant of ) a massive scalar field that was in the ground level during inflation and became squeezed to a classical realization . in the simplest models for inflation this produces @xmath76 in eq .
( [ eq : piso ] ) .
the tilt to @xmath77 requires only modest theoretical ingenuity @xcite .
that is , both models have pedigrees from commonly discussed early universe physics . the lesson from fig .
3 is that at least two families of models , with different relations between @xmath19 and the value of @xmath21 at the peak , come close to the measurements of the cbr fluctuation spectrum , within the still substantial uncertainties .
an estimate of @xmath19 from the cbr anisotropy measurements thus may depend on the choice of the model for structure formation .
programs of measurement of @xmath78 in progress should be capable of distinguishing between the adiabatic and isocurvature models , even given the freedom to adjust the shape of @xmath79 .
the interesting possibility is that some other model for structure formation with a very different value of @xmath19 may give an even better fit to the improved measurements .
i assign a failing grade to the einstein - de sitter model in line 3a because the adiabatic and isocurvature models both prefer low @xmath19 ( @xcite , @xcite ) .
i add question marks to indicate this still is a model - dependent result . in their version of table 1 dekel , burstein , & white
@xcite give the einstein - de sitter model the highest grade on theoretical grounds , and a cosmologically flat model with @xmath1 the next highest grade .
the point is well taken : this is the order most of us would choose .
the issue is whether nature agrees with our ideas of elegance , or maybe prefers physics that produces an open universe ( @xcite - @xcite ) .
full closure of cosmology may come with the discovery of physics that predicts the values of @xmath48 and space curvature ( eq .
[ [ eq : cos_pars ] ] ) in terms of the expansion age of the universe , consistent with all the other constraints in table 1 .
but since we seem to be far from that goal i am inclined to omit entries in line 4 .
we have a secure if still schematic standard model for cosmology , and the prospect for considerable enlargement from the application of the cosmological tests .
the theoretical basis for the tests was discovered seven decades ago . a significant application likely will take a lot less than seven more decades : the constraints in table 1 already are serious , if debatable , and people know how to do better
. application of the tests could yield a set of tightly constrained values of the cosmological parameters and a clear characterization of the primeval departure from homogeneity .
if so cosmology could divide at a fixed point , the situation at @xmath80 , say , when the universe is well described by a slightly perturbed friedmann - lematre model .
one branch of research would analyze evolution from these initial conditions to the present complex structure of the universe .
the other would search for the physics of the very early universe that produced these initial conditions .
but before making any long - term plans based on this scenario i would wait to see whether the evidence really is that the early universe is simple enough to allow such a division of labor . | we have a well - established standard model for cosmology and prospects for considerable additions from work in progress .
i offer a list of elements of the standard model , comments on controversies in the interpretation of the evidence in support of this model , and assessments of the directions extensions of the standard model seem to be taking .
psfig * the standard cosmological model * + + + |
block copolymers ( bcps ) , comprising chemically distinct polymers permanently linked together , are interesting because of the diverse array of ordered phases to which both polymer theory and experiment have been directed.@xcite the phase behavior of diblock copolymer melts is a competition between the entropic tendency to mix the two species into an isotropic melt , and an energetic penalty for having unlike species adjacent , which induces transitions into ordered phases of many symmetries , depending on the topology and composition of the polymers . near the order - disorder transition ( weak incompatibility ) entropy effects dominate , and the individual polymers retain ( within mean field ) their gaussian coil conformation through the transition,@xcite , while at much higher incompatibilities the chains are strongly stretched .
it is this strongly stretched regime which we address here .
leibler developed the first complete theory of ordered phases in bcp melts@xcite , and predicted the by - now classical phases of lamellar ( l ) , cylindrical ( c ) and spherical ( s ) symmetry using the random phase approximation to derive an effective landau free energy in terms of composition modulations in fourier space .
the strong segregation regime was studied by helfand and co - workers @xcite and semenov @xcite , who predicted the same series of phases with increasing asymmetry , denoted by the fraction @xmath1 of polymer a in an @xmath2 diblock .
( in this work we always use a to denote the minority block ) .
this treatment balances the stretching energy of a polymer chain with the interfacial energy between a and b regions . by assuming an incompressible melt , minimization of the free energy gives a preferred domain size which scales as @xmath3 , where @xmath4 is the degree of polymerization .
in the strong segregation limit the free energies of all microphases scale the same way with chain length and interfacial tension , so the phase boundaries become independent of the strength of the repulsion @xmath5 between a and b monomers and depend only on the composition @xmath6 .
semenov s calculation in effect gave a lower bound to the free energy of the l , c , and s phases because the phases he constructed did not fill space , but were micelles of the corresponding topology @xcite .
this approximation treats the @xmath2 interface and outer block surface as having the same circular or spherical shape , and is sufficient for understanding the qualitative aspects of the transitions between the phases .
experiments followed the theories of leibler and semenov and quickly discovered a new phase,@xcite , originally thought to be ordered bicontinuous double diamond ( here denoted d ) , of @xmath7 symmetry , but recently shown to be of @xmath8 symmetry @xcite and related to the minimal surface known as the gyroid ( g).@xcite the g phase occurs for compositions between those of the l and c phases , can occur directly from the disordered phase upon increasing the incompatibility @xmath9 , and is found to be unstable to the l or c phases at high enough @xmath9.@xcite although several groups attempted to describe this transition theoretically,@xcite using variations on leibler s theory , the first successful theory is due to matsen and schick @xcite , who developed a method for computing the free energy of any crystalline structure by expanding the partition function in the basis functions for the symmetry of the desired mesophase , rather than the fourier mode expansion of leibler .
they found a stable gyroid phase for @xmath10 , where the upper limit was determined by extrapolation from the phase boundaries at lower @xmath9.@xcite this was followed by careful application of leibler s method,@xcite to include higher harmonics and calculate the stability of the g phase in weak segregation analytically .
roughly concurrent to the calculations of matsen and schick , methods were developed to calculate the free energy of essentially arbitrary structures in the strong segregation regime ( @xmath11 ) .
these methods use the results for polymer brushes,@xcite , supplemented by an ansatz about the geometry of the relevant phase and an assumption about the chain paths . olmsted and milner assumed straight paths through the @xmath2 interface and locally specified the volume fraction per molecule,@xcite , while likhtman and semenov relaxed the assumption of straight paths @xcite but enforced the constraint of constant @xmath1 per molecule only globally .
the former approach corresponds to an upper bound on the free energy ( see below ) , while it is not clear that the likhtman - semenov calculations corresponds to any bound , or indeed to any systematic approximation , because the local constraint of constant composition is relaxed . by comparing upper bounds between bicontinuous , c , and
l phases ( obtained for the cylindrical phase by assuming hexagonal symmetry and imposing straight paths ) , we showed that the bicontinuous phases are unstable , when comparing upper bounds , to the l and c phases .
later , xi and milner extended this work to calculations with kinked polymer paths , and found an upper bound to the hexagonal phase which lies very close to the lower bound using round unit cells.@xcite experiments have found an additional phase at @xmath5 values between the g and l phases @xcite , a hexagonally - perforated lamellae ( hpl ) phase , which consists of majority lamellae connected through a minority matrix by hexagonal arrays of tubes.@xcite the stacking has been suggested to be @xmath12 @xcite or @xmath13 @xcite .
theoretical attempts to justify this phase have failed in both the strong segregation limit , where fredrickson chose a catenoid as a candidate base surface;@xcite and in the weak - segregation limit by self - consistent field calculations @xcite .
recent experiments @xcite have shown that the hpl phase is not an equilibrium phase in diblock melts , but may be metastable . here
we present the calculations of ref .
@xcite in more detail .
we show that the g geometry is the most stable of the candidate bicontinuous phases , followed by the d and p geometries , and that the g phase can be stable for block - copolymers with sufficient conformational asymmetry .
the outline of this paper is as follows . in section [ sec : general ]
we present the formalism for calculating the free energy in general geometries . in section
[ sec : classical ] we present the results for the classical diblock topologies ( lamellae , cylinders , and spheres ) , extended to include non - round unit cells and , in the case of the cylindrical topology , kinked paths . in section [ sec : bicontinuous ] we present the free energy for a generic `` saddle '' wedge , which is representative of a generic bicontinuous structure as a pie - shaped wedge is representative of cylindrical phases regardless of packing .
we then introduce the geometry necessary for calculating the free energy of the p , d , and g topologies . in section [ sec
: results ] we present our results for both symmetric and non - symmetric stars , and we conclude in section [ sec : summary ] .
we first recall some results for polymer brushes in strong segregation under melt conditions , and then show how to apply this to a general geometry .
we consider a melt of star @xmath14 copolymers , comprising @xmath15 arms of _ a_-blocks of mean square end - to - end distance @xmath16 and similarly for the _ b_-arms .
the volume fraction @xmath1 of _ a _ material is @xcite @xmath17 where @xmath18 is the total chain volume , and @xmath19 and @xmath20 are the volumes of single _ a _ and _ b _ arms .
our calculations are appropriate for strongly - segregated chains , for which interfaces are sharp on the scale of microphase lattice constants . in strong segregation
the free energies of all microphases scale the same way with chain length and interfacial tension , so the phase boundaries become independent of the strength of the repulsion @xmath5 between _ a _ and _ b _ monomers .
= 2.5truein consider an elementary wedge , as in figure [ fig : wedge ] , from which we will construct all of the strong - segregation phases .
our calculations are performed in terms of the ratio @xmath21 of the cross - sectional area @xmath22 at a height @xmath23 relative to that of the outer surface , in an infinitesimal wedge of height @xmath24 .
this function may be easily calculated for wedges of particular shapes by elementary geometry , and is collected in table [ table1 ] for various geometries . since @xmath25 is the projected surface area along the normal vector extending from the wedge point to the flat wedge top , it will be a quadratic function of @xmath26 .
the boundary condition @xmath27 implies that @xmath25 is a sum of @xmath26 and @xmath28 terms , and the boundary condition @xmath29 fixes the sum of the corresponding coefficients to be @xmath30 , leaving a single parameter .
hence we may generally write @xmath31 the location @xmath32 of the `` dividing surface '' separating the two species is determined by equating the relative volume below @xmath23 , denoted @xmath33 , to the volume fraction @xmath1 : @xmath34 where @xmath35 is the ( partial ) volume of the wedge below height @xmath23 ( see figure [ fig : wedge ] ) . .area function @xmath25 ( see eq [ eq : area ] ) for various wedge geometries , where @xmath36 and @xmath24 is the wedge height .
the expression for @xmath37 for the d , p , and g wedges refers to the geometry in figure [ fig : pgwedges ] . [ cols="<,^",options="header " , ] [ table2 ] upon minimizing eq [ eq : ftot ] over the scale of the structure , _
i.e. _ the radius @xmath24 , it is convenient to normalize all energies by a characteristic energy @xmath38 to obtain a compact form for the free energy : @xmath39 where @xcite @xmath40^{1/6 } \left(\ell_{\scriptscriptstyle a } \ell_{\scriptscriptstyle b}\right)^{1/6 } \label{eq : f0}\ ] ] is the free energy of a symmetric @xmath41 lamellar phase .
the procedure above applies to a single wedge . for the classical cylinder and sphere geometries with circular ( spherical ) unit cells ,
each wedge is identical . for non - circular unit cells and for the complex geometries of bicontinuous phases
, we must assemble the structure from many different wedges and minimize over the scale factor for the entire structure .
the average free energy per chain @xmath42 of a structure with many distinct wedges is @xmath43 where @xmath44 is the free energy per molecule in wedge @xmath45 of volume @xmath46 , and @xmath47 is the volume of the structure .
we choose a single scale factor @xmath48 to determine the size of the whole structure .
each wedge @xmath45 has its own area and volume functions @xmath49 and @xmath50 , where the position of the dividing surface of each wedge , @xmath51 , is determined by @xmath52 the dimensionless functions @xmath53 are generalizations of eq [ eq : area ] for each wedge @xmath45 with wedge height @xmath54 .
these functions encode the geometry of the particular structure , and the cross - sectional area at the top of each wedge , @xmath55 , scales as @xmath56 , for a @xmath57-dimensional structure ( _ e.g. _ @xmath58 for lamellae , @xmath59 for cylinders , @xmath60 for spheres ) . expressing eq [ eq : interf],[eq : fstr ] in terms of @xmath61 and @xmath48 , we minimize over @xmath48 to find the following free energy per chain of a particular structure : @xmath62 where @xmath63 where @xmath64 is obtained from eqs [ eq : ia]-[eq : ib ] by substituting @xmath49 in place of @xmath65 , and a similar relation defines @xmath66 . by specifying the volume fraction in each wedge according to eq [ eq : zdloc2 ]
, we locally satisfy the constraint arising from the fixed composition of the copolymers .
in contrast , likhtman and semenov @xcite satisfied this constraint only globally within a particular structure , which would be relevant for mixtures of different diblock copolymers with overall composition @xmath1 @xcite in the strong segregation limit , in which the entropy of mixing of such different copolymers would be negligible
using the results of section [ sec:2a ] we can find the energies of the classical phases of diblock - copolymers : lamellae ( l ) , cylinders ( c ) , and spheres ( s ) in the round unit cell approximation , in which the unit cells are taken to consist of identical wedges .
the corresponding free energies are : @xmath67^{1/3 } \label{eq : felam}\\ { f_{cyl } \over f_0 } & = & \left [ { 2 \varepsilon\phi(1-\phi^{1/2})^3(3+\phi^{1/2 } ) \over ( 1-\phi)^2 } + { 2 \phi \over \varepsilon}\right]^{1/3 } \label{eq : fecyl } \\
{ f_{sph } \over f_0 } & = & 3\left [ { \varepsilon\phi^{4/3}(1-\phi^{1/3})^3 ( \phi^{2/3}+3\phi^{1/3}+6 ) \over 10(1-\phi)^2 } + { \phi
\over 10\varepsilon } \right]^{1/3}\!\!\!\!\!\!\!. \label{eq : fesph}\end{aligned}\ ] ] calculations based on round unit cells @xcite provide lower bounds for the free energy , because they in fact describe the free energy per molecule of micelles.@xcite we may imagine a volume packed with such micelles , the interstitial regions filled with compatible long homopolymer with negligible surface tension against the outsides of the micelles , and negligible entropy of mixing .
then we could do work to deform the micelles into a space - filling array , expelling the homopolymer at no free energy cost . to distinguish between crystal structures within a particular topology , such as between hexagonal and square for the cylindrical topologies
, we must examine the energy for packing the molecules into the particular geometry , which is performed below .
= 3.0truein we can produce an _ upper _ bound for different structures by assembling small pieces of the cylindrical or spherical micelles to fill the appropriate unit cell .
each wedge has a parabolic monomer chemical potential given by eq [ eq : mu ] .
however , each wedge @xmath45 has a slightly different shape and geometry , and thus has a distinct potential @xmath68 .
adjacent wedges are not in equilibrium with each other and will relax if allowed to do so .
hence the calculation yields an upper bound . to construct the unit cell of , _
e.g. _ , hexagonal cylinders , we assume a hexagonal dividing surface scaled down by @xmath69 and assemble the unit cell from tiny pie - shaped wedges extending from the center of the hexagon to the cell boundary .
we make an analogous construction for square arrays of cylinders , or for fcc and bcc packings of spheres .
we calculate the volume - averaged stretching free energy per molecule using eq [ eq : avg ] .
to calculate this in practice we use the following procedure . for cylindrical micelles
we divide a cell of a given symmetry ( say , hexagonal ) into tiny wedges .
each wedge is adjusted slightly by making the segment of the wedge on the dividing surface normal to the bisector of the wedge , which is the path of the polymer .
such an adjustment introduces a negligible volume in the continuum limit of many small wedges .
the surface area used for calculating the surface energy ( eq [ eq : sa ] ) is , of course , the area of the segment in the original hexagonal dividing surface ( before adjusting the wedge to account for straight paths ) .
= 3.5truein for the classical phases ( see figure [ fig : classicala ] ) the ratios of the upper and lower bounds for the free energies are independent of @xmath1 , given by : @xmath70 evidently , the most favorable structures have the `` roundest '' unit cells . the hexagonal phase is favored over the square phase , and bcc is slightly favored over fcc .
= 2.0truein other upper bounds can be obtained by using different prescriptions for the @xmath2 surface .
for example , one could choose a circular @xmath2 surface of radius @xmath71 for the cylindrical phase .
the inner ( @xmath72 ) volume may be divided into wedges , and the outer ( @xmath73 ) volume divided into wedges which each satisfy the volume constraint @xmath1 with a partner @xmath72 wedge @xcite . for a right triangular wedge which subtends an angle @xmath74 , points at angle @xmath75 on the @xmath2 surface map to points @xmath76 on the boundary of the wigner - seitz cell , @xmath77 where @xmath78 , and @xmath79 .
the composition specifies the radius , according to @xmath80 the mapping which obeys the local composition constraint is @xmath81 where @xmath82 and @xmath83 for hexagons and squares ( see figure [ fig : wedgecut ] ) , respectively . minimizing over the scale of the structure , eq [ eq : fsum ] yields , after some calculation , the following free energy : @xmath84\right\}^{1/3},\end{aligned}\ ] ] where @xmath85\end{aligned}\ ] ] and similarly for @xmath86 .
remarkably , the upper bound for the kinked - path - hexagonal ansatz is typically less than @xmath87 above the lower bound of cylindrical micelles , and the transition is shifted to only a slightly smaller @xmath72 fraction @xmath1 ( see figure [ fig : classicalb ] ) .
apparently the extra stretching energy to maintain a hexagonal @xmath2 interface with straight paths is relaxed considerably by allowing the inner block to adopt a more nearly circular dividing surface , which is preferred .
recent accurate numerical self - consistent field calculations of diblock melts have shown that in fact the @xmath2 interface is nearly circular , with a slight hexagonal modulation ( angular modulation with 6-fold symmetry ) of relative amplitude @xmath88 at @xmath89 and @xmath90 @xcite .
before addressing particular symmetries ( p , d , or g ) of bicontinuous phases , we discuss the closest analogue to a round unit cell .
we would like to produce a simple estimate of the free energy , analogous to the cylindrical and spherical micelle calculation , which captures the physics of bicontinuous topologies .
we thus represent a generic bicontinuous phase as a wedge , shown in figure [ fig : chip ] : an infinitesimal patch of `` saddle '' surface , with edges given by the normals , terminating in a small line segment lying along the bond - lattice .
we envision the surface as a minimal surface , which has zero mean curvature.@xcite the stretching free energy per molecule of the symmetric wedge can be calculated as before , given the relative area as a function of relative height along the center normal ( table [ table1 ] ) : @xmath91 , where @xmath92 , @xmath93 is the thin end of the wedge and @xmath94 is the patch of minimal surface . as before ,
the dividing surface location @xmath32 is determined by eq [ eq : zdloc ] .
the resulting free energy is given by applying eqs [ eq : zdloc]-[eq : interf ] , and is shown with the various cylindrical bounds in figure [ fig : classicalb ] .
this estimate misses by a few tenths of one percent the intersection of the lamellar phase and the kinked - path upper bound bound for the hexagonal phase , and is stable with respect to the straight - path upper bound .
= 3.5truein clearly the simple wedge construction captures some important physics .
the structures formed by copolymers at different volume fractions @xmath1 arise from competition between interfacial and stretching free energies .
the different structures present different functions @xmath25 , which determine both the dividing surface area and the stretching energy as a function of volume fraction .
the phases occur in the order they do because the progression of functions @xmath25 from quadratic @xmath28 ( spheres ) to linear @xmath26 ( cylinders ) to @xmath95 ( bicontinuous ) to constant ( lamellae ) gives progressively less volume to the `` outer '' chain to avoid stretching , but uses progressively less area to separate the two species at higher volume fractions of the minority species .
= 3.5truein however , we can not argue as before that our simple estimate for the bicontinuous phase is a lower bound for actual bicontinuous phases , because there is no way to pack together copies of one infinitesimal wedge to produce a `` micelle '' that 1 ) fills some region of space , and 2 ) is bounded by some surface(s ) , the volume outside of which could be filled by homopolymer .
neither can we argue that this estimate is an upper bound , because we certainly can not pack a unit cell of the region bounded by the d or g surfaces with identical copies of one infinitesimal wedge .
bicontinuous phases are assembled from different wedges @xmath45 , with different shape factors @xmath96 in table [ table1 ] .
the shape factor roughly gauges the splay or gaussian curvature of the surface at the top of the wedge , with @xmath97 ( cylinders ) corresponding to zero gaussian curvature , @xmath98 to positive gaussian curvature , and @xmath99 to negative gaussian curvature .
[ the gaussian curvature is the product of the two radii of curvature of a surface ] .
the distribution of wedges must be chosen to pack the desired structure . while the symmetric wedge has @xmath100 ,
this is not an optimum shape .
in fact , the optimum wedge shape depends on composition , as can be seen in figure [ fig : pnwedges ] .
it is evident that there are shape factors @xmath37 which have lower free energy than the straight- and kinked - path hexagonal upper bounds ( figure [ fig : wedgeshape ] ) , so it is not unreasonable to hope that a judicious packing configuration can be a stable thermodynamic phase .
the effect of the shape factor @xmath37 on the topology of the phase diagram emerges upon examining conformationally asymmetric ( @xmath101 ) copolymers .
following ref .
@xcite , we explore the effect of conformational asymmetry on the stability of bicontinuous phases by multiplying the wedge free energy by an additional arbitrary small prefactor @xmath102 which enhances stability .
[ we will see below that for @xmath103 the bicontinuous phases that we can calculate ( g , d , p ) are of order @xmath104 higher than the cylinder - lamellar crossing , depending on which upper bound one compares . ] figure [ fig : asymwedge ] shows ` phase diagrams ' as a function of conformational asymmetry @xmath105 .
recall that for @xmath106 the @xmath73-block is more flexible , while the @xmath72-block is stiffer and better able to stretch .
for the symmetric wedge @xmath107 conformational asymmetry reduces the stability of the stiff - minority wedge phase @xmath108 and enhances the stability of the flexible - minority wedge phase ( @xmath109 ) , and shifts all transitions to greater @xmath1 . for @xmath110
the wedge phases lose stability , as could be guessed from figure [ fig : wedgeshape ] . for @xmath111 ,
for which the wedge is more cylindrical - like ( @xmath97 coresponds to cylinders ) , conformational asymmetry enhances the stiff - minority wedge phase relative to both the lamellar and cylindrical phases , and decreases the stability of the flexible - minority wedge phase .
we emphasize that these are not phase diagrams , for a true phase is a mixture of wedges with different shape factors which fill space .
= 3.5truein = 7.0truein 2 experimentally , the g phase has been observed between the lamellar and cylindrical phases in several strongly - segregated copolymer systems , for @xmath1 around 0.3 ( and , symmetrically , around 0.7 ) .
some groups have argued for phase stability of bicontinuous phases in terms of bending rigidities,@xcite .
however , there is fundamentally no bending energy in the problem ; descriptions in terms of bending energies only arise from a proper accounting of stretching free energy in curved geometries .
our approach is to choose a geometry as an ansatz and compute the corresponding interfacial and stretching free energies in a manner consistent with calculations for cylindrical , spherical , and lamellar phases .
the structure , revealed by scattering and electron microscopy,@xcite studies , can be described as follows,@xcite .
consider first the d geometry , which is easier to visualize .
a skeleton formed of the bonds of a diamond lattice is shown in figure [ fig : skeletons ] .
two such lattices interpenetrate , analogous to the interpenetration of two simple cubic lattices in a bcc structure . now imagine swelling the bonds in these lattices into tubes of a finite diameter .
the walls of these tubes are a rough approximation of the experimentally observed `` dividing surface '' separating the regions containing the two blocks .
the volume contained within the tubes corresponds to the region inhabited by the low volume - fraction monomer . to model the d geometry
, we use a self - dual minimal surface , called the schwartz d ( diamond ) surface , which partitions space into two identical interpenetrating regions , each of which contains and is topologically equivalent to a diamond bond - lattice @xcite . within each of these regions
is a dividing surface , which surrounds a copy of the bond - lattice .
the copolymer chains then have conformations with one ( a ) species stretching towards the bond - lattice , the junction between blocks residing on the dividing surface , and the other ( b ) species stretching towards the minimal surface . in the g
phase the diamond lattice is replaced by a three - fold coordinated lattice , and the surface is replaced by the gyroid minimal surface discovered by schoen in 1970 @xcite , in which the two interpenetrating g volumes are chiral enantiomers of one another . for the p phase the bond lattice is six - fold coordinated ( figure [ fig : skeletons ] ) and the candidate partitioning surface is the schwartz p minimal surface .
there is no compelling reason to choose a minimal surface for the partitioning surface .
however , minimal surfaces solve the variational problem of minimizing surface area with zero pressure across the interface .
if the diblock phase is in fact partitioned into two equivalent connected regions , then by symmetry there can be no net pressure exerted across the dividing surface that separates the two equivalent disjoint connected regions .
so a minimal surface is reasonable , but by no means certain , since there is no obvious area energy to minimize .
= 2.5truein
= 2.5truein = 2.5truein the p , d , and g surfaces may be conveniently calculated using the weierstrass representation.@xcite here , the three - dimensional points @xmath112 of the two - dimensional surface are parametrized by the complex number @xmath113 .
the g , p , and d surfaces are triply - periodic minimal surfaces with space groups @xmath114 and @xmath7 , respectively.@xcite to generate the full surface it is enough to calculate a single patch , to which all the symmetry operations of the space group may be applied to generate the full structure .
a generic surface has two radii of curvature which are generally non - zero and different .
points where the surface is flat are singular points , since both radii of curvature are zero and a direction of the surface can not be determined .
these flat points define the corners of the fundamental patch .
the weierstrass representation is : @xmath115 where @xmath116 and @xmath117 is the domain of integration shown in figure [ fig : weier ] .
the points on the corners correspond to the flat points , and it is evident that the integrand above ( excluding the measure ) is singular at these points .
the angle @xmath75 determines the surface : @xmath118 = 2.1truein = 7.0truein 1.0truecm 2 for other angles the surface intersects itself .
this does not , of course , exhaust the class of triply periodic minimal surfaces either mathematically,@xcite or physically @xcite .
we have chosen the d , p , and g surfaces because they are the most common observed surfactant bicontinuous surfaces , and have been claimed experimentally in block copolymers .
sections of these surfaces are shown in figure [ fig : surfaces ] for fairly accurate calculations ( yielding energies _ lower _ than those for the true surface by of order a few tenths of a percent ) the d surface may be approximated by a simple hyperbolic surface , @xmath119 .
this suggests that the minimal d surface may not the optimal partitioning surface .
however , we have varied the shape of the partitioning surface around the d surface , and found free energy variations of only a few tenths of a percent . because of this
, we have not optimized the free energy with respect to adjustments in the partitioning surfaces . to produce an upper bound on the free energy of the bicontinuous phases we follow a procedure analogous to that used for the classical cylindrical and spherical topologies .
namely , we divide a unit cell of these structures into a large number of wedges ( shown in figure [ fig : pgwedges ] ) , similar to figure 1 , but of varying radii and gaussian curvature , and average the free energy per molecule using eq [ eq : avg]-[eq : sa ] . each structure has a different unit cell ( a big wedge ) from which the entire structure may be generated by applying the symmetry operations of the particular space group .
figure [ fig : gchip ] shows the fundamental cell for the g structure .
this procedure is straightforward .
first we calculate the minimal surface , and then adopt a convenient mapping from the points on this surface to the skeleton .
essentially , we construct an interpolation between those high - symmetry points on the minimal surface with normals that project onto the underlying bond - lattice . for the @xmath120 and @xmath121 phases ,
we optimize the mapping from the partitioning surface to the line segments of the bond lattice to minimize the free energy .
we perform this by a conjugate gradient algorithm that distorts the two dimensional mesh of points on the surface , and gains of order @xmath87 in energy .
= 3.5truein = 3.5truein thus , each small patch on the minimal surface is connected by straight lines to a small line segment on the skeleton , and a set of wedges results .
each wedge is adjusted slightly by making both the top patch and the bottom segment orthogonal to the line segment connecting the center of the patch to the skeleton .
this is analogous to making the outer surface of the wedge in the upper bound for the hexagon phase orthogonal to the line segment connecting the center of the outer surface of the wedge to the center of the hexagon .
such adjustments are negligible in the limit of infinitesimal wedges . as before ,
the area of the @xmath2 interface is the true area , rather than the ( smaller ) area that results from adjusting the wedge to assure a chain path normal to the @xmath2 interface . in this way
, the unit cell of the region bounded by a minimal surface is decomposed into many small wedges , each with a known ( and different ) radius @xmath24 , shape factor @xmath37 , and volume .
the location of the dividing surface within each wedge is fixed by eq [ eq : zdloc2 ] , and the free energy calculated with eqs [ eq : avg]-[eq : sa ] .
we have checked the algorithm and the dependence on the fineness of the mesh by using it to successfully compute the free energies of the hexagonal phase .
= 3.5truein figure [ fig : data2 ] shows the free energy curves as a function of @xmath1 for conformationally symmetric copolymers ( @xmath103 ) .
the upper bound for the @xmath122 phase ( p ) lies @xmath123 above that for the @xmath7 phase ( d ) , which in turn is less than a percent ( @xmath124 at @xmath125 ) above the @xmath8 ( g ) phase .
consistent with experiments and self - consistent field theory , we do not find a stable g phase . at the lamellar kinked - path - hexagons free energy crossing ( @xmath126 )
the free energy of the @xmath121 phase is of order @xmath127 larger , while at the lamellar straight - path - hexagons free energy crossing ( @xmath128 ) the free energy of the @xmath121 phase is only a few tenths of a percent ( @xmath129 ) greater .
this , we have argued , may be the fairer comparison , since both calculations use straight paths .
unfortunately , we do not know how to perform a kinked - paths estimate for the bicontinuous phases .
note , however , that we do not expect to gain as much energy from a kinked - path calculation for the bicontinuous phases as for the cylindrical topologies .
consider the hexagonal calculation .
the boundaries of the wigner - seitz cell , and hence the @xmath2 dividing surface , have sharp corners into which the chains must stretch .
presumably a large part of the gain in the kinked - path calculation comes from relieving the strain associated with this stretch , and relaxing the inner block to its preferred circular structure .
bicontinuous phases , on the other hand , have smooth `` wigner - seitz boundaries '' ( _ i.e. _ the minimal surface ) , and expensive stretching occurs mainly at the junctions of the skeleton lattice .
hence , the anomalous stretching that may be relieved by a kinked - path calculation occurs along points in the structure , rather than along lines .
so we expect that our straight - path estimate is not likely to differ greatly from a kinked - path estimate , and that the free energy of the g phase remains well above that of the kinked - path hexagons .
the stretching of the chains at the junctions presumably contributes to the relative stability of the @xmath130 , @xmath120 , and @xmath121 structures , which have 6- , 4- , and 3-fold coordinated bond lattices .
the more highly - coordinated lattices require more chain stretching to accommodate the space , which suggests that @xmath130 , @xmath120 , and @xmath121 occur in increasing order of stability .
this picture is corroborated by recent work of matsen and bates @xcite , who quantitatively examined the packing frustration in the g , d , and hpl phases .
= 7.0truein 2 our results apply to the strong - segregation limit , which is attained in the limit of large @xmath5 .
because the phase boundaries shift away from @xmath131 as @xmath5 increases from weak segregation @xcite , we expect our phase boundaries to be further from @xmath131 than experimental values , which is indeed the case .
figure [ fig : asym ] shows free energy crossings for various values of the conformational asymmetry parameter @xmath105 .
relative to a conformationally symmetric melt , conformational asymmetry stabilizes phases with a stiff minority species and destabilizes phases with a stiff majority species , moving boundaries to larger @xmath132 for @xmath106 .
recall that for @xmath133 the inner @xmath72-block is more flexible , while the @xmath73-block is stiffer and better able to stretch .
we find that @xmath133 reduces the relative stability of the g phase , while the stability is enhanced for @xmath106 , and becomes stable for rather large asymmetries @xmath134 ( figure [ fig : asymg ] ) .
= 3.5truein previous calculations of phase diagrams of conformationally - asymmetric diblocks have been done in the weak segregation regime,@xcite and , for the generic symmetric wedge as a model bicontinuous structure , in the strong segregation regime.@xcite matsen and bates @xcite found , as we do , that conformational asymmetry stabilizes the g phase with a stiff minority phase , widening the composition window and moving the lamellar - g and g - cylinder boundaries to higher stiff compositions ; and destabilizes the g phase with a stiff majority phase , both narrowing the composition window and shifting it to a higher stiff fraction .
their calculations are limited to @xmath135 , and it is inconclusive whether the @xmath136 limit of this calculation yields a stable g phase .
the same qualitative behavior was found for the generic symmetric wedge in the strong segregation regime.@xcite table [ table2 ] summarizes asymmetry parameters from recently collected data @xcite . while none of these diblocks have the large conformational asymmetry required to test our prediction of a stable phase in strong segregation , @xmath137 starblock copolymers have asymmetry factors larger , by a factor of @xmath138 , than those due to intrinsic chain stiffness effects alone .
for example a @xmath139@xmath140 starpolymer has an asymmetry factor @xmath141 .
we have attempted to calculate energies for the hpl @xcite phase , which is now accepted as a metastable phase @xcite .
the minimal crystal phases d , p , and g have obvious candidate minimal surfaces to act as an intermaterial dividing surface towards which the majority - phase ends stretch ; and the minority - phase ends stretch towards the skeletal bond lattice . on the other hand
, the majority - phase ends in the hpl phase stretch towards a combination of lines ( in the hexagonally - arranged perforating tubes ) and surfaces ( within the majority - phase layer ) .
similarly , it is not obvious how to partition the minority - phase ends between lines and surfaces .
the result is a non - analytic mapping which is difficult to minimize over .
our attempts have thus far yielded quite high energies , of order that of the @xmath130 phase .
we have outlined a general method for computing the free energy of block copolymer phases in the strong segregation regime .
the procedure consists of the following steps : 1 .
choose a candidate geometry and an associated partitioning surface that divides space into disjoint interpenetrating regions ( the majority blocks from the two regions stretch towards this surface ) .
2 . divide the enclosed volume into infinitesimal wedges , defined by straight paths connecting the partitioning surface to a skeleton of bonds ( the minority blocks stretch towards this bond skeleton ) .
the a - b interface in each wedge is located such that the fraction of wedge volume filled by a blocks is locally equal to @xmath1 . the interfacial contribution to the free energy
is the area of the a - b interface times the a - b surface tension .
4 . compute the stretching free energy per chain for each wedge within the approximation of straight paths .
calculations for straight paths involve slight adjustments to the wedges whose contributions vanish in the limit of small wedges .
5 . optimize the free energy per chain with respect to the overall scale of the mesophase ( _ e.g. _ , the dimension of the unit cell ) .
optimize the mapping from the partitioning surface to the bond skeleton to minimize the overall free energy .
in certain structures ( _ e.g. _ hpl ) the majority phase ends lie on both lines and surfaces , in which case the procedure above must be suitably generalized .
we have also shown how to calculate the free energy for geometries where the shape of the @xmath2 interface is specified , for phases of cylindrical topology .
this requires polymer chain paths which are kinked at the @xmath2 interface .
the infinitesimal wedges are described by the relative area function @xmath142 ( eq [ eq : area ] ) which is parametrized by a single scalar @xmath37 ( eq [ eq : p ] and table [ table1 ] ) that roughly gauges the local gaussian curvature of the partitioning surface . for the classical phases
all wedges are identical , while bicontinuous phases have different distributions of shape factors @xmath37 . for symmetric stars
we find a metastable bicontinuous ( gyroid , or g ) phase which is most stable near the lamellae - hexagonal cylinder transition . for sufficiently asymmetric copolymers ( @xmath134 )
we predict a _ stable _ g phase . in a weakly segregated mesophase ,
very long polymer chains are in fact more strongly stretched than gaussian chains [ almdal , k. ; rosedale , j. h. ; bates , f. s. ; wignall , g. d. ; and fredrickson , g. h. _ phys .
lett . _ * 1990 * , _ 65 _ , 1112 ] , which has been explained in terms of concentration fluctuations [ barrat , j .- l . ; and fredrickson , g. h _ j. chem .
phys . _ * 1991 * , _ 95 _ , 1281 ] .
hajduk , d.a . ; harper , p.e . ; gruner , s.m . ; honeker , c.c . ; kim , g. ; thomas , e.l . ; fetters , l.j .
_ macromolecules _ * 1994 * , _ 27 _ , 4063 ; hajduk , d.a . ; gruner , s.m . ; rangarajan , p. ; register , r.a . ; fetters , l.j . ; honeker , c. ; albalak , r.j . ; thomas , e.l .
_ macromolecules _ * 1994 * , _ 27 _ , 490 ; hajduk , d.a . ; harper , p.e . ; gruner , s.m . ; honeker , c.c . ; thomas , e.l . ; fetters , l.j .
_ macromolecules _ * 1995 * , _ 28 _ , 2570 .
the composition variable @xmath1 is defined in ref .
@xcite as the volume fraction of the @xmath73 phase .
similarly , ref .
@xcite defines @xmath105 with @xmath72 and @xmath73 reversed with respect to the present article , so equations such as eqs [ eq : felam]-[eq : fesph ] are identical in the two works ( compare eqs 9 - 11 of ref .
eq 8 of ref .
@xcite has a misprint , and should have a factor of @xmath143 to match eq [ eq : f0 ] in section [ sec : general ] above . | we compute phase diagrams for @xmath0 starblock copolymers in the strong - segregation regime as a function of volume fraction @xmath1 , including bicontinuous phases related to minimal surfaces ( g , d , and p surfaces ) as candidate structures .
we present the details of a general method to compute free energies in the strong segregation limit , and demonstrate that the gyroid g phase is the most nearly stable among the bicontinuous phases considered .
we explore some effects of conformational asymmetry on the topology of the phase diagram .
# 1*#1 * # 1 10truept * [ fig # 1 here ] * 2 |
the nobel prize in physics awarded in 2015 to takaaki kajita and arthur mcdonald for studies of the atmospheric and solar neutrinos @xcite does not leave any doubt that neutrinos oscillate and have nonzero masses .
the latter fact leads to the well - known possibility for neutrino to have non - trivial electromagnetic properties @xcite , which brought forth the research area that was investigated in details by numerous authors ( see recent review @xcite and references therein ) .
in the course of these studies many phenomena that may appear in electromagnetic fields have been recognized and described thoroughly . among them
the neutrino spin - flavor oscillations is the one , featuring both the above mentioned basic neutrino aspects nonzero mass and electromagnetic properties from the one side and mixing from another @xcite . owing to this , in spite of being a longstanding problem
, the spin - flavor oscillations can reveal some new aspects of existence of neutrino mass and electromagnetic properties .
these concern the problems of neutrino parameters relation in neutrino physical ( mass ) and flavor bases and , generally speaking , of an accurate derivation of the formulas used to describe oscillations .
we below give an advanced view on the standard scheme of neutrino spin - flavor oscillations description aiming at the solid determination of parameters involved in the formalism and its rigorous derivations .
we consider neutrino mixing and oscillations in presence of an arbitrary constant magnetic field with nonzero transversal @xmath0 and longitudinal @xmath1 components with respect to the direction of neutrino propagation .
the electromagnetic interaction of neutrinos is determined by neutrino diagonal and transition magnetic moments that are introduced for the neutrino mass states .
explicit expressions for the effective neutrino diagonal and transition magnetic moments for the flavor bases in terms of these values for the mass states are obtained .
the effective evolution hamiltonian for the flavor neutrino and the corresponding oscillation probability are derived .
the role the longitudinal magnetic field component is examined . in particular
, it is shown that : 1 ) @xmath1 coupled to the corresponding magnetic moments shifts the neutrino energy , and 2 ) in case of nonvanishing neutrino transition magnetic moments @xmath1 produces an additional mixing between neutrino states , both in the mass and flovour neutrino bases .
consider two dirac neutrino physical states , @xmath2 and @xmath3 , with masses @xmath4 and @xmath5 . also consider the neutrino electromagnetic interaction via magnetic moment matrix @xmath6 ( @xmath7 ) : @xmath8 where @xmath9 is the electromagnetic field tensor , @xmath10 and @xmath11 being the dirac matrices . in a uniform magnetic field
the hamiltonian ( [ em_hamilt ] ) becomes @xmath12 where @xmath13 @xmath14 are the pauli matrices . for considering the neutrino evolution in the ultrarelativistic limit
we introduce the 4-component basis @xmath15 of states with definite helicity @xmath16 . with the standard column vector notation
, @xmath17 the neutrino evolution equation relevant to electromagnetic interaction has the schr@xmath18dinger - like form , @xmath19 the effective hamiltonian consists of the vacuum and interaction parts @xmath20 where the interaction part @xmath21 is composed of matrix elements of the interaction hamiltonian ( [ b_hamilt ] ) taken over the helicity neutrino states : @xmath22 .
let us calculate the effective interaction hamiltonian under the assumption that neutrino moves along the @xmath23-axis . from the magnetic field interaction hamiltonian ( [ b_hamilt ] )
we have : @xmath24 for the spinors representing the free neutrino states we take @xmath25 where @xmath26 is the neutrino @xmath27 momentum .
the two - component spinors @xmath28 define neutrino helicity states , and are given by @xmath29 recall that in the ultrarelativistic limit these two states correspond to the right - handed @xmath30 and left - handed @xmath31 chiral neutrinos , respectively .
substituting ( [ wave func ] ) into the effective hamiltonian given by ( [ hb ] ) , we get @xmath32 decomposing the magnetic field vector into longitudinal and transversal with respect to neutrino motion components @xmath33 it is possible to show that @xmath34 in the the ultrarelativistic limit @xmath35 one gets : @xmath36 where the quantity @xmath37 , fetched out also in @xcite , we call the transition gamma - factor . similarly , it is also possible to show that @xmath38 introducing an angle @xmath39 between the @xmath40 and @xmath26 vectors and assuming that @xmath41 is aligned along the @xmath42-axis , we further obtain that @xmath43 and , similarly , @xmath44
@xmath45 @xmath46 as it was expected , in neutrino transitions without change of helicity only the @xmath47 component of the magnetic field contributes to the effective potential , whereas in transitions with change of the neutrino helicity the transversal component @xmath48 matters . performing the remaining simple algebra one can readily write out the @xmath21 matrix .
thus , we obtain the evolution equation ( [ schred_eq ] ) in the following form @xmath49 this equation governs all possible oscillations of the four neutrino states with defined masses ( @xmath4 and @xmath5 ) and helicities ( @xmath50 and @xmath51 ) in the presence of a magnetic field . the obtained expression ( [ gen_evol_eq ] ) for the evolution hamiltonian explicitly reveals a particular role of the longitudinal magnetic field component @xmath52 : \1 ) the longitudinal magnetic field component @xmath52 coupled to the corresponding magnetic moment shifts the neutrino energy , \2 ) in case of nonvanishing neutrino transition magnetic moment @xmath53 , the presence of the longitudinal field @xmath52 produces mixing among neutrino species with different masses but with equal helicities . at the same time , from ( [ gen_evol_eq ] ) it is clearly seen that mixings of different helicity neutrino states are due to the corresponding magnetic moment ( or transition magnetic moment ) interactions with the transversal magnetic field @xmath0 .
once having physics in the mass basis in hands , our next step is to bring it to observational terms .
this means that we must elaborate a generalization of the mixing matrix for transitions between neutrino state vectors written in two four - component bases @xmath54 and @xmath55 so that @xmath56 this procedure appears to be not quite direct since we should hold the condition that polarization of the fields must be preserved under transformation of the bases elements .
that is why , keeping in mind that chiral components are almost the helicity ones , we define @xmath57 then , using eqs .
( [ u_def ] ) and ( [ transformations ] ) , it is easy to obtain that @xmath58 given the transition matrix ( [ u ] ) , derivation of the evolution equation in the flavor basis is straightforward : @xmath59 so that the effective magnetic field interaction hamiltonian @xmath60 has the following structure , @xmath61 here we have introduced the following formal notations intended to manifest an analogy with the standard spin - flavor oscillation formalism ( see below ) : @xmath62 @xmath63 the hamiltonian ( [ h_b_f ] ) has been derived within much consistent procedure than is usually given in literature . it should be noted that eqs .
( [ mu_fl ] ) could be also obtained as results of general consideration @xcite of neutrino mixing and oscillations in an arbitrary constant magnetic field ( the final form like that of eqs .
( [ mu_fl ] ) was not established in @xcite ) .
also , an effective neutrino oscillation hamiltonian with terms containing quantities ( [ mu_gammas_fl ] ) , however without account for possibility of spin transitions , was obtained in @xcite .
let us now confront the obtained effective magnetic field interaction hamiltonian ( [ h_b_f ] ) with the one typically written straight in the neutrino flavor basis ( see , for instance , @xcite and @xcite ) : @xmath64 where @xmath65 is the common neutrino gamma - factor .
first of all , we can ascertain that the structure of the obtained expression ( [ h_b_f ] ) is consistent with the standard " hamiltonian ( [ h_b_f_old ] ) . at that , the magnitudes ( [ mu_fl ] ) account for the neutrino mixings and represent the straightforward expressions for the effective magnetic moments in the flavor basis in terms of magnetic moments introduced in the mass basis .
the neutrino effective magnetic moments " ( [ mu_gammas_fl ] ) determine neutrino interactions and mixings due to the longitudinal magnetic field @xmath52 . as in the mass basis , the magnetic field component @xmath52 shifts the neutrino energy
also , it contributes to transitions with change of flavor .
it is interesting to observe that when the transversal field component is set to zero the hamiltonian structure becomes @xmath66 so that neutrino states with different flavors and the same chirality decouple and form subsystems independently mixed by the field . for example , one would have two neutrino species @xmath67 and @xmath68 mixed in accordance with the equation ( here the standard terms for vacuum oscillations are added ) @xmath69 in this way , neutrino oscillations would be influenced by the longitudinal field @xmath52 ( see also @xcite ) and @xmath52 would produce an additional mixing with respect to the usual one described by the vacuum mixing angle @xmath70 . the oscillation probability is just strightforward , @xmath71 where @xmath72 is the denominator of the pre - sine factor and the other quantities have usual for the theory of neutrino oscillations meaning .
it follows that in case the second term in the denominator is much smaller than the first one then the amplitude of the flavor oscillations @xmath73 gets its maximal value .
this can be considered as an effect of neutrinos interaction with the longitudinal magnetic field due to neutrino magnetic moments .
one of the authors ( a.s . ) is thankful to nicolao fornengo and carlo giunti for the kind invitation to participate in the 14th international conference on topics in astroparticle and underground physics .
the work on this paper was partially supported by the russian basic research foundation grants no .
14 - 22 - 03043-ofi - m , 15 - 52 - 53112-gfen and 16 - 02 - 01023-a .
4 fukuda y _ et al _ ( super - kamiokande collaboration ) 1998 _ phys .
lett . _ * 81 * 1562 ahmad q r _ et al _ ( sno collaboration ) 2001 _ phys .
lett . _ * 87 * 071301 ahmad q r _ et al _ ( sno collaboration ) 2002 _ phys .
lett . _ * 89 * 011301 fujikawa k and shrock r 1980 _ phys .
* 45 * 963 giunti c and studenikin a 2015 _ rev .
phys . _ * 87 * 531 ( _ preprint _
arxiv:1403.6344 [ hep - ph ] ) cisneros a 1971 _ astrophys .
space sci . _
* 10 * 87 schechter j and valle j 1981 _ phys .
* d * 24 1883 voloshin m b , vysotsky m i and okun l b 1986 _ sov .
. jetp _ * 64 * 446 akhmedov e k 1988 _ sov
. j. nucl .
phys . _ * 48 * 382 lim c - s and marciano w j 1988 _ phys .
_ d * 37 * 1368 likhachev g and studenikin a 1995 _ sov .
. jetp _ * 81 * 419 akhmedov e k and khlopov m yu 1988 _ sov .
phys . _ * 47 * 689 - 691 akhmedov e k and khlopov m yu 1988 _ mod .
lett . _ * a3 * 451 - 457 | we consider neutrino mixing and oscillations in presence of an arbitrary constant magnetic field with nonzero transversal @xmath0 and longitudinal @xmath1 components with respect to the direction of neutrino propagation .
the electromagnetic interaction of neutrinos is determined by diagonal and transition neutrino magnetic moments that are introduced for the neutrino mass states .
explicit expressions for the effective neutrino diagonal and transition magnetic moments for the flavor basis in terms of these values for the mass states are obtained .
the effective evolution hamiltonian for the flavor neutrino and the corresponding oscillation probability are derived .
the role of the longitudinal magnetic field component is examined . in particular
, it is shown that : 1 ) @xmath1 coupled to the corresponding magnetic moments shifts the neutrino energy , and 2 ) in case of nonvanishing neutrino transition magnetic moments @xmath1 produces an additional mixing between neutrino states , both in the mass and flavor neutrino bases . |
invasion of influenza virus into a host s upper respiratory tract leads to infection of healthy epithelial cells and subsequent production of progeny virions @xcite .
infection also triggers a variety of immune responses . in the early stage of infection a temporary non - specific response ( innate immunity ) contributes to the rapid control of viral growth while in the late stage of infection , the adaptive immune response dominates viral clearance @xcite .
the early immune response involves production of antiviral cytokines and cells , e.g. type 1 interferon ( ifn ) and natural killer cells ( nk cells ) , and is independent of virus type @xcite . in the special case of a first infection in a naive host ,
the adaptive immune response , mediated by the differentiation of naive t cells and b cells and subsequent production of virus - specific t cells and antibodies @xcite , leads to not only a prolonged killing of infected cells and virus but also the formation of memory cells which can generate a rapid immune response to secondary infection with the same virus @xcite .
@xmath0 t cells , which form a major component of adaptive immunity , play an important role in efficient viral clearance @xcite .
however , available evidence suggests they are unable to clear virus in the absence of antibodies @xcite except in hosts with a very high level of pre - existing naive or memory @xmath0 t cells @xcite .
some studies indicate that depletion of @xmath0 t cells could decrease the viral clearance rate and thus prolong the duration of infection @xcite .
furthermore , a recent study of human a(h7n9 ) hospitalized patients has implicated the number of effector @xmath0 t cells as an important driver of the duration of infection @xcite .
this diverse experimental and clinical data , sourced from a number of host - species , indicates that timely activation and elevation of @xmath0 t cell levels may play a major role in the rapid and successful clearance of influenza virus from the host .
these observations motivate our modeling study of the role of @xmath0 t cells in influenza virus clearance .
viral dynamics models have been extensively applied to the investigation of the antiviral mechanisms of @xmath0 t cell immunity against a range of pathogens , with major contributions for chronic infections such as hiv / siv @xcite , htlv - i @xcite and chronic lcmv @xcite . however , for acute infections such as measles @xcite and influenza @xcite , highly dynamical interactions between the viral load and the immune response occur within a very short time window , presenting new challenges for the development of models incorporating @xmath0 t cell immunity .
existing influenza viral dynamics models , introduced to study specific aspects of influenza infection , are limited in their ability to capture all major aspects of the natural history of infection , hindering their use in studying the role of @xmath0 t cells in viral clearance .
some models show a severe depletion of target cells ( i.e healthy epithelial cells susceptible to viral infection ) after viral infection @xcite
. depletion may be due to either infection or immune - mediated protection .
either way , these models are arguably incompatible with recent evidence that the host is susceptible to re - infection with a second strain of influenza a short period following primary exposure @xcite .
furthermore , as reviewed by dobrovolny _
@xcite , target cell depletion in these models strongly limits viral expansion so that virus can be effectively controlled or cleared at early stage of infection even in the absence of adaptive immunity , which contradicts the experimental finding that influenza virus remains elevated in the absence of adaptive immune response @xcite . while a few models do avoid target cell depletion @xcite , they either assume immediate replenishment of target cells @xcite or a slow rate of virus invasion into target cells resulting in a much delayed peak of virus titer at day 5 post - infection ( rather than the observed peak at day 2 ) @xcite .
moreover , models with missing or unspecified major immune components , e.g. no innate immunity @xcite , no antibodies @xcite or unspecified adaptive immunity @xcite , also indicate the need for further model development . for an in - depth review of the current virus dynamics literature on influenza ,
we refer the reader to the excellent article by dobrovolny _
_ @xcite .
t cells and b cells to produce effector @xmath0 t cells and antibodies , responsible for final clearance of virus . ] in this paper , we construct a within - host model of influenza viral dynamics in naive ( i.e. previously unexposed ) hosts that incorporates the major components of both innate and adaptive immunity and use it to investigate the role of @xmath0 t cells in influenza viral clearance . the model is calibrated against a set of published murine data from miao _ et al .
_ @xcite and is then validated through demonstration of its ability to qualitatively reproduce a range of published data from immune - knockout experiments @xcite . using the model
, we find that the recovery time defined to be the time when virus titer first drops below a chosen threshold in the ( deterministic ) model is negatively correlated with the level of effector @xmath0 t cells in an approximately exponential manner . to the best of our knowledge ,
this relationship , with support in both h3n2-infected mice and h7n9-infected humans @xcite , has not been previously identified .
the exponential relationship between @xmath0 t cell level and recovery time is shown to be remarkably robust to variation in a number of key parameters , such as viral production rate , ifn production rate , delay of effector @xmath0 t cell production and the level of antibodies . moreover , using the model , we predict that people with a lower level of naive @xmath0 t cells may receive significantly more benefit from induction of additional effector @xmath0 t cells . such production , arising from immunological memory , may be established through either previous viral infection or t cell - based vaccines .
the model of primary viral infection is a coupled system of ordinary and delay differential equations , consisting of three major components ( see fig . 1 for a schematic diagram ) .
13 describe the process of infection of target cells by influenza virus and are a major component in almost all models of virus dynamics in the literature .
4 and 5 model ifn - mediated innate immunity @xcite .
thirdly , adaptive immunity including @xmath0 t cells and b cell - produced antibodies for killing infected cells and neutralizing influenza virus respectively are described by eqs .
@xmath1 in further detail , eq .
1 indicates that the change in viral load ( @xmath2 ) is controlled by four factors : the production term ( @xmath3 ) in which virions are produced by infected cells ( @xmath4 ) at a rate @xmath5 @xcite ; the viral natural decay / clearance ( @xmath6 ) with a decay rate of @xmath7 ; the viral neutralisation terms ( @xmath8 and @xmath9 ) by antibodies ( both a short - lived antibody response @xmath10 driven by , e.g. igm , and a longer - lived antibody response @xmath11 driven by , e.g. igg and iga @xcite ) , and a consumption term ( @xmath12 ) due to binding to and infection of target cells ( @xmath13 ) . in eq . 2 , the term @xmath14 models logistic regrowth of the target cell pool @xcite .
both target cells ( @xmath13 ) and resistant cells ( @xmath15 , those protected due to ifn - induced antiviral effect ) can produce new target cells , with a net growth rate proportional to the severity of infection , @xmath16 ( i.e. the fraction of dead cells ) .
@xmath17 is the initial number of target cells and the maximum value for the target cell pool @xcite .
target cells ( @xmath13 ) are consumed by virus ( @xmath2 ) due to binding ( @xmath18 ) , the same process as @xmath12 .
note that @xmath19 and @xmath20 have different measurement units due to different units for viral load ( @xmath2 ) and infected cells ( @xmath4 ) .
as already mentioned , the innate response may trigger target cells ( @xmath13 ) to become resistant ( @xmath15 ) to virus , at rate @xmath21 .
resistant cells lose protection at a rate @xmath22 @xcite .
this process also governs the evolution of virus - resistant cells ( @xmath15 ) in eq .
5 . eq .
3 describes the change of infected cells ( @xmath4 ) .
they increase due to the infection of target cells by virus ( @xmath18 ) and die at a ( basal ) rate @xmath23 .
two components of the immune response increase the rate of killing of infected cells .
ifn - activated nk cells kill infected cells at a rate @xmath24 @xcite .
effector @xmath0 t cells ( @xmath25 ) produced through differentiation from naive @xmath0 t cells @xmath26 in eq .
6 kill at a rate @xmath27 . of note
our previous work has demonstrated that models of the innate response containing only ifn - induced resistance for target cells ( state @xmath15 ; eq .
5 ) , while able to maintain a population of healthy uninfected cells , still control viral kinetics through target cell depletion , and therefore can not reproduce viral re - exposure data @xcite .
given our interest in analysing a model that prevents target cell depletion , inclusion of ifn - activated nk cells ( term @xmath24 ) is an essential part of the model construction .
4 models the innate response , as mediated by ifn ( @xmath28 ) .
ifn is produced by infected cells at a rate @xmath29 and decays at a rate @xmath30 @xcite .
6 models stimulation of naive @xmath0 t cells ( @xmath26 ) into the proliferation / differentiation process by virus at a rate @xmath31 ) , where @xmath32 is the maximum stimulation rate and @xmath33 indicates the viral load ( @xmath2 ) at which half of the stimulation rate is achieved .
note that this formulation does not capture the process of antigen presentation and @xmath0 t cell activation , but rather is a simple way to establish the essential coupling between the viral load and the rate of @xmath0 t cell activation in the model @xcite . in eq .
7 , the production of effector @xmath0 t cells ( @xmath25 ) is assumed to be an `` advection flux '' induced by a delayed virus - stimulation of naive @xmath0 t cells ( the first term on the righthand side of eq .
the delayed variables , @xmath34 and @xmath35 , equal zero when @xmath36 .
the introduction of the delay @xmath37 is to phenomenologically model the delay induced by both naive @xmath0 t cell proliferation / differentiation and effector @xmath0 t cell migration and localization to the site of infection for antiviral action @xcite .
the delay also captures the experimental finding that naive @xmath0 t cells continue to differentiate into effector t cells in the absence of ongoing antigenic stimulation @xcite .
the multiplication factor @xmath38 indicates the number of effector @xmath0 t cells produced from one naive @xmath0 t cell , where @xmath39 is the average effector @xmath0 t cell production rate over the delay period @xmath37 .
the exponential form of the multiplication factor is derived based on the assumption that cell differentiation and proliferation follows a first - order advection reaction equation .
effector @xmath0 t cells decay at a rate @xmath40 .
similar to @xmath0 t cells , eqs . 8 and 9
model the proliferation / differentiation of naive b cells , stimulated by virus presentation at rate @xmath41 .
stimulation subsequently leads to production of plasma b cells ( @xmath42 ) after a delay @xmath43 .
the multiplication factor @xmath44 indicates the number of plasma b cells produced from one naive b cell , where @xmath45 is the production rate .
plasma b cells secrete antibodies , which exhibit two types of profiles in terms of experimental observation : a short - lived profile ( e.g. igm lasting from about day 5 to day 20 post - infection ) and a longer - lived profile ( e.g. igg and iga lasting weeks to months ) @xcite .
these two antibody responses are modeled by eqs .
10 and 11 wherein different rates of production ( @xmath46 and @xmath47 ) and consumption ( @xmath48 and @xmath49 ) are assumed .
the model contains 11 equations and 30 parameters ( see table 1 ) .
this represents a serious challenge in terms of parameter estimation , and clearly prevents a straightforward application of standard statistical techniques . to reduce uncertainty ,
a number of parameters were taken directly from the literature , as per the citations in table . 1 .
the rest were estimated ( as indicated in table 1 ) by calibrating the model against the published data from miao _ et al .
_ @xcite who measured viral titer , @xmath0 t cell counts and igm and igg antibodies in laboratory mice ( exhibiting a full immune response ) over time during primary influenza h3n2 virus infection ( see @xcite for a detailed description of the experiment ) .
the approach to estimating the parameters based on miao _ et al . _
s data is provided in the _ supplementary material _ and the estimated parameter values are given in table 1 .
note that the data were presented in scatter plots in the original paper @xcite , while we presented the data here in mean @xmath50 sd at each data collection time point for a direct comparison with our mean - field mathematical model . for model simulation ,
the initial condition is set to be @xmath51 unless otherwise specified . the initial target cell number ( @xmath17 )
was estimated by petrie _
we estimate that of order 100 cells ( resident in the spleen ) are able to respond to viral infection ( @xmath26 ) ( personal communication , n. lagruta , monash university , australia ) .
note that 100 naive @xmath0 t cells might underestimate the actual number of naive precursors that could respond to all the epitopes contained within the virus but does not qualitatively alter the model dynamics and predictions ( see _ results _ where the naive @xmath0 t cell number is varied between 0 to 200 ) . in the absence of further data
, we also use this value for the initial naive b cell number ( @xmath52 ) , but again this choice does not qualitatively alter the model predictions . the numerical method and code ( implemented in matlab , version r2014b , the mathworks , natick , ma ) for solving the model are provided in the _ supplementary material_. ) . note that due to the limit of detection for the viral load ( occurring after 10 days post - infection as seen in viral load data )
, the last three data points in the upper - left panel were not taken into consideration for model fitting . ]
clinical influenza a(h7n9 ) patient data was used to test our model predictions on the relationship between @xmath0 t cell number and recovery time .
the data was collected from 12 surviving patients infected with h7n9 virus during the first wave of infection in china in 2013 ( raw data is provided in _
dataset s1 _ ; see the paper of wang _ et al . _
@xcite for details of data collection ; this study was reviewed and approved by the shaphc ethics committee ) .
note that the clinical data were scarce for some patients .
for those patients , we have assumed that the available data are representative of the unobserved values in the neighboring time period .
for each patient , we took the average @xmath53 @xmath0 t cell number in @xmath54 peripheral blood mononuclear cells ( pbmc ) for the period from day 8 to day 22 ( or the recovery day if it comes earlier ) post - admission as a measure of the effector @xmath0 t cell level .
this period was chosen _ a priori _ as it roughly matches the duration of the @xmath0 t cell profile and clinical samples were frequently collected in this period .
the average @xmath0 t cell count was given by the ratio of the total area under the data points ( using trapezoidal integration ) to the number of days from day 8 to day 22 ( or the recovery day if it comes earlier ) .
for those patients for which samples at days 8 and/or 22 were missing we specified the average @xmath0 t cell level at the missing time point to be equal to the value from the nearest sampled time available .
) , effector @xmath0 t cells ( @xmath25 ) , short - lived antibody response ( @xmath10 ) and long - lived antibody response ( @xmath11 ) have been shown in fig . 2 .
] we first analyze the model behavior in order to establish a clear understanding of the model dynamics .
fig . 2 shows solutions ( time - series ) for the model compartments ( viral load , @xmath0 t cells and igm and igg antibody ) calibrated against the murine data from @xcite .
solutions for the remaining model compartments are shown in fig .
the model ( with both innate and adaptive components active ) prevents the depletion of target cells ( see fig .
3 wherein over 50% of target cells remain during infection ) and results in a minor loss of just 1020% of healthy epithelial cells ( i.e. the sum of target cells ( @xmath13 ) and virus - resistant cells ( @xmath15 ) ; see supplementary fig .
the primary driver for the maintenance of the target cell pool during acute viral infection is a timely activation of the innate immune response , and in particular the natural killer cells ( supplementary fig .
since we have previously shown that the target - cell limited model ( even with the resistant cell compartment ) is unable to reproduce observations from heterologous re - exposure experiments @xcite , our model improves upon previous models where viral clearance was only achieved through depletion of target cells ( a typical solution shown in supplementary fig .
importantly , our result is distinguished from that of saenz et .
@xcite , wherein the healthy cell population was similarly maintained , but primarily through induction of the virus - resistant state , thereby rendering that model incapable of capturing re - infection behavior as established in animal models @xcite .
the modeled viral dynamics exhibits three phases , each dominated by the involvement of different elements of the immune responses ( fig . 4 ) .
immediately following infection ( 02 days post - infection ) and prior to the activation of the innate ( and adaptive ) immune responses , virus undergoes a rapid exponential growth ( fig .
4a ) . in the second phase ( 25 days post - infection )
, the innate immune response successfully limits viral growth ( fig .
4a ) . in the third phase ( 46 days post - infection ) , adaptive immunity ( antibodies and @xmath0 t cells )
is activated and viral load decreases rapidly , achieving clearance .
4b and 4c demonstrate the dominance of the different immune mechanisms at different phases . in fig .
4b models with and without immunity are indistinguishable until day 2 ( shaded region ) , before diverging dramatically when the innate and then adaptive immune responses influence the dynamics . in fig . 4c
, models with and without an adaptive response only diverge at around day 4 as the adaptive response becomes active .
we have further shown that this three - phase property is a robust feature of the model , emergent from its mathematical structure and not a property of fine tuning of parameters ( see supplementary fig .
importantly , it clearly dissects the periods and effect of innate immunity , extending on previous studies of viral infection phases where the innate immune response was either ambiguous or ignored @xcite . and @xmath55 in the model ) .
the trajectories overlap prior to the activation of the innate response , before diverging due to target cell depletion .
the shaded region highlights the first phase ( exponential growth ) . in panel
( c ) , the dashed line shows viral kinetics in the absence of adaptive immunity ( by letting @xmath55 ; innate immunity remains active ) .
the trajectories overlap prior to the activation of the adaptive response .
the shaded region highlights the second phase ( innate response ) .
note : changes in model parameters shifts where the three phases occur , but does not alter the underlying three - phase structure .
i.e. existence of the three phases is robust to variation in parameters ( see _ supplementary material _ and supplementary fig .
s3 in particular ) . ] as reviewed by dobrovolny _
@xcite , a number of _ in vivo _ studies have been performed to dissect the contributions of @xmath0 t cells and antibodies @xcite .
we use the findings of these studies to validate our model , by testing how well it is able to reproduce the experimental findings ( without any further adjustment to parameters ) .
although the determination of the role of @xmath0 t cells is often hindered by co - inhibition of both @xmath0 t cells and the long - lived antibody response ( e.g. using nude mice ) , it is consistently observed that antibodies play a dominant role in final viral clearance while @xmath0 t cells are primarily responsible for the timely killing of infected cells and so indirectly contribute to an increased rate of removal of free virus towards the end of infection @xcite .
furthermore , experimental data demonstrate that a long - lived antibody response is crucial for achieving complete viral clearance , while short - lived antibodies are only capable of driving a transient decrease in viral load @xcite .
we find that our model ( with parameters calibrated against miao _
s data @xcite ) is able to reproduce these observations : * virus can rebound in the absence of long - lived antibody response ( see fig . 5 and supplementary fig .
both the @xmath0 t cell response and short - lived antibody response only facilitate a faster viral clearance , and are incapable of achieving clearance in the absence of long - lived antibody response ( see fig . 5 and supplementary fig .
a lower level of @xmath0 t cells ( modulated by a decreased level of initial naive @xmath0 t cells , @xmath26 ) significantly prolongs the viral clearance ( see supplementary fig .
in addition , the model also predicts a rapid depletion of naive @xmath0 t cells after primary infection ( see fig .
3 ) , which represents a full recruitment of naive @xmath0 t cell precursors .
this result may be associated with the experimental evidence suggesting a strong correlation between the naive @xmath0 t cell precursor frequencies and effector @xmath0 t cell magnitudes for different pmhc - specific t cell populations @xcite .
note that in fig .
5 no adjustments to the model ( e.g. to the vertical scale ) were made ; its behavior is completely determined by the calibration to the aforementioned murine data @xcite and so these findings represent a ( successful ) prediction of the model .
t- , iga- , igg- " was modeled by letting @xmath56 and @xmath57 . external igm " ( in addition to the igm produced by plasma cells ) was modeled by adding a new term @xmath58 to eq . 1 where @xmath59 follows a piecewise function @xmath60 for @xmath61 , @xmath62 for @xmath63 , @xmath64 for @xmath65 and @xmath60 for @xmath66 .
( b ) data is from the paper of iwasaki _ et al .
the data indicates that the long - lasting iga response , but not the long - lasting igg response or the short - lasting igm response , is necessary for successful viral clearance . no long - lived antibody response " was modeled by letting @xmath57 .
note that miao _
only measured igm and igg , but not iga . as such , our model s long - lived antibody response was calibrated against igg kinetics ( see fig . 2 ) .
therefore , we emphasise that we can only investigate the relative contributions of short - lived and long - lived antibodies . ] in summary , we have demonstrated that our model with parameters calibrated against murine data @xcite exhibits three important phases characterized by the involvement of various immune responses .
advancing on previous models , our model does not rely on target cell depletion , and successfully reproduces a multitude of behavior from knockout experiments where particular components of the adaptive immune response were removed .
this provides us with some confidence that each of the major components of the immune response has been captured adequately by our model , allowing us to now make predictions on the effect of the cellular adaptive response on viral clearance .
having established that our model is ( from a structural point of view ) biologically plausible and that our parameterization is capable of reproducing varied experimental data under different immune conditions ( i.e. knockout experiments ) , we now study how the cellular adaptive response influences viral kinetics in detail .
we focus on the key clinical outcome of recovery time , defined in the model as the time when viral titer first falls below @xmath67 , the minimum value detected in relevant experiments ( e.g. fig . 2 ) .
t cells plays an important role in determining recovery time .
recovery time is defined to be the time when viral load falls to 1 @xmath68 .
panel ( a ) shows that the average effector @xmath0 t cell number over days 620 is linearly related to the naive @xmath0 t cell number ( i.e. @xmath69 ) .
panel ( b ) shows that the recovery time is approximately exponentially related to the initial naive @xmath0 t cell number .
combined , these results give panel ( c ) wherein an approximately exponential relationship is observed between the average @xmath0 t cell number and recovery time , both of which are experimentally measurable .
note that the exponential / linear fits shown in the figures are not generated by the viral dynamics model but are used to indicate the trends ( evident visually ) in the model s behavior .
panel ( d ) shows that varying the delay @xmath37 ( in a similar way to that shown in fig .
s5 in the _ supplementary material _ ) , rather than the naive @xmath0 t cell number , does not alter the exponential relationship . in panel
( d ) , the crosses represent the results of varying @xmath37 and the empty circles are the same as those in panel ( c ) for comparison . ]
time series of the viral load show that the recovery time decreases as the initial naive @xmath0 t cell number ( @xmath26 ) increases ( supplementary fig .
s4 ) . with that in mind
, we now examine how recovery time is associated with the clinically relevant measure of effector @xmath0 t cell level during viral infection . with an increasing initial level of naive @xmath0 t cells , the average level of effector @xmath0 t cells over days 620 increases linearly ( fig .
6a ) , while the recovery time decreases in an approximately exponential manner ( fig .
combining these two effects gives rise to an approximately exponential relation between the level of effector @xmath0 t cells and recovery time ( fig .
note that the exponential / linear fits shown in the figures are simply to aid in interpretation of the results .
they are not generated by the viral dynamics model . if varying the delay for naive @xmath0 t cell activation and differentiation , @xmath37 , while keeping the naive @xmath0 t cell number fixed ( at the default value of 100 ) , we find that the average level of effector @xmath0 t cells is exponentially related to the delay , while the recovery time is dependent on the delay in a piecewise linear manner ( see supplementary fig .
s5 ) . nevertheless , the combination still leads to an approximately exponential relationship between the level of effector @xmath0 t cells and recovery time ( supplementary fig .
s6 ) , which is almost identical to that of varying naive @xmath0 t cells ( fig .
we also examine the sensitivity of the exponential relationship to other model parameters generally accepted to be important in influencing the major components of the system , such as the viral production rate @xmath5 , ifn production rate @xmath29 and naive b cell number .
we find that the exponential relationship is robust to significant variation in all of these parameters ( see supplementary figs .
s6 and s7 and fig . 9 ) .
these results suggest that a higher level of effector @xmath0 t cells is critical for early recovery , consistent with experimental findings @xcite .
t cell number and recovery time .
the x - axis is the average level of functional effector @xmath0 t cells ( i.e. @xmath70 cells ) over the period from day 8 to day 22 ( or the recovery day if it comes earlier ) .
spearman s rank correlation test indicates a significant negative correlation between the average @xmath0 t cell numbers and recovery time ( @xmath71 ) . excluding one of the patients ( no .
dataset s1 _ ; discussed in the _ discussion _ ) , all other data points ( solid dots ) are fitted by an offset exponential function @xmath72 , indicating that the best achievable recovery time for individuals with a high @xmath0 t cell response is approximately 17.5356 . ] finally , and perhaps surprisingly given our model has been calibrated purely on data from the mouse , a strikingly similar relationship as shown in fig .
6c is found in clinical data from influenza a(h7n9 ) virus - infected patients ( fig .
7 ) . excluding one patient ( no .
dataset s1 _ ; the exclusion is considered further in the _ discussion _ ) , average @xmath70 cells and recovery time are negatively correlated ( spearman s @xmath73 , @xmath74 ) and well captured by an exponential fit with an estimated offset ( see fig . 7 caption for details ) .
the exponential relationship ( observed in both model and data ) has features of a rapid decay for relatively low / intermediate levels of effector @xmath0 t cells and a strong saturation for relatively high @xmath0 t levels , implying that even with a very high level of naive @xmath0 t cells , recovery time can not be reduced below a certain value ( in this case , estimated to be approximately 17 days ) .
of course , the exponential relationship ( i.e. the scale of @xmath0 t cell level or recovery time ) , is only a qualitative one , as we have no way to determine the scaling between different x - axis measurement units , nor adjust for particular host and/or viral factors that differ between the two experiments ( i.e. h3n2-infection in the mouse @xcite versus h7n9-infeciton in humans @xcite ) .
in addition to naive @xmath0 t cells , memory @xmath0 t cells ( established through previous viral infection ) may also significantly affect recovery time due to both their rapid activation upon antigen stimulus and faster replication rate @xcite . to study the role of memory @xmath0 t cells
, we must first extend our model .
as we are only concerned with how the presence of memory @xmath0 t cells influences the dynamics , as opposed to the development of the memory response itself , the model is modified in a straightforward manner through addition of two additional equations which describe memory @xmath0 t cell ( @xmath75 ) proliferation / differentiation : @xmath76 accordingly the term @xmath77 in eq . 3
is modified to @xmath78 .
the full model and details on the choice of the additional parameters are provided in the _
supplementary material_. note that the model component , @xmath75 , may include different populations of memory @xmath0 t cells , including those directly specific to the virus and those stimulated by a different virus but which provide cross - protection @xcite .
t cells on viral clearance .
recovery time is defined to be the time when the viral load falls to 1 @xmath68 .
panel ( a ) demonstrates that varying the number of memory @xmath0 t cells ( @xmath75 ) reduces the recovery time for any naive @xmath0 t cell number ( i.e. @xmath69 ) .
note that saturation is observed for @xmath79 where the recovery time is about 6 days , independent of the naive cell numbers .
panel ( b ) demonstrates how the presence of pre - existing memory @xmath0 t cells ( solid dots ) leads to a shorter recovery time when compared to the case where no memory @xmath0 t cells are established ( open circles ) .
note the time scale difference in panels ( a ) and ( b ) .
this simulation is based on the assumption that the level of pre - existing memory @xmath0 t cells is assumed to be either 1% or 5% ( as indicated in the legend ) of the maximum effector @xmath0 t cell number due to primary viral infection .
the memory cell number ( which is not shown in this figure ) is about 30 time as many as the naive cell number shown in the figure , i.e. 30 naive cells result in about 900 memory cells before re - infection . ]
8a shows how the pre - existing memory @xmath0 t cell number ( @xmath75 ) changes the exponential relationship between naive @xmath0 t cells and recovery time .
importantly , as the number of memory @xmath0 t cells increases , the recovery time decreases for any level of naive @xmath0 t cells and the exponential relationship remains .
the extent of reduction in the recovery time for a relatively low level of naive @xmath0 t cells is greater than that for a relatively high level of naive @xmath0 t cells .
this suggests that people with a lower level of naive @xmath0 t cells may benefit more through induction of memory @xmath0 t cells , emphasizing the potential importance for taking prior population immunity into consideration when designing @xmath0 t cell - based vaccines @xcite .
the above result is based on the assumption that the initial memory @xmath0 t cell number upon re - infection is independent of the number of naive @xmath0 t cells available during the previous infection .
however , it has also been found that the stationary level of memory @xmath0 t cells is usually maintained at about 510% of the maximum antigen - specific @xmath0 t cell number during primary viral infection @xcite .
this indicates that people with a low naive @xmath0 t cell number may also develop a low level of memory @xmath0 t cells following infection . in consequence
, such individuals may be relatively more susceptible to viral re - infection @xcite .
this alternative and arguably more realistic relationship between the numbers of naive and memory @xmath0 t cells is simulated in fig .
8b where memory @xmath0 t cell levels are set to 5% of the maximum of the effector @xmath0 t cell level .
results suggest that , upon viral re - infection , pre - existing memory @xmath0 t cells are able to significantly improve recovery time except for in hosts with a very low level of naive @xmath0 t cells ( fig .
this is in accordance with the assumption that a smaller naive pool leads to a smaller memory pool and in turn a weaker shortening in recovery time .
although the model suggests that the failure of memory @xmath0 t cells to protect the host is unlikely to be observed ( because of the approximately 30 fold increase in the size of the memory pool relative to the naive pool ) , the failure range may be increased if the memory pool size is much smaller ( modulated by , say , changing 5% to 1% in the model ) .
therefore , for people with a low naive @xmath0 t cell number , the level of memory @xmath0 t cells may be insufficient and prior infection may provide very limited benefit , further emphasizing the opportunity for novel vaccines that are able to induce a strong memory @xmath0 t cell response to improve clinical outcomes . t cell number and recovery time .
recovery time is defined to be the time when viral load falls to 1 @xmath68 .
different antibody levels are simulated by varying the initial number of naive b cells ( i.e. @xmath52 at @xmath80 ) . ]
antibodies appear at a similar time as effector @xmath0 t cells during influenza viral infection and may enhance the reduction in the recovery time in addition to @xmath0 t cells . by varying the naive b cell number @xmath52 ( as a convenient , but by no means unique , way to influence antibody level ) , we find that increasing the antibody level shortens the recovery time regardless of the initial naive @xmath0 t cell number , leaving the exponential relation largely intact ( fig . 9 ) .
a slight saturation occurs for the case in which levels of both naive b cells and @xmath0 t cells are low .
moreover , variation in naive b cell number also results in a wider variation in recovery time for a lower naive @xmath0 t cell level , suggesting that people with a lower level of naive @xmath0 t cells may , once again , receive a more significant benefit ( in terms of recovery time ) through effective induction of an antibody response via vaccination .
in this paper , we have studied the role of @xmath0 t cells in clearing influenza virus from the host using a viral dynamics model .
the model was calibrated on a set of published murine data from miao _ et al .
_ @xcite and has been further shown to be able to reproduce a range of published data from experiments with different immune components knocked out . by avoiding target - cell depletion ,
our model is also compatible with re - infection data @xcite , providing a strong platform on which to examine the role of @xmath0 t cells in determining recovery time from infection .
our primary finding is that the time of recovery from influenza viral infection is negatively correlated with the level of effector @xmath0 t cells in an approximately exponential manner .
this robust property of infection has been identified from the model when calibrated against influenza a(h3n2 ) infection data in mice @xcite , but also observed in clinical case series of influenza a(h7n9 ) infection in humans ( fig .
7 ) @xcite .
our findings , in conjunction with conclusions on the potential role for a t cell vaccine that stimulates and/or boosts the memory response , suggest new directions for research in both non - human species and further studies in humans on the association between @xmath0 t cell levels and clinical outcomes .
further research , including detailed statistical fitting of our model to an extensive panel of infection data ( as yet unavailable ) from human and non - human species , is required to establish the generality of these relationships and provide quantitative insights for specific viruses in relevant hosts .
the non - linear relationship between effector @xmath0 t cell level and recovery time may be useful in clinical treatment .
the saturated property of the relation implies that a linear increase in the effector @xmath0 t cell level may result in diminishing incremental improvements in patient recovery times .
with evidence of a possible age - dependent loss of naive t cells @xcite , our model results imply that boosting the @xmath0 t cell response via t cell vaccination may be particularly useful for those with insufficient naive @xmath0 t cells .
the population - level consequences of such boosting strategies , while beyond the scope of this work , have previously been considered by the authors @xcite .
we also investigated the effect of memory @xmath0 t cell level on viral clearance and found unsurprisingly that a high pre - existing level of memory @xmath0 t cells was always beneficial .
however , our results suggest that pre - existing memory @xmath0 t cells may be particularly beneficial for certain groups of people . for example , if the memory @xmath0 t cell number induced by viral infection or vaccination is assumed to be relatively constant for everyone , people with less naive @xmath0 t cells would benefit more upon viral re - infection ( see fig . 8a ) . on the other hand ,
if assuming pre - existing memory @xmath0 t cell number is positively correlated with the number of naive @xmath0 t cells ( simulated in fig .
8b ) , people with more naive @xmath0 t cells would benefit more upon viral re - infection .
emerging evidence suggests that the relationship between the level of memory @xmath0 t cells and naive precursor frequencies is likely to be deeply complicated @xcite . in that context ,
our model predictions emphasize the importance for further research in this area , and the necessity to take prior population immunity into consideration when designing @xmath0 t cell vaccines @xcite .
we modeled both short - lived and long - lived antibody responses .
experimental data and model predictions consistently show that the short - lived antibody response results in a temporary reduction in virus level whereas the long - lived antibody response is responsible for complete viral clearance ( fig .
we emphasize here that although the model is able to capture the observed short - lived and long - lived antibody responses ( in order to study the virus - immune response interactions ) , it is not designed to investigate the mechanisms inducing different antibody responses .
the observed difference in antibody decay profile may be a result of many factors including the life times of different antibody - secreting cell types @xcite , different antibody life times @xcite and antibody consumption through neutralizing free virions .
detailed study of these phenomena requires a more detailed model and associated data for parameter estimation and model validation , and is thus left for future work .
similarly , @xmath81 t cells are also known to perform a variety of functions in the development of immunity , such as facilitation of the formation and optimal recall of @xmath0 t cells or even direct killing of infected cells during viral infection @xcite .
their depletion due to , say , hiv infection has also been associated with more severe clinical outcomes following influenza infection @xcite .
some of the major functions of @xmath81 t cells may be considered to be implicitly modeled through relevant parameters such as the rate of recall of memory @xmath0 t cells ( modeled by the delay @xmath82 ) in our extended model which includes memory @xmath0 t cells .
however , a detailed viral dynamics study of the role of @xmath81 t cells in influenza infection , including in hiv infected patients with depleted @xmath81 t cells , remains an open and important challenge . in a recent theoretical study
, it was found that spatial heterogeneity in the t cell distribution may influence viral clearance @xcite .
resident @xmath0 t cells in the lungs have a more direct and significant effect on timely viral clearance than do naive and memory pools resident in lymph nodes .
although this factor has been partially taken into consideration in our model by introducing a delay for naive / memory @xmath0 t cells , lack of explicit modeling of the spatial dynamics limits a direct application of our model to investigate these spatial effects . finally , as noted in the results , one of the influenza a(h7n9)-infected patients ( patient a79 )
was not included in our analysis of the clinical data ( fig .
although our model suggests some possibilities for the source of variation due to possible variation in parameter values , large variations in recovery time are only expected to occur for relatively low levels of naive @xmath0 t cells , nominally incompatible with this patient s moderate @xmath0 t cell response but a relatively long recovery time .
however , we note that @xmath83 t cell counts for this patient were only collected at days 10 and 23 , and that the count at day 10 was particularly low and much lower than that at day 23 ( see _ dataset s1 _ ) .
we suspect that delayed , rather than weakened , production ( to at least day 10 ) of the @xmath83 t cell response in this patient substantially contributed to the observed delay in recovery .
further investigation of this patient s clinical course and clinical samples is currently being undertaken .
100 taubenberger , j. k. , and morens , d. m. ( 2008 ) .
the pathology of influenza virus infections .
_ annu rev pathol . _ * 3 , * 499522 .
kreijtz , j. h. _ et al . _
immune responses to influenza virus infection .
_ virus res . _
* 162 , * 1930 .
goodbourn , s. _ et al .
_ ( 2000 ) .
interferons : cell signalling , immune modulation , antiviral response and virus countermeasures . _ j gen virol . _
* 81 , * 23412364 .
sadler , a. j. , and williams , b. r. g. ( 2008 ) . interferon - inducible antiviral effectors .
_ nat rev immunol .
_ * 8 , * 559568 .
biron , c. a. _ et al . _
natural killer cells in antiviral defense : function and regulation by innate cytokines .
_ annu rev immunol .
_ * 17 , * 189220 .
jost , s. , and altfeld , m. ( 2013 ) .
control of human viral infections by natural killer cells .
_ annu rev immunol . _ * 31 , * 163194 .
iwasaki , a. , and pillai , p. s. ( 2014 ) .
innate immunity to influenza virus infection .
_ nat rev immunol . _ * 14 , * 315328 .
murali - krishna , k. _ et al . _
counting antigen - specific cd8 t cells : a reevaluation of bystander activation during viral infection . _ immunity .
_ * 8(2 ) , * 177187 . wherry , e. j. , and ahmed , r. ( 2004 ) .
t - cell differentiation during viral infection .
_ j virol . _ * 78(11 ) , * 55355545 . la gruta , n. l. , and turner , s. j. ( 2014 ) .
t cell mediated immunity to influenza : mechanisms of viral control .
_ trends immunol . _ * 35(8 ) , * 396402 .
zhang , n. , and bevan , m. j. ( 2011 ) .
@xmath84 t cells : foot soldiers of the immune system
. _ immunity . _ * 35(2 ) , * 161168 .
iwasaki , t. , and nozima , t. ( 1977 ) .
the roles of interferon and neutralizing antibodies and thymus dependence of interferon and antibody production .
_ j immunol . _
* 118 , * 256263 .
fang , m. , and sigal , l. j. ( 2005 ) .
antibodies and @xmath85 t cells are complementary and essential for natural resistance to a highly lethal cytopathic virus .
_ j immunol .
_ * 175 , * 68296836 .
graham , m. b. , and braciale , t. j. ( 1997 ) . resistance to and recovery from lethal influenza virus infection in b lymphocyte - deficient mice .
_ j exp med .
_ * 186 , * 20632068 .
moskophidis , d. , and kioussis , d. ( 1998 ) .
contribution of virus - specific @xmath84 cytotoxic t cells to virus clearance or pathologic manifestations of influenza virus infection in a t cell receptor transgenic mouse model . _ j exp med . _ * 188(2 ) , * 223232 .
valkenburg sa .
_ et al . _
. protective efficacy of cross - reactive @xmath85 t cells recognising mutant viral epitopes depends on peptide - mhc - i structural interactions and t cell activation threshold .
_ plos pathog . _ * 6(8 ) , * e1001039 .
doi:10.1371/journal.ppat.1001039 .
yap , k. , and ada , g. ( 1978 ) .
cytotoxic t cells in the lungs of mice infected with an influenza a virus .
_ scand j immunol . _ * 7(1 ) , * 7380 .
wells , m. a. _ et al . _
recovery from a viral respiratory infection : i. influenza pneumonia in normal and t - deficient mice . _
j immunol . _
* 126 , * 10361041 .
hou , s. _ et al . _
delayed clearance of sendai virus in mice lacking class i mhc - restricted @xmath85 t cells .
_ j immunol . _ * 149 , * 13191325
. bender , b. s. _ et al . _
transgenic mice lacking class i major histocompatibility complex - restricted t cells have delayed viral clearance and increased mortality after influenza virus challenge . _
j exp med .
_ * 175 , * 11431145 .
wang , z. _ et al . _
recovery from severe h7n9 disease is associated with diverse response mechanisms driven by @xmath85 t - cells .
_ nat commun .
_ * 6*,7783 ; 10.1038/ncomms7833 .
perelson , a. s. _ et al . _
hiv-1 dynamics in vivo : virion clearance rate , infected cell life - span , and viral generation time . _ science . _
* 217 , * 15821586 .
perelson , a. s. ( 2002 ) .
modelling viral and immune system dynamics .
_ nat rev immunol . _
* 2 , * 2836 .
antia , r. _ et al . _
models of @xmath85 responses : 1 . what is the antigen - independent proliferation program .
_ j theor biol . _
* 221 , * 585598 .
chao , d. l. _ et al . _
( 2004 ) . a stochastic model of cytotoxic t cell responses .
_ j theor biol . _ * 228 , * 227240 .
hraba , t. , and doleal , j. ( 1995 ) .
mathematical modelling of hiv infection therapy .
_ int j immunopharmac . _ * 17(6 ) , * : 523526 .
de boer , r. j. ( 2007 ) . understanding the failure of @xmath85 t - cell vaccination against simian / human immunodeficiency virus .
_ j virol . _ * 81(6 ) , * 28382848 .
lim , a. g. , and maini , p. k. ( 2014 ) .
htlv - i infection : a dynamic struggle between viral persistence and host immunity .
_ j theor biol . _
* 352 , * 92108 .
althaus , c. l. _ et al . _
dynamics of @xmath85 t cell responses during acute and chronic lymphocytic choriomeningitis virus infection .
_ j immunol . _ * 179(5 ) , * 29442951 .
le , d. _ et al . _
mathematical modeling provides kinetic details of the human immune response to vaccination .
_ front cell infect microbiol . _
* 4 , * 177 ; 10.3389/fcimb.2014.00177 .
heffernan , j. m. , and keeling , m. j. ( 2008 ) .
an in - host model of acute infection : measles as a case study .
_ theor popul biol . _ * 73 , * 134147
. bocharov , g. a. , and romanyukha , a. a. ( 1994 ) .
mathematical model of antiviral immune response iii .
influenza a virus infection .
_ j theor biol . _ * 167 , * 323360
. chang , d. b. , and young , c. s. ( 2007 ) .
simple scaling laws for influenza a rise time , duration , and severity .
_ j theor biol . _ * 246 , * 621635 .
hancioglu , b. _ et al . _
( 2007 ) . a dynamical model of human immune response to influenza a virus infection .
_ j theor biol .
_ * 246 , * 7086 .
handel , a. , and antia , r. ( 2008 ) .
a simple mathematical model helps to explain the immunodominance of cd8 t cells in influenza a virus infections . _
_ * 82(16 ) , * 77687772 .
lee , h. y. _ et al . _
simulation and prediction of the adaptive immune response to influenza a virus infection .
_ j virol . _ * 83(14 ) , * 71517165 .
saenz , r. a. _ et al . _
dynamics of influenza virus infection and pathology .
_ j virol . _
* 84 , * 39743983 .
miao , h. _ et al . _
( 2010 ) . quantifying the early immune response and adaptive immune response kinetics in mice infected with influenza a virus . _
j virol . _ * 84 , * 66876698 .
dobrovolny , h. m. _ et al . _
( 2013 ) . assessing mathematical models of influenza infections using features of the immune response . _
plos one . _ * 8(2 ) , * e0057088 . doi:10.1371/journal.pone.0057088 .
reperant , l. a. _ et al . _
the immune response and within - host emergence of pandemic influenza virus .
_ lancet . _
* 384 , * 20772081 .
crauste , f. _ et al . _
( 2015 ) . predicting pathogen - specific cd8 t cell immune responses from a modeling approach . _
j theor biol . _ * 374 , * 6682 .
zarnitsyna , v. i. _ et al . _
. mathematical model reveals the role of memory cd8 t cell populations in recall responses to influenza .
_ front immunol . _ * 7 , * 165 ; 10.3389/fimmu.2016.00165 .
laurie , k. l. _ et al . _
the time - interval between infections and viral hierarchies are determinants of viral interference following influenza virus infection in a ferret model . _ j infect dis . _ * 212(11 ) , * 17011710 .
kris , r. _ et al . _
passive serum antibody causes temporary recovery from influenza virus infection of the nose , trachea and lung of nude mice .
_ immunol . _
* 63 , * 349353 .
pawelek , k. a. _ et al . _
modeling within - host dynamics of influenza virus infection including immune responses .
_ plos comput biol .
_ * 8(6 ) , * e1002588 .
doi:10.1371/journal.pcbi.1002588 .
cao , p. _ et al . _
innate immunity and the inter - exposure interval determine the dynamics of secondary influenza virus infection and explain observed viral hierarchies .
_ plos comput biol . _ * 11(8 ) , * e1004334 .
doi:10.1371/journal.pcbi.1004334 .
baccam , p. _ et al . _
kinetics of influenza a virus infection in humans .
_ j virol .
_ * 80 , * 75907599
. hwang , i. _ et al . _
. activation mechanisms of natural killer cells during influenza virus infection .
_ plos one . _ * 7(12 ) , * e51858 .
doi:10.1371/journal.pone.0051858 .
kaech , s. m. , and ahmed , r. ( 2001 ) .
memory @xmath84 t cell differentiation : initial antigen encounter triggers a developmental program in nave cells .
_ nat immunol . _
* 2 , * 415422 .
cerwenka , a. _ et al . _
naive , effector , and memory cd8 t cells in protection against pulmonary influenza virus infection : homing properties rather than initial frequencies are crucial .
_ j immunol . _ * 163 , * 55355543 .
lawrence , c. w. _ et al . _
frequency , specificity , and sites of expansion of @xmath85 t cells during primary pulmonary influenza virus infection .
_ j immunol . _ * 174 , * 53325340 .
van stipdonk , m. j. _ et al . _
nave ctls require a single brief period of antigenic stimulation for clonal expansion and differentiation .
_ nat immunol . _
* 2 , * 423429 .
petrie , s. m. _ et al . _
( 2013 ) . reducing uncertainty in within - host parameter estimates of influenza infection by measuring both infectious and total viral load . _
plos one . _ * 8(5 ) : * e0064098 .
doi:10.1371/journal.pone.0064098 .
smith , a. m. _ et al . _
an accurate two - phase approximate solution to an acute viral infection model .
_ j math biol . _ * 60 , * 711726
. neff - laford , h. d. _ et al . _
fewer ctl , not enhanced nk cells , are sufficient for viral clearance from the lungs of immunocompromised mice .
_ cellul immunol . _ * 226 , * 5464 .
jenkins , m. k. , and moon , j. j. ( 2012 ) .
the role of naive t cell precursor frequency and recruitment in dictating immune response magnitude .
_ j immunol . _ * 188 , * 41354140 .
marois , i. _ et al . _
the administration of oseltamivir results in reduced effector and memory @xmath85 t cell responses to influenza and affects protective immunity .
_ faseb j. _ * 29 , * pii : fj.14 - 260687 .
asano , m. s. , and ahmed , r. ( 1996 ) .
cd8 t cell memory in b cell - deficient mice .
_ j exp med .
_ * 183 , * 21652174 .
veiga - fernandes , h. _ et al .
_ ( 2000 ) .
response of naive and memory @xmath84 t cells to antigen stimulation _ in vivo_. _ nat immunol . _ * 1(1 ) , * 4753 .
badovinac , v. p. , and harty , j. t. ( 2006 ) .
programming , demarcating , and manipulating @xmath85 t - cell memory .
_ immunol rev . _ * 211 , * 6780 .
dispirito , j. r. , and shen , h. ( 2010 ) .
quick to remember , slow to forget : rapid recall responses of memory @xmath85 t cells .
_ cell res . _
* 20 , * 1323 .
regner , m. ( 2001 ) .
cross - reactivity in t - cell antigen recognition .
_ immunol cell biol . _ * 79(2 ) , * 91100 .
sewell , a. k. ( 2012 ) .
why must t cells be cross - reactive ?
_ nat rev immunol . _ * 12(9 ) , * 669677 .
bolton , k. j. _ et al . _
prior population immunity reduces the expected impact of ctl - inducing vaccines for pandemic influenza control . _
plos one . _ * 10(3 ) , * e0120138 . doi:10.1371/journal.pone.0120138
. harty , j. t. , and badovinac , v. p. ( 2008 ) . shaping and reshaping @xmath85 t - cell memory .
_ nat rev immunol .
_ * 8 , * 107119 .
hou , s. _ et al . _
virus - specific @xmath85 t - cell memory determined by clonal burst . _
* 369 , * 652654 .
lazuardi , l. _ et al . _
age - related loss of nave t cells and dysregulation of t - cell / b - cell interactions in human lymph nodes .
_ immunol .
_ * 114(1 ) , * 3743 . cicin - sain , l. _ et al . _ ( 2007 ) .
dramatic increase in naive t cell turnover is linked to loss of naive t cells from old primates .
_ proc natl acad sci usa . _ * 104(50 ) , * 1996019965 .
cicin - sain , l. _ et al . _
loss of naive t cells and repertoire constriction predict poor response to vaccination in old primates .
_ j immunol . _ * 184 , * 67396745 .
la gruta , n. l. _ et al . _
primary ctl response magnitude in mice is determined by the extent of naive t cell recruitment and subsequent clonal expansion .
_ j clin invest .
_ * 120(6 ) , * 18851894 .
thomas , p. g. _ et al . _
( 2012 ) . ecological analysis of antigen - specific ctl repertoires defines the relationship between naive and immune t - cell populations .
_ proc natl acad sci usa . _ * 110(5 ) , * 18391844 .
tscharke , d. c. _ et al . _
sizing up the key determinants of the @xmath85 t cell response .
_ nat rev immunol . _ * 15(11 ) , * 705716 . ho , f. _ et al . _
. distinct short - lived and long - lived antibody - producing cell populations . _ eur j immunol . _ * 16(10 ) , * 12971301 .
nutt , s. l. _ et al . _
( 2015 ) . the generation of antibody - secreting plasma cells .
_ nat rev immunol . _ * 15(3 ) , * 160171 .
vieira , p. , and rajewsky , k. ( 1988 ) .
the half - lives of serum immunoglobulins in adult mice .
_ eur j immunol . _ * 18(2 ) , * 313316 .
riberdy , j. m. _ et al .
_ ( 2000 ) . diminished primary and secondary influenza virus - specific @xmath85 t - cell responses in cd4-depleted @xmath86 mice .
_ j virol . _ * 74 , * 97629765 .
laidlaw , b. j. _ et al . _
@xmath87 t cell help guides formation of @xmath88 lung - resident memory @xmath84 t cells during influenza viral infection .
_ immunity . _ * 41 , * 633645 . cohen , c. _ et al . _ ( 2015 ) .
mortality amongst patients with influenza - associated severe acute respiratory illness , south africa , 2009 - 2013 .
_ plos one . _ * 10(3 ) , * e0118884 . doi:10.1371/journal.pone.0118884 .
pc is supported by an australian government national health and medical research council ( nhmrc ) funded centre for research excellence in infectious diseases modelling to inform public health policy .
zw is supported by an nhmrc australia china exchange fellowship .
awcy is supported by an australian postgraduate award .
jm and kk are support by nhmrc career development and senior researcher fellowships respectively . jmm is supported by an australian research council future fellowship .
the h7n9 clinical study was supported by national natural science foundation of china ( nfsc ) grants 81471556 , 81470094 and 81430030 .
the authors would like to thank members of the modelling and simulation unit in the centre for epidemiology and biostatistics and the school of mathematics and statistics , university of melbourne , for helpful advice on the study .
we have no competing interests .
.model parameter values obtained by fitting the model to experimental data . @xmath89 $ ] , @xmath90 $ ] and @xmath91 $ ] represent the units of viral load , ifn and antibodies respectively .
@xmath89 $ ] and @xmath91 $ ] are @xmath68 ( 50% egg infective dose ) and @xmath92 , consistent with the units of data .
ifn is assumed to be a non - dimensionalized variable in the model , and therefore @xmath90 $ ] can be ignored .
some parameters are obtained from the literature and the rest are obtained by fitting the model to experimental data in the paper of miao _ et al . _
@xcite , except @xmath93 which is of minor importance when considering a single infection and is thus fixed to reduce uncertainty .
[ cols="<,<,<,<",options="header " , ]
the delay in activation of @xmath0 t cells and b cells directly results in the delay of production of effector @xmath0 t cells and antibody - producing cells . by a time - shift transform @xmath94 ,
7 in the main text becomes @xmath95 where @xmath96 and @xmath97 starts from @xmath98 . as defined in the main text , @xmath99 for any negative time , i.e. @xmath100 for @xmath101 .
moreover , @xmath102 for @xmath103 .
thus , the equation is trivial for @xmath104 . therefore ,
for @xmath104 , we replace @xmath97 back to @xmath105 and eq .
s1 becomes @xmath106 where @xmath96 .
similarly , eqs . 9 - 11 in the main text can be changed to @xmath107 in this way , the model presented in the main text is equivalent to the following model : @xmath108 for variables whose independent variable is not explicitly specified , they are all functions of @xmath105 , i.e. @xmath2 reads @xmath109 .
this model avoids negative time and is solved by the following steps : * firstly choosing a time step size @xmath110 and using it to discretise the time domain to be @xmath111 .
the results in the main text are generated using @xmath112 ( day ) , the choice of which is based on the result that further decreasing @xmath110 does not improve the solution ( results not shown ) . *
given initial condition + @xmath113 , + we solve the model iteratively .
for iteration from @xmath114 to @xmath115 , @xmath116 , @xmath117 and @xmath118 in eqs .
s6 and s8 are already known and thus treated as parameters .
the system becomes an ode system and can be easily solved by using a built - in ode solver _ ode15s _ in matlab(r2014b ) with default settings .
all the variables at time @xmath115 are then updated for use in the next iteration . .... clear dt=0.1 ; % time step size time=0:dt:100 ; % parameters t0=7e+7;gt=0.8;pv=210;deltav=5.0 ; beta=5e-7;betap=3e-8 ; deltai=2;kappan=2.5;kappae=5e-5 ; phi=0.33;rho=2.6;pf=1e-5;deltaf=2 ; betacn=1;betabn=0.03 ; kappas=0.8;kappal=0.4 ; pc=1.2;pb=0.52;deltae=0.57;deltap=0.5 ; ps=12;pl=4;deltas=2;deltal=0.015 ; tauc=6;taub=4;hc=1e+4;hb=1e+4 ; % index indicating when delayed process starts indc = round(tauc / dt+1 ) ; % for tauc indb = round(taub / dt+1 ) ; % for taub % variable vectors and initial conditions v = zeros(1,length(time));v(1)=1e+4 ; t = v;t(1)=7e+7 ; i = t;i(1)=0 ; r = i ; f = i ; cn=100*ones(1,length(time ) ) ; bn=100*ones(1,length(time ) ) ; e = zeros(1,indc+length(time ) ) ; p = i ; as = zeros(1,indb+length(time ) ) ; al = as ; init=[v(1),t(1),i(1),r(1),f(1),cn(1),e(1),bn(1),p(1),as(1),al(1 ) ] ' ; options = odeset('reltol',1e-3,'abstol',1e-6 ) ; for i=2:length(time ) [ ~,y]=ode15s(@odemodel,[0 dt],init , options , e(i),al(i),as(i),tauc , taub , phi , rho , deltaf , gt , pf , pv , beta , betap , kappan , deltav , deltai , betacn , betabn , kappae , kappas , pl , ps , deltal , deltas , deltap , deltae , pc , pb , kappal , hc , hb , t0 ) ; v(i)=y(end,1);t(i)=y(end,2);i(i)=y(end,3);r(i)=y(end,4 ) ; f(i)=y(end,5);cn(i)=y(end,6);bn(i)=y(end,8);p(i)=y(end,9 ) ; e(indc+i)=y(end,7 ) ; as(indb+i)=y(end,10 ) ; al(indb+i)=y(end,11 ) ; init = y(end , : ) ' ; % initial condition for next iteration end .... .... function ynew = odemodel(~,y , e , al , as , tauc , taub , phi , rho , deltaf , gt , pf , pv , beta , betap , kappan , deltav , deltai , betacn , betabn , kappae , kappas , pl , ps , deltal , deltas , deltap , deltae , pc , pb , kappal , hc , hb , t0 ) % v : viral load % t : target cell % i : infected cell % r : resistant cell % f : ifn % cn : naive cd8 + t cells % e : effector cd8 + t cells % bn : naive b cells % p : plasma b cells % as : short - lived antibodies % al : long - lived antibodies % y=[v , t , i , r , f , cn , e , bn , p , as , al ] ynew = zeros(11,1 ) ; ynew(1)=pv*y(3)-deltav*y(1)-kappas*y(1)*as - kappal*y(1)*al - beta*y(1)*y(2 ) ; ynew(2)=gt*(y(2)+y(4))*(1-(y(2)+y(3)+y(4))/t0)-betap*y(1)*y(2)+rho*y(4)-phi*y(2)*y(5 ) ; ynew(3)=betap*y(1)*y(2)-deltai*y(3)-kappan*y(3)*y(5)-kappae*y(3)*e ; ynew(4)=phi*y(2)*y(5)-rho*y(4 ) ; ynew(5)=pf*y(3)-deltaf*y(5 ) ; ynew(6)=-betacn*y(1)/(y(1)+hc)*y(6 ) ; ynew(7)=betacn*y(1)/(y(1)+hc)*y(6)*exp(pc*tauc)-deltae*y(7 ) ; ynew(8)=-betabn*y(1)/(y(1)+hb)*y(8 ) ; ynew(9)=betabn*y(1)/(y(1)+hb)*y(8)*exp(pb*taub)-deltap*y(9 ) ; ynew(10)=ps*y(9)-deltas*y(10 ) ; ynew(11)=pl*y(9)-deltal*y(11 ) ; .... the model contains 11 equations and 30 parameters ( see table 1 in the main text ) .
this represents a serious challenge in terms of parameter estimation , and clearly prevents a straightforward application of standard statistical techniques .
however , based on an extensive survey of the experimental literature , we have been able to identify plausible , but by no means unique , combinations of parameters that successfully explain the available data .
a number of parameters were taken directly from the literature , as per the citations in table 1 . the rest ( 18 parameters )
were estimated by calibrating the model to the published data from miao _ et al .
_ @xcite who measured viral titre , @xmath0 t cell counts and igm and igg antibodies in laboratory mice ( exhibiting a full immune response ) over time during primary influenza h3n2 virus infection ( shown in fig . 2 in the main text ) .
note that the data were presented in scatter plots in the original paper @xcite , while we presented the data in mean @xmath50 sd at each data collection time point ( as shown in fig . 2 in the main text ) and
fit our mean - field mathematical model to the means .
* we first manually determined a set of the 18 parameters which produced a model solution that reasonably matched the experimental data shown in fig . 2 in the main text .
in detail , the main criteria include : 1 ) the viral load starts from about @xmath119 , reaches a peak of about @xmath120 at 12 days p.i .
, then declines rapidly from about 5 days p.i .
( note that the last three data points were not considered due to the limit of detection ) ; 2 ) the @xmath0 t cell count starts to increase rapidly at about 6 days p.i .
, reaches a peak of about @xmath121@xmath54 at 810 days p.i .
and returns back to zero after about 20 days p.i . ;
3 ) the igm level starts to increase rapidly at 45 days p.i .
, reaches a peak of about 200300
pg / ml at about 10 days p.i . and returns back to baseline after about 20 days p.i .
; 4 ) the igg level starts to increase rapidly at 45 days p.i . , reaches a peak of about 8001000 pg / ml at about 20 days p.i . and decays slowly .
given the high - dimensionality of the parameter space and limited experimental data , the procedure was essential in allowing us to identify a candidate parameter set which was not far from generating a local minimum in the following optimisation process .
* the candidate parameter set was then used as an initial estimate for optimization using matlab s built - in function _ fmincon _ with default settings .
the target of optimization was to minimize the least - squares error ( lse ) : @xmath122 where the four lse components for viral load , effector @xmath0 t cells , igm and igg were given respectively by @xmath123 ^ 2 , \tag{s18}\ ] ] @xmath124 ^ 2 , \tag{s19}\ ] ] @xmath125 ^ 2 , \tag{s20}\ ] ] @xmath126 ^ 2 .
\tag{s21}\ ] ] @xmath127 , @xmath128 , @xmath129 and @xmath130 indicate the model solution for variables @xmath2 , @xmath25 , @xmath10 and @xmath11 evaluated at time @xmath131 respectively .
@xmath132 , @xmath133 , @xmath134 and @xmath135 indicate the associated data .
@xmath136 is the index of the time point and the @xmath131 may differ for the four components .
@xmath137 , @xmath138 , @xmath139 and @xmath140 were used to scale the errors to the same order of magnitude . due to the fact that data were not collected at a fixed frequency (
i.e. the time interval between adjacent data points is not constant ) , the errors at different time points were assigned different weights based on the length of time intervals between adjacent points .
for example , if viral load was measured at time @xmath141 , the weight function @xmath142 for interior points was given by @xmath143 and for boundary points by @xmath144 this error weighting was used to weaken the domination of dense data points on model fits .
it is evident that a lot of measurements were done within the first 10 days post - infection but only a few were performed after day 20 post - infection , in particular for igg data ( see fig . 2 in the main text ) .
we found that using equally weighted lses led to a model fit that manifestly failed to capture those sparse data points which we believe are equally , if not more important from a more biological perspective , in determining the igg kinetics ( see fig .
s8 , compared with fig . 2 in the main text ) .
* the parameter constraints when using _ fmincon _ were set to @xmath145 $ ] , @xmath146 $ ] , @xmath147 $ ] , @xmath148 $ ] , @xmath149 $ ] , @xmath150 $ ] , @xmath151 $ ] , @xmath152 $ ] , @xmath153 $ ] , @xmath154 $ ] , @xmath155 $ ] , @xmath156 $ ] , @xmath157 $ ] , @xmath158 $ ] , @xmath159 $ ] , @xmath160 $ ] , @xmath161 $ ] and @xmath162 $ ] . * after obtaining a locally optimized solution for the candidate parameter set , we then checked the solution generated and evaluated its biological plausibility ( based on the criteria mentioned above ) .
this step was essential as given the over - specification of the model ( in a statistical sense ) , it was possible for good fitting solutions to be identified by matlab s optimization algorithm , which were nonetheless biologically implausible .
for example , oscillatory solutions for quantities such as igg , while providing a good - fit " to data , were not deemed acceptable on biological grounds ( see fig .
s9 for such an example ) .
if the optimised solution failed our ( qualitative ) evaluation , we returned to the first step and redetermined a new set of parameters as new initial estimates to be optimized using matlab .
this entire process was repeated to arrive at the default parameter set shown in table 1 in the main text . to guarantee that the default parameter set was a good choice , we further randomly generated 10,000 sets of parameter samples near the default parameter set ( within @xmath5050% from the default values ) and used them as initial estimates with _
fmincon _ to search for locally optimized solutions . of these
, 31 generated a better lse but all failed to meet the criteria mentioned above ( results not shown ) .
2 in the main text shows how the model reproduces the key dynamic behavior shown in the data .
we also show in the _ results _ section in the main text , that the model behavior is robust to perturbation of model parameters and that model predictions are reasonably consistent with a range of other experimental data , demonstrating the plausibility , if not uniqueness ( of course ) , of the parameter set .
we emphasize that , although the default parameter set is successful in reproducing a multitude of experimental observations ( e.g. full immune , knockout , re - infection ) as presented in the main text , we by no means claim that this parameter set is unique .
it remains an open and challenging problem to reliably identify a biologically plausible and statistically identifiable solution for what is a highly complex system , where we are severely limited by available experimental data . incorporating memory @xmath0 t cells into the model in the main text ,
we only make two changes .
the first is adding two equations to describe the memory @xmath0 t cell ( @xmath75 ) proliferation / differentiation , similar to eqs .
6 and 7 in the main text @xmath163 then we change the term @xmath77 in eq . 3 in the main text to @xmath78 .
hence , similar to the approach mentioned above that moving the delayed term from viral load to effector cells , we write down the model in an equivalent form , @xmath164 , \tag{s28}\\ \frac{df}{dt } & = p_fi-\delta_ff , \tag{s29}\\ \frac{dr}{dt } & = \phi ft-\rho r , \tag{s30}\\ \frac{dc_{n}}{dt } & = -\beta_{cn}(\frac{v}{v+h_c})c_{n } , \tag{s31}\\ \frac{de(t+\tau_c)}{dt } & = \beta_{cn}(\frac{v}{v+h_c})c_{n}e^{(p_c\tau_c ) } - \delta_ee(t+\tau_c ) , \tag{s32}\\ \frac{db_n}{dt } & = -\beta_{bn}(\frac{v}{v+h_b})b_{n } , \tag{s33 } \\
\frac{dp(t+\tau_b)}{dt } & = \beta_{bn}(\frac{v}{v+h_b})b_{n}e^{(p_b\tau_b ) } - \delta_pp(t+\tau_b ) , \tag{s34}\\ \frac{da_s(t+\tau_b)}{dt } & = p_sp(t+\tau_b)-\delta_sa_s(t+\tau_b ) , \tag{s35 } \\
\frac{da_l(t+\tau_b)}{dt } & = p_lp(t+\tau_b)-\delta_la_l(t+\tau_b ) .
\tag{s36}\\ \frac{dc_{m}}{dt } & = -\beta_{cm}(\frac{v}{v+h_{cm}})c_{m } , \tag{s37}\\ \frac{de_m(t+\tau_{cm})}{dt } & = \beta_{cm}(\frac{v}{v+h_{cm}})c_{m}e^{(p_{cm}\tau_{cm } ) } - \delta_ee_m(t+\tau_{cm } ) .
\tag{s38}\end{aligned}\ ] ] for variables whose independent variables are not explicitly specified , they are all functions of @xmath105 , i.e. @xmath2 reads @xmath109 .
memory @xmath0 t cells show a shorter delay and faster proliferation than naive @xmath0 t cells @xcite .
the shortened delay may be caused by a shortened lag time to the first division and/or a reduced delay for effector cells migrating from the lymphatic compartment to the lung @xcite .
the former reduction is about 15 hours @xcite and the latter is less than about 12 hours @xcite .
thus , we choose @xmath165 ( days ) , correspond to a one day reduction in the delay time compared to the delay of naive cells ( @xmath166 ) .
memory @xmath0 t cells show a higher division rate and a lower loss rate than naive @xmath0 t cells @xcite , based on which the net production rate of effector cells for memory @xmath0 t cells is estimated to be about 1.5 times of that for naive cells .
thus , we choose @xmath167 ( @xmath168 ) . in the absence of data , we assume @xmath169 and @xmath170 .
the initial number of memory @xmath0 t cells is varied as specified in the main text or figures .
we assume that the effector @xmath0 t cells produced by either naive or memory @xmath0 t cells are functionally identical ( i.e. then have the same decay rate @xmath40 and killing rate @xmath171 ) . note that we do not model the process of differentiation of effector @xmath0 t cells into memory cells but use a memory cell pool as an initial condition to simulate viral re - infection . ....
clear dt=0.1 ; % time step size time=0:dt:100 ; % parameters t0=7e+7;gt=0.8;pv=210;deltav=5.0 ; beta=5e-7;betap=3e-8 ; deltai=2;kappan=2.5;kappae=5e-5 ; phi=0.33;rho=2.6;pf=1e-5;deltaf=2 ; betacn=1;betabn=0.03 ; kappas=0.8;kappal=0.4 ; pc=1.2;pb=0.52;deltae=0.57;deltap=0.5 ; ps=12;pl=4;deltas=2;deltal=0.015 ; tauc=6;taub=4;hc=1e+4;hb=1e+4 ; betacm=1;pcm=1.8;taucm=5 ; % index indicating when delayed process starts indc = round(tauc / dt+1 ) ; % for tauc indcm = round(taucm / dt+1 ) ; % for taucm indb = round(taub / dt+1 ) ; % for taub % variable vectors and initial conditions v = zeros(1,length(time));v(1)=1e+1 ; t = v;t(1)=7e+7;i = t;i(1)=0;r = i;f = i ; cn=100*ones(1,length(time ) ) ; bn=100*ones(1,length(time ) ) ; e = zeros(1,indc+length(time ) ) ; p = i;as = zeros(1,indb+length(time));al = as ; cm=5000*ones(1,length(time ) ) ; em = zeros(1,indcm+length(time ) ) ; init=[v(1),t(1),i(1),r(1),f(1),cn(1),e(1),bn(1),p(1),as(1),al(1 ) , cm(1),em(1 ) ] ' ; options = odeset('reltol',1e-3,'abstol',1e-6 ) ; for i=2:length(time ) [ ~,y ] = ode15s(@odemodel_with_memory,[0 dt],init , options , e(i),al(i),as(i),em(i),tauc , taub , phi , rho , deltaf , gt , pf , pv , beta , betap , kappan , deltav , deltai , betacn , betabn , kappae , kappas , pl , ps , deltal , deltas , deltap , deltae , pc , pb , kappal , hc , hb , t0,betacm , pcm , taucm ) ; v(i)=y(end,1);t(i)=y(end,2);i(i)=y(end,3);r(i)=y(end,4 ) ; f(i)=y(end,5);cn(i)=y(end,6);bn(i)=y(end,8);p(i)=y(end,9 ) ; e(indc+i)=y(end,7 ) ; as(indb+i)=y(end,10);al(indb+i)=y(end,11 ) ; cm(i)=y(end,12);em(indcm+i)=y(end,13 ) ; init = y(end , : ) ' ; % initial condition for next iteration end .... .... function ynew = odemodel_new_with_memory(~,y , e , al , as , em , tauc , taub , phi , rho , deltaf , gt , pf , pv , beta , betap , kappan , deltav , deltai , betacn , betabn , kappae , kappas , pl , ps , deltal , deltas , deltap , deltae , pc , pb , kappal , hc , hb , t0,betacm , pcm , taucm ) % v : viral load % t : target cell % i : infected cell % r : resistant cell % f : ifn % cn : naive cd8 + t cells % e : effector cd8 + t cells % bn : naive b cells % p : plasma b cells % as : short - lived antibodies % al : long - lived antibodies % cm : memory cd8 + t cells % em : effector cd8 + t cells produced from memory cells % y=[v , t , i , r , f , cn , e , bn , p , as , al , cm , em ] ynew = zeros(13,1 ) ; ynew(1)=pv*y(3)-deltav*y(1)-kappas*y(1)*as - kappal*y(1)*al - beta*y(1)*y(2 ) ; ynew(2)=gt*(y(2)+y(4))*(1-(y(2)+y(3)+y(4))/t0)-betap*y(1)*y(2)+rho*y(4)-phi*y(2)*y(5 ) ; ynew(3)=betap*y(1)*y(2)-deltai*y(3)-kappan*y(3)*y(5)-kappae*y(3)*(e+em ) ; ynew(4)=phi*y(2)*y(5)-rho*y(4 ) ; ynew(5)=pf*y(3)-deltaf*y(5 ) ; ynew(6)=-betacn*y(1)./(y(1)+hc)*y(6 ) ; ynew(7)=betacn*y(1)./(y(1)+hc)*y(6)*exp(pc*tauc)-deltae*y(7 ) ; ynew(8)=-betabn*y(1)./(y(1)+hb)*y(8 ) ; ynew(9)=betabn*y(1)./(y(1)+hb)*y(8)*exp(pb*taub)-deltap*y(9 ) ; ynew(10)=ps*y(9)-deltas*y(10 ) ; ynew(11)=pl*y(9)-deltal*y(11 ) ; ynew(12)=-betacm*y(1)./(y(1)+hc)*y(12 ) ; ynew(13)=betacm*y(1)./(y(1)+hc)*y(12)*exp(pcm*taucm)-deltae*y(13 ) ; .... | myriad experiments have identified an important role for @xmath0 t cell response mechanisms in determining recovery from influenza a virus infection .
animal models of influenza infection further implicate multiple elements of the immune response in defining the dynamical characteristics of viral infection . to date , influenza virus models , while capturing particular aspects of the natural infection history , have been unable to reproduce the full gamut of observed viral kinetic behaviour in a single coherent framework . here , we introduce a mathematical model of influenza viral dynamics incorporating all major immune components ( innate , humoral and cellular ) and explore its properties with a particular emphasis on the role of cellular immunity . calibrated against a range of murine data ,
our model is capable of recapitulating observed viral kinetics from a multitude of experiments .
importantly , the model predicts a robust exponential relationship between the level of effector @xmath0 t cells and recovery time , whereby recovery time rapidly decreases to a fixed minimum recovery time with an increasing level of effector @xmath0 t cells .
we find support for this relationship in recent clinical data from influenza a(h7n9 ) hospitalised patients .
the exponential relationship implies that people with a lower level of naive @xmath0 t cells may receive significantly more benefit from induction of additional effector @xmath0 t cells arising from immunological memory , itself established through either previous viral infection or t cell - based vaccines . |
continuous - variable quantum - key distribution ( cvqkd ) , as an unconditionally secure communication scheme between two legitimate parties alice and bob , has achieved advanced improvements in theoretical analysis and experimental implementation in recent years @xcite .
practical implementation systems , such as fiber - based gaussian - modulated @xcite and discrete - modulated @xcite coherent - state protocol qkd systems over tens of kilometers , have been demonstrated in a few groups .
the unconditional security of such systems with prepare - and - measure ( pm ) implementation has been confirmed by the security analysis of the equivalent entanglement - based ( eb ) scheme @xcite .
however , the traditional security analysis of the eb scheme of cvqkd just includes the signal beam and not the local oscillator ( lo ) , which is an auxiliary light beam used as a reference to define the phase of the signal state and is necessary for balanced homodyne detection .
this will leave some security loopholes for eve because lo is also unfortunately within eve s manipulating domain .
the necessity of monitoring lo intensity for the security proofs in discrete qkd protocols embedded in continuous variables has been discussed @xcite .
moreover , in @xcite , the excess noise caused by imperfect subtraction of balanced homodyne detector ( bhd ) in the presence of lo intensity fluctuations has been noted and quantified with a formulation .
however , in the practical implementation of cvqkd , shot noise scaling with lo power measured before keys distribution is still assumed to keep constant if the fluctuations of lo intensity are small . and in this circumstance , pulses with large fluctuation are just discarded as shown in @xcite . unfortunately , this will give eve some advantages in exploiting the fluctuation of lo intensity . in this paper , we first describe bob s measurements under this fluctuation of lo intensity , and propose an attacking scheme exploiting this fluctuation .
we consider the security of practical cvqkd implementation under this attack and calculate the secret key rate with and without bob monitoring the lo for reverse and direct reconciliation protocol .
and then , we give a qualitative analysis about the effect of this lo intensity fluctuation on the secret key rate alice and bob hold .
we find that the fluctuation of lo could compromise the secret keys severely if bob does not scale his measurements with the instantaneous lo intensity values .
finally , we briefly discuss the accurate monitoring of lo intensity to confirm the security of the practical implementation of cvqkd .
generally , in practical systems of cvqkd , the local oscillator intensity is always monitored by splitting a small part with a beam splitter , and pulses with large lo intensity fluctuation are discarded too . however , even with such monitoring , we do not yet clearly understand how fluctuation , in particular small fluctuation , affects the secret key rate . to confirm that the secret key rate obtained by alice and bob is unconditionally secure , in what follows , we will analyze the effects of this fluctuation on the secret key rate only , and do not consider the imperfect measurement of bhd due to incomplete subtraction of it in the presence of lo intensity fluctuations , which has been discussed in @xcite . ideally , with a strong lo , a perfect pulsed bhd measuring a weak signal
whose encodings are @xmath0 will output the results@xcite , @xmath1 where _ k _ is a proportional constant of bhd , @xmath2 is the amplitude of lo , @xmath3 is the relative phase between the signal and lo except for the signal s initial modulation phase . so scaling with lo power or shot noise ,
the results can be recast as @xmath4 with @xmath3 in eq .
( [ eq : x0 ] ) is 0 or @xmath5 . here
the quadratures @xmath6 and @xmath7 are defined as @xmath8 and @xmath9 , where @xmath10 is the quadrature of the vacuum state .
however , in a practical system , the lo intensity fluctuates in time during key distribution . with a proportional coefficient @xmath11 ,
practical lo intensity can be described as @xmath12 , where @xmath2 is the initial amplitude of lo used by normalization and its value is calibrated before key distribution by alice and bob . if we do not monitor lo or quantify its fluctuation @xcite ,
especially just let the outputs of bhd scale with the initial intensity or power of lo , the outputs then read @xmath13 unfortunately , this fluctuation will open a loophole for eve , as we will see in the following sections . in conventional security analysis , like the eb scheme equivalent to the usual pm implementation depicted in fig .
[ fig:1](a ) , lo is not taken into consideration and its intensity is assumed to keep unchanged .
however , in practical implementation , eve could intercept not only the signal beam but also the lo , and she can replace the quantum channel between alice and bob with her own perfect quantum channel as shown in figs .
[ fig:1](b ) and [ fig:1](c ) . in so doing ,
eve s attack can be partially hidden by reducing the intensity of lo with a variable attenuator simulating the fluctuation without changing lo s phase , and such an attack can be called a lo intensity attack ( loia ) . in the following analysis
, we will see that , in the parameter - estimation procedure between alice and bob , channel excess noise introduced by eve can be reduced arbitrarily , even to its being null , just by tuning the lo transmission .
consequently , alice and bob would underestimate eve s intercepted information and eve could get partial secret keys that alice and bob hold without being found under this attack .
figure [ fig:1](b ) describes the loia , which consists of attacking the signal beam with a general gaussian collective attack @xcite and attacking the lo beam with an intensity attenuation by a non - changing phase attenuator * a * , such as a beam splitter whose transmission is variable .
this signal - beam gaussian collective attack consists of three steps : eve interacts her ancilla modes with the signal mode by a unitary operation * u * for each pulse and stores them in her quantum memory , then she makes an optimal collective measurement after alice and bob s classical communication . figure [ fig:1](c )
is one practical loia with * u * being a beam - splitter transformation .
its signal attack is also called an entangling cloner attack , which was presented first by grosshan @xcite and improved by weedbrook @xcite . in appendix
[ sec : security ] , we will demonstrate that with this entangling cloner , eve can get the same amount of information as that shown in fig .
[ fig:1](b ) .
we analyze a practical cvqkd system with homodyne protocol to demonstrate the effect of loia on the secret key rate , and for simplicity we do not give the results of the heterodyne protocol , which is analogous to the homodyne protocol . in the usual pm implementation , alice prepares a series of coherent states centered on @xmath0 with each pulse , and then she sends them to bob through a quantum channel which might be intercepted by eve . here ,
@xmath14 and @xmath15 , respectively , satisfy a gaussian distribution independently with the same variance @xmath16 and zero mean . this initial mode prepared by alice
can be described as @xmath17 , and @xmath18 is the quadrature variable . here
@xmath9 describes the quadrature of the vacuum mode .
note that we denote an operator with a hat , while without a hat the same variable corresponds to the classical variable after measurement .
so the overall variance of the initial mode prepared by alice is @xmath19 .
when this mode comes to bob , bob will get a mode @xmath20 , @xmath21 where @xmath22 describes eve s mode introduced through the quantum channel whose quadrature variance is @xmath23 .
bob randomly selects a quadrature to measure , and if eve attenuates the lo intensity during the key distribution , but bob s outputs still scale with the initial lo intensity , just as in eq .
( [ eq : x0 ] ) , he will get the measurement @xmath24 where @xmath25 is @xmath26 ( or equivalently @xmath27 ) .
however , if bob monitors lo and also scales with the instantaneous intensity value of lo with each pulse , he will get @xmath25 without any loss of course .
note that , for computation simplicity , hereafter we assume the variable transmission rate @xmath28 ( or attenuation rate @xmath29 ) of each pulse of lo is the same without loss of generality .
thus the variance of bob s measurements and conditional variance on alice s encodings with and without monitoring ( in what follows , without monitoring specially indicates that bob s measurement is obtained just by scaling with the initial lo intensity instead of monitoring instantaneous values , and vice versa ) can be given by @xmath30\;,\label{eq : vbm}\\ v_{b|a}&=t+(1-t)n\ ; , \\
v_{b|a}^w&=\eta\left[t+(1-t)n\right]\;,\end{aligned}\ ] ] where the superscript @xmath31 indicates without monitoring " and all variances are in shot - noise units ; the conditional variance is defined as @xcite @xmath32 hence , the covariance matrix of alice s and bob s modes can be obtained as @xmath33\mathbb{i } \end{pmatrix}\label{eq : rab},\end{split } \\ & \gamma_{ab}^w=\begin{pmatrix } v\mathbb{i } & \sqrt{\eta t(v^2 - 1)}\sigma_z\\ \sqrt{\eta t(v^2 - 1)}\sigma_z & \eta[tv+(1-t)n]\mathbb{i } \end{pmatrix},\end{aligned}\ ] ] where @xmath34 is the pauli matrix and @xmath35 is a unit matrix . from eqs .
( [ eq : vb ] ) and ( [ eq : vbm ] ) we can derive that the channel transmission and excess noise are @xmath36 , @xmath37 with monitoring , and @xmath38 , @xmath39 without monitoring .
hence , by attenuating the lo intensity as fig . [ fig:1 ] shows , to make @xmath40 , eve could arbitrarily reduce @xmath41 to zero , thus she will get the largest amount of information permitted by physics . in the following numerical simulation
we always make @xmath42 , namely , @xmath43 .
thus , the covariance matrix @xmath44 and eve s introducing noise @xmath23 [ in an entangling cloner it is eve s epr state s variance as fig .
[ fig:1](c ) shows ] should be selected to be @xmath45 to estimate the secret key rate , without loss of generality , we first analyze the reverse reconciliation then consider the direct reconciliation . from alice and
bob s points of view , the secret key rate for reverse reconciliation with monitoring or not are given , respectively , by @xmath46 where the mutual information between alice and bob with and without monitoring are the same , and that is @xmath47 this is because bob s measurements in these two cases are just different with a coefficient @xmath28 , and they correspond with each other one by one , so they are equivalent according to the data - processing theorem @xcite .
however , the mutual information between eve and bob given by the holevo bound @xcite in these two cases is not identical . as the previous analysis showed , in bob s point of view , channel transmission and the excess noise estimation are different .
but from eve s point of view , they are identical according to the data - processing theorem because she estimates bob s measurements in these two cases just by multiplying a coefficient @xmath28 .
we ll calculate the real information intercepted by eve first .
it can be given by @xmath48 where @xmath49 is the von neumann entropy @xcite .
for a gaussian state @xmath50 , this entropy can be calculated by the symplectic eigenvalues of the covariance matrix @xmath51 characterizing @xmath50 @xcite . to calculate eve s information
, first eve s system @xmath52 can purify @xmath53 permitted by quantum physics , so that @xmath54 .
second , after bob s projective measurement , the system @xmath55 is pure , so that @xmath56 .
designating @xmath57 , @xmath58 , and @xmath59 , the symplectic eigenvalues of @xmath60 are given by @xmath61 where @xmath62 and @xmath63 .
similarly , the entropy @xmath64 is determined by the symplectic eigenvalue @xmath65 of the covariance matrix @xmath66 @xcite , namely , @xmath67 where @xmath68 and @xmath69 stands for the moore - penrose inverse of a matrix .
then @xmath70 , and the holevo bound reads @xmath71 where @xmath72 . however , eve s information estimated by bob without monitoring is given by @xmath73 by substituting eqs .
( [ eq : xbe ] ) and ( [ eq : xbem ] ) into eqs .
( [ eq : krr ] ) and ( [ eq : krrm ] ) , respectively , the secret key rate with and without bob s monitoring can be obtained .
however , the secret key rate in eq .
( [ eq : krrm ] ) without monitoring is unsecured in evidence .
eve s interception of partial information from @xmath74 is not detected , in other words , alice and bob underestimate eve s information without realizing it .
actually , the real or unconditionally secure secret key rate @xmath74 , which we called a truly secret key rate , should be available by replacing @xmath75 in eq .
( [ eq : krrm ] ) with eq .
( [ eq : xbe ] ) .
note that it is identical with the monitoring secret key rate in eq .
( [ eq : krr ] ) due to eq .
( [ eq : iab ] ) .
we investigate the secret key rate @xmath74 bob measured without monitoring and the true one or equivalently monitoring one @xmath76 for reverse reconciliation under eve attacking the intensity of lo during key distribution . as fig .
[ fig:2 ] shows , with various values of transmission of lo that can be controlled by eve , the truly secret key rate alice and bob actually share decreases rapidly over long distances or small channel transmissions .
( color online ) reverse reconciliation pseudosecret key rate and the truly secret one vs channel transmission @xmath36 under loia .
solid lines are secret key rate estimated by bob without monitoring lo intensity and dashed lines are the truly secret ones .
colored lines correspond to the lo transmissions @xmath28 as labeled . here
alice s modulation variance @xmath77 .
] additionally , because the mutual information between alice and bob with and without monitoring is identical as eq.([eq : iab ] ) shows , subtracting eq .
( [ eq : krr ] ) from eq .
( [ eq : krrm ] ) we can estimate eve s intercepted information @xmath78 which is plotted in fig .
[ fig:3 ] .
we find that eve could get partial or full secret keys which alice and bob hold by controlling the different transmissions of lo . taking a 20-km transmission distance as an example , surprisingly , just with lo intensity fluctuation or attenuating rate 0.08 , eve is able to obtain the full secret keys for reverse reconciliation without bob s monitoring the lo .
we now calculate the secret key rate for direct reconciliation , which is a little more complicated , and we investigate the effect of lo intensity attack by eve on cvqkd .
the secret key rate estimated by bob with and without lo monitoring is given , respectively , by @xmath79 note that we have already calculated @xmath80 and @xmath81 in eq.([eq : iab ] ) and they are identical for direct and reverse reconciliation .
for eve we have @xmath82 where @xmath54 has been already computed in the previous section , and @xmath83 using the fact that after alice s projective measurement on modes @xmath84 and @xmath85 obtaining @xmath86 in the eb scheme shown in fig .
[ fig:1](a ) , the system @xmath87 is pure . to calculate @xmath88
, we have to compute the symplectic eigenvalues of covariance matrix @xmath89 , which is obtained by @xmath90 where @xmath91 and @xmath92 can be read in the decomposition of the matrix @xmath93 which is available by elementary transformation of the matrix [ see fig . [ fig:1](a ) ] @xcite @xmath94 where @xmath95 is a unit matrix .
it is obtained by applying a homodyne detection on mode a after mixing @xmath84 and @xmath85 with a balanced beam - splitter transformation @xmath96 .
the matrix @xmath97 and @xmath98 actually is @xmath60 in eq .
( [ eq : rab ] ) , @xmath99 is a unit matrix .
so we can get @xmath100 and the symplectic eigenvalues of it @xmath101 where @xmath102 and @xmath103 .
the holevo bound then reads @xmath104 substituting eqs .
( [ eq : xae ] ) and ( [ eq : xaem ] ) into eqs .
( [ eq : kdr ] ) and ( [ eq : kdrm ] ) , respectively , the secret key rates in these two cases are obtained . in fig .
[ fig:4 ] , we plotted them for channel transmission @xmath36 with various values of @xmath28 and find that the difference between the pseudosecret key rate with bob not monitoring lo and the truly secret one is still increasing with the channel transmission @xmath36 becoming smaller . for eve , when bob does not monitor lo , she will get the partial or total secret key rate @xmath105 without being found by subtracting eq .
( [ eq : kdr ] ) from eq .
( [ eq : kdrm ] ) when she reduces the intensity of lo . in fig
. [ fig:5 ] , we plotted the pseudosecret key rate for direct reconciliation and the mutual information overestimated by alice and bob .
we find that for short distance communication ( less 15 km or 3 db limit ) , a small fluctuation of lo intensity could still hide eve s attack partially or totally .
note that in the above estimation we assume each pulse s transmission rate @xmath28 ( or attenuation rate @xmath29 ) is identical .
however , when @xmath28 is different for each pulse ( eve simulates the fluctuation of lo to hide her dramatic attack on lo ) , eve still could intercept as much as or even more secret key rates than above for reverse and direct reconciliation , as long as the largest value of @xmath28 among all pulses ( or approximately most pulse transmission rates ) is smaller than the above constant value .
our analysis shows that reverse reconciliation is more sensitive than direct reconciliation about the fluctuation of lo intensity , and , even with a small attenuation of lo intensity , eve can get full secret keys but not be found .
this is consistent with the fact that channel excess noise has a more severe impact on reverse reconciliation than on direct reconciliation . of course , when the intensity of lo fluctuates above the initial calibrated value ( i.e. , @xmath106 ) , eve could not get any secret keys , but alice and bob would overestimate eve s intercepted information due to the overestimation of channel excess noise . however , when lo fluctuates around the initial calibrated value , how to quantify eve s information is still an open question , because the distribution of the fluctuation of lo ( or @xmath28 ) is not a normal distribution and unclear for alice and bob due to eve s arbitrary manipulation .
but in this circumstance , eve still could intercept partial secret keys if she increases the channel excess noise of one part of the signal pulses when she controls @xmath107 and decreases it for the other part when controlling @xmath106 , i.e. , making the overall estimated excess noise by alice and bob lower than the real one . remarkably , lo intensity fluctuation opens a loophole for eve to attack the practical system , especially in the case of communication with low channel transmission or over long distance .
consequently , in the practical implementation of cvqkd , we must monitor the lo fluctuation carefully and in particular scale the measurements with instantaneous intensity values of lo .
alternatively , we can also scale with the lowest intensity value of lo if the fluctuations are very small , but it will estimate the secret key rate pessimistically thus leading to the reduction of the efficiency of the key distribution .
however , we can not use the average intensity value of lo to normalize the measurements as most current implementations do , because it still could overestimate the secret key rate for alice and bob .
additionally , for reverse reconciliation communication over long distance , very small fluctuation of lo might compromise the secret key rate completely , which presents a big challenge for accurately monitoring lo intensity .
finally , we point out that in this paper we do not consider the imperfections of bhd such as detection efficiency , electronic noise , and incomplete subtraction , which may make lo intensity fluctuation have a more severe impact on estimating the secret key rate for alice and bob . in conclusion
, we have analyzed the effect of lo intensity fluctuation on the secret key rate estimation of alice and bob for reverse and direct reconciliation .
incredibly , bob s estimation of the secret key rate will be compromised severely without monitoring lo or if his measurements do not scale with lo instantaneous intensity values even with monitoring but just discard large fluctuation pulses like in @xcite .
furthermore , we have shown that eve could hide her attack partially by reducing the intensity of lo and even could steal the total secret keys alice and bob share without being found by a small attenuation of lo intensity , especially for reverse reconciliation .
finally , we have also briefly discussed the monitoring of lo and pointed out that it would be a challenge for highly accurate monitoring .
this work is supported by the national natural science foundation of china , grant no .
is supported by the program for new century excellent talents .
c.m . is supported by the hunan provincial innovation foundation for postgraduates .
acknowledge support from nudt under grant no . kxk130201 .
in this appendix , we calculate the holevo bound obtained by eve for direct and reverse reconciliation using weedbrook s entangling cloner model @xcite , and then give the secret key rate shared by alice and bob under loia .
we begin the analysis by calculating the von neumann entropy of eve s intercepting state first . as fig .
[ fig:1](c ) shows , the entangling cloner consists of eve replacing the gaussian quantum channel between alice and bob with a beam splitter of transmission @xmath36 and an epr pair of variance @xmath23 .
half of the epr pair mode @xmath108 is mixed with alice s mode in the beam splitter and is sent to bob to match the noise of the real channel by tuning n. the other half mode @xmath109 is kept by eve to reduce the uncertainty on one output of the beam splitter , the mode @xmath110 , which can be read as @xmath111 where @xmath112 is the quadrature of mode @xmath108 .
thus , the variance of mode @xmath110 is given by @xmath113 and the conditional variance @xmath114 can be calculated as , using eq .
( [ eq : cvar ] ) , @xmath115 hence , eve s covariance matrix can be obtained as @xmath116 where @xmath117 and the notation @xmath118 stands for a matrix with the arguments on the diagonal elements and zeros everywhere else .
the symplectic eigenvalues of this covariance matrix are given by @xmath119 where @xmath120 , and @xmath121 .
hence , the von neumann entropy of eve s state is given by @xmath122 for the direct reconciliation protocol of cvqkd , the holevo bound between eve and alice is given by eq .
( [ eq : xae0 ] ) , where @xmath123 has been calculated by eq .
( [ eq : se ] ) .
@xmath124 can be obtained by the conditional covariance matrix @xmath125 and its symplectic eigenvalues are given by @xmath126 where @xmath127 , and @xmath128 .
thus , the conditional entropy is @xmath129 substituting eqs .
( [ eq : se ] ) and ( [ eq : sea ] ) into eq .
( [ eq : xae0 ] ) , we can get the mutual information between alice and eve , @xmath130 under loia , bob s estimation of the holevo bound without monitoring lo intensity then reads , using eq .
( [ eq : xae1 ] ) , @xmath131 with eqs .
( [ eq : xae1 ] ) and ( [ eq : xaem1 ] ) , the secret key rates in eqs .
( [ eq : kdr ] ) and ( [ eq : kdrm ] ) then can be calculated respectively , and the calculation numerically demonstrates that they are perfectly consistent with the fig . [ fig:4 ] .
the calculation of the holevo bound between eve and bob for reverse reconciliation is a bit more complicated . using eq .
( [ eq : xbe0 ] ) , we only need to calculate the conditional entropy @xmath132 , which is determined by the symplectic eigenvalues @xmath133 of the covariance matrix @xmath134 , @xmath135 where @xmath136 and @xmath137 , @xmath138 .
then , @xmath134 can be recast as @xmath139 where + @xmath140 , @xmath141 , + and + @xmath142 .
+ hence , its symplectic eigenvalues are given by @xmath143 where @xmath144 , @xmath145 , and @xmath146 is the determinant of a matrix .
so , we get the conditional entropy @xmath147 and then the holevo bound @xmath148 consequently , without monitoring lo intensity , alice and bob will give eve the holevo bound @xmath149 substituting eqs .
( [ eq : xbe1 ] ) and ( [ eq : xbem1 ] ) into eqs .
( [ eq : krr ] ) and ( [ eq : krrm ] ) , respectively , the secret key rates with and without bob s monitoring can be obtained , and for channel transmission @xmath36 with various values of @xmath28 , they are numerically demonstrated to be perfectly consistent with fig .
[ fig:2 ] , too .
hence , it also indirectly confirms that either for direct or reverse reconciliation , the entangling cloner could reach the holevo bound against the optimal gaussian collective attack . in this paper ,
lo intensity fluctuation indicates the deviation of each pulse s intensity from the initial calibrated value during the key distribution .
it does not mean the quantum fluctuation of each pulse itself , because lo is a strong classical beam whose quantum fluctuation is very small relative to itself and can be neglected . | we consider the security of practical continuous - variable quantum key distribution implementation with the local oscillator ( lo ) fluctuating in time , which opens a loophole for eve to intercept the secret key .
we show that eve can simulate this fluctuation to hide her gaussian collective attack by reducing the intensity of the lo .
numerical simulations demonstrate that , if bob does not monitor the lo intensity and does not scale his measurements with the instantaneous intensity values of lo , the secret key rate will be compromised severely . |
it is by now standard to parameterize transverse momentum distributions with functions having a power law behaviour at high momenta .
this has been done by the star @xcite and phenix @xcite collaborations at rhic and by the alice @xcite , atlas @xcite and cms @xcite collaborations at the lhc . in this talk
we would like to pursue the use of the tsallis distribution to describe transverse momentum distributions at the highest beam energies .
+ in the framework of tsallis statistics @xcite the entropy @xmath1 , the particle number , @xmath2 , the energy density @xmath3 and the pressure @xmath4 are given by corresponding integrals over the tsallis distribution : @xmath5^{-\frac{1}{q-1 } } .\label{tsallis}\ ] ] it can be shown ( see e.g. @xcite ) that the relevant thermodynamic quantities are given by : @xmath6 , \label{entropy } \\ n & = & gv\int\frac{d^3p}{(2\pi)^3 } f^q , \label{number } \\
\epsilon & = & g\int\frac{d^3p}{(2\pi)^3}e f^q , \label{epsilon}\\ p & = & g\int\frac{d^3p}{(2\pi)^3}\frac{p^2}{3e } f^q\label{pressure } .\end{aligned}\ ] ] where @xmath7 and @xmath8 are the temperature and the chemical potential , @xmath9 is the volume and @xmath10 is the degeneracy factor .
we have used the short - hand notation @xmath11 often referred to as q - logarithm .
it is straightforward to show that the relation @xmath12 ( where @xmath13 refer to the densities of the corresponding quantities ) is satisfied .
the first law of thermodynamics gives rise to the following differential relations : @xmath14 since these are total differentials , thermodynamic consistency requires the following maxwell relations to be satisfied : @xmath15 this is indeed the case , e.g. for eq .
this follows from @xmath16^{-\frac{q}{q-1 } } \nonumber \\ & = & - g\int\frac{d^3p}{(2\pi)^3}\frac{p^2}{3 } \frac{d}{pdp}\left [ 1 + ( q-1 ) \frac{e-\mu}{t}\right]^{-\frac{q}{q-1 } } \nonumber \\ & = & g\int\frac{d\cos\theta d\phi dp}{(2\pi)^3 } \left [ 1 + ( q-1 ) \frac{e-\mu}{t}\right]^{-\frac{q}{q-1 } } \frac{d}{dp}\frac{p^3}{3 } \nonumber \\ & = & n \nonumber\end{aligned}\ ] ] after an integration by parts and using @xmath17 .
+ following from eq . , the momentum distribution is given by : @xmath18^{-q/(q-1 ) } , \label{tsallismu}\ ] ] or , expressed in terms of transverse momentum , @xmath19 , the transverse mass , @xmath20 , and the rapidity @xmath21 @xmath22^{-q/(q-1 ) } .
\label{tsallismu1}\ ] ] at mid - rapidity , @xmath23 , and for zero chemical potential , as is relevant at the lhc , this reduces to @xmath24^{-q/(q-1)}. \label{tsallisfit1}\ ] ] in the limit where the parameter @xmath0 goes to 1 it is well - known that this reduces to the standard boltzmann distribution : @xmath25 the parameterization given in eq .
is close to the one used by various collaborations @xcite : @xmath26^{-n } , \label{alice}\ ] ] where @xmath27 and @xmath28 are fit parameters .
this corresponds to substituting @xcite @xmath29 and @xmath30 after this substitution eq .
becomes @xmath31^{-q/(q-1)}\nonumber\\ & & \left [ 1 + ( q-1)\frac{m_t}{t } \right]^{-q/(q-1 ) } .
\label{alice2}\end{aligned}\ ] ] at mid - rapidity @xmath32 and zero chemical potential , this has the same dependence on the transverse momentum as eq .
apart from an additional factor @xmath33 on the right - hand side of eq . .
however , the inclusion of the rest mass in the substitution eq . is not in agreement with the tsallis distribution as it breaks @xmath33 scaling which is present in eq . but not in eq . .
the inclusion of the factor @xmath33 leads to a more consistent interpretation of the variables @xmath0 and @xmath7 . +
a very good description of transverse momenta distributions at rhic has been obtained in refs @xcite on the basis of a coalescence model where the tsallis distribution is used for quarks .
tsallis fits have also been considered in ref . @xcite but with a different power law leading to smaller values of the tsallis parameter @xmath0 .
+ interesting results were obtained in refs .
@xcite where spectra for identified particles were analyzed and the resulting values for the parameters @xmath0 and @xmath7 were considered . +
the transverse momentum distributions of identified particles , as obtained by the alice collaboration at 900 gev in @xmath34 collisions , are shown in figure fig : positive .
the fit for positive pions was made using @xmath35^{-q/(q-1)}. \label{tsallisfitpi}\ ] ] with @xmath0 , @xmath7 and @xmath9 as free parameters .
+ + in figure strange we show fits to the transverse momentum distributions of strange particles obtained by the alice collaboration @xcite in @xmath34 collisions at 900 gev .
+ similarly we show fits to the transverse momentum distributions obtained by the cms collaboration @xcite in figure cms and by the atlas collaboration in figure chargedatlas .
+ the transverse momentum distributions of charged particles were fitted using a sum of three tsallis distributions , the first one for @xmath36 , the second one for @xmath37 and the third one for protons @xmath38 .
the relative weights between these were determined by the corresponding degeneracy factors , i.e. 1 for for @xmath36 and @xmath37 and 2 for protons .
the fit was taken at mid - rapidity and for @xmath39 using the following expression was used @xmath40^{-\frac{q}{q-1}},\ ] ] where @xmath41 and @xmath42 , @xmath43 and @xmath44 .
the factor @xmath45 in front of the right hand side of this equation takes into account the contributions of the antiparticles @xmath46 .
the tsallis distribution also describes the transverse momentum distributions of charged particles in @xmath47 collisions in all pseudorapidity intervals as shown in figure ppb .
+ collisions obtained by the alice collaboration @xcite using the tsallis distribution.,height=377 ] collisions obtained by the alice collaboration @xcite using the tsallis distribution.,height=377 ] obtained from fits to transverse momentum spectra described in the text.,height=377 ] obtained from fits to transverse momentum spectra described in the text.,height=377 ]
the tsallis distribution described here in eq . leads to excellent fits to the transverse momentum distributions in high energy @xmath34 and @xmath47 collisions .
the values obtained for the tsallis parameter @xmath0 are truly remarkably consistent , a feature which does not become apparent when using the parametrization of eq . .
90 b. i. abelev et al .
( star collaboration ) , phys .
c * 75 * , 064901 ( 2007 ) .
a. adare et al .
( phenix collaboration ) , phys .
c * 83 * , 052004 , ( 2010 ) ; phys . rev .
c * 83 * , 064903 ( 2011 ) .
alice collaboration , eur .
j. c * 71 * 1594 ( 2011 ) ; eur .
j. c * 71 * 1655 ( 2011 ) ; phys .
b * 693 * ( 2010 ) 53 ; phys .
( 2013 ) 082302 .
atlas collaboration , new j. phys .
* 13 * ( 2011 ) 053033 .
cms collaboration , phys .
* 105 * ( 2010 ) 022002 ; eur .
j. c * 72 * ( 2012 ) 2164 .
c. tsallis , j. statist . phys .
* 52 * , 479 ( 1988 ) .
t.bir , g. purcsel , k. rmssy , eur .
j. a * 40 * ( 2009 ) 325 .
j. m. conroy , h. g. miller , a. r. plastino , phys .
a * 374 * , 4581 ( 2010 ) .
j. cleymans and d. worku , j. phys .
g * 39 * ( 2012 ) 025006 .
j. cleymans and d. worku , eur .
j. a * 48 * ( 2012 ) 160 .
k. rmssy , t.s .
bir , phys .
b * 689 * ( 2010 ) 14 .
k. rmssy , t.s .
bir , j. phys .
g * 36 * ( 2009 ) 064044 .
cheuk - yin wong , g. wilk , acta physica polonica , 43 ( 2012 ) 2047 .
t. wibig , j. phys .
g : nucl . part .
phys . * 37 * 115009 ( 2010 ) .
t. wibig , i. kurp , jhep * 0312 * 039 ( 2003 ) .
l. marques , e.andrade-ii , a. deppman , arxiv:1210.1725[hep - ph ] i. sena , a. deppman , eur .
j. a 49 ( 2013 ) 17 ; arxiv:1209.2367[hep - ph ] k. rmssy , arxiv:1212.0260[hep - ph ] .
j. cleymans , g.i .
lykasov , a.s .
parvan , a.s .
sorin , o.v .
teryaev , d. worku phys .
b * 723 * ( 2013 ) 351 .
[ arxiv:1104.0620 [ hep - ph ] ] . | an overview is presented of transverse momentum distributions of particles at the lhc using the tsallis distribution .
the use of a thermodynamically consistent form of this distribution leads to an excellent description of charged and identified particles .
the values of the tsallis parameter @xmath0 are truly remarkably consistent .
= by -1 |
magnetic superconductivity has attracted much research attention since it was reported in the strongly - correlated rusr@xmath13rcu@xmath13o@xmath14 ru-1212 cuprate system ( r = sm , eu , gd , y ) with the tetragonal space group p4/mbm .
@xcite the ru magnetic moments order weak - ferromagnetically ( wfm ) with ordering temperature t@xmath15 130 k. high - t@xmath16 superconductivity occurs in the quasi-2d cuo@xmath13 bi - layers from doped holes with maximum superconducting transition onset t@xmath16(onset ) @xmath6 60 k for r = gd and coexists with the wfm order . a crossover from anisotropic 2d - like to less - anisotropic 3d - like structure
was observed near r = sm , along with a metal - insulator transition .
no superconductivity can be detected for the mott insulators r = pr and nd .
since the oxygen content for all samples is close to eight after oxygen annealing , the variation of t@xmath16 with rare - earth ions indicates a self - doping of electrons from cuo@xmath13 layers to ruo@xmath17 layers .
such self - doping creates hole carriers in cuo@xmath13 layers and conduction electrons in ruo@xmath17 layers . the ru l@xmath18-edge x - ray absorption near - edge spectrum ( xanes ) of rusr@xmath13gdcu@xmath13o@xmath14 indicates that ru valence is basically at ru@xmath19 ( 4d - t@xmath20 , s = 3/2 ) state with small amount ( @xmath620 @xmath21 ) of ru@xmath22 ( 4d - t@xmath23 , s = 1 in low spin state ) which corresponds to doped electrons . @xcite the strong antiferromagnetic superexchange interaction between ru@xmath19 moments is responsible for the basic g - type antiferromagnetic order observed in the neutron diffraction study .
@xcite the weak ferromagnetic component observed from magnetic susceptibility and nmr spectrum is probably due to weak - ferromagnetic double - exchange interaction through doped conduction electrons in the metallic ruo@xmath17 layers .
since the magnetic superexchange and double - exchange interaction is anisotropic in general , the study of anisotropic physical properties is crucial for this quasi-2d system . in this report
, we align the microcrystalline rusr@xmath13rcu@xmath13o@xmath14 ( r = rare earths ) powder ( @xmath61 - 10 @xmath7 m ) in magnetic field to investigate the anisotropic properties .
the stoichiometric rusr@xmath13gdcu@xmath13o@xmath14 bulk sample was synthesized by the standard solid - state reactions .
high - purity ruo@xmath13 ( 99.99 @xmath21 ) , srco@xmath18 ( 99.9 @xmath21 ) , gd@xmath13o@xmath18 ( 99.99 @xmath21 ) and cuo ( 99.99 @xmath21 ) preheated powders with the nominal composition ratio of ru : sr : gd : cu = 1:2:1:2 were well mixed and calcined at 960@xmath24c in air for 16 hours .
the calcined powders were then pressed into pellets and sintered in flowing n@xmath13 gas at 1015@xmath24c for 10 hours to form sr@xmath13gdruo@xmath17 and cu@xmath13o precursors .
the sintered pellets were then heated at 1060 - 1065@xmath24c in flowing o@xmath13 gas for 7 days to form the ru-1212 phase and slowly furnace cooled to room temperature with a rate of 15@xmath24c per hour . for powder alignment in magnetic field
, samples were ground into powders with an average microcrystalline grain size of 1 - 10 @xmath7 m and mixed with epoxy ( 4-hour curing time ) in a quartz tube ( @xmath25 = 8 mm ) with the ratio of powder : epoxy = 1:5 then immediately put into the alignment field environments ( simple field or rotation - field alignment ) .
for simple powder alignment , the mixture was placed in a 14-t superconducting magnet at room temperature in flowing n@xmath13 gas and slowly hardened overnight as shown in figure 1 .
the powder x - ray diffraction pattern of three typical aligned powder - in - epoxy samples rusr@xmath13rcu@xmath13o@xmath14 ( r = sm , eu , gd@xmath2dy@xmath2 ) are shown collectively in figure 2 . for r = sm
( as well as for r = pr and nd ) , no magnetic alignment can be achieved .
the lack of magnetic anisotropy may closely relate to the variation of tetragonal lattice parameters where @xmath5/3 @xmath26/@xmath27 for r = sm with @xmath28 = 0.5448 nm and @xmath5 = 1.1560 nm ( space group p4/mbm ) as shown in figure 3 . for r = eu ( as well as for r = gd ) , partial ( @xmath6 90@xmath21 ) @xmath29-plane aligned along alignment magnetic field b@xmath8 is observed through the appearance of enhanced ( @xmath300 ) diffraction lines .
a small amount of srruo@xmath18 impurity is presented .
the @xmath29-plane alignment may be due to the fact that @xmath5/3 @xmath31/@xmath27 for r = eu ( @xmath28 = 0.5435 nm , @xmath5 = 1.1572 nm ) . for metastable compound r = gd@xmath2dy@xmath2 near the phase boundary with some unreacted precursor sr@xmath13rruo@xmath17 ,
partially @xmath5-axis alignment along b@xmath8 is detected with enhanced ( 00@xmath32 ) lines due to @xmath5/3 @xmath33/@xmath27 in this compound ( @xmath28 = 0.5426 nm , @xmath5 = 1.1508 nm ) .
schematic diagram for the magnetic field powder alignment method in a 14 t superconducting magnet at 300 k. ] powder x - ray diffraction patterns for rusr@xmath13rcu@xmath13o@xmath14 aligned powder .
( a ) r = sm , ( b ) r = eu , ( c ) r = gd@xmath2dy@xmath2 . ]
the variation of superconducting transition t@xmath16 and tetragonal lattice parameters @xmath28 , @xmath5 with rare earth ionic radius r@xmath34 for rusr@xmath13rcu@xmath13o@xmath14 system ( r = pr - dy ) . ]
the phase diagram in figure 3 indicates a structural crossover from less - anisotropic 3d - like ( @xmath5/3 @xmath26 ) to anisotropic 2d - like structure ( @xmath5/3 @xmath35 @xmath28/@xmath27 ) near r = sm , along with an insulator - to - metal transition .
superconductivity appears only in the quasi-2d metallic region with resistivity onset transition temperature t@xmath36 0 for r = sm , t@xmath16 = 36 k for r = eu , t@xmath16 = 56 k for gd , and t@xmath16 = 55 k for metastable r = gd@xmath2dy@xmath2 . for r = eu with @xmath29-plane aligned along b@xmath8
, @xmath5-axis can be in any direction within the plane perpendicular to b@xmath8 . to obtain the @xmath5-axis aligned powder ,
a field - rotation alignment method is used as shown in figure 4 . since @xmath29-plane is fixed along b@xmath8 , the rotation of quartz tube ( 10 rpm ) perpendicular to b@xmath8 forces the microcrystalline @xmath5-axis to have no choice but to be aligned along the rotation axis .
the powder x - ray diffraction patterns of rusr@xmath13eucu@xmath13o@xmath14 random powder ( @xmath37 ) , partially @xmath29-plane aligned along b@xmath8 ( @xmath300 ) , and highly @xmath5-axis aligned along the rotation axis ( 00@xmath32 ) are shown collectively in figure 5 .
the relative intensity of the enhanced ( 00@xmath32 ) lines and ( 113 ) line indicates a @xmath6 90@xmath21 c - axis alignment in this field - rotation aligned powder - in - epoxy sample .
schematic diagram for the field - rotation powder alignment method with @xmath5-axis perpendicular to aligned magnetic field and along the rotation axis . ]
powder x - ray diffraction patterns for rusr@xmath13eucu@xmath13o@xmath14 .
( a ) random powder , ( b ) @xmath29-plane aligned along b@xmath8 , and ( c ) @xmath5-axis aligned along the rotation axis . ]
figure 6 shows the field dependence of paramagnetic moment of rusr@xmath13gdcu@xmath13o@xmath14 aligned powder up to 7 t at 300 k. since @xmath29-plane is aligned along the magnetic field in the alignment procedure at room temperature , magnetic anisotropy of @xmath38 at 300 k is expected . however , 300 k m - b@xmath39 data showed weak magnetic anisotropy with linear paramagnetic magnetic moment m@xmath40 0.95 m@xmath16 or susceptibility @xmath41 = m / b@xmath8 with @xmath42 .
the anisotropic temperature dependence of logarithmic molar magnetic susceptibility of rusr@xmath0gdcu@xmath0o@xmath1 c - axis aligned powder in 1-t applied magnetic field is shown in figure 7 . a crossover from @xmath42 at 300 k to @xmath43 at lower temperature
was observed around 185 k , followed by a weak - ferromagnetic ordering at t@xmath44(ru ) = 131 k. the field dependence of paramagnetic moment of rusr@xmath13gdcu@xmath13o@xmath14 aligned powder up to 7 t at 300 k. linear paramagnetic magnetic moment . ]
temperature - dependence of logarithmic molar magnetic susceptibility @xmath45 and @xmath46 . ]
the magnetic anisotropy of @xmath42 observed at 300 k is mainly due to the contribution of magnetic gd@xmath34 ions ( j = 7/2 ) .
the anisotropy of @xmath45(gd ) @xmath47 @xmath46(gd ) is from the tetragonal gdo@xmath14 cage with anisotropic @xmath48-factor @xmath49 , but with little 4@xmath50 wavefunction overlap with the neighbor oxygen 2@xmath51 orbital . although there are three types of magnetic moments in this magnetic superconductor : ru@xmath19 ( s = 3/2 ) with doped electrons or ru@xmath22 ( s = 1 ) , cu@xmath52 ( s = 1/2 ) with doped holes , and gd@xmath34 moment ( j = 7/2 ) , not all moments have the same contribution in powder alignment . in the aligned magnetic field ,
anisotropic orbital wavefunction is tied to the spin direction , and a strong spin - orbital related short - range anisotropic exchange interaction at 300 k should dominate the magnetic alignment . in the present case , it is believed that ru moment with the strong short - range anisotropic double - exchange / superexchange interaction along the @xmath29-plane due to the jahn - teller distortion of ruo@xmath17 octahedron with @xmath45(ru ) @xmath53 @xmath46(ru ) is the dominant factor for @xmath29-plane alignment along b@xmath8 at 300 k. the shorter ru - o(1 ) bond length in the tetragonal @xmath29-basal plane provides strong 4d@xmath54(ru)-2p@xmath55(o(1))-4d@xmath54(ru ) wavefunction overlap .
this exchange interaction increases with decreasing temperature , and eventually total @xmath43 was observed below 185 k as expected .
the weak - ferromagnetic state below 131 k is due to the long range order of this anisotropic double - exchange / superexchange interaction .
reciprocal molar magnetic susceptibility 1/@xmath45 and 1/@xmath46 for aligned rusr@xmath13gdcu@xmath13o@xmath14 powder . ] the reciprocal molar magnetic susceptibility 1/@xmath45 and 1/@xmath46 of rusr@xmath13gdcu@xmath13o@xmath14 aligned powder are shown in figure 8 .
a curie - weiss behavior @xmath41= c/(t - @xmath56 ) was observed in the high temperature paramagnetic region above 200 k with a curie - weiss intercept @xmath56 = 60 k for both field orientations and the anisotropic effective magnetic moment @xmath57 = 7.44 @xmath58 per formula unit along the @xmath5-axis , and @xmath59 = 7.26 @xmath58 per formula unit along the @xmath29-plane .
diamagnetic superconducting transition t@xmath16(dia ) at 39 k and gd ordering temperature t@xmath44(gd ) = 2.5 k are not clearly seen in this high applied field .
the temperature - dependence of logarithmic molar magnetic susceptibility @xmath45 and @xmath46 of aligned rusr@xmath13eucu@xmath13o@xmath14 powder in 1 t applied magnetic field . ]
the temperature dependence of logarithmic molar magnetic susceptibility of aligned rusr@xmath0eucu@xmath0o@xmath1 powder along @xmath5-axis and @xmath29-plane in 1-t applied magnetic field are shown in figure 9 .
similar to r = gd compound , weak paramagnetic anisotropy of @xmath45 = 0.95 @xmath46 was observed at 300 k , and a crossover to @xmath60 was detected below 190 k with a weak - ferromagnetic ordering temperature t@xmath44(ru ) = 133 k. superconducting diamagnetic signal t@xmath16(dia ) at 21 k ( resistivity zero point ) is very weak in large 1-t applied field .
low temperature , low field ( 1-g field - cooled ( fc ) and zero - field - cooled ( zfc ) ) anisotropic magnetic and superconducting properties of rusr@xmath13gdcu@xmath13o@xmath14 aligned powder . ] low temperature superconducting properties of aligned powder ( dispersed microcrystallines in epoxy ) , random powder and bulk rusr@xmath13gdcu@xmath13o@xmath14 sample . ] low temperature , low field ( 1-g field - cooled ( fc ) and zero - field - cooled ( zfc ) ) anisotropic magnetic and superconducting properties of rusr@xmath13gdcu@xmath13o@xmath14 aligned powder are shown in figure 10 . a clear weak - ferromagnetic ordering t@xmath44(ru ) at 131 k
was observed in all data with @xmath43 in the weak - ferromagnetic state for both fc and zfc measurements .
a superconducting diamagnetism setting in at the vortex melting temperature t@xmath4(dia ) of 39 k , same as the t@xmath4(dia ) in the bulk sample , can be clearly seen in the zfc measurements but with a weaker diamagnetic signal . in the fc measurements ,
the onset of diamagnetic signal can be recognized as a kink at t@xmath4(dia ) , the abrupt increase of magnetic signal at 30 k was attributed to the magnetic field profile inside the squid magnetometer @xcite and corresponds to the spontaneous vortex state temperature t@xmath61 , the same as the one observed in bulk samples.@xcite the antiferromagnetic gd@xmath34 order is observed at t@xmath44(gd ) = 2.5 k. the electrical resistivity data of bulk sample in figure 11 indicates a high superconducting onset temperature of 56 k , with a much lower t@xmath4(zero ) = t@xmath16(dia ) at the vortex melting temperature of 39 k. slightly larger diamagnetic signal for random powder is probably due to partially intergrain supercurrent shielding through partial grain contact .
the weak diamagnetic signal observed in aligned powder - in - epoxy samples below t@xmath16(dia ) is due to pure intragain shielding with , in addition , long penetration depth @xmath62 ( @xmath11 0.55 @xmath7 m , @xmath12 0.66 @xmath7 m ) in comparison with powder grain size ( @xmath6 1 - 10 @xmath7 m ) and the two - dimensional ( 2d ) character of cuo@xmath0 layers . the anisotropic high - field ( @xmath637 t ) isothermal superconducting hysteresis loops m - b@xmath8 at 100 k ( fig .
12 ) indicate the initial m@xmath64 m@xmath16 as expected from the susceptibility .
since the magnetization curves for both field orientations showed weak - ferromagnetic behavior , the weak - ferromagnetic ordered m(ru ) dominates over the paramagnetic m(gd ) in the magnetic response and should be responsible for the complex metamagnetic - like behavior around 1 t for field applied along the @xmath29-plane .
anisotropic high - field hysteresis loop m@xmath16(b@xmath8 ) and m@xmath65(b@xmath8 ) at 100 k for aligned rusr@xmath13gdcu@xmath13o@xmath14 sample . ] anisotropic high - field hysteresis loop m@xmath16(b@xmath8 ) and m@xmath65(b@xmath8 ) at 10 k for aligned rusr@xmath13gdcu@xmath13o@xmath14 sample . ]
the magnetization curves with applied field b@xmath8 along @xmath5-axis , m@xmath16(b@xmath8,t ) , and @xmath29-plane , m@xmath65(b@xmath8,t ) , for aligned rusr@xmath13gdcu@xmath13o@xmath14 powder at 10 k are shown in fig . 13 .
a strongly anisotropic magnetization was observed for different field orientations .
a superconducting hysteresis loop with a weak paramagnetic background was observed in b@xmath39 @xmath66 @xmath5-axis with a maximum diamagnetic signal at magnetization peak field @xmath67h@xmath68(peak ) of 11 g. however , no superconducting diamagnetic signal was directly observed for the initial magnetization curve for field applied along @xmath29-plane due to the much stronger paramagnetic - like background . after subtracting the paramagnetic - like background ,
the maximum diamagnetic signal for m@xmath65 can be obtained as @xmath67h@xmath69(peak ) = 9 g. the great difference of magnetization anisotropy between 100 k and 10 k observed can not be explained by simple superposition of magnetic and superconducting components that suggests complicated interplay between the doped electrons in weak - ferromagnetically ordered ruo@xmath70 layers and superconducting holes in cuo@xmath0-layers .
anisotropic powder alignment is achieved for rusr@xmath13rcu@xmath13o@xmath14 weak - ferromagnetic superconductors ( r = eu , gd , and gd@xmath2dy@xmath2 ) . due to spin - orbital related short - range anisotropic exchange interaction , paramagnetic susceptibility @xmath45(ru / cu )
@xmath71(ru / cu ) at 300 k in rusr@xmath13gdcu@xmath13o@xmath14 and rusr@xmath13eucu@xmath13o@xmath14 , @xmath5-axis aligned powder can be achieved only using field - rotational method .
total @xmath42 at room temperature is dominated by r@xmath34 .
due to long - range ru anisotropic exchange interaction , total @xmath43 were observed below crossover temperature @xmath6 185 k , with weak - ferromagnetic t@xmath44(ru ) = 131 k , superconducting t@xmath16(dia ) = 39 k and t@xmath44(gd ) = 2.5 k. weak diamagnetic signal observed below t@xmath16(dia ) was due to pure intragain shielding with long penetration depth @xmath62 ( @xmath11 0.55 @xmath7 m , @xmath12 0.66 @xmath7 m ) and the two - dimensional ( 2d ) character of cuo@xmath0 layers . | the powder alignment method is used to investigate the anisotropic physical properties of the weak - ferromagnetic superconductor system rusr@xmath0rcu@xmath0o@xmath1 ( r = pr , nd , sm , eu , gd , gd@xmath2dy@xmath2 ) . the rusr@xmath0gdcu@xmath0o@xmath1 ru-1212 cuprate is a weak - ferromagnetic superconductor with a magnetic ordering of ru moments at t@xmath3(ru ) = 131 k , a superconducting transition in the cuo@xmath0 layers at t@xmath4 = 56 k , and a low temperature gd antiferromagnetic ordering at t@xmath3(gd ) = 2.5 k. due to weak magnetic anisotropy of this tetragonal system , highly @xmath5-axis aligned microcrystalline powder ( diameter @xmath6 1 - 10 @xmath7 m ) in epoxy can be obtained only for r = eu and gd through the field - rotation powder alignment method where @xmath5-axis is perpendicular to the aligned magnetic field b@xmath8 = 0.9 t and parallel to the rotation axis . for smaller rare
earth compound r = gd@xmath2dy@xmath2 , powder alignment can be achieved using the simple field powder alignment method where @xmath5-axis is partially aligned along the aligned magnetic field .
no powder alignment can be achieved for larger rare earths r = pr , nd or sm due to the lack of magnetic anisotropy in these compounds .
the anisotropic temperature dependence of magnetic susceptibility for the @xmath5-axis aligned powders exhibit weak anisotropy with @xmath9 at room temperature due to anisotropic rare earth , eu and gd , contribution and crossover to @xmath10 below 190 k where strong ru anisotropic short - range exchange interaction overtakes the rare earth contribution .
anisotropic diamagnetic superconducting intragrain shielding signal of aligned microcrystalline rusr@xmath0gdcu@xmath0o@xmath1 powder - in - epoxy below vortex lattice melting temperature at 39 k in 1-g field is much weaker than the intergrain polycrystalline bulk sample signal due to the small grain size ( d @xmath6 1 - 10 @xmath7 m ) , long penetration depth ( @xmath11 0.55 @xmath7 m , @xmath12 0.66 @xmath7 m ) and the two - dimensional ( 2d ) character of cuo@xmath0 layers . |
observations of nearby star - forming regions show that the vast majority of pre - main - sequence ( pms ) stars are either accreting classical t tauri stars ( cttss ) with optically thick inner disks extending inward to the dust sublimation radius and showing a narrow range of infrared ( ir ) colors or , more evolved , non - accreting weak - line t tauri stars ( wttss ) with bare stellar photospheres . the few
transition objects " that are caught between the typical cttss and wttss stages present very diverse ir seds associated with a wide range of disk structures , but they usually have little or no excess at near - ir wavelength and much larger excesses at longer wavelengths ( see williams & cieza , 2011 for a recent review ) . ever since they were discovered by the _ infrared astronomical satellite _
( strom et al . 1989 ; wolk & walter 1996 ) , the deficit of near - infrared excess in transition disks has been interpreted as a diagnostic of inner disk clearing , possibly connected to planet formation .
however , in addition to planet formation , at least three other disk evolution processes can produce the inner opacity holes that are characteristic of transition objects : grain growth , photoevaporation , and dynamical interactions with ( sub)stellar companions .
the four different mechanisms potentially responsible for the holes in transition disks might become distinguishable when , in addition to seds , disk masses , accretion rates and multiplicity information are available ( najita et al . 2007 ) . in our recent multi - wavelength study of ophiuchus transition disks ( cieza et al .
2010 , hereafter paper i ) , we have estimated disk masses ( from submillimeter photometry ) , accretion rates ( from optical echelle spectroscopy ) , and multiplicity information ( from near - ir adaptive optics imaging ) of 26 transition objects in order to shed some light on the origin of their inner opacity holes . of these 26 transition disks , 4 were classified as
planet - forming disks candidates " based on their sed morphologies , multiplicity , accretion rates , and disk mass measurements ( objects # 11 , 21 , 31 , and 32 in the study ) .
these 4 objects have seds consistent with sharp , dynamically induced inner holes ( as opposed to the smooth decrease in opacity expected from grain growth ) , yet our adaptive optics imaging showed that they lack _ stellar _ companions beyond @xmath210 au . also , their large disks masses and/or high accretion rates disfavor photoevaporation as the disk clearing mechanism .
overall , the properties of these objects are those expected for protoplanetary disks with embedded giant planets .
the recent discoveries of what appear to be forming planets embedded within the disks of the transition objects t cha ( huelamo et al .
2011 ) and lkca 15 ( kraus & ireland , 2012 ) give credence to this interpretation and encourages detailed studies of similar objects .
while sed modeling can be a powerful tool to estimate the physical properties of circumstellar disks , it has severe limitations .
it is a well known problem that different combination of parameters can reproduce the photometry data equally well , even when _ very _ well sampled seds are available ( e.g. , see cieza et al .
2011 for the t cha case ) .
fortunately , resolved ( sub)millimeter images can break many of the degeneracies between model parameters and help constrain disk properties much better than it is possible with the sed alone ( andrews et al . 2009 ; brown et al .
2009 ; isella et al .
2010 ) .
rxj1633.9 - 2442 is the most massive of the 4 planet forming disk " candidates identified in paper i. its sed was recently modeled by orellana et al .
( 2012 ) as a 17 m@xmath5 disk with a 7.9 au radius cavity . here
we present high resolution ( 0.3@xmath1 @xmath2 35au ) sma continuum observations at 340 ghz ( 880 @xmath0 m ) of rxj1633.9 - 2442 and use the mcfost radiative transfer code ( pinte et al .
2006 ) to simultaneously model the entire optical to millimeter wavelength sed and sma visibilities in order to constrain the structure of its disk . in 2 ,
we present our sma data and the photometry from the literature that we use to construct to the full sed , as well as the keck aperture masking data we have obtained to search for low - mass companions and the magellan optical spectroscopy data we use to better measure the spectral type of the central star . our disk model and the degree to which each of the disk parameters can be constrained
are discussed in 3 . in 4 , we analyze our results in the broader context of disk evolution and planet formation models .
a summary of our main conclusions is presented in 5 .
submillimeter interferometric observations of our target were conducted in service mode with the sma ( ho et al .
2004 ) , on mauna kea , hawaii , on february 9@xmath6 and february 22@xmath7 , 2010 .
the receivers were tuned to a local oscillator frequency of 340 ghz ( 880 @xmath0 m ) .
both the upper and lower sideband data were used , providing a total bandwidth of 4ghz .
the observations were obtained with seven of the eight 6-meter antennas in the
very extended configuration " , resulting on 21 baselines from 120 to 510 meter in length and an elongated synthesized beam , 0.26@[email protected]@xmath1 in size ( i.e. , 31 au @xmath8 37 au ) .
the zenith opacities during both nights were @xmath9 0.07 .
for each target , the observations cycled rapidly between the target and two gain calibrators , 1625 - 254 and 1626 - 298 , located at 2.0 and 5.3 degrees from rxj1633.9 - 2442 , respectively . in order to ensure the appropriate calibration of short - timescale phase variations
, we adopted integration times of 5 minutes on target and 3 minutes on each calibrator . in each of the two nights
, our target was observed between hour angles of @xmath103.2 and @xmath113.4 , amounting to a combined _ integration _ time of 7.5 hs .
the raw visibility data were calibrated with the mir reduction packagecqi / mircook.html ] .
the passband was flattened using @xmath260 to 90 min scans of the bright quasar 0854 + 201 and the solutions for the antenna - based complex gains were obtained using the primary calibrator 1625 - 254 .
these gains , applied to our secondary calibrator , 1626 - 298 , served as a consistency check for the solutions . the absolute flux scale was determined through observations of vesta and is estimated to be accurate to 15@xmath12 .
the visibilities were fourier transformed , deconvolved with the @xmath13 algorithm , and restored with the synthesized beam using the standard @xmath14 software package ( sault et al .
each track was first processed independently to check for consistency , but both were later combined to increase the signal to noise of the final image . the final image is shown in figure 1 and has a rms noise of 1.1 mjy / beam .
an inspection of the image gives a first - order approximation to some disk properties .
first , the image shows two clear peaks , implying that the inner hole has been resolved in one direction ( east - west ) , but not in the other .
also , the aspect ratio of the image suggests the disk is highly inclined ( i.e. , @xmath15 45 deg from face - on ) .
fortunately , the major axes of the disk and the synthesized beam are almost perpendicular to each other , maximizing the spatial resolution along the disk .
the disk is @xmath21@xmath1 ( 120 au ) across , and the diameter of the inner hole seems to be slightly larger than the 0.3@xmath1 ( 40 au ) beam . the disk diameter should be considered a lower limit as the sma observations are insensitive to the low optical depths of the outermost parts of the disk ( see [ modeling ] ) . finally , it is clear that the disk is not located at the exact center of the field . for interferometric observations , the
_ uv_-plane provides more direct means than the image plane to derive quantitative constraints on disk parameters .
a convenient way to encapsulate the information from all physical scales sampled by the interferometer is to deproject the visibilities to 0.0 deg inclination and 0.0 deg position angle ( pa ) for the major axis ( andrews et al .
2009 ; brown et al . 2009 ) .
the deprojected _ uv_-distances are given by @xmath16 , where @xmath17 and @xmath18 , @xmath19 , and @xmath20 is the inclination ( lay et al .
a first approximation to the offsets in @xmath21 and @xmath22 can be obtained by examining the image itself , but more accurate numbers can be calculated by assuming radial symmetry and minimizing the imaginary flux components of the deprojected visibilities ( a face - on radially symmetric disk at the center of the field should have zero imaginary flux components at all spatial frequencies ) . using this latter approach
, we obtained positional offsets of @xmath23 and @xmath24 in @xmath21 and @xmath22 , respectively .
we estimate the pa to be 100@xmath255 deg west of north from the orientation of the continuum image and the inclination to be 50@xmath255 deg from the aspect ratio of simulated images ( see section [ empty ] ) .
the deprojected visibility profile , shown in figure 2 , has a null characteristic of disks with sharp inner holes ( i.e. , a large drop in the surface density over a small radial distance ) .
the location of this null is set by the spatial frequency associated with the inner hole and directly constrains its radius .
we binned the data using 40 k@xmath26 bins .
this bin size is somewhat arbitrary , but it results in 13 visibility values , a number similar to that of the sed points .
we have verified that the shape of the visibility profile and , in particular , the location of the null are robust to the choice of bin size . in [ modeling ] , we model the deprojected visibility profile , together with the sed , to place constraints on several disk structure parameters .
we constructed the optical to millimeter wavelength sed for rxj1633.9 - 2442 from the following sources .
the r - band flux comes from the usno - b1 catalog ( monet et al .
2003 ) , and the ground based near - ir fluxes are from the _ 2mass _ survey ( skrutskie et al .
the 12 and 22 @xmath0 m fluxes come from the wide - field infrared survey explorer ( wise ) .
the _ spitzer _ and 1.3 mm fluxes are from paper i , while the 850 @xmath0 m flux is from nutter et al .
unfortunately , no _ spitzer _
spectrum is available for this source .
our target is embedded in the ophiuchus molecular cloud , and thus is strongly affected by extinction .
taking advantage of the lack of near - ir excess , we estimate an extinction a@xmath27 = 5.6 mag from the j - k@xmath28 color excess , adopting a@xmath29 = 5.88 @xmath8 ( ( j - k@xmath28)-(j - k@xmath28)@xmath30 ) , where ( j - k@xmath30 ) is the expected color of a dwarf main sequence star ( kenyon @xmath31 hartmann , 1995 ) of the same spectral type as rxj1633.9 - 2442 , a k5 star ( see [ magellan ] ) .
the extinction at other wavelengths were estimated from the extinction relations listed in cieza et al .
extinction becomes negligible at 24 @xmath0 m and beyond , so the long wavelength fluxes have not been corrected for it .
the photometry data from the literature and the adopted uncertainties , are listed in table 1 .
the resulting sed is shown in figure 3 and is characterized by the complete lack of detectable excess emission at _
2mass _ ( 1.2 to 2.2 @xmath0 m ) and _ spitzer_-irac ( 3.6 - 8.0 @xmath0 m ) wavelengths .
since a very small amount of dust ( @xmath32 m@xmath33 ) is needed to produce detectable near - ir excess , the sed alone indicates extreme levels of dust depletion in the inner disk ( r @xmath3 1 au ) of rxj1633.9 - 2442 .
the 22/24 @xmath0 m fluxes are somewhat below the lower quartile of the ctts population ( furlan et al .
2006 ) , but the 70 @xmath0 m and ( sub)millimeter fluxes are typical of a massive primordial disk . non - redundant aperture masking ( nrm ) has been well - established as a means of achieving the full diffraction limit of a single telescope ( tuthill et al .
2000 , 2006 ; kraus et al .
nrm uses a pupil - plane mask to block most of the light from a target , resampling the primary mirror into set of smaller subapertures that form a sparse interferometric array .
nrm allows for superior calibration of the stellar primary s point spread function and elimination of speckle noise by the application of interferometric analysis techniques , specifically the measurement of closure phases .
we observed rx j1633.9 - 2442 on april 22 - 23 , 2011 , using the keck - ii 10-m telescope with laser - guide star adaptive optics .
all observations were conducted with the facility ao imager , nirc2 , which has aperture masks installed in the cold filter wheel near the pupil stop .
we used a 9-hole aperture mask , which yields 28 independent baseline triangles about which closure phases are measured .
all nrm observations operate in a subarray mode of the narrow camera , which has a pixel scale of 9.963 mas / pix , and we conducted our observations using the broadband @xmath34 filter . each observing sequence consisted of multiple `` visits '' of rx j1633.9 - 2442 , alternating with observations of independent calibrator stars .
many of these calibrators were other transitional disk hosts in ophiuchus listed in paper i. each visit consisted of a sequence of 12 exposures that were each 20s , and there were 14 and 17 visits per night .
the data analysis was identical to that used in previous papers ( e.g. , ireland et al .
2008 ; kraus et al .
2008 , 2011 ) , combined with the new calibration technique described in kraus & ireland ( 2012 ) . to briefly summarize ,
the images were flat - fielded and bad pixels were removed by interpolating between neighboring pixels .
the image was then multiplied by a super - gaussian window function of the form @xmath35 , with @xmath36 the radius in pixels from the center of the interferogram .
a two - dimensional fourier transform was then made of each exposure in a visit , and this fourier transform was point - sampled at the positions corresponding to the baseline vectors in the aperture mask . for each visit we then computed the vector of mean uncalibrated closure - phases and the standard error of the mean .
finally , we calibrated the closure - phases for each visit using an optimal linear combination of the calibrators observed in the same sequence of visits .
our analysis found no statistically significant signal in the calibrated closure phases for rx j1633.9 - 2442 , and hence that it is single to within the detection limits of the observations . using the same procedures as in our previous nrm work mentioned above ( i.e. , a monte - carlo method that simulates random closure phase datasets of a point source with closure - phase errors and covariances that match those of the real data ) , we found contrast limits ( @xmath37 ) of 5.9 mag at 20 - 40 mas , 6.9 mag at 40 - 80 mas , and 6.8 mag at @xmath3880 mas .
the corresponding mass detection limits , based on the 1 myr dusty models of chabrier et al .
( 2000 ) and the assumed distance of 120 pc , are 6 @xmath39 at 2.4 - 4.8 au , and 3.5 @xmath39 at @xmath40 4.8 au . if any planetary companions are brightened by significant accretion luminosity , as seems likely ( see 4.4 ) , then the mass detection limits could be even lower . as part of our recent survey of ophiuchus transition disks ( paper i )
, we obtained high resolution spectra of rx j1633.9 - 2442 using the 2.5-m du pont telescope in las campanas observatory . from these data , we derived a k7 spectral type and an accretion rate of @xmath210@xmath41(m@xmath42yr@xmath43 ) .
we reobserved rx j1633.9 - 2442 with the magellan inamori kyocera echelle ( mike ) spectrograph on the 6.5-m clay telescope , also at las campanas observatory , on june 25@xmath6 , 2011 in order to obtain a second - epoch accretion rate .
we used the red arm of the spectrograph and a 1@xmath1 slit to obtain the complete optical spectrum between 4900 and 9500 at a resolution of 22,000 .
this resolution corresponds to @xmath20.3 at the location of the h@xmath44 line and to a velocity dispersion of @xmath214 km / s .
we obtained a set of 3 spectra , with a exposure time of 3 minutes each .
the data were reduced using the standard iraf packages imred : cddred and echelle : doecslit .
since the final clay spectrum has a better signal to noise ratio than our previous du pont observations , we have revisited the spectral type classification of rx j1633.9 - 2442 by comparing the new data against the elodie high - resolution spectral library ( prugniel & soubiran , 2001 ) .
figure 4 ( * left panel * ) shows the rx j1633.9 - 2442 spectra in the narrow 6197 to 6203 region , containing v i and fe i lines , whose shapes , depths , and relative strengths are highly sensitive to effective temperature ( padgett et al .
we find that the k5 template is a much better match to the rx j1633.9 - 2442 spectrum than the k7 template is and hence adopt this new and slightly revised spectral type for the modeling work .
figure 4 ( * right panel * ) shows the continuum subtracted velocity profile of the h@xmath44 line for our du pont and clay observations .
the vertical lines mark a velocity width ( @xmath45 , measured at 10@xmath12 of the peak value ) of 200 km / s , the boundary between accreting and non - accreting objects suggested by jayawardana et al .
the h@xmath44 line is broader than 200 km / s , asymmetric and variable in both shape and intensity as expected from magnetospheric accretion .
we find @xmath45 @xmath2300 km / s from our du pont observations and @xmath2230 km / s in our new clay data . for accreting objects
, @xmath45 correlates with accretion rates derived from models of the magnetospheric accretion process .
the relation given by natta et al .
( 2004 ) , and adopted in paper i , @xmath46 , translate the @xmath45 values into accretion rates of 10@xmath41 and 10@xmath47 m@xmath42yr@xmath43 , for the du pont and clay data , respectively .
in order to constrain the structure of the rx j1633.9 - 2442 disk based on the observed sma visibilities and sed , we use the mcfost radiative transfer code ( pinte et al .
mcfost adopts a monte carlo approach to follow `` photon packets '' propagated through the disk ( i.e. , a parametrized dust density structure ) .
mcfost outputs synthetic seds and monochromatic raytraced images .
the raytraced images can be used to simulate synthetic sma visibilities with the same _ u - v _ sampling as the actual observations .
the deprojected synthetic visibilities can then be directly compared against the real data , using the same radial binning , as described in the following section .
we follow the model fitting procedure described by mathews et al .
( 2012 ) , which uses the levenberg - marquardt @xmath48 minimization algorithm to perform an efficient exploration of the parameter space .
we start by using mcfost to create a small grid of models and the corresponding seds and 880 @xmath0 m raytraced images .
the raytraced images are input to the ft ( fourier transform ) task in the common astronomy software applications ( casa ) package , which outputs visibility data sets with the same _ u - v _ sampling as the real sma observations .
then , for each model , the @xmath49 of the model is calculated as the sum of the @xmath49 of the sed and the @xmath49 of the visibility profile .
we use the idl routine mpfit ( markwardt , 2009 ) to implement the levenberg - marquardt @xmath48 minimization algorithm to calculate the numerical gradients of the @xmath49 function and determine the next point in the parameter space to be sampled until the algorithm converges to a @xmath49 minimum . to better sample the parameter space and in order to avoid local minima
, we carried out the search algorithm several times using different starting values .
each of the runs provide a set of best - fit parameters .
the distribution of the best - fit values for each parameter can be used to calculate the mean and an associated uncertainty ( see 4.2 in mathews et al .
2012 for a discussion on estimating the uncertainties ) .
transition disks are complex systems that have been modeled with a wide range of structures .
some objects such as coku tau/4 and dm tau have inner holes that seem to be completely depleted of ir emitting grains ( dalessio et al .
2005 ; calvet et al . 2005 ) , while others such as gm aur have cavities filled with optically thin dust ( calvet et al . 2005 ) . yet
other systems , like t cha and lkca 15 are best described as having optically thin gaps separating optically thick inner and outer disk components ( olofsson et al .
2011 ; espaillat et al .
2010 ) . in what follows
, we explore all 3 possibilities for the structure of the rx j1633.9 - 2442 disk .
we begin our modeling by adopting the simplest possible structure for a transition disk : a disk with an empty cavity . motivated by physical models of viscous accretion disks ( e.g. , hartmann et al .
1998 , lynden - bell & pringle 1974 ) and the discrepancy in disk sizes obtained from continuum and co line images ( hughes et al . 2008 ; isella et al . 2007 ; petu , guilloteau & dutrey , 2005 ) , we follow mathews et al .
( 2012 ) and adopt the following description for the surface density profile and the size of the disk : @xmath50 $ ] where @xmath51 is the surface density at a characteristic radius , @xmath52 , and @xmath53 is the radial dependence of the disk viscosity , @xmath54 . in this prescription ,
the surface density is proportional to r@xmath55 in the inner disk , but it quickly becomes dominated by the exponential taper at large radii .
@xmath56 = 0 within the cavity of radius @xmath57 .
we set the outer radius of the disk , @xmath58 , to 200 au .
however , we note that our model is not very sensitive to the exact value of @xmath58 because the exponential taper of the outer disk implies very low surface densities ( below the sensitivity of our sma observations ) for radii @xmath59 @xmath60
. the vertical distribution of the dust is given by a gaussian with a scale height @xmath61 , where @xmath62 is the scale height at @xmath63 , and @xmath64 is the power law describing the flaring of the disk .
the dust content is described by a differential power - law for the grain size distribution ( d@xmath65(@xmath66 ) @xmath67 @xmath68d@xmath66 ) , between 0.005 @xmath0 m and 3900 @xmath0 m .
the size of the longest wavelength in the sed , which approximates the maximum grain sizes the data are sensitive to ( draine 2006 ) .
] we adopt the porous grains from mathis @xmath31 whiffen ( 1989 ) for the grain composition .
the stellar parameter are those of the best matched photospheric model ( kurucz 1979 , 1993 ) for a k5 star ( t@xmath69=4350 k , log g = 4.0 ) at the relatively well established distance of 120 pc to the ophiuchus molecular cloud ( loinard et al .
we initially created seds and visibility profiles assuming an inclination of 60 deg for the disk , but quickly found that adopting an inclination of 50 deg resulted in synthetic images with aspect ratios that are closer to the observed image and hence use this latter value for all models .
table 2 lists all the parameters that are fixed in our model .
the 6 free parameters for the empty cavity " model are listed in table 3 .
m@xmath70 is not a free parameter .
it is obtained by integrating the surface density profile over radius and assuming a gas to dust mass ratio of 100 . for this parameterization , we ran the search algorithm 5 times . the visibility profile and
the sed of the overall best - fit model from all 5 runs are indicated in figures 2 and 3 , respectively .
the probability weighted means and uncertainties ( calculated as in mathews et al .
2012 ) for each parameter are listed in table 3 in the empty cavity " column . while the model reproduces the visibility profile very well , the observed 22 and 24 @xmath0 m fluxes are factors of @xmath22 - 3 ( i.e. , @xmath210 to 20@xmath71 ) higher then predicted by the model . since the 22 and 24 @xmath0 m measurements are independent ( the former is from wise and the latter is from _ spitzer _ ) , photometric problems can be ruled out .
the empty cavity " model does not reproduce the significant 12 @xmath0 m excess either .
the large discrepancy between the observations and the best - fit model in the mid - ir can be understood considering that the search algorithm was most likely driven by the @xmath49 of the visibility profile , which is very sensitive to the value of r@xmath72 .
in other words , the visibility profile very strongly constrains the size of the inner cavity to be [email protected] au , but such a large cavity is incompatible with the observed 12 and 22/24 @xmath0 m excesses .
we find that the sed and sma data can _ not _ be reconciled even adopting a hotter k1-type central star ( which is clearly ruled out by the optical spectrum , see figure 4 ) .
it is thus unavoidable to conclude that the @xmath223 au cavity imaged at submillimeter wavelengths is not completely depleted of mid - ir emitting grains . to try to reproduce the observed mid - ir excesses
, we partially fill the cavity by incorporating 2 additional free parameters that result in a 2-component disk : @xmath73 and @xmath74 . following andrews et al .
( 2011a ) and mathews et al .
( 2012 ) , the surface density profile of the disk is modified such that @xmath75 = @xmath73 @xmath56 , for @xmath74 @xmath76 r @xmath76 @xmath57 .
that is , @xmath56 no longer drops to zero at @xmath57 , but is sharply reduced to a lower value between @xmath57 and @xmath74 .
this sharp reduction in the surface density profile at @xmath57 is meant to reproduce the inner hole seen at submillimeter wavelengths .
@xmath56 remains zero for r @xmath76 r@xmath77 .
we ran the search algorithm 5 times for this new parameterization . the visibility profile and
the sed of the overall best - fit 2-component model are indicated in figures 2 and 3 , respectively .
the weighted means and uncertainties for each parameter are listed in table 3 in the 2-component disk " column .
the inner component of this model is characterized by a surface density reduction of @xmath2100 with respect to the outer disk and an inner radius of @xmath27 au .
as shown by figures 2 and 3 , the 2-component parametrization allow us to simultaneously obtain satisfactory fits for both the visibility profile and the sed .
however , as the properties of the inner disk component are effectively controlled by only 3 sed points ( i.e , the 8.0 , 12 , and 22/24 @xmath0 m fluxes ) , the solution is unlikely to be unique .
we next explore an alternative geometry for the inner disk , a narrow ring . as an alternative of the 2-component model
, we modify the empty cavity model with two additional parameters , @xmath78 and @xmath79 , describing the location and width of a ring within the cavity .
the description of the outer disk ( i.e. , beyond r@xmath72 ) remains unchanged . with this parametrization
, we also ran the search algorithm 5 times .
the results are shown in figures 2 , 3 and listed in table 3 , together with the other 2 earlier parameterizations .
we find that a narrow ring at @xmath210 au fits the visibility profile and the sed almost as well as the 2-component model does ( see also @xmath49 of the visibilities and the sed in table 3 ) . the small width of the ring ( 2 au ) can be understood from the fact that the ring is optically thick and its mid - ir emission is dominated by the inner rim facing the star .
[ mod - results ] table 3 shows that the outer disk parameters for all three model structures agree remarkably well .
all the values are well within 1-@xmath80 or 2-@xmath80 , which give us confidence in both our modeling results and uncertainty estimates . also , because the sma data is only sensitive to the properties of the outer disk , the visibility profiles of the three models match the observations comparably well ( see figure 2 and @xmath49 of the visibilities in table 3 ) .
we find that the disk is relatively massive ( @xmath215 m@xmath5 ) and rather flat .
the scale height at r@xmath81 is given by @xmath82 , corresponding to @xmath3 0.8 au at a radius of @xmath240 au for all 3 models .
this flat geometry is in agreement with the result by orellana et al .
( 2012 ) , who found a scale height of 2 au at 100 au for rx j1633.9 - 2442 , and is most likely due to significant grain growth and dust settling . our models , driven by the ( sub)millimeter colors , do favor a grain size distribution extending beyond the millimeter size scale .
as for the the inner disk , its structure is not well constrained by current data .
the narrow optically thick ring and the more extended optically thin region are only a subset of possible solutions , and the radial symmetry assumed is not necessarily correct ( see [ dynamical ] ) .
however , we can say that the mid - ir emission originates beyond @xmath25 au , from a dust component that is distinct from the outer disk imaged at submillimeter wavelengths . similarly ,
the extreme depletion of dust grains within a few au of the star is well established by the lack of detectable excess at irac wavelengths ( 3.6 to 8.0 @xmath0 m ) .
as mentioned in the introduction , orellana et al .
( 2012 ) successfully modeled the sed of rx j1633.9 - 2442 with a 7.9 au inner hole , which was virtually empty .
in other words , the larger submillimeter cavity imaged by the sma is _ not _ detectable from the sed alone .
the surface density profile of the 3 models we considered are shown in figure [ sigma ] . by simultaneously modeling the visibility and sed data ,
we are able to demonstrate that the circumstellar environment of rx j1633.9 - 2442 presents _ at least _ three distinct radial regions . the innermost region ( r @xmath3 5 au ) is depleted of small grains . the middle region contains _ some _
mid - ir emitting grains , with an unknown configuration .
the outermost region of the disk starts at @xmath225 au , with a sudden increase in the surface density . this complex structure is certainly intriguing , but is not unique to rx j1633.9 - 2442 .
similar structures have already been proposed to reconcile the submillimeter images and seds of several other transition disks , including dm tau , rx j1615.3 - 3255 ( andrews et al .
2011a ) , and rx j1604.3 - 2130 ( mathews et al .
2012 ) . in the next section
, we discuss the physical processes that could potentially explain our modeling results and the overall properties of rx j1633.9 - 2442 .
most of our conclusions are also applicable to the 3 other similar objects listed above .
as discussed in [ intro ] , multiple mechanisms have been proposed to explain the inner opacity holes of transition disks , including photoevaporation , grain growth , and dynamical clearing . in what follows ,
we consider how the predictions from models of each process compare to the properties of rx j1633.9 - 2442 . toward the end of the section
, we assess the likelihood of a connection between transition disks and each of the two leading theories of giant planet formation , core accretion and gravitational instability .
photoevaporation by the central star is currently believed to play an important role on the dissipation of circumstellar disks .
photoevaporation can be driven by energetic photons in the fuv ( 6ev @xmath83 13.6ev ) , euv ( 13.6ev @xmath83 0.1kev ) and x - ray ( @xmath84kev ) energy range .
photons in each energy domain operate in different ways .
euv photons can not penetrate far into the disk and drive relatively weak photovaporation winds ( 10@xmath41 m@xmath42/yr ) .
thus , euv photoevaporation only becomes important once most of the disk mass has been depleted and the accretion rate drops below @xmath210@xmath41 m@xmath42/yr ( alexander et al .
2006a,2006b ) , at which point the outer disk is no longer able to resupply the inner disk with material and the inner disk drains on a viscous timescale ( @xmath85yr ) .
once the inner disk is drained , a hole is formed , the disk edge is directly exposed to the euv radiation , and the disk rapidly photoevaporates from the inside out . the very large disk mass of rx j1633.9 - 2442 and the presence of accretion are inconsistent with euv - induced photoevaporation being the formation mechanism for its inner hole .
more recent studies have incorporated x - ray ( owen et al . 2011 ; 2012 ) and/or fuv irradiation ( gorti , dullemond & hollenbach 2009 ; gorti & hollenbach 2009 ) into photoevaporation models . according to these models
, x - ray and fuv photons can penetrate deeper into the disk , and drive much higher photoevaporation rates ( @xmath210@xmath86 m@xmath42/yr ) . as a result
, the hole is expected to form earlier in the evolution of the disk , while the disk is still relatively massive .
since the surface density of the inner disk at the time it starts to drain is high , x - ray / fuv photoevaporation models can in principle explain the presence of massive disks with inner holes and moderate accretion ( i.e. , they would represent the inner disk draining stage ) .
however , these models also predict that accretion onto the star should quickly drop as the size of the inner cavity grows . in the context of these models , whether the inner hole of a transition disk can potentially be explained by x - ray / fuv photoevaporation depends on the size of the inner hole , the observed accretion rate , the stellar mass , and the x - ray luminosity . in particular
, the region in the hole size versus accretion rate plane that is consistent with x - ray / fuv photoevaporation is a very strong function of stellar mass , m@xmath87 , ( see fig .
17 in owen et al .
2012 ) because : 1 ) the location at which the x - ray heated gas becomes unbound and opens a gap in the disk is proportional to m@xmath87 and 2 ) the x - ray luminosity is also a strong function of stellar mass ( l@xmath88 @xmath67 m@xmath89 ; preibisch et al .
we thus estimate the stellar mass of rx j1633.9 - 2442 by comparing its temperature and luminosity to the predictions of theoretical evolutionary tracks .
a t@xmath69 of 4350 k is directly derived from the k5 spectral type .
we estimate a luminosity of 0.70 l@xmath42 applying a bolometric correction ( from hartigan et al .
1994 ) to the extinction corrected j - band magnitude and adopting a distance of 120 pc . according to the models by d@xmath90antona @xmath31 mazzitelli ( 1998 ) and siess et al .
( 2000 ) the temperature and luminosity of rx j1633.9 - 2442 correspond to those of a 2 myr old 0.7 m@xmath42 star and a 6 myr 1.0 m@xmath42 star , respectively .
the large discrepancy in stellar age highlights the uncertainty of these models .
however , with that caveat , rx j1633.9 - 2442 can be considered to be a @xmath3 1 m@xmath42 star , for which disk holes larger than 20 au around accreting objects can not be explained by photoevaporation .
also , photoeveporation can not easily account for the 3-region structure we found in 3 .
we thus conclude that the inner hole of rx j1633.9 - 2442 is unlikely to be due to any kind of photoevaporation process .
dust opacity , @xmath91 ( @xmath92/g ) , is a very strong function of particle size .
as soon as primordial sub - micron dust grains grow into larger bodies ( r @xmath59 @xmath26 ) , most of the solid mass never interacts with the radiation , and @xmath91 plunges .
observational support for grain growth in disks is robust and comes from at least two independent lines of evidence : the shapes of the silicate features around 10 and 20 @xmath0 m ( kessler - silacci ; 2006 , olofsson et al . 2010 ) and the spectral slopes of disks at ( sub)millimeter wavelengths ( andrews & williams , 2005 , 2007 ; wilner et al .
2005 ; ricci et al .
grain growth has been proposed as one of the possible explanations for the opacity holes of transition disks because it might be a strong function of radius ( it is expected to be more efficient in the inner regions where the surface density is higher and the dynamical timescales are shorter ) .
idealized dust coagulation models , ignoring fragmentation and radial drift , do in fact predict extremely efficient grain growth in the inner disk and can produce seds similar to those of rx j1633.9 - 2442 ( dullemond & dominik , 2005 ) . however , dust fragmentation and radial drift result in the efficient replenishment of micron size grains ( brauer , dullemond & henning 2008 ; birnstiel , ormel & dullemond 2011 ) and a smooth and _ shallow _ dependence of @xmath91 on disk radius .
in contrast , both the sed and sma visibilities of rx j1633.9 - 2442 are consistent with a _
steep _ discontinuity in the optical depth ( i.e. , @xmath93 ) at r@xmath72 and r@xmath77 ( or r@xmath94 ) .
furthermore , if the optical depth discontinuities are mainly due to a reduction in @xmath91 instead of @xmath95 , this would favor the onset of the magneto - rotational instability ( chiang & murray - clay , 2007 ) and would exacerbate accretion .
the very low accretion rate of the rx j1633.9 - 2442 disk ( @xmath3 10@xmath41 m@xmath42yr@xmath43 ; see [ magellan ] ) contradicts this scenario .
we thus consider grain growth to be a very unlikely explanation for the inner opacity reductions .
unlike grain growth , the dynamical interaction of a ( sub)stellar companion embedded within the disk can produce a sharp inner hole ( artymowicz & lubow ( 1994 ) .
ireland @xmath31 kraus ( 2008 ) showed that the famous transition disk coku tau 4 is in fact a near - equal mass binary system , which naturally explains the hole that had been inferred from its sed .
this immediately raised the question of whether most transition disks were close binaries .
however , it is now clear that most sharp holes are _ not _ due to binarity .
the transition objects dm tau , gm aur , lkca 15 , ux tau , and ry tau have all been observed with the keck interferometer ( pott et al .
2010 ) . for these objects
, stellar companions with flux ratios @xmath76 20 can be ruled out down to sub - au separations .
our keck aperture masking observations discussed in [ masking ] rule out the presence of even a brown dwarf companion down to a projected separation of 2.4 au ( corresponding to a maximum physical separation of 3.8 au ) .
therefore , a stellar ( or brown dwarf ) companion can not explain neither the @xmath225 au hole seen in our sma image nor the inner most cavity indicated by the sed . instead , the 3-region structure of the rx j1633.9 - 2442 disk discussed in [ mod - results ] suggests the presence of _ at least _ two low - mass objects dynamically sculpting the disk : an object at @xmath96 au from the star that creates the discontinuity in the surface density seen in the submillimeter image and another object at @xmath3 7 au that clears up the inner disk of material .
we thus argue that the dynamical interaction of multiple planets embedded within the disk is the most likely explanation for the overall properties of rx j1633.9 - 2442 .
the need for multiple planets to explain the properties of rx j1633.9 - 2442 is very strongly supported by recent hydrodynamical simulations of giant planets embedded in primordial disks by dodson - robinson & salyk ( 2011 ) and zhu et al .
( 2011 ) showing that multiple planets are in fact _ required _ to produce inner holes and gaps wide enough to have a noticeable effect in the emerging sed .
both studies find that a single giant planet can not explain wide optically thin gaps and holes .
multiple forming planets also help explaining the low accretion rates _ onto the star _ ( for a given disk mass ) of many transition objects ( najita et al .
2007 ; espaillat et al . 2012 ) as each planet accretes a significant fraction of the material being transported across the disk . _ kepler _ observations have demonstrated that systems with multiple planets with small semi - major axes are common ( lissauer et al .
the lack of near - ir excess ( @xmath97 8.0 @xmath0 m ) in the sed of rx j1633.9 - 2442 combined with variable and detectable levels of accretion ( @xmath210@xmath4110@xmath98 m@xmath42yr@xmath43 ) could be an indication of a densely packed planetary system resulting in multiple optically thick tidal streams that transport a significant amount of material onto the star but cover a small area of the inner disk .
we ran some tests and found that _
axisymmetrically distributed _ circumstellar material would produce a detectable near - ir excess unless the surface density of the inner disk ( r @xmath3 5 au ) is reduced by a factor of 10@xmath99 with respect to that of a typical " ctts disk . in the absence of planets ,
the accretion rate onto the star should be proportional to the surface density of the inner disk .
therefore , if the surface density is reduced by a 10@xmath99 factor from typical levels , one would expect an accretion rate of 10@xmath10010@xmath101 m@xmath42yr@xmath43 for rx j1633.9 - 2442 .
such low rates are undetectable and 2 orders of magnitude lower than the observed value .
nevertheless , the lack of near - ir excess and the accretion rate could be reconciled if the inner disk contains optically thick tidal streams with a geometric filling factor of a few percent , which is in agreement with the results of hydrodynamic simulations ( dodson - robinson & salyk , 2011 ) . since these hydrodynamic simulations predict very complex disk structures , the 2-component " and ring within cavity " models presented herein are by necessity a crude oversimplification that reflects the lack of resolved data at the appropriate resolution ( i.e. , at a few au scale ) .
the notion that the properties of some transition disks are signposts of ongoing planet formation ( e.g. , najita et al .
2007 ; paper i ; dodson - robinson & salyk 2011 ; espaillat et al . 2012 ) has gained credence from the recently identified companions to t cha and lkca 15 . using the aperture masking technique on the very large telescope , hulamo et al .
( 2011 ) detected a faint object within the inner cavity of the t cha disk .
the object is located at 62 mas ( @xmath27 au ) from the primary and has a luminosity ratio of 5.1 mag in the l@xmath90-band ( 3.8 @xmath0 m ) .
based on the upper limits from similar k@xmath102-band ( 2.2 @xmath0 m ) observations , the authors derived a k@xmath103-l@xmath90 color @xmath38 1.25 for the companion , suggesting the object must be surrounded by dust . also using the aperture masking technique , but in the keck telescope , kraus & ireland ( 2012 ) , identified a similar object inside the inner hole of lkca 15 . in this case
, the companion has been detected in multiple epochs and at multiple wavelengths .
the object seems to be a point source at 2.1 @xmath0 m ( 6.8 mag fainter than the primary ) , but is extended at 3.7 @xmath0 m . since the inclination of the lkca 15 disk is known from resolved submillimeter observations ( andrews et al .
2011b ) , a deprojected separation of @xmath215 au from the primary can be derived for the companion , assuming it is coplanar with the disk .
kraus & ireland interpreted their observations as a young planet surrounded by warm dust .
rx j1633.9 - 2442 shares an intriguing property with t cha and lkca 15 : a very low accretion rate for a given disk mass .
lkca 15 has a disk mass of @xmath255 m@xmath5 ( andrews et al .
2011b ) and an accretion rate of @xmath210@xmath104 m@xmath42/yr
( hartmann et al . 1998 ) .
similarly , t cha has a disk mass of @xmath217 m@xmath5 ( olofsson et al .
2011 ) and it seems to be accreting only very weakly and sporadically onto the star ( alcala et al . 1993 ; schisano et al .
2009 ) . in the absence of a planet
, the mass accretion rate onto the star should be roughly proportional to the mass of the disk ( najita et al .
2007 ) ; however , a planet massive enough to open a gap in the disk is expected to divert most of the material accreting from the outer disk onto itself . as a result , in the presence of a jupiter mass planet ,
the accretion _ onto the star _ is reduced by a factor of @xmath210 with respect to the mass transport across the outer disk ( lubow @xmath31 d@xmath90angelo , 2006 ) . in [ modeling ]
we derived a disk mass of @xmath215 m@xmath5 for rx j1633.9 - 2442 . despite this large disk mass
, we estimate an accretion rate of @xmath3 10@xmath105 m@xmath42yr@xmath43 based on the velocity dispersion of its h@xmath106 line ( see [ magellan ] ) . for comparison ,
the median disk mass and accretion rate for classical t tauri stars in ophiuchus are 5 m@xmath5 ( andrews @xmath31 williams , 2005 ) and 10@xmath107 m@xmath42yr@xmath43 ( natta et al .
2006 ) , respectively .
as discussed in the previous section , the unusually low accretion rate onto the star for the given disk mass observed in rx j1633.9 - 2442 is consistent with the presence of actively accreting objects embedded within the disk .
moderate to low accretion rates seem to be a general feature of transition disks .
while the distributions of mass accretion rates of transition and non - transition disks are very wide and clearly overlap , the accretion rates of disks with evidence for holes and gaps tend to be a factor of 5 lower than those of full `` disks '' ( espaillat et al .
all things considered , the properties of rx j1633.9 - 2442 , t cha , and lkca 15 are best explained by the runaway gas accretion phase in models of giant planet formation through core accretion ( lissauer @xmath31 stevenson , 2007 ) . according to the core accretion model , gas giant planets form by first accreting a solid core , which later attracts a massive gaseous envelope .
initially the gas accretion rate onto the core is much lower than the accretion rate of solids . as the mass of the core increases , the rate of gas accretion accelerates .
once the mass of gas in the core matches the mass of solids , runaway gas accretion occurs and is sustained for as long as there is material in the gas feeding zone of the planet ( i.e. , until the forming planet clears a gap in the disk ) . due to accretion shock luminosity
, the envelope accretion phase corresponds to the highest luminosity state a giant planet will ever have ( marley et al .
this luminosity spike could be in fact what makes the detection of forming planets even possible with current instrumentation .
the duration of the runaway gas accretion phase is estimated to be of the order of 10@xmath108 years for a 1 m@xmath5 planet and somewhat longer for a larger planet ( marley et al .
this timescale is also in agreement with the incidence of transition disks sharing the properties of rx j1633.9 - 2442 ( combining a sharp inner hole , a large disk mass , and a low accretion rate ) .
we note that , while the occurrence rate of objects that can be broadly defined as transition disks is of the order of 20@xmath12 , they are _ not _ all consistent with giant planet formation . accreting objects with steeply rising mid - ir seds , such as rx j1633.9 - 2442 , t cha , and lkca 15 are an order of magnitude less common ( cieza et al . 2010 , 2012 ; romero et al .
2012 ) .
if our interpretation is correct , massive accreting transition disks with rising mi - ir seds around single stars are _ by far _ the best places for direct imaging searches of forming planets since they are not only the sites of ongoing planet formation , but also the places where forming planets should be the brightest ! as shown by hulamo et al .
( 2011 ) and kraus @xmath31 ireland ( 2012 ) , non - redundant aperture masking is the most promising technique for such searches as it delivers the highest contrast ratio at the diffraction limit of the telescope .
in addition to core accretion , gravitational instability ( gi ) has also been proposed as a formation mechanism for giant planets ( boss 1997 ; durisen et al .
nevertheless , gi seems to be less relevant to the transition disks discussed herein for several reasons .
first , gi planets are expected to form at evolutionary stages much earlier ( age @xmath76 1 myr ) than those of transition disks , when the disk is still extremely massive ( m@xmath70/m@xmath109 @xmath15 0.1 ) and deeply embedded within an extended envelope , while the age distribution of transition disks likely to harbor forming planets favors a @xmath15 2 - 3 myr formation timescale ( cieza et al .
second , gi should operate mostly at large radii where the cooling times are shorter than the local orbital periods , a condition needed for fragmentation ( gammie 2001 , rafikov 2007 ) .
gi is believed to be much less effective at disk radii @xmath3 40 au ( boley @xmath31 durinsen , 2008 ) or even @xmath3 100 au ( boley 2009 ) , which makes gi less consistent with the inner hole sizes of most transition disks .
finally , since the timescale for the formation of planets through gi is @xmath210@xmath110 yr , it is statistically unlikely that the formation event itself would be observed in a nearby molecular cloud with an age of few million years and hundreds , not thousands , of young stellar objects .
since gi models do not run long enough to predict the long term evolution of the disk after the formation of the planet , it could be argued that a gi planet could remain embedded in a massive disk for a relative long period of time after it has formed .
however , as discussed above , the high disk masses and low accretion rates of rx j1633.9 - 2442 , tcha and lkca 15 , as well as the properties of the tcha , and lkca 15 companions , suggest that the putative planets are accreting most of their mass at the current epoch ( i.e. , we are watching them form ! ) .
even though some planets might form through gi , the properties and incidence of the transition objects mentioned here are in much better agreement with planet formation through core accretion .
we present continuum high resolution sma observations of the transition disk rxj1633.9 - 2442 and simultaneously model the entire optical to millimeter wavelength sed and sma visibilities in order to constrain the structure of its disk .
the submillimeter image reveals that the disk is highly inclined ( @xmath11150 deg ) and has an inner cavity @xmath225 au in radius .
this cavity is not empty as some warm dust is needed to explain the excess emission observed at 12 , 22 , and 24 @xmath0 m .
the mid - ir excess can be reproduced with either a narrow , optically thick ring at @xmath210 au or an optically thin region extending from @xmath27 to 25 au .
the lack of near - ir excess emission indicates that the inner disk ( r @xmath3 5 au ) is mostly depleted of ir - emitting grains . since rx j1633.9 - 2442 is a single star , the properties of the disk ( a complex structure , a relative massive outer disk , and weak accretion ) favors dynamical clearing by multiple planets as the hole formation mechanism .
this conclusion can be extended to objects with similar properties and disk structures , such as the three transition objects mentioned in section [ mod - results ] : dm tau , rx j1615.3 - 3255 , and rx j1604.3 - 2130 .
the filamentary structures predicted by hydrodynamical models of multiple planets embedded within a disk can reconcile the accretion rate and sed of rx j1633.9 - 2442 as they can transport significant amounts of material to the inner disk without overproducing the observed ir excess .
the properties and occurrence rate of objects such as rx j1633.9 - 2442 , t cha , and lkca 15 ( and those of the companions recently identified to these latter objects ) are in good agreement with the runaway gas accretion phase of the core accretion model , when giant planets gain their envelopes and suddenly become massive enough to dynamically clear a gap in the disk .
if the inner holes of rx j1633.9 - 2442 , dm tau , rx j1615.3 - 3255 , rx j1604.3 - 2130 , t cha , and lkca 15 are in fact due to ongoing giant planet formation through core accretion , these types of systems would represent ideal laboratories to study this complex process in detail and place much needed observational constraints .
for instance , the location of the lkca 15 companion and the age of the system would already imply that core accretion can actually form giant planets at @xmath215 au within @xmath23 myr , which is a difficult challenge for current models ( dodson - robinson @xmath31 bodenheimer , 2010 ) .
similarly , the sizes of the inner holes in the disks of rx j1633.9 - 2442 , rx j1615.3 - 3255 , and rx
j1604.3 - 2130 also suggest the presence of young giant planets at @xmath15 20 au orbital separations . in the near future
, the unprecedented sensitivity and resolution of the atacama large millimeter / submillimeter array ( alma ) will revolutionize the fields of disk evolution and planet formation .
most studies of nearby circumstellar disks will soon be based on high resolution images of both thermal emission and molecular gas tracers rather than on sed observations .
alma will provide new insights on the structure of disks and their dynamics and teach us about turbulence , grain growth and dust settling , and the evolution of the dust to gas mass ratio , the understanding of all of which are key to planet formation theory .
detailed alma studies of disks hosting forming planets is the most direct and promising approach to learn about the planet formation process and the conditions in which planets form .
support for this work was provided by nasa through the _ sagan _ fellowship program under an award from the california institute of technology .
g.s.m . and
acknowledge nasa / jpl and nsf for funding support through grants rsa-1369686 and ast08 - 08144 respectively .
a.l.k . was supported by nasa through the _ hubble _ fellowship program . m.r.s and f.d.m acknowledge support from millenium science initiative , chilean ministry of economy , nucleus p10 - 022-f .
crrlrr 0.65 & 3.53e+00 & 14.85 & 30@xmath12 & usno - b1 & 1 + 1.2 & 1.04e+02 & 10.46 & 15@xmath12 & 2mass & 2 + 1.6 & 1.85e+02 & 9.36 & 15@xmath12 & 2mass & 2 + 2.2 & 2.02e+02 & 8.80 & 15@xmath12 & 2mass & 2 + 3.6 & 9.67e+01 & 8.66 & 10@xmath12 & _ spitzer _ & 3 + 4.5 & 7.10e+01 & 8.51 & 10@xmath12 & _ spitzer _ & 3 + 5.8 & 5.12e+01 & 8.37 & 10@xmath12 & _ spitzer _ & 3 + 8.0 & 3.28e+01 & 8.23 & 10@xmath12 & _ spitzer _ & 3 + 12 & 2.41e+01 & 7.80 & 10@xmath12 & wise & 4 + 22 & 2.54e+02 & 3.79 & 10@xmath12 & wise & 4 + 24 & 2.28e+02 & 3.74 & 10@xmath12 & _ spitzer _ & 3 + 70 & 7.13e+02 & & 15@xmath12 & _ spitzer _ & 3 + 850 & 2.10e+02 & & 15@xmath12 & jcmt & 5 + 1300 & 8.18e+01 & & 15@xmath12 & sma & 3 + lr + stellar t@xmath69 [ k ] & 4350 + log g & 4 + distance [ pc ] & 120 + inclination [ deg ] & 50 + grain size distribution slope , p & @xmath103.5 + a@xmath112 [ @xmath0 m ] & 0.005 + a@xmath113 [ @xmath0 m ] & 3900 + [ fix_table ] @xmath60 [ au ] & 39.7 @xmath25 5.6 & 38.2 @xmath25 5.3 & 42.7 @xmath25 3.7 + @xmath57 & 22.7 @xmath25 1.6 & 25.6 @xmath25 1.4 & 27.3 @xmath25 2.1 + log(@xmath114)[g/@xmath92 ] & @xmath100.23 @xmath116 & @xmath100.31 @xmath25 0.31 & @xmath100.39 @xmath25 0.24 + @xmath53 & -0.49 @xmath25 0.24 & @xmath100.30 @xmath25 0.41 & @xmath100.81@xmath25 0.16 + @xmath117 & 0.014 @xmath25 0.003 & 0.018 @xmath25 0.002 & 0.019 @xmath25 0.002 + @xmath64 & 0.37 @xmath25 0.04 & 0.30 @xmath25 0.09 & 0.22 @xmath25 0.14 + @xmath74 [ au ] & & 6.9 @xmath25 1.6 & + log(@xmath73 ) & & @xmath102.0 @xmath25 0.4 & + @xmath78 [ au ] & & & 10.0 @xmath25 2.3 + @xmath79 [ au ] & & & 2.2 @xmath25 0.15 + @xmath118m@xmath5]@xmath120 & 19 & 13 & 14 + @xmath49 visibility ( range)@xmath121 & 1438 & 1227 & 2650 + @xmath49 sed ( range ) & 89109 & 2433 & 2733 + @xmath49 visibility ( adopted)@xmath122 & 36 & 21 & 28 + @xmath49 sed ( adopted ) & 103 & 24 & 31 + @xmath49 total ( adopted ) & 139 & 45 & 59 + | we present continuum high resolution submillimeter array ( sma ) observations of the transition disk object rxj1633.9 - 2442 , which is located in the ophiuchus molecular cloud and has recently been identified as a likely site of ongoing giant planet formation .
the observations were taken at 340 ghz ( 880 @xmath0 m ) with the sma in its most extended configuration , resulting in an angular resolution of 0.3@xmath1 ( 35 au at the distance of the target ) .
we find that the disk is highly inclined ( _ i _
@xmath2 50 deg ) and has an inner cavity @xmath225 au in radius , which is clearly resolved by our observations .
we simultaneously model the entire optical to millimeter wavelength spectral energy distribution ( sed ) and sma visibilities of rxj1633.9 - 2442 in order to constrain the structure of its disk .
we find that an empty cavity @xmath225 au in radius is inconsistent with the excess emission observed at 12 , 22 , and 24 @xmath0 m . instead
, the mid - ir excess can be modeled by either a narrow , optically thick ring at @xmath210 au or an optically thin region extending from @xmath27 au to @xmath225 au .
the inner disk ( r @xmath3 5 au ) is mostly depleted of small dust grains as attested by the lack of detectable near - ir excess .
we also present deep keck aperture masking observations in the near - ir , which rule out the presence of a companion up to 500 times fainter than the primary star ( in @xmath4-band ) for projected separations in the 5 - 20 au range .
we argue that the complex structure of the rxj1633.9 - 2442 disk is best explained by _ multiple _
planets embedded within the disk .
we also suggest that the properties and incidence of objects such as rxj1633.9 - 2442 , t cha , and lkca 15 ( and those of the companions recently identified to these two latter objects ) are most consistent with the runaway gas accretion phase of the core accretion model , when giant planets gain their envelopes and suddenly become massive enough to open wide gaps in the disk . |
in this section the definitions of the basic concepts and the notation to be used throughout this contribution shall be presented .
differentiable manifolds are denoted by italic capital letters @xmath0 and , to our purposes , all such manifolds will be connected causally orientable lorentzian manifolds of dimension @xmath1 .
the signature convention is set to @xmath2 .
@xmath3 and @xmath4 will stand respectively for the tangent and cotangent spaces at @xmath5 , and @xmath6 ( resp .
@xmath7 ) is the tangent bundle ( cotangent bundle ) of @xmath8 .
similarly the bundle of @xmath9-contravariant and @xmath10-covariant tensors of @xmath11 is denoted @xmath12 .
if @xmath13 is a diffeomorphism between @xmath11 and @xmath14 , the push - forward and pull - back are written as @xmath15 and @xmath16 respectively .
the hyperbolic structure of the lorentzian scalar product naturally splits the elements of @xmath3 into timelike , spacelike , and null , and as usual we use the term _ causal _ for the vectors ( or vector fields ) which are non - spacelike . to fix the notation
we introduce the sets : @xmath17 the simplest example ( leaving aside $ \r^+$ ) of causal tensors are the causal 1-forms ( $ \equiv \dp_{1}(v)$ ) \cite{s - e}\footnote{see also bergqvist 's and senovilla 's contributions to this volume . } , while a general characterization of $ \dp^{+}_r\equiv \dp^{+}_r(v)$ is the following ( see \cite{sup } for a proof)$^1 $ : \begin{prop } $ { \bf t}\in \dp^{+}_r$ if and only if the components $ t_{i_1\dots i_r}$ of $ { \bf t}$ in all orthonormal bases fulfill $ t_{0\dots 0}\geq|t_{i_1\dots i_r}|$ , $ \forall i_1\dots i_r$ , where the $ 0$-index refers to the temporal component .
\label{orthonormal } \end{prop } % \p see \cite{sup}.\n we are now ready to present our main concept , which tries to capture the notion of some kind of relation between the causal structure of $ v$ and $ w$. \begin{defi } let $ \f : v\rightarrow w$ be a global diffeomorphism between two lorentzian manifolds .
we shall say that $ w$ is properly causally related with $ v$ by $ \f$ , denoted $ v\prec_{\f}w$ , if for every $ \x\in\z^{+}(v)$ we have that $ \f^{'}\x$ belongs to $ \z^{+}(w)$. $ w$ is said to be properly causally related with $ v$ , denoted simply as $ v\prec w$ , if $ \exists \f$ such that $ v\prec_{\f}w$. \label{prec } \end{defi } { \bf remarks } \begin{enumerate } \item this definition can also be given for any set $ \zeta\subseteq v$ by demanding that $ ( \f^{'}\x)_{\f(x)}\in \z^{+}(\f(x))\,\,\,\ , \forall \x\in\z^{+}(x)$ , $ \forall x\in\zeta$. \item two diffeomorphic lorentzian manifolds may fail to be properly causally related as we shall show later with explicit examples .
\end{enumerate } \begin{defi } two lorentzian manifolds $ v$ and $ w$ are called causally isomorphic if $ v\prec w$ and $ w\prec v$. this shall be written as $ v\sim w$. \label{equiv } \end{defi } we claim that if $ v\sim w$ then their causal structure are somehow the same
. let $ \g$ and $ \tilde{\g}$ be the lorentzian metrics of $ v$ and $ w$ respectively . by using \be \tilde{\g}(\f^{'}\x,\f^{'}\y)=\f^{*}\tilde{\g}(\x,\y ) ,
\label{pull - back } \ee we immediately realize that $ v\prec_{\f } w$ implies that $ \f^{*}\tilde{\g}\in\dp^{+}_2(v)$. conversely , if $ \f^{*}\t\in\dp^{+}_2(v)$ then for every $ \x\in\z^{+}(v)$ we have that $ ( \f^{*}\t)(\x,\x)=\t(\f^{'}\x,\f^{'}\x)\geq 0 $ and hence $ \f^{'}\x\in\z(w)$. however , it can happen that $ \z^{+}(v)$ is actually mapped to $ \z^-(w)$ , and $ \z^{-}(v)$ to $ \z^+(w)$. this only means that $ w$ { \em with the time - reversed orientation } is properly causally related with $ v$. keeping this in mind , the assertion $ \f^{*}\t\in\dp^{+}_2(v)$ will be henceforth taken as equivalent to $ v\prec_{\f}w$. \section{mathematical properties } let us present some mathematical properties of proper causal relations .
\begin{prop } if $ v\prec_{\f}w$ then : \begin{enumerate } \item $ \x\in\z^{+}(v)$ is timelike $ \longrightarrow$ $ \f^{'}\x\in \z^{+}(w)$ is timelike .
\item $ \x\in\z^{+}(v)$ and $ \f^{'}\x\in \z^{+}(w)$ is null $ \longrightarrow$ $ \x$ is null .
\end{enumerate } \label{caus } \end{prop } \p for the first implication , if $ \x\in\z^{+}(v)$ is timelike we have , according to equation ( \ref{pull - back } ) , that $ \f^{*}\tilde{\g}(\x,\x)=\t(\f^{'}\x,\f^{'}\x)$ which must be a strictly positive quantity as $ \f^{*}\tilde{\g}\in\dp^{+}_2(w)$ \cite{s - e}. for the second implication , equation ( \ref{pull - back } ) implies $ 0=\f^{*}\t(\x,\x)$ which is only possible if $ \x$ is null since $ \f^{*}\t\in\dp^{+}_2(v)$ and $ \x\in\z^{+}(v)$ ( see again \cite{s - e}).\n \begin{prop } $ v\prec_{\f}w \hspace{2 mm } \longleftrightarrow \hspace{2 mm } \f^{'}\x\in\z^{+}(w)$ for all null $ \x\in\z^{+}(v)$. \label{null }
\end{prop } \p for the non - trivial implication , making again use of ( \ref{pull - back } ) we can write : \
[ \f^{'}\x\in\z^{+}(w)\ \forall\x\ \mbox{null in}\ \z^{+}(v)\leftrightarrow\f^{*}\tilde{\g}(\x,\y)\geq 0\
\ \forall\ \x,\y\ \mbox{null in}\ \z^{+}(v ) \ ] which happens if and only if $ \f^{*}\tilde{\g}$ is in $ \dp^{+}_2(v)$ ( see \cite{s - e } property 2.4).\n \begin{prop}[transitivity of the proper causal relation]\hspace{0.1 cm } \\ if $ v\prec_{\f } w$ and $ w\prec_{\psi } u$ then $ v\prec_{\psi\circ\f}u$ \label{order } \end{prop } \p consider any $ \x\in \z^{+}(v)$. since $ v\prec_{\f } w$ , $ \f^{'}\x\in \z^{+}(w)$ and since $ w\prec_{\psi } u$ we get $ \psi^{'}[\f^{'}\x]\in\z^{+}(u)$ so that $ ( \psi\circ\f)^{'}\x\in\z^{+}(u)$ from what we conclude that $ v\prec_{\psi\circ\f } u$.\n therefore , we see that the relation $ \prec$ is a preorder .
notice that if $ v\sim w$ ( that is $ v\prec w$ and $ w\prec v$ ) this does not imply that $ v = w$. nevertheless , one can always define a partial order for the corresponding classes of equivalence .
next , we identify the part of the boundary of the null cone which is preserved under a proper causal relation .
a lemma is needed first .
recall that $ \x$ is called an `` eigenvector '' of a 2-covariant tensor $ { \bf t}$ if $ { \bf t}(\cdot , \x ) = \lambda \g ( \cdot , \x ) $ and $ \lambda$ is then the corresponding eigenvalue .
\begin{lem } if $ { \bf t}\in \dp^{+}_2(x)$ and $ \x\in\z^{+}(x)$ then $ { \bf t}(\x,\x)=0\ \longleftrightarrow\x$ is a null eigenvector of $ { \bf t}$. \label{null - eigen } \end{lem } \p
let $ \x\in\z^{+}(x)$ and assume $ 0={\bf t}(\x,\x)=t_{ab}x^{a}x^{b}$. then since $ t_{ab}x^{b}\in\dp^{+}_1(x)$ \cite{s - e } we can conclude that $ x_a$ and $ t_{ab}x^{b}$ must be proportional which results in $ x^{a}$ being a null eigenvector of $ t_{ab}$. the converse is straightforward.\n \begin{prop } assume that $ v\prec_{\f } w$ and $ \x\in\z^{+}(x),\ x\in v$. then $ \f^{'}\x$ is null at $ \f(x)\in w$ if and only if $ \x$ is a null eigenvector of $ \f^{*}\tilde{\g}(x)$. \label{cone } \end{prop } \p let $ \x$ be in $ \z^{+}(x)$ and suppose $ \f^{'}\x$ is null at $ \f(x)$. then , according to proposition \ref{caus } , $ \x$ is also null at $ x$. on the other hand we have $ 0=\tilde{\g}(\f^{'}\x,\f^{'}\x)=\f^{*}\tilde{\g}(\x,\x)$ and since $ \f^{*}\t|_{x}\in\dp^{+}_2(x)$ , lemma \ref{null - eigen } implies that $ \x$ is a null eigenvector of $ \f^{*}\t$ at $ x$.\n the vectors which remain null under the causal relation $ \f$ are called its { \em canonical null directions}. on the other hand , the null eigenvectors of $ { \bf t}\in\dp^{+}_2 $ can be used to classify this tensor , as proved in \cite{s - e}. as a result we have \begin{prop } if the relation $ v\prec_{\f}w$ has n linearly independent canonical null directions then $ \f^{*}\t=\lambda\g$. \label{conf } \end{prop } \p if there exist $ n$ independent canonical null directions , then $ \f^{*}\t$ has $ n$ independent null eigenvectors which is only possible if $ \f^{*}\t$ is proportional to the metric tensor $ \g$ ( \cite{s - e , sup}.)\n proposition \ref{conf } has an interesting application in the following theorem \begin{theo } suppose that $ v\prec_{\f}w$ and $ w\prec_{\f^{-1}}v$. then $ \f^{*}\t=\lambda\g$ and $ ( \f^{-1})^{*}\g=\fr{1}{(\f^{-1})^{*}\lambda}\t$ for some positive function
$ \lambda$ defined on $ v$. \label{inv } \end{theo } \p under these hypotheses , using proposition \ref{caus } , we get the following intermediate results \bea
\f^{'}\x\in \z^{+}(w)\ \mbox{null and $ \x\in\z^{+}(v)$}\longrightarrow\x\ \mbox{is null,}\nonumber\\ ( \f^{-1})^{'}\y\in\z^{+}(v)\ \mbox{null and $ \y\in\z^{+}(w)$}\longrightarrow\y\ \mbox{is null.}\nonumber \eea now , let $ \x\in \z^{+}(v)$ be null and consider the unique $ \y\in t(v)$ such that $ \x=(\f^{-1})^{'}\y$. then $ \y=\f^{'}\x$ and $ \y\in\z^{+}(w)$
as $ \f$ sets a proper causal relation and $ \x$ is in $ \z^{+}(v)$. hence , according to the second result above $ \y$ must be null and we conclude that every null $ \x\in \z^{+}(v)$ is push - forwarded to a null vector of $ \z^{+}(w)$ which in turn implies that $ \f^{*}\t=\lambda\g$. in a similar fashion , we can prove that $ ( \f^{-1})^{*}\g=\mu\t$ and hence $ ( \f^{-1})^{*}\lambda = 1/\mu$.\n \begin{coro } $ v\prec_{\f}w$ and $ w\prec_{\f^{-1}}v \,\ , \longleftrightarrow \,\ , \f$ is a conformal relation . \end{coro } \section{applications to causality theory } in this section we will perform a detailed study of how two lorentzian ma-\\ nifolds $ v$ and $ w$ such that $ v\prec_{\f}w$ share common causal features . to begin with
, we must recall the basic sets used in causality theory , namely $ i^{\pm}(p)$ and $ j^{\pm}(p)$ for any point
$ p\in v$ ( these definitions can also be given for sets ) .
one has $ q\in j^+(p)$ ( respectively $ q\in i^+(p)$ ) if there exists a continuous future directed causal ( resp.\ timelike ) curve joining $ p$ and $ q$. recall also the cauchy developments $ d^{\pm}(\zeta)$ for any set
$ \zeta\subseteq v$ \cite{ff , w , cond}. another important concept is that of future set : $ \a\subset v$ is said to be a future set if $ i^{+}(\a)\subseteq \a$. for example $ i^{+}(\zeta)$ is a future set for any $ \zeta$. all these concepts are standard in causality theory and are defined in many references , see for instance \cite{ff , w , cond}. \begin{prop } if $ v\prec_{\f } w$ then , for every set $ \zeta\subseteq v$ ,
we have $ \f(i^{\pm}(\zeta))\subseteq i^{\pm}(\f(\zeta))$ and $ \f(j^{\pm}(\zeta))\subseteq j^{\pm}(\f(\zeta))$. \label{set } \end{prop } \p it is enough to prove it for a single point $ p\in v$ and then getting the result for every $ \zeta$ by considering it as the union of its points . for the first relation , let $ y$ be in $ \f(i^{+}(p))$ arbitrary and take $ x\in i^{+}(p)$ such that $ \f(x)=y$. choose a future - directed timelike curve $ \g$ joining $ p$ and $ x$. from proposition \ref{caus } , $ \f(\g)$ is then a future - directed timelike curve joining $ \f(p)$ and $ y$ , so that $ y\in i^{+}(\f(p))$. the second assertion is proven in a similar way using again proposition \ref{caus}. the proof for the past sets is analogous.\n the converse of this proposition does not hold in general unless we impose some causality conditions on the spacetime .
\begin{defi } a lorentzian manifold $ v$ is said to be distinguishing if for every neighbourhood $ u_p$ of $ p\in v$ there exist another neighbourhood $ b_p\subset u_p$ containing $ p$ which intersects every causal curve meeting $ p$ in a connected set .
\label{distinguishing } \end{defi } we need some concepts of standard causality theory . for
any $ p\in v$ one can introduce normal coordinates in a neighbourhood $ { \cal n}_p$ of $ p$ ( see , e.g. \cite{cond } ) . then the exponential map provides a diffeomorphism $ \exp:{\cal o}\subset t_p(v)\rightarrow { \cal n}_p$ where $ { \cal o}$ is an open neighbourhood of $ \vec{0}\in t_p(v)$. the interior of the future ( past ) light cone of $ p$ is defined by $ c^{\pm}_p=\exp(\mbox{int}(\z^{\pm}(p))\cap { \cal o})$ , and obviously $ c^{\pm}_p\subseteq i^{\pm}(p)$ \cite{cond}. other important issue deals with the chronology relation $ < \!\!<$ between two points .
we have $ p<\!\!<q$ if there exist a future timelike curve joining $ p$ and $ q$. see \cite{kronheimer } for an axiomatic study of this relation .
\begin{prop } let $ \g$ be a piecewise continuous curve of a distinguishing lorentzian manifold $ ( v,\g)$. then , $ \g$ is total with respect to $ < \!\!<$ if and only if $ \g$ is timelike .
\end{prop } \p clearly if $ \g$ is timelike then $ \g$ must be a total set for the relation $ <
\!\!<$ ( this is true for every spacetime ) . for the converse consider a curve $ \g$ which is total with respect to $ < \!\!<$ and
let $ q\in\g$ be an arbitrary point of the curve .
if we take a normal neighbourhood of $ q$ , $ { \cal n}_q$ then we can find a neighbourhood $ u_q$ of $ q$ which is intersected in a connected set by every causal curve meeting $ q$. now , if we pick up a point $ z\in\g\cap u_q$ we have that either $ q<\!\!<z$ or $ z<\!\!<q$. assuming the former we deduce that there exists a timelike curve $ \tilde{\g}$ joining $ q$ and $ z$ which implies that $ \tilde{\g}\cap u_q$ is a connected set . this property
together with the distinguishability of $ v$ implies that $ \tilde{\g}$ must be a subset of $ u_q$ and hence $ \tilde{\g}\subset { \cal n}_q$ from what we conclude that $ \tilde{\g}\subset c_p$ ( \cite{cond } ) and hence $ z\in c_p$ $ \forall
z\in\g\cap u_q$ which is only possible if $ \g\cap u_q$ is timelike . by covering $ \g$ with sets of the form $ \g\cap u_q$ , $
q\in \g$ we arrive at the desired result.\n \begin{prop } let $ \f : v\rightarrow w$ be a diffeomorphism with the property $ \f(i^{+}(p))\subseteq i^{+}(\f(p))\ \forall p\in v$. then if $ w$ is distinguishing , $ \f$ is a proper causal relation .
a similar result holds replacing $ i^{+}$ by $ i^{-}$. \label{chronological } \end{prop } \p from the statement of this proposition is clear that $ \forall$ $ p , q$ of $ v$ such that $ p<\!\!<q$ then $ \f(p)<\!\!<\f(q)$. therefore every timelike curve $ \g$ of $ v$ is mapped onto a continuous curve in $ w$ total with respect to $ < \!\!<$ and hence timelike due to the distiguishability of $ w$. furthermore if the curve $ \g$ is future directed then $ \f(\g)$ must be also future directed since $ < \!\!<$ is preserved which is only possible if every timelike future - pointing vector is mapped onto a future - pointing timelike vector . as a consequence ,
if $ \k$ is a null vector , $ \f^{'}\k$ must be a causal vector ( to see it just construct a sequence of timelike future directed vectors converging to $ \k$ ) which proves that $ \f$ is a proper causal relation.\n the results for the cauchy developments are the following : \begin{prop } if $ v\prec_{\f } w$ then $ d^{\pm}(\f(\zeta))\subseteq \f(d^{\pm}(\zeta ) ) , \,\ , \forall \zeta\subseteq v$. \label{cauch } \end{prop } \p it is enough to prove the future case .
let $ y\in d^{+}(\f(\zeta))$ arbitrary and consider any causal past directed curve $ \g^{-}_{\f^{-1}(y)}\subset v$ containing $ \f^{-1}(y)$. since the image curve by $ \f$ of $ \g^{-}_{\f^{-1}(y)}$ is a causal curve passing through $ y$ , ergo meeting $ \f(\zeta)$ , we have that $ \g^{-}_{\f^{-1}(y)}$ must meet $ \zeta$ from what we conclude that $ y\in\f(d^{+}(\zeta))$ due to the arbitrariness of $ \g^{-}_{\f^{-1}(y)}$.\n \begin{coro } if $ \s\subset w$ is a cauchy hypersurface then $ \f^{-1}(\s)$ is also a cauchy hypersurface of $ v$. \label{hyp } \end{coro } \p if $ \s$ is a cauchy hypersurface then $ d(\s)=w$ , and from proposition \ref{cauch } $ d(\s)\subseteq\f(d(\f^{-1}(\s)))$. since $ \f$ is a diffeomorphism the result follows.\n one can prove the impossibility of the existence of proper causal relations sometimes .
for instance , from the previous corollary we deduce that $ v\prec w$ is impossible if $ w$ is globally hyperbolic but $ v$ is not .
other impossibilities arise as follows .
let us recall that , for any inextendible causal curve $ \g$ , the boundaries $ \partial i^{\pm}(\g)$ of its chronological future and past are usually called its future and past event horizons , sometimes also called particle horizons \cite{ff , w , cond}. of course these sets can be empty ( then one says that $ \g$ has no horizon ) .
\begin{prop } suppose that every inextendible causal future directed curve in $ w$ has a non - empty $ \partial i^{-}(\g)$ ( $ \partial i^{+}(\g)$ ) .
then any v such that $ v\prec w$ can not have inextendible causal curves without past ( future ) event horizons .
\label{hor } \end{prop } \p if there were a future - directed curve $ \g$ in $ v$ with $ \partial i^{-}(\g)=\emptyset$ , $ i^{-}(\g)$ would be the whole of $ v$. but according to proposition \ref{set } $ \f(i^{-}(\g))\subseteq i^{-}(\f(\g))$ from what we would conclude that $ i^{-}(\f(\g))=w$ against the assumption .
\n the class of future ( or past ) sets characterize the proper causal relations for distinguishing spacetimes as it is going to be shown next ( every statement for future objects has a counterpart for the past ) .
\begin{lem } if $ \a$ is a future set then $ p\in\overline{\a}\longleftrightarrow i^{+}(p)\subseteq\a$. \label{closure } \end{lem } \p suppose $ i^{+}(p)\subseteq a$. then since $ c_p^{+}\subseteq i^{+}(p)$ and $ p\in \overline{c_p^{+}}$ we have that $ u_p\cap c_p^{+}\neq\emptyset$ for every neighbourhood $ u_p$ of $ p$ which in turn implies that $ u_p\cap\a\neq\emptyset$ and hence $ p\in\overline{\a}$. conversely , let $ p$ be any point of $ \overline{\a}$ then $ i^+(p)\subseteq i^+(\overline{a})=i^+(a)\subseteq a$.\n \begin{theo } suppose that $ ( w,\t)$ is a distinguishing spacetime .
then a diffeomorphism $ \f:(v,\g)\rightarrow ( w,\t)$ is a proper causal relation if and only if $ \f^{-1}(\a)$ is a future set for every future set . $ \a\subseteq w$. \label{key } \end{theo } \p suppose $ \a\subseteq w$ is a future set , $ v\prec_{\f } w$ and take $ \f^{-1}(\a)\subseteq v$. proposition \ref{set } implies $ \f(i^{+}(\f^{-1}(\a)))\subseteq i^{+}(\f(\f^{-1}(\a)))=i^{+}(\a)\subseteq\a$ which shows that $ i^{+}(\f^{-1}(\a))\subseteq \f^{-1}(\a)$. conversely , for any $ p\in v$ take the future set $ i^{+}(\f(p))$ and consider the future set $ \f^{-1}(i^{+}(\f(p)))$. as $ \f(p)\in\overline{i^{+}(\f(p))}$ then $ p\in\overline{\f^{-1}(i^{+}(\f(p)))}$ and according to lemma \ref{closure } $ i^{+}(p)\subseteq\f^{-1}(i^{+}(\f(p)))$ so that $ \f(i^{+}(p))\subseteq i^{+}(\f(p))$. since this holds for every $ p\in v$ and $ w$ is distinguishing , proposition \ref{chronological } ensures that $ \f$ is a proper causal relation.\n this theorem has important consequences .
\begin{prop } if $ v\sim w$ and both manifolds are distinguishing , then there is a one - to - one correspondence between the future ( and past ) sets of $ v$ and $ w$. \label{conserv } \end{prop } \vspace{-0.5 cm } \p if $ v\sim w$ then $ v\prec_{\f}w$ and $ w_{\prec\psi}v$ for some diffeomorphisms $ \f$ and $ \psi$. by denoting with $ { \cal f}_v$ and $ { \cal f}_w$ the set of future sets of $ v$ and $ w$ respectively , we have that $ \f^{-1}({\cal f}_w)\subseteq{\cal f}_v$ and $ \psi^{-1}({\cal f}_v)\subseteq{\cal f}_w$ , due to theorem \ref{key}. since both $ \f$ and $ \psi$ are bijective maps we conclude that $ { \cal f}_v$ is in one - to - one correspondence with a subset of $ { \cal f}_w$ and vice versa which , according to the equivalence theorem of bernstein \cite{bern } , implies that $ { \cal f}_v$ is in one - to - one correspondence with $ { \cal f}_w$.\n \section{causal transformations } in this section we will see how the concepts above generalize , in a natural way , the group of conformal transformations in a lorentzian manifold $ v$. \begin{defi } a transformation $ \f : v\longrightarrow v$ is called { \em causal } if $ v\prec_{\f}v$. \label{gg } \end{defi } the set of causal transformations of $ v$ will be denoted by $ \c(v)$. this is a subset of the group of transformations of $ v$ which is closed under the composition of diffeomorphisms , due to proposition \ref{order } , and contains the identity map .
this algebraic structure is well - known , see e.g. \cite{semigroup } , and called subsemigroup with identity or submonoid .
thus , $ \c(v)$ is a { \em submonoid } of the group of diffeomorphisms of $ v$. nonetheless , $ \c(v)$ usually fails to be a group .
in fact we have , \begin{prop } every subgroup of causal transformations is a group of conformal transformations .
\label{group } \end{prop } \p let $ g\subseteq \c(v)$ be a subgroup of causal transformations and consider any $ \f\in g$ , so that both $ \f$ and $ \f^{-1}$ are causal transformations .
then $ \f$ is necessarily a conformal transformation as follows from theorem \ref{inv}.\n from standard results , see \cite{semigroup } , we know that $ \c(v)\cap\c(v)^{-1}$ is just the group of conformal transformations of $ v$ and there is no other subgroup of $ \c(v)$ which contains $ \c(v)\cap\c(v)^{-1}$. the causal transformations which are not conformal transformations are called proper causal transformations .
it is now a natural question whether one can define infinitesimal generators of one - parameter families of causal transformations which generalize the `` conformal killing vectors '' , and in which sense .
notice , however , that if $ \{\f_{s}\}_{s\in \r}$ is a one - parameter group of causal transformations , from the previous results the only possibility is that $ \{\f_{s}\}$ be in fact a group of conformal motions . on the other hand ,
things are more subtle if there are no conformal transformations in the family $ \{\f_s\}$ other than the identity , in which case it is easy to see that the ` best ' one can accomplish is that either $ g^{+}\equiv \{\f_{s}\}_{s\in \r^{+}}$ or $ g^{-}\equiv \{\f_{s}\}_{s\in \r^{-}}$ is in $ \c(v)$. if this happens one talks about maximal one - parameter submonoids of proper causal transformations .
of course , it is also possible to define local one - parameter submonoids of causal transformations $ \{\f_s\}_{s\in i}$ for some interval $ i=(-\epsilon,\epsilon)$ of the real line assuming that $ \{\f_s\}_{s\in ( 0,\epsilon)}$ consists of proper causal transformations . in any of these cases
, we can define the infinitesimal generator of $ \{\f_s\}$ as the vector field $ \xiv = d\f_{s}/ds|_{s=0}$. given that $ \f_{s}^{*}\g\in\dp_{2}$ for all $ s\geq 0 $ ( or for all non - positive $ s$ ) , one can somehow control the lie derivative of $ \g$ with respect to $ \xiv$. for instance , it is easy to prove that $ \lie\g ( \k,\k ) \geq 0 $ ( or $ \leq 0 $ ) for all null $ \k$ , clearly generalizing the case of conformal killing vectors .
an explicit example of this will be shown in the next section .
\section{examples } { \bf example 1 einstein static universe and de sitter spacetime . }
let us take $ v$ as the einstein static universe \cite{ff } and $ w=\ss$ as de sitter spacetime . in both cases
the manifold is $ \r\times s^{3}$ and hence they are diffeomorphic . by proposition \ref{hor } we know that $ v\not\prec w$ because every causal curve in de sitter spacetime possesses event horizons .
however , the proper causal relation in the opposite way does hold as can be shown by constructing it explicitly .
the line element of each spacetime is ( with the notation $ d\o^2=d\z^2+\sin^2\z d\phi^2 $ ) : \bea v : & & ds^{2}=dt^{2}-a^2(d\x^{2}+\sin^{2}\x d\o^{2})\ \nonumber\\ w : & & d\tilde{s}^{2}=d\bar{t}^{2}-\alpha^{2}\cosh^{2}(\bar{t}/\alpha ) ( d\bar{\x}^{2}+\sin^{2}\bar{\x } d\bar{\o}^{2})\ \nonumber \eea where $ \x,\z,\phi$ ( and their barred versions ) are standard coordinates in $ s^{3}$ and $ a,\alpha$ are constants .
the diffeomorphism $ \p : w\rightarrow v$ is chosen as $ \{t = b\bar{t } , \x={\bar\x } , \z={\bar\z } , \phi={\bar\phi}\}$ for a constant $ b$. one can readily get $ \p^{*}\g$ \begin{eqnarray * } ( \p^{*}\g)_{ab}dx^{a}dx^{b}= b^{2}d\bar{t}^2-a^2(d\bar{\x}^{2}+\sin^{2}\bar{\x } d\bar{\o}^{2})\end{aligned}\ ] ] which on using proposition [ orthonormal ] shows that @xmath18 if @xmath19 and therefore @xmath20 is proper causal relation for those @xmath21 . + consider the following spacetimes : @xmath22 is the region of lorentz - minkowski spacetime with @xmath23 in spherical coordinates @xmath24 ; @xmath25 is the outer region of schwarzschild spacetime with @xmath26 in schwarzschild coordinates @xmath27 .
define the diffeomorphism @xmath28 given by @xmath29 for an appropriate positive constant @xmath21 , so that we have @xmath30 by choosing @xmath21 and @xmath31 one can achieve @xmath32 _ whenever _ @xmath33 , while for @xmath34 @xmath13 fails to be a proper causal relation .
actually @xmath35 due to corollary [ hyp ] as @xmath36 is globally hyperbolic but @xmath22 is not .
take now the diffeomorphism @xmath37 defined by @xmath38 , so that @xmath39 reads @xmath40 from where we immediately deduce that @xmath41 for every @xmath42 as long as @xmath43 .
we have thus proved that @xmath44 if @xmath33 , but not for @xmath34 .
this is quite interesting and clearly related to the null character of the event horizon @xmath45 in schwarzschild s spacetime .
+ let us take as @xmath46 the flat friedman - robertson - walker ( frw ) spacetimes in standard frw coordinates @xmath47 with line element given by @xmath48 and assume that the source of einstein s equations is a perfect fluid with equation of state given by @xmath49 ( @xmath50 pressure , @xmath51 density , @xmath52 constant ) .
then the scale factor is @xmath53 with constant @xmath54 . by straightforward calculations
, it can be proven the following causal equivalences : w~_0 + w~v these causal equivalences are rather intuitive if we have a look at the penrose diagram of each spacetime ( figure [ steady ] ) .
, b ) @xmath55 and c ) @xmath56 .
notice the similar shape of diagram c ) with that of @xmath57 , and of the steady state part of @xmath58 with a ) and b ) @xcite . ]
* example 4 ( vaidya s spacetime . )
* let us show finally an example of a submonoid of causal transformations .
consider the vaidya spacetime whose line element is @xcite @xmath59 where @xmath60 is a null coordinate ( that is , @xmath61 is a null 1-form ) , and @xmath62 is a non - increasing function of @xmath60 interpreted as the mass .
take the diffeomorphisms @xmath63
. then @xmath64 can be cast in the form @xmath65 hence , @xmath66 iff @xmath67 , which implies that @xmath68 are causal transformations , so that @xmath69 is a maximal submonoid of causal transformations . the differential equation for the infinitesimal generator @xmath70 of this submonoid
is easily calculated and reads @xmath71 this is a particular case of a proper kerr - schild vector field , recently studied in @xcite .
notice that schwarzschild spacetime is included for the case @xmath72const .
, in which case @xmath73 is a killing vector .
this may lead to a natural generalization of symmetries .
in this work a new relation between lorentzian manifolds which keeps the causal character of causal vectors has been put forward . with the aid of this relation ,
we have introduced the concepts of causal relation and causal isomorphism of lorentzian manifolds which allow us to establish rigorously when two given lorentzian manifolds are causally indistinguishable regardless their metric properties .
this tools could be also useful in order to find out the global causal structure of a given spacetime by just putting it in causal equivalence with another known spacetime .
finally a new transformation for lorentzian manifolds , called causal transformation has been defined .
these transformations are a natural generalization of the group of conformal transformations and their actual relevance is one of our main lines of future research .
this research has been carried out under the research project upv 172.310-g02/99 of the university of the basque country .
99 g. bergqvist and j. m. m. senovilla _ null cone preserving maps , causal tensors and algebraic rainich theory .
quantum grav.*18 * , 5299 - 5326 , ( 2001 ) j. m. m. senovilla _ super - energy tensors . _
quantum grav .
* 17 * , 2799 - 2842 , ( 2000 ) .
s. w. hawking and g. f. r. ellis _ the large scale structure of spacetime . _ cambridge university press , cambridge , ( 1973 ) .
r. m. wald _ general relativity . _ the university of chicago press , ( 1984 ) .
j. m. m. senovilla _ singularity theorems and their consequences . _
. grav . * 30 * , 701 - 848 , ( 1998 ) .
e. h. kronheimer and r. penrose proc .
. soc . * 63 * 481 - 501 ( 1967 ) see e. g. f. hausdorff _ set theory _ chelsea n.y .
j. hilgert , k. h. hofmann and j. d. lawson _ lie groups , convex cones and semigroups .
_ oxford sciencie publications , ( 1989 ) .
vaidya _ proc .
indian acad .
_ * a*33 , 264 ( 1951 ) .
b. coll , s. r. hildebrandt and j. m. m. senovilla .
_ kerr schild symmetries .
* 33 * , 649 - 670 , ( 2001 ) . | in this work we define and study the relations between lorentzian manifolds given by the diffeomorphisms which map causal future directed vectors onto causal future directed vectors . this class of diffeomorphisms , called _ proper causal relations _ , contains as a subset the well - known group of conformal relations and are deeply linked to the so - called causal tensors of ref.@xcite .
if two given lorentzian manifolds are in _ mutual _ proper causal relation then they are said to be causally isomorphic : they are indistinguishable from the causal point of view .
finally , the concept of causal transformation for lorentzian manifolds is introduced and its main mathematical properties briefly investigated .
alfonso garca - parrado and jos m. m. senovilla 0.5 cm _ departamento de fsica terica .
universidad del pas vasco .
+ apartado 644 , 48080 bilbao , spain _ |
a crucial ingredient in quantum information processing based on solid state systems is the transfer of quantum information . assuming that there are quantum registers for computing and storing information , the ability to transfer this information reliably and efficiently from one register to another is vital for the construction of larger , distributed and networked systems .
a solution to this challenge has been proposed through the use of spin chains @xcite .
the mathematical framework underpinning spin chains can be applied to various physical devices ; these could be made of any components whose states can be mapped onto spin @xmath0 particles interacting with their neighbors .
electrons or excitons trapped in nanostructures form explicit examples @xcite , as do nanometer scale magnetic particles @xcite or a string of fullerenes @xcite .
another representation is the encoding into a soliton - like packet of excitations @xcite . within spin chains ,
a single - site excitation is defined as an up " spin in a system that is otherwise prepared to have all spins down " .
a discussion about unmodulated spin chains has been given in @xcite whereas in @xcite the couplings were chosen to be unequal .
there has also been research on wire - like chains with controlled coupling strength at either end @xcite and transfer through parallel spin chains @xcite , to name but a few closely related areas .
here we only consider linear spin chains whose coupling strength @xmath1 between two neighboring sites @xmath2 and @xmath3 has been pre - engineered to ensure perfect state transfer ( pst ) along the chain @xcite . for a chain of length @xmath4 with characteristic coupling constant @xmath5
, the pst coupling strength sequence is defined as @xcite @xmath6 for devices based on excitons in self - assembled quantum dots , @xmath5 is mainly governed by frster coupling @xcite , which in turn depends on the distance between the dots as well as the overlap between the electron and hole wavefunctions in each dot . in gate - defined quantum dots , however , @xmath5 will depend on tunnelling and thus on parameters such as the width and height of the barriers which separate the different dots , as well as on the overlap of electronic wavefunctions centered in different dots .
for chains of fullerenes or actual atoms @xmath5 will represent some hopping " parameter describing the propensity of the excitation to transfer from one site to the other .
the natural dynamics of a spin chain can then be described by a time independent hamiltonian as follows @xmath7.\end{aligned}\ ] ] in a perfect system ( to which perturbations will then be applied ) we will assume the single excitation energies @xmath8 to be independent of the site @xmath2 , and therefore only concentrate on the second term of eq .
( [ hami ] ) . in some physical systems such as quantum dot strings
, @xmath8 could naturally differ according to position , but may be tuned to be the same at all sites via application of local fields @xcite . the fidelity @xmath9 , corresponding to mapping the initial state @xmath10 over a time @xmath11 into the desired state @xmath12 by means of the chain natural dynamics , is given by @xmath13 and pst
is realized when the evolution is arranged to achieve @xmath14 .
we use the fidelity of state vectors to determine the transfer quality of information for unentangled states , as detailed for example in @xcite .
for entangled states , we measure instead the entanglement of formation ( eof ) as defined in ref .
+ the time evolution of a system is dependent on its characteristic coupling constant @xmath5 .
in particular , the time scale for pst from one end of a chain to the other , also known as _ mirroring time _ , is @xmath15 so that the periodicity of the system evolution is given by @xmath16 .
as the hamiltonian ( [ hami ] ) preserves the excitation number , the evolution of the initial state will remain within the original excitation subspace .
we will now consider the influence of general fabrication defects on linear spin chains with multiple excitations .
+ * ( a ) random noise * + we model the effect of fabrication errors ( random , but fixed in time ) for the energies and couplings in the system by adding to all non - zero entries in the hamiltonian matrix a random energy @xmath17 for @xmath18,@xmath19 .
the scale is fixed by @xmath20 which we set to 0.1 and for each @xmath21 the different random number @xmath22 is generated with a flat distribution between zero and unity . for the other side of the diagonal with @xmath23 , @xmath24 , preserving the hermiticity of the hamiltonian .
this method of including fabrication defects means that we could observe effects of a reasonable magnitude although clearly other distributions could also be modeled ; for specific tests , the weight of the noise would have to be determined according to the individual experiment being simulated .
+ * ( b ) site - dependent `` single - particle '' energies * + as a further possible fabrication defect , we consider the effect of the first term of ( [ hami ] ) that we previously dismissed under ideal conditions @xmath25 @xmath26 may represent external perturbations , such as local magnetic fields , or additional single site fabrication imperfections .
we thus assume here that @xmath8 is not independent of the site _
i _ any more .
+ * ( c ) excitation - excitation interactions * + in spin chains with multiple excitations , we also consider a perturbation term @xmath27 which represents the interaction between excitations in nearby sites .
for example , this may correspond to a biexcitonic interaction in quantum dot - based chains @xcite .
+ * ( d ) next - nearest neighbor interactions * + finally , we also investigate the effect of unwanted longer range interactions , which could be an issue when considering pseudospins based on charge degrees of freedom . for this
we add to ( [ hami ] ) the perturbative term @xmath28.\end{aligned}\ ] ] the expression for @xmath29 will depend on the type of interaction between spin chain sites . here
we explicitly consider three cases .
the first and more general approximates the next - nearest neighbor interaction as proportional to the average of the related interactions between nearest neighbor sites , @xmath30 with the parameter @xmath31 defining the strength of the interaction .
this expression simulates the original coupling modulation of the chain .
secondly we explicitly consider dipole - dipole interactions , which are relevant , e.g. to chains of quantum dots with exciton qubits and frster coupling @xcite . in this case
the coupling between sites scales as @xmath32 , with @xmath33 the distance between the two sites considered . for roughly equidistant sites
, we then expect the next - nearest neighbor couplings to be about a tenth of the nearest neighbor couplings . by using this and eq .
( [ pst ] ) we obtain @xmath34^{-\frac{1}{6}}+[(i+1)(n - i-1)]^{-\frac{1}{6}}\right\}^{-3}.\ ] ] finally we consider the case of coupling due to tunnelling , relevant e.g. to graphene quantum dots with spin qubits @xcite . here
the coupling scales as @xmath35 , with @xmath36 the on - site coulomb energy and @xmath37 the tunnelling ( hopping ) parameter , with @xmath38 the forbidden momentum in the barrier @xcite . in this case
there is no explicit expression for the next - nearest neighbor couplings in terms of @xmath39 , but @xmath40 can be determined numerically by using the expression for @xmath11 and eq .
( [ pst ] ) .
we note that we expect @xmath40 to be very small , as the interaction decays exponentially with the distance .
if we consider chains of quantum dots with exciton qubits , we can assume @xmath41 , where @xmath42 , and @xmath43 .
results in fig . [
fig : nnnicomp ] then show that this is in very good agreement with considering eq .
( [ ji+2 ] ) with @xmath44 , a value that one would expect from dipole - dipole interaction and basically equidistant sites ( see discussion above ) . in this case
the effect of next - nearest neighbor interaction is extremely detrimental to the system as even the fidelity of the first state transfer at @xmath45 is reduced by almost 50% . on the example of an 8-spin chain with initial input state @xmath46 , fidelity vs. rescaled time @xmath47 , according to both eq .
( [ ji+2 ] ) and eq .
( [ ji+2dip - dip ] ) ( dipole - dipole coupling ) .
the peak at @xmath48 is approximately 0.52 . as the model of eq .
( [ ji+2 ] ) matches the data derived from eq .
( [ ji+2dip - dip ] ) extremely well , the lines of the respective plots are nearly indistinguishable . + inset : as for main panel but for graphene quantum dots and spin qubits , with tunnelling coupling ( thin lines ) and coupling according to eq .
( [ ji+2 ] ) , @xmath49 ( bold lines ) .
again results from the two models for the coupling constants are almost indistinguishable from each other.,scaledwidth=45.0% ] by contrast , the tunnelling mechanism case leads to very different results . using the parameters in @xcite for graphene dots and spin qubits
, we obtain the results shown in the inset of fig .
[ fig : nnnicomp ] : after @xmath50 the system has lost less than 2% of its fidelity .
these results are in very good agreement with eq .
( [ ji+2 ] ) with @xmath51 as can be seen in the inset of fig.1 , where the plots are virtually indistinguishable .
it is very encouraging that in this case realistic parameters point to such small values of @xmath31 and thus generate very high fidelity transfer . in the following
we will use the expression in eq .
( [ ji+2 ] ) to further discuss the effects of @xmath52 . as values of @xmath31
smaller than 0.01 have a minor detrimental effect on the system , we will from here onwards focus on the range @xmath53 , where the upper bound may be of interest to some experimental implementations , e.g. as discussed in the dipole - dipole interaction case .
we will now consider the effect of the fabrication defects ( a ) to ( d ) first on the transfer of factorisable states , i.e. unentangled chains , and then on entanglement creation and entanglement transfer along a spin chain .
for unentangled states , we will consider a 6-spin chain for all investigations of fabrication defects while in the case of entanglement , we also consider an 8-spin chain . as we will explicitly see later for
@xmath26 , the influence of fabrication defects does depend on the chain length , but not on the parity of the chain .
the device we consider in this paper is a linear spin chain with couplings fixed such that the conditions for pst are satisfied .
one of the properties of these devices , which we make heavy use of , is the mirroring rule @xcite .
the mirroring rule is such that any state injected into a linear spin chain , subject to the constraints of eq . ( [ pst ] ) and site - independent @xmath8 , evolves into its `` mirror state '' , where the symmetry center of a chain is the middle point in chains with @xmath4 even , or middle site @xmath54 in chains with @xmath4 odd .
this mirroring property is independent of the number of excitations a state comprises and also of the length of the chain .
furthermore the mirroring rule holds for states spread across the whole chain : all excited sites are mirrored across to their `` twin '' with respect to the chain center of symmetry .
we call this mirrored state the `` twin state '' .
when investigating the quality of a device with fabrication defects of any sort , we are primarily interested in the effect on the `` twin state '' at @xmath55 but it is also important to consider the effects on the next few periods as some quantum information protocols may require periodicity of their systems @xcite . as can be seen in fig .
[ fig:6spinnoise ] , the effect of random noise leads to a continuous and definite decline in the transfer fidelity of both the input state @xmath56 and its twin state
@xmath57 at the mirroring times .
naturally , this means an increased probability of the occurrence of other possible states ( not shown on the graph ) but this happens on a relatively unpredictable basis , with no one state ever becoming and remaining particularly prominent . on a 6-spin chain with two excitations ,
fidelity vs. rescaled time @xmath47 .
states other than the input state and its twin state are not shown .
the first peak at @xmath58 is 0.9975.,scaledwidth=45.0% ] ( eq .
( [ interc ] ) ) on a 6-spin chain , fidelity vs. rescaled time @xmath47 .
+ ( a ) @xmath59 : the first peak at @xmath58 is 0.9988 . + ( b ) @xmath60 : the first peak at @xmath58 is 0.9954.,scaledwidth=45.0% ] in comparison , when looking at the evolution over a few periods , even a relatively large value of @xmath59 in eq .
( [ interc ] ) has less effect on the 6-spin chain , as shown in fig .
[ fig : interc ] ( a ) .
similarly , the on - site energies represented by eq .
( [ enerc ] ) also lead to an unrecoverable decay in the state transmission fidelity , with the excitations being ultimately entirely spread out along the chain ( not shown ) .
a combination of any of these three perturbation factors simply accelerates the decay trend of the desired states .
+ however , if we simply analyze the first revival peak at @xmath61 , we see that for varying @xmath62 and @xmath63 , even when perturbing the system by as much as 20% of @xmath5 , the system suffers less than a 10% loss in fidelity ( fig . [ fig:3d ] ) . here , to represent the fact that @xmath8 is site - dependent , we used a value @xmath64 for all sites _ i _ , with @xmath65 a random number generated from a flat distribution for each @xmath2 .
an average over 200 realizations for every value of @xmath63 was then taken . in a linear 6-spin chain , measured on the first revival peak at time @xmath61 , vs @xmath62 and @xmath63 .,scaledwidth=45.0% ]
finally , we also note that next - nearest neighbor interactions perturb the system similarly to noise .
this can be seen in fig .
[ fig:110000nnni ] where even for a reasonable value of @xmath66 , the transfer peaks of both the initial state of a 6-spin chain and its mirror twin quickly decay . in order to achieve the same fidelity loss at @xmath67 as for @xmath66 ( fig .
[ fig:110000nnni ] ) , we have to consider a value of @xmath62 _ one order of magnitude bigger _ , as is shown in fig .
[ fig : interc ] ( b ) .
( eq.([hami2 ] ) ) with @xmath66 : fidelity of the state @xmath68 in a 6-spin chain vs. rescaled time @xmath47 .
the first peak at @xmath58 is 0.9966.,scaledwidth=45.0% ] in our simulations we have kept the value of @xmath69 constant as @xmath4 is varied : this models the physical constraint that in any realistic system the coupling strength is capped by a maximum characteristic value . as a result @xmath70 for even ( odd ) chains . to avoid this implicit dependence on @xmath4
, we have here set @xmath71 ( where @xmath72 is a random number from a uniform distribution within 0 and 1 ) and averaged every point on the graph from 200 random realizations . as we see in part
( a ) of fig .
[ fig:110000_nnnienerc ] , the effect of @xmath52 becomes very detrimental to the system even for the relatively small value of @xmath73 , although the loss of fidelity for @xmath66 is very small even for long chains . on the other hand ,
the effect of on - site energies may be tolerable for values of @xmath63 up to @xmath74 , where long chains of @xmath75 suffer less than a 20% loss in fidelity .
a simple estimate of the effect of errors in the energy level spectrum @xcite suggests an overall error , or loss in fidelity for pst that scales as an exponential decay in @xmath4 with gaussian dependence on the characteristic noise parameters .
we compare this with numerical results in fig .
[ fig:110000_nnnienerc ] , where the loss of fidelity due to @xmath26 scales as @xmath76 and the loss due to @xmath52 scales as @xmath77 with increasing chain length @xmath4 , where @xmath78 and @xmath79 so that @xmath80 and @xmath81 characterize the impact of the noise .
the comparison in fig .
[ fig:110000_nnnienerc ] shows that this simple analytical form can reproduce the numerical results to a high degree of accuracy : the loss of fidelity scales indeed as an exponential decay with gaussian damping in the noise parameters . at @xmath61 vs. chain length @xmath4 for three values of @xmath31 , as labelled .
fits are according to @xmath82 with @xmath83 .
+ ( b ) average fidelity of initial vector @xmath84 at @xmath61 vs. chain length @xmath4 for three values of @xmath63 , as labelled .
fits are according to @xmath85 with @xmath86.,scaledwidth=48.0% ] we conclude from these results that the transfer of unentangled states across the device is very robust against the perturbations @xmath26 and @xmath87 , and less so for @xmath52 .
the effect of noise in the system , as we have implemented it , is slightly more noticeable at the transfer time @xmath88 but still allows for excellent transfer at a loss rate of just over 10% over the course of 4 periods . with regards to longer term periodicity , it is the next - nearest neighbor interaction term @xmath52 which perturbs the system most for the values of @xmath8 , @xmath62 and @xmath31 shown .
one of the most outstanding properties of spin chains is their ability to not just transfer reliably factorisable states , but also to transfer information encoded as entangled states .
this is again based on the mirroring rule @xcite and was first mentioned and discussed under various aspects in refs .
entanglement is one of the key resources in quantum computing , and is crucial to some quantum cryptography protocols , and to quantum teleportation .
being able to reliably transfer entanglement from one place to another is therefore a core interest that the device we are analyzing should be able to respond to .
this property follows from the fact that the set of states @xmath89 , @xmath90 to which the mirroring rule applies , is a basis set so that any state , and in particular entangled states , can be written as the superposition @xmath91 . here
@xmath92 is the total number of excitations in the state , and @xmath93 the ensemble of indices @xmath94 different from zero .
the above relation implies that after @xmath88 has passed the following twin state is reached : @xmath95 , so that in particular any entangled state is transferred into its mirror entangled state . when considering spin chains with ( i ) an entangled initial state or ( ii ) an initial state that leads to entanglement , it is more useful to observe the evolution of the eof in the system .
we use chains with an initial bell state on spins 1 and 2 to represent scenario ( i ) , as for example @xmath96 .
accordingly , we monitor the eof in the reduced density operator of spins @xmath97 , tracing out the rest of the chain . to monitor the
twin state " entanglement we calculate the eof for the reduced density operator of spins @xmath98 . scenario ( ii ) is different in that the initial state of the chain is not entangled , but will lead to entanglement through natural dynamics .
as an example of this , we use a linear chain with input @xmath99 on both spins 1 and @xmath4 .
this is equivalent to an initial state @xmath100 , where the subscripts designate the spin site .
a system set up in this way will then show a maximally entangled state in spins 1 and @xmath4 at time @xmath55 @xcite .
the effect of noise on entangled chains shows a similar trend to that of unentangled chains .
however , we notice that the loss of eof in fig . [
fig : gatenoise ] which represents case ( ii ) is nearly 10% bigger than the loss of eof in fig .
[ fig:0110noise ] ( case ( i ) ) over the course of the shown 7 periods . as a comparison , the unentangled state of fig .
[ fig:6spinnoise ] loses a similar amount of fidelity over the same amount of time as the chain in fig .
[ fig : gatenoise ] . on the entanglement of formation between the end spins of a linear 8-spin chain with input state
@xmath101 vs. rescaled time @xmath47 .
the first peak at @xmath58 is 0.9983.,scaledwidth=45.0% ] on the entanglement of formation between the end spins of an 8-spin chain with input state @xmath102 vs. rescaled time @xmath47 .
the first peak at @xmath58 is 0.9993.,scaledwidth=45.0% ] there is no effect from perturbation @xmath87 , ( eq .
( [ interc ] ) ) , in case ( i ) , as there is only one excitation in the system in both amplitudes .
obviously this would not be so for a bell state with a doubly excited amplitude .
similarly , the effect of @xmath87 in case ( ii ) is restricted to the two excitation subspace populated by the evolution of the @xmath103 component of @xmath10 only and is thus not very prominent , as is shown in fig .
[ fig : gateinterc ] .
( eq.([interc ] ) ) with @xmath59 on the entanglement of formation between the end spins of a linear 8 spin chain with input state @xmath101 vs. rescaled time @xmath47 .
the first peak at @xmath58 is 0.9996.,scaledwidth=45.0% ] the effect of on - site energies on the other hand remains , regardless of the system .
we demonstrate this in part ( b ) of figs .
[ fig:10 + 01_nnnienerc ] and [ fig : gate_nnnienerc ] , which show the detrimental effect of @xmath26 on entangled systems ( types ( i ) and ( ii ) respectively ) .
similar to fig .
[ fig:110000_nnnienerc ] , the loss in eof scales as an exponential in @xmath4 with gaussian damping in the noise parameters .
again , the influence of @xmath26 has been averaged over 200 random realization using random numbers @xmath8 from a flat distribution , such that @xmath104 and @xmath65 . ): + ( a ) eof of qubits 1 and 2 at @xmath61 vs. chain length @xmath4 for three values of @xmath31 , as labelled .
fits are according to @xmath82 with @xmath105 .
+ ( b ) average eof of qubits 1 and 2 at @xmath61 vs. chain length @xmath4 for three values of @xmath63 , as labelled .
fits are according to @xmath85 with @xmath106.,scaledwidth=48.0% ] ): + ( a ) eof of qubits 1 and @xmath4 at @xmath55 vs. chain length @xmath4 for three values of @xmath31 , as labelled .
fits are according to @xmath82 with @xmath107 .
+ ( b ) average eof of qubits 1 and @xmath4 at @xmath55 vs. chain length @xmath4 for three values of @xmath63 , as labelled .
fits are according to @xmath85 with @xmath108.,scaledwidth=48.0% ] figs .
[ fig:10 + 01_nnnienerc ] ( b ) and [ fig : gate_nnnienerc ] ( b ) also show that for @xmath109 , eof close to unity can still be achieved for all chain lengths considered . however , we note that for larger values of @xmath110 , chains with type ( ii ) entanglement suffer significantly more entanglement loss than those of type ( i ) which are already entangled at @xmath111 . as with unentangled states ,
next - nearest neighbor interaction is a relevant issue for entangled states . in fig .
[ fig:110000nnni ] , we showed that a 6-spin chain suffered serious fidelity loss ( @xmath112 ) for relatively small @xmath113 after about 4 periods , but still reached a fidelity of 0.9966 at @xmath88 ; similarly we see in fig .
[ fig : gate_nnnienerc ] that for the same value of @xmath31 , a 6-spin chain with case ( ii ) entanglement would reach eof of 0.99 at @xmath55 and performs thus equally well .
eof of almost unity persists for all @xmath4 considered and small @xmath31 .
more than 90% of the eof is maintained even for @xmath31 as large as 0.05 for long chains with type ( ii ) entanglement while chains with type ( i ) suffer significantly more and long chains lose over 30% of their eof for the same value of @xmath31 .
we note that for values of @xmath114 , in chains with type ( ii ) entanglement , whilst eof is fairly well maintained at @xmath115 , _ its subsequent periodicity is completely lost for any chain length _
[ fig : nnnicon ] , for a demonstration of the same effect in an unentangled chain see fig .
[ fig : nnnicomp ] ) . :
+ ( a ) eof of @xmath10 vs rescaled time @xmath47 for @xmath73 : the periodicity of the entanglement of formation is completely lost after the second peak at @xmath116 . + ( b ) eof of @xmath10 vs rescaled time @xmath47 for @xmath117 : the periodicity of the entanglement of formation is completely lost after the first peak at @xmath45.,scaledwidth=45.0% ] as noted above the effect of @xmath52 on entangled states is different from case ( i ) to case ( ii ) . in fig .
[ fig:0110nnni ] we see that for @xmath66 a state that is initially already entangled does not suffer very much and retains an eof of over 90% at @xmath118 , after 4 periods , whereas a chain where the entanglement is created through the system dynamics suffers a loss in eof of about 15% after 4 periods , at @xmath119 , as shown in fig .
[ fig : gatennni ] .
( [ hami2 ] ) ) with @xmath66 on the entanglement of formation between the end spins of an 8 spin chain with input state @xmath120 vs. rescaled time @xmath47 .
the first peak at @xmath58 is 0.9986.,scaledwidth=45.0% ] ( eq .
( [ hami2 ] ) ) with @xmath66 on the entanglement of formation between the end spins of a linear 8 spin chain with input state @xmath101 vs. rescaled time @xmath47 .
the first peak at @xmath58 is 0.9976.,scaledwidth=45.0% ] overall , we observe that the perturbative influence of next - nearest neighbor interaction ( eq .
( [ hami2 ] ) ) is the main cause for loss in fidelity in both unentangled states as well as states whose entanglement results only from their dynamics and which are initially unentangled , although unentangled chains suffer slightly more .
entangled chains with an initially entangled state on the other hand have been shown to be more robust against this type of defect . despite these variations , our study clearly demonstrates that next - nearest neighbor interactions are the most damaging form of perturbation overall , as seen in state transfer fidelity or eof .
the reason for this is that the next - nearest neighbor interaction is the only fabrication defect or limitation in the set ( a)-(d ) that effectively opens up new ` channels ' for the system dynamics .
the hamiltonian @xmath52 of eq .
( [ hami2 ] ) connects chain sites which would otherwise be disconnected ( at the same order in perturbation ) .
it therefore facilitates a more efficient ( in a detrimental sense ) ` spread ' of excitations .
the general consequence of this is more damage to transfer fidelity or eof , when compared to defects ( a)-(c ) with the same level of noise .
numerical simulations ( not shown ) support this explanation , as addition of new perturbative ` channels ' for the dynamics ` by hand ' ( as opposed to via @xmath52 ) into the full hamiltonian can lead to similar results to those in , for example , fig . [ fig : nnnicon ] .
it has been shown in ref .
@xcite that opening new channels even beyond next - nearest neighbor interaction can be compensated for if local control within the chain is possible by adjusting the nearest neighbor coupling .
this degree of local control may not always be available and/or might not be desirable , so in our work we consider the longer range interactions as a potential perturbation on nearest neighbor systems designed to produce pst .
in this section we will discuss the effect of imperfect excitation injection into the device when multiple excitation states are considered .
a possible device configuration for input / output of multiple excitations is sketched in fig .
[ diag ] , where each site in the chain is associated with a register which can act as input / output device .
we assume that there exists a clock to which the machinery at both ends of the chain have reference .
( without such a clock even simple pst could not operate , as extraction has to be timed with respect to injection . ) the timing errors we consider are with respect to this reference clock . and
@xmath121 registers and the information is transferred to the @xmath122 and @xmath123 via the mirroring rule.,scaledwidth=44.5% ] we will first consider the case of unentangled input states @xmath124 and then analyze the effects on entangled input states .
let us consider the injection and transfer of a state @xmath124 .
an important question to ask is how and to what extent a time delay in input operations would alter the transmission fidelity .
this question is particularly important for multiple excitations : for a single excitation a delayed injection would not alter the overall state evolution , but there would be merely an overall time shift . for multiple excitations
however , it may occur that during the preparation of a state @xmath124 , with @xmath125 , not all input sites are accessed at exactly the same time . as a consequence
a system which is supposed to be prepared in a two - excitation state may start evolving as a one excitation system if the two required excitations are not injected in synchrony .
+ let us focus on the latter case .
when considering such a delayed input there is a finite probability that the second spin @xmath38 we want to inject an excitation into is already occupied .
the result of this scenario is dependent on the injection mechanism , so we will consider the two main possibilities : ( i ) spin excitation via a rabi - flopping control pulse on spin @xmath38
applicable for example to systems in which excitations correspond to ground state excitons confined within a quantum dot chain or to flipping the spin of an electron already confined within the chain and ( ii ) injection via swap operation or injection of an additional particle in the chain .
the latter may correspond e.g. to the scenario in which the state in the qubit of register @xmath38 closest to the spin chain ( see fig . [ diag ] ) is swapped with the state in the chain site @xmath38
e.g. via a train of laser pulses in the case of exciton qubits ( see @xcite ) or to the scenario in which the main computation occurs via coherent electron transport in quantum wires ( such as in @xcite ) , each connected to a spin chain site . in case ( i )
injecting an excitation corresponds to applying a @xmath126-pulse using the hamiltonian @xmath127 with @xmath128 the rabi frequency ( we assume that qubits can be manipulated on an individual basis ) .
accordingly , at the delayed injection of the second excitation , the system state will evolve as @xmath129 with @xmath130 for @xmath131 and @xmath132 .
here we have explicitly displayed the set of indices @xmath93 .
the last term in eq .
( [ evol_1 ] ) corresponds to the error induced by having a non - zero probability @xmath133 of an excitation present in spin @xmath38 , and translates , after injection , in the probability of having _ no - excitations _
at all present in the chain ( see fig . [
fig : delay_6_rs](a ) ) .
if the desired dynamics is such that a certain site @xmath121 has unit probability to contain an excitation at a later time @xmath134 though , the system can be refocused by measuring site @xmath121 at @xmath134 : a result of excitation present " would collapse the system state into two - excitation dynamics , which is the closer to the desired dynamics the smaller the delay ; a result of no - excitation present " would imply that the chain contains indeed no excitation at all .
we underline that the latter result could be used as a protocol to reinitialize the chain itself . in case ( ii )
the scenario is very different .
should the first excitation already have a non - zero probability of occupying site @xmath38 , the second excitation will have a related probability of remaining in the register .
the register would then become entangled with the spin chain . the injection process can in fact
be described as follows ( where we assume that the presence of the excitation in site @xmath38 is the only cause of failed injection ) @xmath135 with @xmath130 for @xmath131 .
the simplest way to destroy this unwanted entanglement is to measure , after injection , whether the second excitation is still present in the register or wire . by this measure
, we remove the entanglement between register and device , but we also get to know exactly what state the device itself is in : if the measure outcome is no excitation in the register " the injection has been successful and the spin chain now follows a two excitation evolution which is the closer to the desired dynamics the smaller the injection delay has been . this is described in fig . [
fig : delay_6_rs](b ) .
if the outcome of the measurement is excitation in the register " the chain will continue to evolve as a one excitation system .
however in the latter case we have also re - set the chain ready for trying again the injection of the second excitation at the earliest convenient time .
delay vs. rescaled time @xmath47 : + ( a ) shows injection by rabi - flopping , where the resulting error remains in the system as the zero vector @xmath136 . at @xmath61 the state of the first spin is measured and an excitation is found .
this allows for refocusing the system in the two excitation subspace only .
the peak of @xmath56 at multiples of @xmath137 plus delay is 0.7807 .
if at @xmath61 the second spin is measured instead , the peak fidelity becomes 0.7203 .
+ ( b ) shows injection by swap operation where the error remains in the system in the one excitation subspace . at @xmath61
the state of the register is measured and no excitation is found .
this projects the system in the two excitation subspace only and disentangles the chain from the register .
the peak of @xmath56 at multiples of @xmath137 plus delay is 0.9870.,scaledwidth=45.0% ] in fig .
[ fig:6spin2exdelays ] , we give an overview of various possible delay scenarios , assuming input by swap operation with error correction by measurement of the injecting register or wire performed immediately after the attempted injection of the second excitation . in panel ( a ) , the delay of the second excitation is equal to @xmath138 .
even with the error correction measurement we see an impact on the system as the delay between the two injections puts a cap on the fidelity of the revivals of the desired two - excitation input state .
the larger the delay ( up to the mirror time @xmath88 ) , the more serious the impact .
frame ( b ) demonstrates that even if timely simultaneous injection were not possible , the ability to inject the second excitation an integer number of periods @xmath137 after the first one would allow for perfect revivals .
in contrast , the third frame ( c ) shows that a second injection exactly @xmath88 away from the ideal case would result in the total decay of the desired state .
however , we have to keep in mind that at @xmath55 where @xmath139 has the lowest fidelity , the corresponding twin state will be achieving perfect fidelity and that therefore a new state will emerge , we illustrate this in fig .
[ fig : newborn ] : at a time equal to an odd multiple of @xmath88 , the initial input state @xmath140 has zero fidelity , while its twin state @xmath141 has unit fidelity .
injection of the second excitation into the second site ( which would have resulted in the ideal state @xmath56 if there had been no delay ) results in @xmath142 .
this state then continues to evolve with its twin state @xmath143 , both of them alternately reaching perfect fidelity .
the additional features on the plots of vectors with two excitations , as well as the narrowing of the main peak are due to excitation of the particular superpositions of energy eigenstates in the two - excitation subspace that correspond to our initial state .
this shows how we can use delayed input to transfer states which are different from the `` twin '' sites of the input sites while still assuring pst
. this may be useful if e.g. not all sites are accessible for input operations . .
+ ( a ) input delayed by @xmath138 . + ( b ) input delayed by an integer number of @xmath137 ( here : @xmath144 ) .
+ ( c ) input delayed by an odd multiple of @xmath88 ( here : @xmath145 ) .
+ for ( a ) the maximum recurring fidelity of @xmath57 is 0.9313 , at odd integer multiples of @xmath146 plus the delay . for ( b ) , due to the complete period delay , @xmath57 emerges with unit fidelity , whereas for ( c ) the occurrence is negligible.,scaledwidth=45.0% ] ( here : @xmath147 ) vs. rescaled time @xmath47 .
the peak of @xmath148 at @xmath149 reaches unity.,scaledwidth=45.0% ] imperfect injection can occur even without delays between different excitations injections due to other device imperfections .
a typical example could be the possibility that in the swap injection scenario one of the two excitations is partially reflected into the wire .
this may occur in a device in which the spin chain is formed by gate - defined quantum dots where electrons are injected via ( computational ) wires , with the reflection being caused by imperfect lowering of a potential barrier between the wire and the selected spin chain site @xmath38 . in this case the injection would be described by @xmath150 where the first spin has been perfectly injected at site @xmath121 , while there is a reflection probability @xmath151 associated with spin @xmath38 . in this scenario measuring the absence ( presence ) of excitation in the register after injection would ensure that the chain is undergoing exactly the desired dynamics ( or that the chain is ready for re - injecting the second excitation at the closest suitable time ) .
the above scenario can be straightforwardly extended to the case in which a finite probability of reflection is associated with both injection sites .
+ while in this paper we are only discussing delays at the input stage , similar problems may arise of course at read - out , should the extraction of a state covering multiple sites not be as timely as we would hope it to be .
it is also worth noting that any peak following delayed input is shifted forward in time linearly with increasing delay . when considering imperfect or non - synchronous input and entanglement transfer , we see a similar effect to that on unentangled states .
again , as the overall desired system state is more complicated , we monitor the evolution of the amount of entanglement in the relevant sub - system , instead of individual vectors . in fig .
[ fig : gatedelay ] we show the evolution of the first eof peak after the delayed input has taken place , for case ( ii ) where entanglement is created from an initial product state . in this setting it
is not possible to implement the refocusing protocol for the rabi flopping , but we are assuming that the confirmed injection protocol for the swap operation is implemented instantaneously , just after the injection attempt . due to the fact that the two injection sites in the state described in this figure are at opposite chain ends , there is virtually no difference between injection by rabi flopping and injection by swap operation .
as long as the second injection is only delayed by a small fraction of @xmath88 , the first excitation will not have spread out far enough yet to affect the second injection site .
we see that even for delays as large as @xmath152 , the system retains over 90% of its possible entanglement . vs. input delay for an 8-spin chain with initial input state @xmath101 for both rabi flopping and swap operation type of injections.,scaledwidth=45.0% ]
when we consider two adjacent injection sites , there is instead a clear discrepancy between the decay in entanglement of formation in a chain with input via rabi flopping or swap operation ( fig [ fig:0110delay ] ) .
again , we assume that there is an instantaneous swap correction protocol applied , while refocussing after the rabi - flopping is not possible . as in the case of non - synchronous input in unentangled states , a delay in a swap type injection perturbs the system far less than using rabi flopping , and shows virtually no loss in the amount of entanglement in the system , even for delays as large as @xmath153 .
if instead rabi flopping is used , the decay in eof is similar to that of fig .
[ fig : gatedelay ] . for both swap operation and rabi flopping type injections.,scaledwidth=45.0% ]
in this paper we have considered a variety of physically relevant perturbative factors in spin chains .
we have investigated in detail their effects on information transfer and entanglement generation and transfer . for different forms of perturbation , the results of our extensive numerical studies on the quality of transfer can be captured with a straightforward analytic expression that demonstrates exponential damping with the number of spins on the chain and gaussian dependence on the relevant perturbation parameter .
this expression provides a simple tool for estimating the efficiency of a chain under the action of perturbations .
we have also related our dimensionless perturbation parameter scales to actual parameters for candidate spin chain realizations such as quantum dots with exciton qubits and graphene dots with spin qubits , providing a calibration of our estimator tool for these experimental systems .
we have considered the transport of both unentangled states as well as entangled states via spin chains , subject to various forms of perturbation , and a general conclusion is that next - nearest neighbor interactions are the most damaging to transfer fidelity or eof .
the reason for this can be traced back to the fact that the next - nearest neighbor interaction is the only perturbation that effectively opens up new ` channels ' for the system dynamics . by connecting sites which would otherwise be disconnected ( at the same order in perturbation )
, it allows a more efficient ( in a detrimental sense ) ` spread ' of excitations , and consequently drastically diminishes the occurrence of quantum coherent effects such as revivals and pst . in particular
, we have seen that the introduction of next - nearest neighbor coupling may lead to non - negligible quality loss after only a few periods .
systems with simpler input states are generally slightly less affected than those of a slightly more complex nature but as different perturbation factors affect different excitation subspaces , there is no clear advantage of one particular type of state in terms of robustness . however , for all fabrication defects considered we have found that the transport during the first period remains of high quality for perturbation amplitudes of the order of few percent , while the periodicity itself may be destroyed for next - nearest neighbor couplings above about @xmath154 .
this is of particular concern for schemes based on dipole - dipole interactions , where , for roughly equidistant sites , the next - nearest neighbor interactions are of the order of @xmath155 .
nevertheless , due to the gaussian dependence of transport efficiency on perturbation amplitudes , provided that the perturbations are kept below these few percent thresholds , spin chains are demonstrated to be very good candidates for the implementation of solid state quantum information processing devices , robust against various forms of perturbation .
additionally , we have seen that the effect of imperfect input or injection operations leads to a permanent ( but constant ) loss in information transport quality and entanglement generation . for input delayed by a time
approaching the mirroring time @xmath88 the transport of the intended state is replaced by a new input state , which is then subject to the system dynamics as per usual .
furthermore , there is a fundamental difference between the two input methods we have considered .
if rabi flopping is used and the injection of the second excitation is delayed , the error induced by the imperfect input remains in the system , except if a site is expected to be in state @xmath156 at a known time , when measurement of this site can recover dynamics close to the ideal case . in the case of input by swap operation
, the error can be dramatically reduced by subsequent measuring of the environment .
when considering swap type injection , unentangled states and type ( i ) entanglement states are remarkably robust against non - synchronous injection , with hardly any loss in fidelity even for large delays . type ( ii )
dynamically - generated entangled states are instead more affected due to their more complex set - up . even so the loss in eof remains less than 10% for delay values up to 10% of the mirroring time @xmath88 .
our studies demonstrate quantitatively the criteria that need to be met , in terms of perturbation scales and injection errors , for imperfect spin chains to work efficiently as quantum information transfer and entanglement transfer / generation devices . with modest errors at or below the few percent level , spin chains prove to be good and robust devices , which is very encouraging for future experiments on these systems .
+ rr was supported by epsrc - gb and hp .
ida acknowledges partial support by hp .
ida and rr acknowledge the kind hospitality of the hp research labs bristol . | spin chains have been proposed as a reliable and convenient way of transferring information and entanglement in a quantum computational context .
nonetheless , it has to be expected that any physical implementation of these systems will be subject to several perturbative factors which could potentially diminish the transfer quality . in this paper , we investigate a number of possible fabrication defects in the spin chains themselves as well as the effect of non - synchronous or imperfect input operations , with a focus on the case of multiple excitation / qubit transfer .
we consider both entangled and unentangled states , and in particular the transfer of an entangled pair of adjacent spins at one end of a chain under the mirroring rule and also the creation of entanglement resulting from injection at both end spins . |
it is now established that the collision of heavy - ions at the relativistic heavy - ion collider have led to the formation of an entirely new form of matter @xcite .
while the underlying degrees of freedom prevalent in the hot plasma are , as yet , unknown @xcite , various constraints may be imposed through a study of its partonic substructure .
the foremost tool in this study is the modification of the hard jets , usually referred to as jet quenching @xcite .
the number of hadrons with transverse momentum @xmath0 gev ( which , necessarily originate in the fragmentation of hard jets ) is reduced by almost a factor of 5 in central @xmath1-@xmath1 collisions , compared to that expected from elementary nucleon nucleon encounters enhanced by the number of expected binary collisions @xcite .
jet modification is a probe with a wide range of complexity in terms of experimental observables . by now
, measurements on single inclusive observables have been extended to very high @xmath2 ( @xmath3 gev ) .
there also exist , a large number of multi - particle jet - like correlation observables , photon - jet , jet - medium and heavy flavor observables @xcite . in these proceedings ,
we attempt a very brief review of the underlying theory and some of the new jet correlation observables which may be used to understand the underlying space - time and momentum space structure of the produced matter .
most current calculations of the in - medium modification of light partons may be divided into four major schemes , often referred to by the names of the original authors .
all schemes utilize a factorized approach , where the final cross section to produce a hadron @xmath4 with high transverse momentum @xmath2 and a pseudo - rapidity between @xmath5 and @xmath6 may be expressed as an integral over the product of the nuclear structure functions [ @xmath7 , to produce partons with momentum fractions @xmath8 , a hard partonic cross section to produce a hard parton with a transverse momentum @xmath9 and a medium modified fragmentation function for the final hadron [ @xmath10 , the modification of the partonic jet is encoded in the calculation of the medium modified fragmentation function .
the four schemes of energy loss are in principle a different set of approximation schemes to estimate this quantity from perturbative qcd calculations .
the reaction operator approach in opacity , often referred to as the gyulassy - levai - vitev ( glv ) scheme @xcite , assumes the medium to be composed of heavy , almost static , colour scattering centers ( with debye screened yukawa potentials ) which are well separated in the sense that the mean free path of a jet @xmath11 , the colour screening length of the medium .
the opacity of the medium @xmath12 quantifies the number of scattering centers seen by a jet as it passes through the medium , _
i.e. _ , @xmath13 , where @xmath14 is the thickness of the medium . at
leading order in opacity , a hard jet , produced locally in such a plasma with a large forward energy @xmath15 , scatters off one such potential and in the process radiates a soft gluon .
multiple such interactions in a poisson approximation are considered to calculate the probability for the jet to lose a finite amount of its energy .
the path integral in opacity approach , referred to as the armesto - salgado - wiedemann ( asw ) approach @xcite , also assumes a model for the medium as an assembly of debye screened heavy scattering centers .
a hard , almost on shell , parton traversing such a medium will engender multiple transverse scatterings of order @xmath16 .
it will in the process split into an outgoing parton and a radiated gluon which will also scatter multiply in the medium . the propagation of the incoming ( outgoing ) partons as well as that of the radiated gluon in this background colour field may be expressed in terms of effective green s functions , which are obtained in terms of path integrals over the field . also similar to the glv approach , a poisson approximation
is then used to obtain multiple emissions and a finite energy loss . in
the finite temperature field theory scheme referred to as the arnold - moore - yaffe ( amy ) approach @xcite , the energy loss of hard jets is considered in an extended medium in equilibrium at asymptotically high temperature @xmath17 ( and as a result @xmath18 ) . in this limit
, one uses the effective theory of hard - thermal - loops ( htl ) to describe the collective properties of the medium .
a hard on - shell parton undergoes soft scatterings with momentum transfers @xmath19 off other hard partons in the medium .
such soft scatterings induce collinear radiation from the parton , with a transverse momentum of the order of @xmath20 .
multiple scatterings of the incoming ( outgoing ) parton and the radiated gluon need to be considered to get the leading order gluon radiation rate .
this is obtained from the imaginary parts of infinite order ladder diagrams .
these rates are then used to evolve an initial distribution of hard partons through the medium in terms of a fokker - plank equation . in the higher - twist scheme @xcite ,
one directly computes the modification to the fragmentation functions due to multiple scattering in the medium by identifying and re - summing a class of higher twist contributions which are enhanced by the length of the medium .
the initial hard jet is assumed to be considerably virtual , with @xmath21 .
the propagation and collinear gluon emissions from such a parton are influenced by the multiple scattering in the medium .
one assumes that , on exiting the medium , the hard parton has a small , yet perturbative scale @xmath22 .
one evolves this scale back up to the hard scale of the original produced parton , @xmath23 , by including the effect of multiple emissions in the medium .
the multiple scatterings introduce a destructive interference for radiation at very forward angles and as such modify the evolution of the fragmentation functions in the medium .
in any scheme , the magnitude of the modification is controlled by a single space - time dependent parameter which may be related to the well known transport coefficient @xmath24 .
this is defined as the mean transverse momentum squared per unit length , transferred by the medium to the hard jet . in actual computations , a model of the space - time dependence
is invoked , and a maximum for @xmath24 is set to best fit with experimental results .
values of the maximum of @xmath24 , in the vicinity of the center of a central collision , at a time of @xmath25fm / c , range from 1 gev@xmath26/fm up to 20 gev@xmath26/fm @xcite , depending on the scheme , as well as , the model of the medium used .
while the suppression of single inclusive hadrons may be used to determine the maximum value of @xmath24 , corrrelations between the leading hadron and the medium may be used to test the space - time profile which is used as the ansatz @xcite . once considers the nuclear modification factor @xmath27 as a function of the angle with the reaction plane .
the results of such an analysis , from the higher twist approach , for a medium profile taken from a full @xmath28 hydrodynamics simulation at @xmath29 % centrality are presented in the left panel of fig .
the plots clearly demonstrate that the modification is maximal when the jet propagates through the thickest part of the medium in a direction perpendicular to the reaction plane .
the spread between the lines is directly dependent on the particular space - time ansatz chosen for the evolution of the produced matter . to determine the momentum structure of the plasma , one generalizes @xmath24 from a scalar to a tensor @xcite , where the scalar @xmath30 .
an example where such a situation may arise is in the presence of large turbulent colour fields , which may be generated in the early plasma due to anisotropic parton distributions @xcite .
these large fields , transverse to the beam , tend to deflect radiated gluons from a transversely traveling jet , preferentially , in the longitudinal directions .
while such effects influence the solid angle distributions of the radiated gluons around the originating parton , they do not have a considerable effect on the total energy lost .
such phenomena may yield an explanation for the ridge like structure seen in the near side correlations @xcite . in the right panel of fig .
[ fig1 ] , results from a quantitative estimate are presented where the trigger quark ( with @xmath31 gev ) radiates a 4 gev gluon , which is then subjected to such transverse fields ( over a distance of 3 fm ) . using a time dependent @xmath24 , with an initial value consistent with fits to experiment , one obtains a noticeable ridge like structure , _
i.e. _ , a considerable broadening in @xmath5 but not in @xmath32 .
99 k. adcox _ et al .
_ , nucl .
phys . a * 757 * , 184 ( 2005 ) ; j. adams _ et al . _ , nucl . phys . a * 757 * , 102 ( 2005 )
. e. v. shuryak and i. zahed , phys .
c * 70 * , 021901 ( 2004 ) ; phys .
d * 70 * , 054507 ( 2004 ) .
v. koch _ et .
* 95 * , 182301 ( 2005 ) ; a. majumder and b. muller , phys .
c * 74 * , 054901 ( 2006 ) .
m. gyulassy and m. plumer , phys .
b * 243 * , 432 ( 1990 ) ; x. n. wang and m. gyulassy , phys .
lett . * 68 * , 1480 ( 1992 ) ; r. baier _ et .
b * 483 * , 291 ( 1997 ) ; nucl .
b * 484 * , 265 ( 1997 ) .
k. adcox _ et al .
_ , phys . rev
. lett . * 88 * , 022301 ( 2002 ) ; c. adler _ et al .
_ , phys .
lett . * 89 * 202301 ( 2002 ) .
x. f. guo and x. n. wang , phys .
lett . * 85 * , 3591 ( 2000 ) ; nucl .
a * 696 * , 788 ( 2001 ) ; b. w. zhang and x .-
wang , nucl .
a * 720 * , 429 ( 2003 ) ; a. majumder _ et .
lett . * 99 * , 152301 ( 2007 ) .
a. majumder , j. phys .
g * 34 * , s377 ( 2007 ) .
a. majumder , phys .
c * 75 * , 021901 ( 2007 ) ; a. majumder _ et .
c * 76 * , 041902 ( 2007 ) ; t. renk _ et .
al . _ , phys .
c * 75 * , 031902 ( 2007 ) ; g. y. qin _ et .
c * 76 * , 064907 ( 2007 ) .
a. majumder _ et .
al . _ , phys .
lett . * 99 * , 042301 ( 2007 ) .
m. asakawa _ et .
lett . * 96 * 252301(2006 ) .
p. arnold _ et .
_ , phys .
d * 72 * , 054003 ( 2005 ) ; p. romatschke and m. strickland , phys .
d * 68 * , 036004 ( 2003 ) ; s. mrowczynski , phys .
b * 314 * ( 1993 ) 118 . | the use of jet modification to study the properties of dense matter is reviewed .
different sets of jet correlations measurements which may be used to obtain both the space - time and momentum space structure of the produced matter are outlined . |
chemical elements heavier than lithium are synthesized in stars .
such `` metals '' are observed at times when the universe was only @xmath6% of its current age in the inter galactic medium ( igm ) as absorption lines in quasar spectra ( see ellison et al . 2000 , and references therein ) .
hence , these heavy elements not only had to be synthesized but also released and distributed in the igm within the first billion years
. only supernovae of sufficiently short lived massive stars are known to provide such an enrichment mechanism .
this leads to the prediction that _ the first generation of cosmic structures formed massive stars ( although not necessarily only massive stars ) . _ in the past 30 years it has been argued that the first cosmological objects form globular clusters ( ) , super massive black holes ( ) , or even low mass stars ( ) . this disagreement of theoretical studies might at first seem surprising .
however , the first objects form via the gravitational collapse of a thermally unstable reactive medium , inhibiting conclusive analytical calculations .
the problem is particularly acute because the evolution of all other cosmological objects ( and in particular the larger galaxies that follow ) will depend on the evolution of the first stars .
nevertheless , in comparison to present day star formation , the physics of the formation of the first star in the universe is rather simple . in particular : * the chemical and radiative of processes in the primordial gas are readily understood . * strong magnetic fields
are not expected to exist at early times .
* by definition no other stars exist to influence the environment through radiation , winds , supernovae , etc . *
the emerging standard model for structure formation provides appropriate initial conditions . in previous work
we have presented three dimensional cosmological simulations of the formation of the first objects in the universe ( , ) including first applications of adaptive mesh refinement ( amr ) cosmological hydrodynamical simulations to first structure formation ( , , abn hereafter ) . in these studies we achieved a dynamic range of up to @xmath7 and
could follow in detail the formation of the first dense cooling region far within a pre
galactic object that formed self
consistently from linear density fluctuation in a cold dark matter cosmology . here
we report results from simulations that extend our previous work by another 5 orders of magnitude in dynamic range . for the first time it is possible to bridge the wide range between cosmological and stellar scale .
we employ an eulerian structured adaptive mesh refinement cosmological hydrodynamical code developed by bryan and norman ( , ) .
the hydrodynamical equations are solved with the second order accurate piecewise parabolic method (; ) where a riemann solver ensures accurate shock capturing with a minimum of numerical viscosity .
we use initial conditions appropriate for a spatially flat cold dark matter cosmology with 6% of the matter density contributed by baryons , zero cosmological constant , and a hubble constant of 50 km / s / mpc ( ) . the power spectrum of initial density fluctuations in the dark matter and the gas are taken from the computation by the publicly available boltzmann code cmbfast ( ) at redshift 100 ( assuming an harrison zeldovich scale invariant initial spectrum ) .
we set up a three dimensional volume with 128 comoving kpc on a side and solve the cosmological hydrodynamics equations assuming periodic boundary conditions .
this small volume is adequate for our purpose , because we are interested in the evolution of the first pre galactic object within which a star may be formed by a redshift of @xmath8 .
first we identify the lagrangian volume of the first proto galactic halo with a mass of @xmath9 in a low resolution
pure n body simulation .
then we generate new initial conditions with four initial static grids that cover this langrangian region with progressively finer resolution . with a @xmath10 top grid and a refinement factor of 2
this specifies the initial conditions in the region of interest equivalent to a @xmath11 uni
grid calculation . for the adopted cosmology
this gives a mass resolution of @xmath12 for the dark matter ( dm , hereafter ) and @xmath13 for the gas .
the small dm masses ensure that the cosmological jeans mass is resolved by at least ten thousand particles at all times .
smaller scale structures in the dark matter will not be able to influence the baryons because of their shallow potential wells .
the theoretical expectation holds , because the simulations of abn which had 8 times poorer dm resolution led to identical results on large scales as the simulation presented here . during the evolution ,
refined grids are introduced with twice the spatial resolution of the parent ( coarser ) grid .
these child ( finer ) meshes are added whenever one of three refinement criteria are met .
two langrangian criteria ensure that the grid is refined whenever the gas ( dm ) density exceeds 4.6 ( 9.2 ) its initial density . additionally , the local jeans length is always covered by at least 64 grid cells cells . ]
( 4 cells per jeans length would be sufficient , ) .
we have also carried out the simulations with identical initial conditions but varying the refinement criteria . in one series of runs
we varied the number of mesh points per jeans length .
runs with 4 , 16 , and 64 zones per jeans length are indistinguishable in all mass weighted radial profiles of physical quantities .
no change in the angular momentum profiles could be found , suggesting negligible numerical viscosity effects on angular momentum transport .
a further refinement criterion that ensured the local cooling time scale to be longer than the local courant time also gave identical results .
this latter test checked that any thermally unstable region was identified .
the simulation follows the non equilibrium chemistry of the dominant nine species species ( h , h@xmath14 , h@xmath15 , e@xmath15 , he , he@xmath14 , he@xmath16 , h@xmath17 , and h@xmath18 ) in primordial gas .
furthermore , the radiative losses from atomic and molecular line cooling , compton cooling and heating of free electrons by the cosmic background radiation are appropriately treated in the optically thin limit ( , ) . to extend our previous the studies to higher densities three essential modifications to the code
were made .
first we implemented the three body molecular hydrogen formation process in the chemical rate equations . for temperatures below 300 k we fit to the data of orel ( ) to get @xmath19 . above 300 k we then match it continuously to a powerlaw ( ) @xmath20 .
secondly , we introduce a variable adiabatic index for the gas ( ) . the dissipative component ( baryons ) may collapse to much higher densities than the collisionless component ( dm ) .
the discrete sampling of the dm potential by particles can then become inadequate and result in artificial heating of the baryons ( cooling for the dm ) once the gas density becomes much larger than the local dm density . to avoid this , we smooth the dm particles with a gaussian of width 0.05 for grids with cells smaller than this length . at this scale ,
the enclosed gas mass substantially exceeds the enclosed dm mass . the standard message passing library ( mpi )
was used to implement domain decomposition on the individual levels of the grid hierarchy as a parallelization strategy .
the code was run in parallel on 16 processors of the sgi origin2000 supercomputer at the national center for supercomputing applications at the university of illinois at urbana champaign .
we stop the simulation at a time when the molecular cooling lines reach an optical depth of ten at line center because our numerical method can not treat the difficult problem of time
dependent radiative line transfer in multi dimensions . at this time the code utilizes above 5500 grids on 27 refinement levels with @xmath21 computational grid cells .
an average grid therefore contains @xmath22 cells .
[ colorplate ] our simulations ( fig . [ colorplate ] , fig . [ 5panel ] ) ,
identify at least four characterisic mass scales . from the outside
going in , one observes infall and accretion onto the pre galactic halo with a total mass of @xmath23 , consistent with previous studies ( , , , abn , and for discussion and references ) .
at a mass scale of about 4000 solar mass ( @xmath24 ) rapid cooling and infall is observed .
this is accompanyed by the first of three valleys in the radial velocity distribution ( fig .
[ 5panel]e ) . the temperature drops and the molecular hydrogen fraction increases .
it is here , at number densities of @xmath25 , that the high redshift analog of a molecular cloud is formed . although the molecular mass fraction is not even 0.1% it is sufficient to cool the gas rapidly down to @xmath26 .
the gas can not cool below this temperature because of the sharp decrease in the cooling rate below @xmath27 . at redshift 19 ( fig .
[ 5panel ] ) , there are only two mass scales ; however , as time passes the central density grows and eventually passes @xmath28 , at which point the ro - vibrational levels of are populated at their equilibrium values and the cooling time becomes independent of density ( instead of inversely proportional to it ) .
this reduced cooling efficiency leads to an increase in the temperature ( fig . [
5panel]d ) .
as the temperature rises , the cooling rate again increases ( it is 1000 times higher at 800 k than at 200 k ) , and the inflow velocities slowly climb . in order to better understand what happens next , we examine the stability of an isothermal gas sphere . the critical mass for gravitational collapse given an external pressure @xmath29 ( be mass hereafter ) is given by ebert ( ) and bonnor ( ) as : @xmath30 here @xmath29 is the external pressure and @xmath31 , @xmath32 , and @xmath33 the gravitational constant , the boltzmann constant and the sound speed , respectively .
we can estimate this critical mass locally if we set the external pressure to be the local pressure to find @xmath34 where @xmath35 is the mean mass per particle in units of the proton mass . using an adiabatic index @xmath36
, we plot the ratio of the enclosed gas mass to this modified be mass in figure [ bemass ] .
our modeling shows ( fig .
[ bemass ] ) , that by the fourth considered output time , the central 100 exceeds the be mass at that radius , indicating unstable collapse . this is the third mass scale and corresponds to the second local minimum in the radial velocity curves ( fig .
[ 5panel]e ) .
the inflow velocity is @xmath37 is still subsonic .
although this mass scale is unstable , it does not represent the smallest scale of collapse in our simulation .
this is due to the increasing molecular hydrogen fraction .
when the gas density becomes sufficiently large ( @xmath38 ) , three - body molecular hydrogen formation becomes important .
this rapidly increases the molecular fraction ( fig .
[ 5panel]c ) and hence the cooling rate .
the increased cooling leads to lower temperatures and even stronger inflow and . at a mass scale of @xmath39 , not only is the gas nearly completely molecular , but the radial inflow has become supersonic ( fig . [
5panel]e ) .
when the mass fraction approaches unity , the increase in the cooling rate saturates , and the gas goes through a radiative shock .
this marks the first appearance of the proto
stellar accretion shock at a radius of about 20 astronomical units from its center . when the cooling time becomes independent of density the classical criterion for fragmentation @xmath40 ( )
can not be satisfied at high densities .
however , in principal the medium may still be subject to thermal instability .
the instability criterion is @xmath41 where @xmath42 denotes the cooling losses per second of a fluid parcel and @xmath43 and @xmath44 are the gas temperature and mass density , respectively . at densities above the critical densities of molecular hydrogen the cooling time is independent of density , i.e. @xmath45 where @xmath46 is the high density cooling function ( e.g. ) .
fitting the cooling function with a power - law locally around a temperature @xmath47 so that @xmath48 one finds @xmath49 . hence , under these circumstances the medium is thermally stable if @xmath50 . because , @xmath51 for the densities and temperatures of interest , we conclude that the medium is thermally stable .
the above analysis neglects the heating from contraction , but this only strengthens the conclusion . if heating balances cooling one can neglect the @xmath52 term in equation ( [ eq : ic ] ) and find the medium to be thermally stable for @xmath53 .
[ bemass ] however , here we neglected the chemical processes .
the detailed analysis for the case when chemical processes occur on the collapse time scale is well known ( ) . this can be applied to primordial star formation ( ) including the three
body formation of molecular hydrogen ( ) which drives a chemo thermal instability .
evaluating all the terms in this modified instability criterion ( , equation 36 ) one finds the simple result that for molecular mass fractions @xmath54 the medium is expected to be chemo thermally unstable .
these large molecular fractions illustrate that the strong density dependence of the three body formation dominates the instability .
examining the three dimensional temperature and density field we clearly see this chemo thermal instability at work .
cooler regions have larger fractions .
however , no corresponding large density inhomogeneities are found and fragmentation does not occur .
this happens because of the short sound crossing times in the collapsing core . when the formation time scale becomes shorter than the cooling time the instability originates . however , as long as the sound crossing time is much shorter than the chemical and cooling time scales the cooler parts are efficiently mixed with the warmer material .
this holds in our simulation until the final output where for the first time the formation time scale becomes shorter than the sound crossing time .
however , at this point the proto
stellar core is fully molecular and stable against the chemo thermal instability .
consequently no large density contrasts are formed .
because at these high densities the optical depth of the cooling radiation becomes larger than unity the instability will be suppressed even further .
interestingly , rotational support does not halt the collapse .
this is for two reaons .
the first is shown in panel a of fig .
[ angtrans ] , which plots the specific angular momentum against enclosed mass for the same seven output times discussed earlier . concentrating on the first output ( fig .
[ angtrans ] ) , we see that the central gas begins the collapse with a specific angular momentum only @xmath55% as large as the mean value .
this type of angular momentum profile is typical of halos produced by gravitational collapse ( e.g. ) , and means that the protostellar gas starts out without much angular momentum to lose . as a graphic example of this , consider the central one solar mass of the collapsing region .
it has only an order of magnitude less angular momentum at densities @xmath56 than it had at @xmath57 although it collapsed by over a factor 100 in radius .
the remaining output times ( fig .
[ angtrans ] ) indicate that there is some angular momentum transport within the central @xmath58 ( since l plotted as a function of enclosed mass should stay constant as long as there is no shell crossing ) . in panel c
, we divide @xmath42 by @xmath59 to get a typical rotational velocity and in panels b and d compare this velocity to the keplerian rotational velocity and the local sound speed , respectively .
we find that the typical rotational speed is a factor two to three below that required for rotational support .
furthermore , we see that this azimuthal speed never significantly exceeds the sound speed , although for most the mass below @xmath58 it is comparable in value .
we interpret this as evidence that it is shock waves during the turbulent collapse that are responsible for much of the transported angular momentum .
a collapsing turbulent medium is different from a disk in keplarian rotation . at any radius
there will be both low and high angular momentum material , and pressure forces or shock waves can redistribute the angular momentum between fluid elements .
lower angular momentum material will selectively sink inwards , displacing higher angular momentum gas .
this hydrodynamic transport of angular momentum will be suppressed in situation where the collapse proceeds on the dynamical time rather on the longer cooling time as in the presented case .
this difference in cooling time and the widely different initial conditions may explain why this mechanism has not been observed in simulations of present day star formation ( e.g. , and references therein ) .
however , such situations may also arise in the late stages of the formation of present day stars and in scenarios for the formation of super massive black holes .
to ensure that the angular momentum transport is not due to numerical shear viscosity ( ) we have carried out the resolution study discussed above .
we have varied the effective spatial resolution by a factor 16 and found identical results .
furthermore , we have run the adaptive mesh refinement code with two different implementations of the hydrodynamics solver .
the resolution study and the results presented here were carried out with a direct piecewise parabolic method adopted for cosmology (; ) .
we ran another simulation with the lower order zeus hydrodynamics ( ) and still found no relevant differences .
these tests are not strict proof that the encountered angular momentum transport is not caused by numerical effects ; however , they are reassuring .
the strength of magnetic fields generated around the epoch of recombination is minute .
in contrast , phase transitions at the qantum chromo
dynamic ( qcd ) and electro weak scales may form even dynamically important fields . while there is a plethora of such scenarios for primordial magnetic field generation in the early universe they are not considered to be an integral part of our standard picture of structure formation .
this is because not even the order of these phase transitions is known ( ) , and references therein ) .
unfortunately , strong primordial small - scale ( @xmath60 comoving ) magnetic fields are poorly constrained observationally ( ) .
the critical magnetic field for support of a cloud ( ) allows a rough estimate up to which primordial magnetic field strengths we may expect our simulation results to hold . for this
we also assume a flux frozen flow with no additional amplification of the magnetic field other than the contraction ( @xmath61 ) . for a comoving b field of @xmath62 on scales
@xmath63 the critical field needed for support may be reached during the collapse possibly modifying the mass scales found in our purely hydrodynamic simulations . however , the ionized fraction drops rapidly during the collapse because of the absence of cosmic rays ionizations .
consequently ambipolar diffusion should be much more effective in the formation of the first stars even if such strong primordial magnetic fields were present .
previously we discussed the formation of the pre galactic object and the primordial `` molecular cloud '' that hosts the formation of the first star in the simulated patch of the universe ( ) .
these simulations had a dynamic range of @xmath64 and identified a @xmath0 core within the primordial `` molecular cloud '' undergoing renewed gravitational collapse .
the fate of this core was unclear because there was the potential caveat that three body formation could have caused fragmentation .
indeed this further fragmentation had been suggest by analytic work ( ) and single zone models ( ) . the three dimensional simulations described here were designed to be able to test whether the three body process will lead to a break up of the core . _
no fragmentation due to three body formation is found .
_ this is to a large part because of the slow quasi
hydrostatic contraction found in abn which allows sub sonic damping of density perturbations and yields a smooth distribution at the time when three body formation becomes important . instead of fragmentation a single fully molecular proto
star of @xmath65 is formed at the center of the @xmath66 core .
however , even with extraordinary resolution , the _ final _ mass of the first star remains unclear .
whether all the available cooled material of the surroundings will accrete onto the proto star or feedback from the forming star will limit the further accretion and hence its own growth is difficult to compute in detail . within @xmath67
about @xmath68 may be accreted assuming that angular momentum will not slow the collapse ( fig .
[ acrete ] ) .
the maximum of the accretion time of @xmath69 is at @xmath70 .
however , stars larger than @xmath71 will explode within @xmath72 .
therefore , it seems unlikely ( even in the absence of angular momentum ) that there would be sufficient time to accrete such large masses . a one solar mass proto
star will evolve too slowly to halt substantial accretion . from the accretion time profile ( fig .
[ acrete ] ) one may argue that a more realistic minimum mass limit of the first star should be @xmath73 because this amount would be accreted within a few thousand years .
this is a very short time compared to expected proto stellar evolution times .
however , some properties of the primordial gas may make it easier to halt the accretion .
one possibility is the destruction of the cooling agent , molecular hydrogen , without which the acreting material may reach hydrostatic equilibrium .
this may or may not be sufficient to halt the accretion .
one may also imagine that the central material heats up to @xmath74 k , allowing lyman-@xmath75 cooling from neutral hydrogen .
that cooling region may expand rapidly as the internal pressure drops because of infall , possibly allowing the envelope to accrete even without molecular hydrogen as cooling agent .
additionally , radiation pressure from ionizing photons as well as atomic hydrogen lyman series photons may become significant and eventually reverse the flow .
the mechanisms discussed by haehnelt ( 1993 ) on galactic scales will play an important role for the continued accretion onto the proto star .
this is an interplay of many complex physical processes because one has a hot ionized strmgren sphere through which cool and dense material is trying to accrete .
in such a situation one expects a raleigh taylor type instability that is modified via the geometry of the radiation field . at the final output time
presented here there are @xmath76 hydrogen molecules in the entire protogalaxy . also the formation time scale is long because there are no dust grains and the free electrons ( needed as a catalyst ) have almost fully recombined .
hence , as soon as the the first uv photons of lyman werner band frequencies are produced there will be a rapidly expanding photo dissociating region ( pdr ) inhibiting further cooling within it .
this photo - dissociation will prevent further fragmentation at the molecular cloud scale .
i.e. no other star can be formed within the same halo before the first star dies in a supernova .
the latter , however , may have sufficient energy to unbind the entire gas content of the small pre galactic object it formed in ( ) .
this may have interesting feedback consequences for the dispersal of metals , entropy and magnetic field into the intergalactic medium ( , )
. smoothed particle hydrodynamics ( sph , e.g. ) , used extensively in cosmological hydrodynamics , has been employed ( ) to follow the collapse of solid body rotating uniform spheres .
the assumption of coherent rotation causes these clouds to collapse into a disk which developes filamentary structures which eventually fragment to form dense clumps of masses between @xmath77 and @xmath78 solar masses .
it has been argued that these clumps will continue to accrete and merge and eventually form very massive stars .
these sph simulation have unrealistic initial conditions and much less resolution then our calculations .
however , they also show that many details of the collapse forming a primordial star are determined by the properties of the hydrogen molecule .
we have also simulated different initial density fields for a lambda cdm cosmology .
there we have focused on halos with different clustering environments .
although we have not followed the collapse in these halos to proto - stellar densities , we have found no qualitative differences in the `` primordial molecular cloud '' formation process as discussed in abn . also other amr simulations ( )
give consistent results on scales larger than @xmath79 .
in all cases a cooling flow forms the primordial molecular cloud at the center of the dark matter halo . we conclude that the molecular cloud formation process seems to be independent of the halo clustering properties and the adopted cdm type cosmology .
also the mass scales for the core and the proto star are determined by the local bonnor
ebert mass .
consequently , we expect the key results discussed here to be insensitive to variations in cosmology or halo clustering .
the picture arising from these numerical simulations has some very interesting implications .
it is possible that all metal free stars are massive and form in isolation .
their supernovae may provide the metals seen in even the lowest column density quasar absorption lines ( , and references therein ) .
massive primordial stars offer a natural explanation for the absence of purely metal free low mass stars in the milky way .
the consequences for the formation of galaxies may be even more profound in that the supernovae provide metals , entropy , and magnetic fields and may even alter the initial power spectrum of density fluctuations of the baryons .
interestingly , it has been recently argued , from abundance patterns , that in low metallicity galactic halo stars seem to have been enriched by only one population of massive stars ( ) .
these results , if confirmed , would represent strong support for the picture arising from our ab initio simulations of first structure formation . to end on a speculative note
there is suggestive evidence that links gamma ray bursts to sites of massive star formation ( e.g. ) .
it would be very fortunate if a significant fraction of the massive stars naturally formed in the simulations would cause gamma ray bursts ( e.g. ) .
such high redshift bursts would open a remarkably bright window for the study of the otherwise dark ( faint ) ages . 1 . [ pd68 ] peebles , p. j. e. & dicke , r. h. 1968 , apj * 154 * , 891 2 .
[ h69 ] hirasawa , t. 1969 , prog .
. phys . * 42 * , 523 3 . [ pss83 ] palla , f. , salpeter , e.e . ,
stahler , s.w .
1983 , apj * 271 * , 632 4 . [ a95 ] abel , t. 1995 , thesis , university of regensburg .
[ aanz ] abel , t. , anninos , p. , norman , m.l . , zhang , y. 1998 , apj * 508 * , 518 .
[ abn99 ] abel , t. , bryan , g. l. , & norman , m. l. 1999 , in `` evolution of large scale structure : from recombination to garching '' , eds .
banday , t. , sheth , r. k. and costa , l. n. 7 .
[ abn00 ] abn : abel , t. , bryan , g.l . , norman , m.l . 2000 , apj * 540 * , 39 8 .
[ cosmo ] friedmann models with a cosmological constant which currently seem to fit various observational test better differ from the standard cdm model considered here only slightly at the high redshifts modeled .
[ bn97 ] bryan , g.l . , norman , m.l .
1997 , in _ computational astrophysics _ , eds .
clarke and m. fall , asp conference # 123 10 .
[ bn99 ] bryan , g.l . , norman , m.l . 1999 , in _ workshop on structured adaptive mesh refinement grid methods _ , i m a volumes in mathematics no . 117 , ed . n. chrisochoides , p. 165
[ wc84 ] woodward , p. r. , & colella , p. 1984
, j. comput
54 * , 115 12 .
[ betal95 ] bryan , g.l . ,
cen , r. , norman , m.l . ,
ostriker , j.p . & stone , j.m .
1994 , apj * 428 * , 405 13 . [ sz96 ] seljak , u. , & zaldarriaga , m. 1996 , apj * 469 * , 437 14 . [ tkmh97 ]
truelove , j. k. , klein , r. i. , mckee , c. f. , holliman , j. h. , howell , l. h. & greenough , j. a. 1997 , apjl * 489 * , l179 15 . [ aazn97 ] abel , t. , anninos , p. , zhang , y. , norman , m.l .
1997 , newa * 2 * , 181 16 .
[ azan97 ] anninos , p. , zhang , y. , abel , t. , and norman , m.l .
1997 , newa * 2 * , 209 .
[ orel ] orel , a.e .
1987 , j.chem.phys .
* 87 * , 314 18 .
[ on98 ] omukai , k. & nishi , r. 1998 , apj * 508 * , 141 19 .
[ tegmark97 ] tegmark , m. , silk , j. , rees , m.j . ,
blanchard , a. , abel , t. , palla , f. 1997 , apj * 474 * , 1 .
[ bcl99 ] bromm , v. , coppi , p. s. , & larson , r. b. 1999 , apjl * 527 * , l5 21 .
[ ebert ] ebert , r. 1955 , zs .
[ bonnor ] bonnor , w. b. 1956 , mnras * 116 * , 351 23 .
[ field65 ] field , g. b. 1965 , apj * 142 * , 531 24 .
[ gp98 ] galli , d. & palla , f. 1998 , a&a * 335 * , 403 25 .
[ sy77 ] sabano , y. & yoshi , y. 1977 , pasj * 29 * , 207 26 .
[ silk83 ] silk , j. 1983 , mnras * 205 * , 705 27 .
[ qz88 ] quinn , p.j . &
zurek , w.h .
1988 , apj * 331 * , 1 28 .
[ bb93 ] burkert , a. & bodenheimer , p. 1993
, mnras * 264 * , 798 29 .
[ nwb80 ] norman , m.l . , wilson , j.r . &
barton , r.t .
1980 , apj * 239 * , 968 30 . [ sn92 ] stone , j.m . &
norman , m.l .
1992 , apjs * 80 * , 791 31 .
[ soj97 ] sigl , g. , olinto , a. v. & jedamzik , k. 1997 , phys.rev.d * 55 * , 4582 32 .
[ bfs97 ] barrow , j. d. , ferreira , p. g. and silk , j. 1997 , physical review letters * 78 * , 3610 33 .
[ ms76 ] mouschovias , t. ch . , spitzer , l. 1976 , apj * 210 * , 326 34 .
[ haehnelt ] haehnelt , m.g .
1995 , mnras * 273 * , 249 35 .
[ mf99 ] mac low , m. & ferrara , a. 1999 , apj * 513 * , 142 36 .
[ f98 ] ferrara , a. 1998 , apjl * 499 * , l17 37 .
[ cb00 ] cen , r. & bryan , g.l .
2000 , apjl * 546 * , l81 38 . [ m92 ] monaghan , j. j. 1992 , araa * 30 * , 543 39 .
[ mba01 ] machacek , m.e . , bryan , g.l .
, & abel , t 2001 , apj * 548 * , 509 40 . [ essp00 ]
ellison , s. l. , songaila , a. , schaye , j. & pettini , m. 2000 , , 1175 41 .
[ wq00 ] wasserburg , g. j. & qian , y. 2000 , apjl * 529 * , l21 42 .
[ r99 ] reichart , d. e. 1999 , apjl * 521 * , l111 43 .
[ cl00 ] ciardi , b. & loeb , a. 2000 , apj * 540 * , 687 44 .
t.a . happily acknowledges stimulating and insightful discussions with martin rees and richard larson .
glb was supported through hubble fellowship grant hf-0110401 - 98a from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . under nasa contract nas5 - 26555 . | we describe results from a fully self consistent three dimensional hydrodynamical simulation of the formation of one of the first stars in the universe .
dark matter dominated pre - galactic objects form because of gravitational instability from small initidal density perturbations . as they assemble via hierarchical merging , primordial gas cools through ro - vibrational lines of hydrogen molecules and sinks to the center of the dark matter potential well .
the high redshift analog of a molecular cloud is formed .
when the dense , central parts of the cold gas cloud become self - gravitating , a dense core of @xmath0 undergoes rapid contraction . at densities
@xmath1 a @xmath2 proto - stellar core becomes fully molecular due to three body formation .
contrary to analytical expectations this process does not lead to renewed fragmentation and only one star is formed .
the calculation is stopped when optical depth effects become important , leaving the final mass of the fully formed star somewhat uncertain . at this stage
the protostar is acreting material very rapidly ( @xmath3 ) .
radiative feedback from the star will not only halt its growth but also inhibit the formation of other stars in the same pre galactic object ( at least until the first star ends its life , presumably as a supernova ) .
we conclude that at most one massive ( @xmath4 ) metal free star forms per pre galactic halo ,
consistent with recent abundance measurements of metal poor galactic halo stars .
# 1_[#1 ] _
# 110^#1 # 1n _ # 1 # 1k_#1 # 1@xmath5 # 1_#1 _ # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
the early calculations by svane and gunnarson showed that when self - interaction corrections were included in the local density approximation , density functional theory ( dft ) predicted a first order phase transition located near the wigner - seitz radius @xmath1 where @xmath2 , @xmath3 is the density , and @xmath4 is the bohr radius @xcite . on the contrary , dft calculations using either
the generalized gradient approximation ( gga ) or local spin density approximation ( lsda ) without the self - interaction correction have predicted a second - order phase transition at @xmath5 and @xmath6 and an itenerant anti - ferromagnetic phase up to @xmath7 and @xmath8 respectively @xcite .
g@xmath9w@xmath9 , using the lda or gga orbitals to compute the initial green s function , finds the same transition order as their underlying dft functionals , though the phase transition density is shifted upwards to @xmath10 @xcite .
the most recent set of g@xmath9w@xmath9 calculations begin with lda+u and gga+u single particle orbitals for the initial green s function @xcite .
the `` + u '' methods include an on - site repulsion for the two different spin densities to penalize double occupancy and pushes the system towards an anti - ferromagnetic state .
using g@xmath9w@xmath9 on top of these methods , researchers find a continuous metal to insulator phase transition and locate it close to @xmath11 .
this phase transition has also been investigated using dynamical mean field theory ( dmft ) by approximating the coulomb interaction as a strictly short ranged on - site interaction between two electrons on the same hydrogen ion @xcite . using this method
it was found to be a first - order phase transition at @xmath12 .
this transition location is an extrapolation from their finite temperature data to the ground state @xcite .
a highly accurate benchmark is required to disambiguate these results .
previous efforts to produce such a benchmark have been performed using variational quantum monte carlo@xcite .
this calculation was consistent with either a very weak first order or a second order transition at @xmath13 .
the error estimates in these measurements are sufficiently large to include a number of the previous results .
our goal in this work is to provide a benchmark with improved accuracy .
in this section we will discuss the method we use , the hamiltonian for the system , and some computational aspects particular to our calculation . in this work we use dmc to generate all of our results .
this method has been used to produce benchmark results for light elements such as hydrogen and the electron gas and has been increasingly used for solid state systems @xcite .
this variational stochastic projector method filters out the ground state component of a trial wave function to sample the ground state probability distribution @xcite . by using a trial wave function we are able to avoid the notorious `` sign problem '' which plagues exact monte carlo calculations of fermions but introduce error which raises the energy .
the nodes or phase of the trial wave function serves as a boundary condition on the the random walk .
the error introduced by this approximation is referred to as the `` fixed - node error '' @xcite . in rydberg units ,
the hamiltonian for hydrogen is , @xmath14 where capital letters , @xmath15 , correspond to ion coordinates and lower case letter , @xmath16 , correspond to electronic coordinates .
this is a zero temperature calculation and does not include the kinetic energy of the protons ; they are clamped to the bcc lattice . in this work we will refer to the two atoms in the bcc unit cell as the a and b simple cubic sublattices .
our trial wave function is a single slater jastrow wave function , @xmath17 where @xmath18 where @xmath19 and similarly for the down spin electrons , @xmath20 . for the ground state it is always the case that @xmath21 . for the quasiparticle calculation
they differ by 1 .
the jastrow consists of two terms : a one - body term , @xmath22 , and a two - body term , @xmath23@xcite and are of the form , @xmath24 where @xmath25 refer to ionic coordinates , @xmath16 refer to electron coordinates , @xmath26 and @xmath27 are the electron spins , and @xmath28 and @xmath29 are bspline@xcite functions whose parameters are variational degrees of freedom .
both the one body and two body terms include a cusp condition which , in conjunction with the determinant , exactly cancels the divergent coulomb potential energy contribution when an ion and electron or two electrons coincide@xcite .
we optimize the parameters in the trial wave function using a variant of the linear method of umrigar and coworkers@xcite . instead of rescaling the eigenvalues found during the generalized eigenvalue problem , we perform a line minimization on them using a @xmath30-point fit to a quadratic function .
we find that this can increase the rate of convergence to the optimal set of variational parameters@xcite .
we parameterize the two - body jastrow function so that it is symmetric under exchange of up and down electron labels .
this requires the same parameterization for @xmath23 between up - up and down - down pairs , @xmath31 , but allows for a separate set of parameters for up - down @xmath23 terms,@xmath32 .
the one - body jastrow is parameterized differently in the paramagnetic and anti - ferromagnetic phases . in the paramagnetic phase we use a one body jastrow which is not a function of electron spin or ion sublattice . in the anti - ferromagnetic phase
we use a jastrow that is the same for up - a / down - b , @xmath33 , and for up - b / down - a , @xmath34 , electron spin - ion sublattice pairs .
this ensures that the wave function is unchanged if up and down electron labels are swapped at the same time as the a and b sublattice labels are . for a slater - jastrow wave function
, the magnitude of the fixed node error is affected by the quality of the single particle orbitals used in the slater determinant .
there are many different orbitals that can be used in the determinant : analytic forms such as plane waves or gaussians , orbitals from density function theory , or orbitals derived from another quantum chemistry method .
it has been found that , for solids , the best trial wave functions are often obtained using orbitals from dft .
however , different functionals produce different orbitals , so a careful choice of functional is necessary @xcite .
in this work we consider single particle orbital sets generated using three different techniques all of which are based on dft . in the paramagnetic phase we use orbitals generated using the local density approximation ( lda)@xcite . in the anti - ferromagnetic phase , as we will describe below , we use two different sets of orbitals , one generated using a lda+u functional , and the other generated by performing two lda calculations , each of which includes only the ions of one sublattice .
we generate these orbitals using quantum espresso @xcite and a lda pseudopotential generated using opium @xcite . ) .
the horizontal line is the energy obtained using the split sublattice orbitals , and dashed line is a single standard deviation in energy .
the minimum energy orbitals are generated using u@xmath35 .
the line passing through the lda+u energies is a quadratic fit for all u@xmath36.[fig : pol_u ] ] .
the minimum energy lda+u orbitals ( circles ) are slightly less polarized than the split sublattice ones ( diamond ) .
the line is a guide to the eye .
[ fig : pol_e ] ] for the anti - ferromagnetic phase , we obtain the split sublattice orbitals by performing a dft calculation on the simple cubic lattice and translating the orbitals according to the bcc basis vector , @xmath37 where @xmath38 is the lattice constant .
we choose to use this basis because , for the ground state , it explicitly breaks spin symmetry and creates a set of anti - ferromagnetic orbitals centered on each sublattice and has no adjustable parameters .
we also use lda+u orbitals to be able to tune the degree of spin polarization in the system . the lda+u functional can interpolate between the paramagnetic and anti - ferromagnetic phases by tuning u. for small values of u the spin polarization is small and for large u
it is complete .
though this is a simple single parameter single particle orbital optimization technique , it can be very expensive to converge .
the sensitivity of the orbitals to the choice of u is illustrated in figure [ fig : pol_u ] .
the resulting energy vs. polarization curve is presented in figure [ fig : pol_e ] .
to find the optimal value of u we must perform qmc calculations on large systems , 432 electrons , for several values of u and orbital occupations .
the orbitals may be occupied according to the lda+u energy prediction , or by filling the lowest band regardless of lda+u occupation .
these partially polarized phases exhibit large finite size effects , and require large simulation cells to accurately determine the optimal u , its polarization and energy . near the phase transition
the lda+u basis provides a better description of the anti - ferromagnetic state as evidenced by lower energies .
because the dmc uses the full coulomb potential for the electron - ion potential , the only effect the choice of pseudopotential has on our dmc results is through the quality of the single particle orbitals . to check the effect of the pseudopotential
, we compared the dmc energy of the single particle orbitals generated using our lda pseudopotential with those generated using the trail - needs pseudopotential in the paramagnetic phase at @xmath39 @xcite . for the @xmath40 hydrogen supercell we found no statistical difference in their energies .
as described above , we use trial wave functions for the anti - ferromagnetic phase that explicitly break spin symmetry .
we compute the polarization of the phase using the sublattice spin density , @xmath41 where @xmath42 is the spin density for the up or down electrons , @xmath43 , on the a sublattice whose protons are located at @xmath44 , and the step function is zero outside a sphere of radius @xmath45 .
@xmath46 is defined identically but with a and b indices exchange .
we are free to adjust @xmath45 between zero and @xmath47 , the wigner - seitz radius . as @xmath48 the statistics of the observable improve due to increased sampling volume . as @xmath49
the estimator projects out just the spin at the sublattice proton positions .
we find that @xmath50 is a reasonable compromise and use it for all densities .
using these spin densities we define the a sublattice polarization , @xmath51 , as , @xmath52 in this work we only investigate phases where the polarization is symmetric , @xmath53 . in the paramagnetic phase @xmath54 and in the anti - ferromagnetic phase @xmath55 .
this estimator works in trial wave function based qmc when the up / down symmetry of the state has explicitly been broken . if the trial wave function is a linear combination of up - a / down - b and up - b / down - a spin electrons assigned to sublattices , then the previous estimator yields zero even for anti - ferromagnetic phases . in that case
it is necessary to compute the correlation function using the polarization operator ,
@xmath56 we converge the dft calculations using a @xmath57 k - point grid for the metallic phases and @xmath58 grid for the insulating or partially polarized phases .
we find that using an energy cutoff of @xmath59 rydberg is sufficient to converge the dft energy as well as the orbitals used in the quantum monte carlo . in the paramagnetic phase , we perform dmc for a @xmath60 supercell of @xmath61 twists and @xmath62 total ions , and a @xmath63 supercell of @xmath64 twists and @xmath65 total ions using both the ewald and modified periodic coulomb ( mpc ) potential @xcite for the electrons @xcite .
these runs were performed with more than two thousand configurations and extrapolated to zero time step using a three point extrapolation .
the smallest time steps had accept ratios greater than @xmath66 .
the energy is then averaged over twist points and the two supercells are extrapolated linearly as volume@xmath67 to infinite volume .
we find that both the ewald and mpc extrapolate to similar values , and forgo larger simulation cells in the extrapolation .
we put more effort into computing the properties of the anti - ferromagnetic phase because it is relatively more complex than the paramagnetic phase .
because the anti - ferromagnetic phase is insulating , we use a smaller number of twists . the finite size and brillouin zone sampling
are done using a @xmath60 supercell of @xmath63 twists ( @xmath62 protons ) , a @xmath63 supercell of @xmath60 twists ( @xmath65 protons ) , and a @xmath64 supercell of @xmath40 twists ( @xmath68 protons ) .
we locate the transition density between the paramagnetic and anti - ferromagnetic phases by computing the energy of each phase as a function of wigner - seitz radius and finding the crossing point . by carefully extrapolating finite size effects , sampling the fermi surface using twist averaging , and controlling for time step error
, we determine the transition density within a small statistical error @xcite .
we control for systematic error , due to the fixed node approximation , by comparing the energies for several different sets of single particle orbitals in the anti - ferromagnetic phase and by using a slater - jastrow - backflow wave function in the paramagnetic phase .
.fixed - phase dmc energy per hydrogen atom in rydberg for bcc hydrogen in the anti - ferromagnetic ( afm ) , and paramagnetic ( pm ) phases.[tbl : h_eos ] [ cols="^,^,^",options="header " , ] + the gaps we compute and reference values from previous studies are presented in table [ table : gaps ] .
the excitonic gap extrapolates to zero gap while the quasiparticle gap projects to a small but finite gap , @xmath69 .
both of these gaps are consistent with a phase transition location closer to @xmath39 than previous g@xmath9w@xmath9 studies .
we note that the quasiparticle gap suffers from much larger finite size effects than the excitonic gap which results in a more difficult extrapolation and larger error bar estimate . at this density ,
an estimate for the excitonic binding energy , the difference between the excitonic and quasiparticle gap extrapolated to the thermodynamic limit , is @xmath70 .
we have presented benchmark calculations on the phase transition density and order for bcc hydrogen .
as shown in table [ tbl : heos_prev ] , the phase transition is likely to be continuous and is located at @xmath71 .
the most recent g@xmath9w@xmath9 calculations@xcite show the same transition order and agree very well with our transition location .
if these calculations were to choose a slightly different value of u , they may be brought into perfect agreement .
previous publications@xcite using g@xmath9w@xmath9 resulted in slightly less accurate transition densities , likely due to a worse dft starting point .
our results provide confidence in the g@xmath9w@xmath9 method when starting from a high quality dft calculation .
previous dft results identify the same phase transition order , but greatly overestimate the transition location@xcite .
previous qmc results@xcite were performed on small systems , of @xmath62 ions , which is under converged with respect to system size .
they are also performed using vmc which is much more sensitive to the form and optimization of the trial wave function . however , these results predict a similar transition density as the current ones .
the remainder of the methods , dmft @xcite and sic - dft @xcite get the phase transition order and location incorrect .
jm acknowledges useful conversations with sarang gopalakrishnan , norm tubman , and lucas wagner .
this work was performed in part under the auspices of the u.s .
department of energy ( doe ) by llnl under contract de- ac52 - 07na27344 .
jm and dmc were supported by nsf oci-0904572 .
jm , dmc , mm , and jk were supported by the network for ab initio many - body methods supported by predictive theory and modeling program of basic energy sciences , department of energy ( doe ) .
this work used the extreme science and engineering discovery environment ( xsede ) , which is supported by national science foundation grant number oci-1053575 , and resources provided by the innovative and novel computational impact on theory and experiment ( incite ) at the oak ridge leadership computing facility at the oak ridge national laboratory , which is supported by the office of science of the u.s .
department of energy under contract no .
de - ac05 - 00or22725 .
llnl release number llnl - jrnl-638420 .
47ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) link:\doibase 10.1016/0009 - 2614(84)85513-x [ * * , ( ) ] link:\doibase 10.1021/ct301044e [ * * , ( ) ] link:\doibase " 10.1016/0038 - 1098(90)90641-n [ * * , ( ) ] link:\doibase 10.1103/physrevlett.65.2414 [ * * , ( ) ] link:\doibase 10.1103/physrevb.58.12680 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.195123 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.155114 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.035102 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.73.33 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.086403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.70.1952 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1073/pnas.1007309107 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.68.13 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.45.566 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.84.1607 [ * * , ( ) ] link:\doibase 10.1063/1.443766 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.71.2777 [ * * , ( ) ] link:\doibase 10.1103/physrev.98.1479 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1002/cpa.3160100201 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.110201 [ * * , ( ) ] link:\doibase 10.1063/1.2437215 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.3665391 [ * * , ( ) ] link:\doibase 10.1021/ct3003404 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.185502 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.73.33 [ * * , ( ) ] link:\doibase 10.1103/physrevb.45.13244 [ * * , ( ) ] http://www.quantum-espresso.org [ * * , ( ) ] @noop `` '' link:\doibase 10.1063/1.1888569 [ * * , ( ) ] link:\doibase 10.1063/1.1829049 [ * * , ( ) ] link:\doibase 10.1103/physrevb.53.1814 [ * * , ( ) ] link:\doibase 10.1103/physreve.64.016702 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.125106 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.72.2438 [ * * , ( ) ] @noop * * , ( ) _ _ , @noop ph.d .
thesis , ( ) link:\doibase 10.1103/physrevb.51.4014 [ * * , ( ) ] | solid atomic hydrogen is one of the simplest systems to undergo a metal - insulator transition . near the transition
, the electronic degrees of freedom become strongly correlated and their description provides a difficult challenge for theoretical methods . as a result ,
the order and density of the phase transition are still subject to debate . in this work
we use diffusion quantum monte carlo to benchmark the transition between paramagnetic and anti - ferromagnetic body centered cubic atomic hydrogen in its ground state .
we locate the density of the transition by computing the equation of state for these two phases and identify the phase transition order by computing the band gap near the phase transition .
these benchmark results show that the phase transition is continuous and occurs at a wigner - seitz radius of @xmath0 .
we compare our results to previously reported density functional theory , hedin s gw approximation , and dynamical mean field theory results . the metal - insulator transition in body centered cubic ( bcc ) hydrogen
was first considered by mott @xcite . at high density , the system is metallic : the electrons are in the paramagnetic , conducting phase . as the lattice spacing
is increased the electrons localize on the hydrogen ions forming an anti - ferromagnetic , insulating phase .
predicting the order of the phase transition allows us to identify the qualitative nature of the transition : whether the transition is due to strong particle correlations , a mott insulator transition , or due to long range coulomb interactions , the band - insulator transition @xcite .
locating the precise transition density is a difficult test for an electronic structure method . since the original study ,
several closely related systems , finite lattices or 1d chains of hydrogen atoms , have become a standard test used to benchmark quantum chemistry methods @xcite .
unfortunately , current electronic structure methods such as density functional theory , dynamical mean field theory , variational quantum monte carlo , and hedin s gw disagree over both the phase transition order and density @xcite . in this work
, we determine the phase transition density and order using diffusion quantum monte carlo ( dmc ) .
dmc is one of the most accurate methods for computing electronic structure properties of atomic , molecular , and condensed phases @xcite .
it has been used extensively to characterize hydrogen and the alkali metals at ambient and at high pressure @xcite .
we use dmc to compute the ground state equation of state in both the paramagnetic and anti - ferromagnetic phase as well as the quasiparticle and excitonic band gaps near the transition .
we begin by reviewing previous calculations and the details of the current calculation .
next we discuss our results for the density of the phase transition and the magnitude of the band gap .
finally , we conclude with a comparison of results obtained with other methods and how they compare to those obtained using dmc . |
hera is an @xmath0 collider which has especially high sensitivity to new particles coupling to lepton - quark pairs . in 1994
- 97 hera collided 27.5 gev positrons on 820 gev protons . in 1998
the proton energy was raised to 920 gev increasing the center - of - mass energy @xmath1 from 300 gev to 318 gev . in 1998 and in the first months of 1999 , hera ran with electrons . in
may 1999 hera switched back to @xmath2 collisions .
the three main colliding periods as well as the corresponding luminosities for each experiment are summarized in table [ tab : lumit ] . * * . _
_ luminosities collected by h1 and zeus for each colliding period .
_ _ [ cols="^,^,^,^,^",options="header " , ] hera resulting limits on @xmath3 coupling are the following : * h1 : @xmath4 * zeus : @xmath5 figure [ fig : final ] summarizes the limits on the anomalous coulings @xmath6 ( @xmath7 vectorial coupling ) and @xmath8 ( @xmath3 magnetic coupling ) obtained at hera , lep and tevatron .
hera sensitivity to @xmath8 is competitive with other colliders .
hera is the unique collider to test direct @xmath10 interactions . about @xmath11 of @xmath2 data and @xmath12 of @xmath13 data
have been collected per experiment at hera-1 .
no evidence of new physics has been observed in various models in inclusive analyses ( contact interactions , extra - dimensions , leptoquarks ) and exclusive analyses ( lepton - flavour violation , @xmath14-violating susy , excited fermions ) , therefore new constraints have been set .
hera limits are seen to be competitive with and complementary to the lep and tevatron searches .
the status of isolated lepton events with missing @xmath15 is still intriguing and will become clearer with the new hera-2 data .
hera has been shutdown since fall 2000 for a general upgrade : new focussing magnets have been installed in order to increase the luminosity and many improvements have been performed in the detectors in order to increase their sensitivity . moreover ,
the lepton beam will be longitudinally polarised in the h1 and zeus interaction regions .
the first luminosity runs are predicted for beginning 2002 and hera-2 is expected to accumulate 1 @xmath16 in the next 5 years .
the anticipated factor of ten increase in the integrated luminosity will give an outstanding discovery potential for hera .
i which to thank my h1 and zeus colleagues who contributed to the results presented here as well as people who helped me in preparing this talk . | recent results on searches for physics beyond the standard model obtained by the h1 and zeus experiments are reported here .
after a brief introduction to the hera collider , indirect searches for contact interactions and extra - dimensions are presented as well as direct searches for new physics including leptoquarks , lepton - flavour violation , squarks produced by r - parity violation and excited fermions .
new results from isolated lepton events and single top searches are also presented .
finally the future prospects of hera-2 are shown .
* search for new particles at hera * + mireille schneider + cppm , 163 , avenue de luminy , case 907 , + 13288 marseille cedex 9 , france |
the in - spiral and coalescence of binary neutron star systems is a topic of increasingly intensive research in observational and theoretical astrophysics .
it is anticipated that the first direct detections of gravitational wave ( gw ) will be from compact binary mergers .
binary neutron star ( bns ) mergers are also thought to produce short - hard gamma - ray bursts ( sgrb s ) @xcite .
simultaneous detections of a prompt gravitational wave signal with a spatially coincident electromagnetic ( em ) counterpart dramatically increases the potential science return of the discovery .
for this reason , there has been considerable interest as to which , if any , detectable em signature may result from the merger @xcite .
other than sgrbs and their afterglows , including those viewed off - axis @xcite , suggestions include optical afterglows associated with the radio - active decay of tidally expelled r - process material@xcite ( though detailed calculations indicate they are faint @xcite ) , radio afterglows following the interaction of a mildly relativistic shell with the interstellar medium @xcite , and high - energy pre - merger emission from resistive magnetosphere interactions @xcite .
merging neutron stars possess abundant orbital kinetic energy ( @xmath4ergs ) .
a fraction of this energy is certain to be channelled through a turbulent cascade triggered by hydrodynamical instabilities during merger .
turbulence is known to amplify magnetic fields by stretching and folding embedded field lines in a process known as the small - scale turbulent dynamo @xcite .
amplification stops when the magnetic energy grows to equipartition with the energy containing turbulent eddies @xcite .
an order of magnitude estimate of the magnetic energy available at saturation of the dynamo can be informed by global merger simulations .
these studies indicate the presence of turbulence following the nonlinear saturation of the kelvin - helmholtz ( kh ) instability activated by shearing at the ns surface layers @xcite .
the largest eddies produced are on the @xmath5 km scale and rotate at @xmath6 , setting the cascade time @xmath7 and kinetic energy injection rate @xmath8 at @xmath9ms and @xmath10 respectively .
when kinetic equipartition is reached , each turbulent eddy contains @xmath11 of magnetic energy , and a mean magnetic field strength @xmath12 whether such conditions are realized in merging neutron star systems depends upon the dynamo saturation time @xmath13 and equipartition level @xmath14 . in particular , if @xmath15 then turbulent volumes of neutron star material will contain magnetar - level fields throughout the early merger phase .
once saturation is reached , a substantial fraction of the injected kinetic energy , @xmath16 , is resistively dissipated @xcite at small scales . magnetic energy dissipated by reconnection in optically thin surface layers
will accelerate relativistic electrons @xcite , potentially yielding an observable electromagnetic counterpart , independently of whether the merger eventually forms a relativistic outflow capable of powering a short gamma - ray burst . in this letter
we demonstrate that the small - scale turbulent dynamo saturates quickly , on a time @xmath15 , and that @xmath17 g magnetic fields are present throughout the early merger phase .
this implies that the magnetic energy budget of merging binary neutron stars is controlled by the rate with which hydrodynamical instabilities randomize the orbital kinetic energy .
our results are derived from simulations of the small scale turbulent dynamo operating in the high - density , trans - relativistic , and highly conductive material present in merging neutron stars .
we have carefully examined the approach to numerical convergence and report grid resolution criteria sufficient to resolve aspects of the small - scale dynamo .
our letter is organized as follows .
the numerical setup is briefly described in section 2 .
section 3 reports the resolution criterion for numerical convergence of the dynamo completion time and the saturated field strength . in section 4
we asses the possibility that magnetic reconnection events may convert a sufficiently large fraction of the magnetic energy into high energy photons to yield a prompt electromagnetic counterpart detectable by high energy observatories including _ swift _ and _ fermi_. and @xmath18 .
lower resolutions are shown in red and graduate to black with higher resolution .
_ top _ : the root mean square magnetic field strength in units of @xmath19 . when a turbulent volume is resolved by @xmath20 zones , the small - scale dynamo proceeds so slowly that almost no amplification is observed in the first 1ms .
_ middle _ : the magnetic energy in units of the rest mass @xmath21 shown on logarithmic axes .
it is clear that the linear growth rate increases at each resolution . _
bottom _ : the kinetic energy ( upper curves ) shown again the magnetic energy ( lower curves ) again in units of @xmath21 .
for all resolutions , the kinetic energy saturates in less than 1 @xmath7.,width=326 ]
the equations of ideal relativistic magnetohydrodynamics ( rmhd ) have been solved on the periodic unit cube with resolutions between @xmath20 and @xmath18 .
[ eqn : rmhd - system ] @xmath22 here , @xmath23 is the magnetic field four - vector , and @xmath24 is the total specific enthalpy , where @xmath25 is the total pressure , @xmath26 is the gas pressure and @xmath27 is the specific internal energy .
the source term @xmath28 includes injection of energy and momentum at the large scales and the subtraction of internal energy ( with parameter @xmath29mev ) to permit stationary evolution .
vortical modes at @xmath30 are forced by the four - acceleration field @xmath31 which smoothly decorrelates over a large - eddy turnover time , as described in @xcite .
we have employed a realistic micro - physical equation of state ( eos ) appropriate for the conditions of merging neutron stars .
it includes contributions from high - density nucleons according a relativistic mean field model @xcite , a relativistic and degenerate electron - positron component , neutrino and anti - neutrino pairs in beta equilibrium with the nucleons , and radiation pressure . for our conditions ,
all the components make comparable contributions to the pressure . we have also employed a simpler gamma - law eos and found close agreement for the conditions explored in this paper , indicating that the mhd effects are insensitive the eos for trans - sonic conditions .
the models presented in our resolution study use the far less expensive gamma - law equation of state .
all of the simulations presented in this study use the hlld approximate riemann solver @xcite , which has been demonstrated as crucial in providing the correct spatial distribution of magnetic energy in mhd turbulence @xcite .
the solution is advanced with an unsplit , second - order muscl - hancock scheme .
spatial reconstruction is accomplished with the piecewise linear method configured to yield the smallest possible degree of numerical dissipation .
the divergence constraint on the magnetic field is maintained to machine precision at cell corners using the finite volume ct method of @xcite .
full details of the numerical scheme may be found in @xcite .
magnetic fields are amplified in our simulations by the small - scale turbulent dynamo .
turbulent fluid motions stretch and fold the magnetic field lines , causing exponential growth of magnetic energy ( e.g. * ? ? ?
this growth is attributed to the advection and diffusion of @xmath32 through the mhd induction equation ( eq . [ eqn : rmhd - system]c ) .
when the magnetic field is weak ( @xmath33 ) @xmath32 evolves passively , and the turbulence is hydrodynamical .
this limit is referred to as small - scale kinematic turbulent dynamo , and is well described by kazentzev s model @xcite .
this model predicts that the power spectrum of magnetic energy peaks at the resistive scale @xmath34 and obeys a power law @xmath35 at longer wavelengths .
the kinematic phase ends when the magnetic field acquires sufficient tension to modify the hydrodynamic motions , after which time a dynamical balance between kinetic and magnetic energy is established .
numerical simulations of mhd turbulence are typically limited to magnetic prandtl numbers @xmath36 .
however , neutron star material is characterized by @xmath37 , with the viscous cutoff due to neutrino diffusion occurring at around 10 cm , while the resistive scale is significantly smaller @xcite .
however , the disparity between true and simulated magnetic prandlt number does not influence our conclusions .
this is because dynamos are generically easier to establish in the high pm regime than the small @xcite .
, together with the empirical model ( equation [ eqn : magnetic - fit ] ) with best - fit parameters .
the horizontal dashed line indicates the magnetic energy , @xmath38 at the dynamo completion . from left to right , the vertical dashed lines mark the end of the startup , kinematic , and saturation phases.,width=326 ] we use an initially uniform , pulsar - level ( @xmath39 g ) seed magnetic field .
this field is sub - dominant to the kinetic energy by 10 orders of magnitude , so that the initial field amplification is expected to be well - described by kazentsev s model .
indeed , we find that during this phase the power spectrum of magnetic energy follows @xmath35 ( fig .
[ fig : pspec - time - devel - res ] ) , peaking at around 10 grid zones , which we identify as the effective scale of resistivity .
the saturation process begins at ever - earlier times with increasing numerical resolution .
this reflects the fact that during the kinematic phase , magnetic energy exponentiates on a time scale controlled by shearing at the smallest scales . in numerically converged runs ,
full saturation occurs with @xmath40 and is characterized by scale - by - scale super - equipartition , with @xmath41 at all but the largest scale .
the same driven turbulence model was run through magnetic saturation at resolutions @xmath20 , @xmath42 , @xmath43 , @xmath44 , @xmath45 , and @xmath46 .
another model at @xmath18 was run through the end of the kinematic phase , but further evolution was computationally prohibitive .
[ fig : energy - growth ] shows the development of @xmath47 , @xmath48 , and @xmath49 as a function of time at each resolution .
we find that sufficiently resolved runs ( @xmath50 ) attain _ mean _ magnetic field strengths of @xmath51 g within two large eddy rotations .
all models with resolutions @xmath52 eventually attain mean fields of @xmath53 g .
the saturated field strength increases until resolution @xmath45 .
we find that the kinematic growth rate is higher at each higher resolution , while the time - scale for the non - linear saturation converges at @xmath45 to roughly five large - eddy rotation times , or about 0.5ms for the physical parameters of binary neutron star mergers .
lcc @xmath54 & startup time - scale & none , artifact @xmath55 & hydrodynamic cascade fully developed & @xmath7 @xmath56 & kinematic growth time - scale & none , @xmath57 @xmath58 & end of kinematic phase & @xmath59 @xmath60 & non - linear saturation time - scale & @xmath61 , @xmath45 @xmath38 & saturated magnetic energy & @xmath62 , @xmath46 [ tab : magnetic - fit ] in order to quantitatively describe the time development of magnetic energy @xmath63 , we describe it with an empirical model , @xmath64 where the 6 parameters ( summarized in table [ tab : magnetic - fit ] ) are obtained by a least - squares optimization .
[ fig : magnetic - fit ] shows the empirical model given in equation [ eqn : magnetic - fit ] applied to a representative run at @xmath44 . the first phase , @xmath65 is a startup transient , and lasts until the hydrodynamic cascade is fully developed at @xmath66 .
the kinematic dynamo phase lasts between @xmath55 and @xmath58 , during which the magnetic energy exponentiates on the time scale @xmath56 . at @xmath58 ,
the smallest scales reach kinetic equipartition and the growth rate slows . in the final stage ,
@xmath48 asymptotically approaches @xmath38 on the time - scale @xmath60 .
we define the dynamo completion time @xmath13 as @xmath67 .
[ fig : convergence ] shows the best - fit @xmath38 , @xmath13 , and @xmath56 as a function of the resolution .
the magnetic energy @xmath38 at dynamo completion shows signs of converging to a value of @xmath62 by resolution @xmath46 .
the time scale @xmath60 on which the magnetic energy asymptotically approaches @xmath38 is consistently @xmath68 at different resolutions .
the dynamo completion time @xmath13 is numerically converged at @xmath69 by @xmath45 .
the best - fit kinematic growth time follows a power law in the resolution , @xmath70 .
this is consistent with the value of @xmath71 expected if the dynamo time is controlled by shearing at the smallest scale , the cascade is kolmogorov ( i.e. @xmath72 ) , and the viscous cutoff @xmath73 occurs at a fixed number of grid zones . in that case ,
@xmath74 . ( _ blue _ ) and the dynamo completion time @xmath13(_green _ ) defined as @xmath75 .
_ bottom _ : convergence study of the best - fit model parameter @xmath38 expressed as the ratio of magnetic to kinetic energy @xmath76 .
the converged value of the volume - averaged @xmath77 .
nevertheless , at intermediate wavelengths @xmath78 .
as shown in figure [ fig : pspec - time - devel - res ] , the largest scales remain kinetically dominated .
this indicates the suppression of coherent magnetic structure formation near the integral scale of turbulence.,width=326 ] the time development of kinetic and magnetic energy power spectra has been studied for a single run with resolution @xmath46 .
we present three - dimensional , spherically integrated power spectra with the dimensions of @xmath79 , defined as @xmath80\right|^2}\\ p_m(k_i ) & = \frac{1}{\delta k_i}\sum _ { { \mathbf k } \in \delta k_i } { \left|\mathcal{f } _ { { \mathbf k } } \left [ { \mathbf b } /\sqrt{8\pi}\right]\right|^2 } \end{aligned}\ ] ] [ eqn : pspec ] where the newtonian versions of kinetic and magnetic energy are appropriate since the conditions are only mildly relativistic .
the definitions in equations [ eqn : pspec ] satisfy @xmath81 for @xmath82 and @xmath83 .
figure [ fig : pspec - time - devel - res ] shows the power spectrum of kinetic and magnetic energy at various times throughout the growth and saturation of magnetic field . during the kinematic phase ,
the kinetic energy has a power spectrum @xmath84 consistent with the kolmogorov theory for incompressible hydrodynamical turbulence , while @xmath85 consistent with kazenstev s model .
@xmath86 maintains the same shape , but exponentiates in amplitude at the time scale @xmath56 which is controlled by shearing at the resistive scale .
according to kazentsev s model , @xmath86 should peak at the resistive scale .
this is consistent with the observed peak in the magnetic energy at roughly 10 grid zones , the same scale at which we observe the viscous cutoff .
this is also consistent with @xmath87 expected from the numerical scheme employed .
when the magnetic energy at the resistive scale surpasses the level of the kinetic energy at that scale , @xmath86 changes shape .
the equipartition scale @xmath88 moves into the inertial range , and migrates to larger scale until full saturation occurs with @xmath89 .
the movement of @xmath88 to larger scale is associated with the formation of coherent and dynamically substantial magnetic structures of increasing size .
the time - dependence has been suggested to be @xmath90 @xcite . in the fully saturated state ,
the magnetic field is in scale - by - scale super - kinetic equipartition throughout the inertial range , with @xmath91 .
the largest scales remain kinetically dominated so that the numerically converged saturation level is @xmath92 .
in this letter we have determined the time scale and saturation level of the small - scale turbulent dynamo operating in the conditions of binary neutron star mergers .
we have presented numerically converged simulations showing that magnetic fields are amplified to the @xmath93 g level within a small fraction of the merger dynamical time ( @xmath15 ) , independently of the seed field strength . if hydrodynamical instabilities create fluctuating velocities on the order of @xmath94 as indicated by global simulations , then each @xmath95 turbulent volume dissipates @xmath96 erg of magnetic energy per 0.1 ms .
if @xmath97 represents the fraction of the merger remnant that contains such turbulence , the magnetic energy dissipated during the merger is at least @xmath98 a fraction of that dissipation will occur through magnetic reconnection in optically thin surface layers , supplying relativistic electrons which synchrotron radiate in the merger remnant magnetosphere .
if 5% of that magnetic energy dissipation creates radiation in the 15 - 150 kev band , then the fluence at 200 mpc would be @xmath3 , potentially rendering most merging neutron stars in the advanced ligo and virgo detection volumes detectable by _ swift _ bat .
if so , then merging neutron stars are accompanied by a prompt electromagnetic counterpart , independently of whether a later merger phase yields a collimated outflow capable of powering a short gamma - ray burst .
we suggest that merger flares may be present in the current sample of short grbs and may be roughly isotropic on the sky since they are seen to distances where the cosmological matter distribution becomes homogeneous .
searches for merger flares should seek to identify short flares , not unlike soft - gamma repeaters , among the short burst population . if mergers also produce short grbs short - hard grbs , then merger flares may constitute a precursor component of the emission .
the presence of strong magnetic fields may also aid in the ejection of neutron - rich material from surface layers of the merger remnant , possibly enhancing the enrichment of inter - stellar medium by r - process nuclei @xcite .
enhanced production of r - process nuclei also increases the likelihood of em detection by radio - active decay powered afterglows , or `` kilonovae '' @xcite .
finally , it has been shown that magnetic fields will significantly influence the gravitational wave signature and remnant disk mass , if they exist at the @xmath99 g level @xcite .
such strong fields are unlikely in older neutron star binaries , but our results suggest they may be revived , albeit at small scales , during the merger . to have significant influence , those fields would have to fill a considerable fraction of the merger volume .
as we have shown in this letter , the overall magnetic energy budget is controlled by the prevalence ( @xmath97 ) and vigor ( @xmath100 ) of the turbulent volumes .
this fact motivates the use of higher resolution global simulations aimed at measuring @xmath97 and @xmath100 .
this research was supported in part by the nsf through grant ast-1009863 and by nasa through grant nnx10af62 g issued through the astrophysics theory program .
resources supporting this work were provided by the nasa high - end computing ( hec ) program through the nasa advanced supercomputing ( nas ) division at ames research center . | the simultaneous detection of electromagnetic and gravitational wave emission from merging neutron star binaries would aid greatly in their discovery and interpretation . by studying turbulent amplification of magnetic fields in local high - resolution simulations of neutron star merger conditions ,
we demonstrate that magnetar - level ( @xmath0 g ) fields are present throughout the merger duration .
we find that the small - scale turbulent dynamo converts 60% of the randomized kinetic energy into magnetic fields on a merger time scale .
since turbulent magnetic energy dissipates through reconnection events which accelerate relativistic electrons , turbulence may facilitate the conversion of orbital kinetic energy into radiation .
if @xmath1 of the @xmath2 erg of orbital kinetic available gets processed through reconnection , and creates radiation in the 15 - 150 kev band , then the fluence at 200 mpc would be @xmath3 , potentially rendering most merging neutron stars in the advanced ligo and virgo detection volumes detectable by _ swift _ bat . |
oscillatory integrals play an important role in the theory of pseudodifferential operators .
they are also a useful tool in mathematical physics , in particular in quantum field theory , where they are used to give meaning to formal fourier integrals in the sense of distributions . for phase functions which are homogeneous of order one
, this also leads to a characterization of the wave front set of the resulting distribution , as it is known to be contained in the manifold of stationary phase of the phase function . in these applications ,
the restriction to phase functions that are homogeneous of order one is often obstructive . in many cases ,
this restriction can be overcome by shifting a part of the would - be phase function to the symbol , cf .
example [ ex : delta+ ] below .
however , such a shift is not always possible , for instance if the would - be phase function contains terms of order greater than one .
such phase functions are present in the twisted convolutions that occur in quantum field theory on moyal space , cf .
examples [ ex : ncqft ] and [ ex : ncqftb ] below . up to now
, a rigorous definition of these twisted convolution integrals could be given only in special cases and in such a way that the information on the wave front set is lost .
thus , it is highly desirable to generalize the notion of oscillatory integrals to encompass also phase functions that are not homogeneous of order one .
such generalizations were proposed by several authors . however , to the best of our knowledge , the wave front sets of the resulting distributions were not considered , except for one very special case .
we comment on these settings below , cf .
remark [ rem : asadafujiwara ] .
it is shown here that the restriction to phase functions that are homogeneous of order one can indeed be weakened , without losing information on the wave front set .
the generalization introduced here not only allows for inhomogeneous phase functions , but also for phase functions that are symbols of any positive order .
however , one has to impose a condition that generalizes the usual nondegeneracy requirement .
it is also shown that the wave front sets of the distributions thus obtained are contained in a set that generalizes the notion of the manifold of stationary phase .
we conclude with a discussion of some applications . throughout , we use the following notation : for an open set @xmath0 , @xmath1 means that @xmath2 is a compact subset of @xmath0 .
@xmath3 stands for @xmath4 . for a subset @xmath5 , @xmath6 stands for the projection on the first component
@xmath7 denotes the @xmath8 times continuously differentiable functions supported on @xmath0 and @xmath9 the set of elements of @xmath7 with compact support in @xmath0 .
the dual space of @xmath10 is denoted by @xmath11 and @xmath12 .
the pairing of a distribution @xmath13 and a test function @xmath14 is denoted by @xmath15 .
the dot @xmath16 stands for the scalar product on @xmath17 .
@xmath18 denotes the angle between two vectors @xmath19 .
as usual , cf .
@xcite , we define a symbol as follows : let @xmath20 be an open set .
a function @xmath21 is called a _ symbol of order @xmath22 _
if for each @xmath1 and multiindices @xmath23 , we have @xmath24 the set of all such functions , equipped with these seminorms will be denoted by @xmath25 .
furthermore , we denote @xmath26 and @xmath27 . for simplicity , we restrict ourselves to these symbols . the generalization to the symbols
@xmath28 is straightforward .
one then has to restrict to @xmath29 , @xmath30 , where @xmath31 is the order of the phase function introduced below .
also the generalization to asymptotic symbols as discussed in @xcite is straightforward .
the following proposition is a straightforward consequence of the definition of @xmath25 : [ prop : cont ] the maps @xmath32 , @xmath33 and the multiplication @xmath34 are continuous .
the following proposition is proven in ( * ? ? ?
1.7 ) : [ prop : dense ] if @xmath35 , then @xmath36 is dense in @xmath25 for the topology of @xmath37 .
now we introduce our new definition of a phase function .
[ def : phase ] a _ phase function _ of order @xmath31 on @xmath38 is a function @xmath39 such that 1 .
@xmath40 is a symbol of order @xmath41 .
[ enum : phase ] for each @xmath1 there are positive @xmath42 such that @xmath43 [ rem : phase ] condition [ enum : phase ] generalizes the usual nondegeneracy requirement and ensures that @xmath40 oscillates rapidly enough for large @xmath44 .
in particular it means that @xmath40 is not a symbol of order less than @xmath31 .
it also means that one can choose @xmath45 such that @xmath46 is well - defined and a symbol of order @xmath47 . here
@xmath48 can be chosen such that @xmath49 is compact for each @xmath50 .
[ rem : asadafujiwara ] our definition of a phase function is a generalization of a definition introduced by hrmander ( * ? ? ?
2.3 ) in the context of pseudodifferential operators .
he considered phase functions of order 1 ( in the nomenclature introduced above ) and characterized the singular support of the resulting distribution , but not its wave front set .
our characterization of the singular support ( cf .
corollary [ cor : m ] ) coincides with the one given by hrmander ( * ? ? ?
inhomogeneous phase functions were also considered by asada and fujiwara @xcite in the context of pseudodifferential operators on @xmath51 . in their setting , @xmath52 , @xmath53 and
there must be a positive constant @xmath54 such that @xmath55 furthermore , all the entries of this matrix ( and their derivatives ) are required to be bounded .
thus , the phase function is asymptotically at least of order 1 and at most of order 2 .
the admissible amplitudes are at most of order 0 . the wave front set of such operators on @xmath51
is not considered by asada and fujiwara .
the same applies to the works of boulkhemair @xcite and ruzhansky and sugimoto @xcite , who work in a similar context .
coriasco @xcite considered a special case of hrmander s framework , where again @xmath52 , @xmath53 and @xmath56 with @xmath57 , a subset of the symbols of order 1 .
furthermore , he imposed growth conditions on @xmath58 that are more restrictive than condition [ enum : phase ] .
the resulting operators on @xmath59 can then be extended to operators on @xmath60 .
if a further condition analogous to is imposed , then also the wave front set , which is there defined via @xmath61-microregularity , can be characterized ( at least implicitly , by the change of the wave front set under the action of the operator ) .
[ prop : diff ] if @xmath40 is a phase function of order @xmath31 and @xmath62 and there is a @xmath63 such that @xmath64 , then @xmath65 and the map @xmath66 is continuous . for @xmath67
we have @xmath68 so that @xmath69 is continuous
. differentiation gives @xmath70 the expression in curly brackets is a symbol of order @xmath71 . with the same argument as before one
can thus differentiate @xmath72 times .
we formulate the main theorem of this section analogously to ( * ? ? ? * thm .
the proof is a straightforward generalization of the proof given there .
[ thm : osc ] let @xmath73 be a phase function of order @xmath31 on @xmath38 .
then there is a unique way of defining @xmath74 for @xmath75 such that @xmath76 coincides with when @xmath62 for some @xmath67 and such that , for all @xmath22 , the map @xmath77 is continuous .
moreover , if @xmath63 and @xmath78 , then the map @xmath79 is continuous . to prove this , we need the following lemma : [ lemma : v ] let @xmath40 be a phase function of order @xmath31 on @xmath38 .
then there exist @xmath80 , @xmath81 and @xmath82 such that for the differential operator @xmath83 with adjoint @xmath84 we have @xmath85 furthermore , @xmath86 is a continuous map from @xmath25 to @xmath87 .
we choose @xmath88 as in definition [ def : phase ] and @xmath48 as in remark [ rem : phase ] .
we set @xmath89 then we have @xmath90 it is easy to see that @xmath91 , @xmath92 and @xmath93 are symbols in the required way .
the last statement follows from proposition [ prop : cont ] .
the uniqueness is a consequence of proposition [ prop : dense ] . for @xmath94 and @xmath95
, we have , with @xmath86 as in lemma [ lemma : v ] , @xmath96 { \mathrm{d}}^nx { \mathrm{d}}^s\theta,\end{aligned}\ ] ] for any @xmath63 and thus @xmath97 \rvert } { \mathrm{d}}^nx { \mathrm{d}}^s\theta.\ ] ] now the multiplication @xmath98 is continuous .
thus , if @xmath62 , then @xmath99 $ ] is a symbol of order @xmath100 and in particular we have , for each @xmath1 , @xmath101 \rvert } ( 1+{\lvert \theta \rvert})^{p\mu - m } \leq f_{p , k}(a ) \sum_{{\lvert \alpha \rvert } \leq p } \sup_{x \in k } { \lvert d^\alpha f \rvert},\ ] ] where @xmath102 is a seminorm on @xmath25 . for @xmath62 , we may thus choose @xmath72 such that @xmath103 and define @xmath104 { \mathrm{d}}^nx { \mathrm{d}}^s\theta.\ ] ] as the sum on the r.h.s . of is a seminorm on @xmath105 , and due to the continuity properties discussed above , the map @xmath106 is continuous . for @xmath107
, we have @xmath108 by .
thus , we can unambiguously define @xmath109 .
we may now further characterize the distributions that result from a generalized oscillatory integral .
[ def : sp ] let @xmath40 be a phase function of order @xmath31 on @xmath38 .
we define @xmath110 we call @xmath111 the _ asymptotic manifold of stationary phase_. by definition , it is conic .
[ lemma : sp ] @xmath112 is a closed conic subset of @xmath113 .
@xmath111 is a closed subset of @xmath114 . from the definition of @xmath112 it follows that if @xmath115 , then @xmath116 for all @xmath117 , so @xmath112 is conic .
we now show that @xmath118 is open in @xmath113 .
let @xmath119 be such that there are positive @xmath42 such that @xmath120 we set @xmath121 .
by taylor s theorem we have @xmath122 where @xmath123 and @xmath124 fulfill
the bounds @xmath125 here @xmath126 are chosen such that @xmath127 , @xmath128 .
furthermore , we restrict to @xmath129
. then we may use that @xmath13 is a symbol of order @xmath130 to conclude that there are positive constants @xmath131 , which are bounded for @xmath132 , for which @xmath133 holds . as the zeroth order term in the taylor expansion grows faster than @xmath134 for a fixed positive constant @xmath54
, we can make @xmath135 so small that @xmath136 grows faster than @xmath137 for some positive @xmath138 .
thus , @xmath112 is closed in @xmath113 . in order to prove the closedness of @xmath111
, we first note that if @xmath139 , then by the above there is a neighborhood of @xmath140 that does not intersect @xmath141 .
thus , it suffices to show that for @xmath142 , @xmath143 there is a neighborhood that does not intersect @xmath111 . by condition [ enum :
phase ] of definition [ def : phase ] , there must be positive constants @xmath42 such that @xmath144 by the same argument as above , such a bound holds true in a conic neighborhood @xmath145 of @xmath146 . by the definition of @xmath111 , there are posititve @xmath147 such that @xmath148 we now want to show that one can choose a conic neighborhood @xmath149 of @xmath119 , contained in @xmath145 , such that an analogous bound holds , i.e. , there are positive @xmath147 so that @xmath150 by the above construction , @xmath151 grows as @xmath152 in @xmath145 . the deviations that occur by varying @xmath50 and @xmath44 also scale as @xmath152 , as @xmath40 is a symbol of order @xmath31 .
recalling @xmath153 and again using taylor s theorem , one shows that by making @xmath149 small enough , one still retains an inequality of the form . by suitably restricting @xmath149 in @xmath0
, we can ensure that for @xmath154 we have @xmath155 whenever @xmath115 .
then no new direction @xmath44 for which we would have to check the bound can appear while varying @xmath50 in @xmath156 .
given , it is clear that we can also take a conic neighborhood @xmath157 of @xmath158 , by tilting it by angles less than @xmath159 .
choosing @xmath160 gives a neighborhood of @xmath161 that does not intersect @xmath111 .
[ prop : m ] if the support of the symbol @xmath162 does not intersect @xmath112 , then @xmath76 is smooth .
we choose a neighborhood @xmath163 of @xmath164 whose closure does not intersect @xmath112 . we choose a smooth function @xmath48 that is equal to one in a neighborhood of @xmath112 and vanishes on @xmath165 .
we choose another smooth function @xmath166 on @xmath38 with support in @xmath163 which is identical to one whenever @xmath167 for @xmath168 . by definition of @xmath112 , the set @xmath169 is bounded , so we can choose @xmath166 such that @xmath170 is compact for each @xmath50
. then we define @xmath171 by the definition of @xmath112 and @xmath48 , we have @xmath80 and @xmath82 . with these definitions , we have @xmath172 here we used that @xmath48 and @xmath166 have nonoverlapping supports . as @xmath173 and @xmath86 differentiates only w.r.t .
@xmath44 , we thus have @xmath174 for arbitrary integer @xmath72 . as @xmath86 maps symbols of order @xmath22 to symbols of order @xmath175 , @xmath76 is smooth by proposition [ prop : diff ] .
[ cor : m ] the singular support of @xmath76 is contained in @xmath141 .
[ thm : sp ] the wave front set of @xmath76 is contained in @xmath111 .
[ lemma : sp1 ] let @xmath176 , @xmath143 .
then there is a conic neighborhood @xmath145 of @xmath119 , a conic neighborhood @xmath86 of @xmath158 and positive constants @xmath177 such that @xmath178 furthermore , there is a positive @xmath138 such that @xmath179 condition is fulfilled for @xmath119 by condition [ enum : phase ] of definition [ def : phase ] . that it is also fulfilled in a neighborhood of @xmath119
can be shown analogously to the proof of the closedness of @xmath112 in lemma [ lemma : sp ] .
condition is fulfilled for @xmath180 .
that it is also fulfilled in a neighborhood of @xmath180 can again be shown as in lemma [ lemma : sp ] . in order to prove the last statement , we note that by we have @xmath181 where @xmath182 has length @xmath183 and lies on the cone with angle @xmath159 around @xmath8 ( see the figure , where @xmath184 is denoted by @xmath72 ) . for the distance of @xmath8 and @xmath185
we have the bound ( see the dashed lines in the figure ) @xmath186 using , we then obtain the above statement . by corollary [ cor : m ]
, it suffices to consider points @xmath187 .
let @xmath143 .
due to proposition [ prop : m ] , we may assume that @xmath162 is supported in an arbitrarily small closed conic neighborhood @xmath145 of @xmath112 .
we thus need to show that there is a @xmath188 , identically one near @xmath140 and a conic neighborhood @xmath86 of @xmath158 such that for each @xmath189 there is a seminorm @xmath190 on @xmath25 such that @xmath191 for all @xmath162 supported in @xmath145 . as in the proof of theorem [ thm : osc ] , it suffices to construct such a bound for @xmath94 and then make use of proposition [ prop : dense ] .
let @xmath192 be as in lemma [ lemma : sp1 ] .
choose @xmath193 that is identically one near @xmath140 and whose support is contained in @xmath194 .
choose a @xmath195 that is identical to one on @xmath196 .
now we set @xmath197 then for @xmath198 we have @xmath199 now we choose @xmath200 , identical to one near @xmath140 and with support in @xmath201
. we also choose @xmath202 which is identically one on @xmath203 .
we consider @xmath204 . then by proposition
[ prop : diff ] , the second term yields a smooth function , so that the above bound is fulfilled .
it remains to consider @xmath205 { \mathrm{d}}^n x { \mathrm{d}}^s \theta \rvert } \\ & \leq \int { \lvert v_k^p[\psi(x ) ( 1-\xi(\theta ) ) a(x , \theta ) ] \rvert } { \mathrm{d}}^n x { \mathrm{d}}^s \theta.\end{aligned}\ ] ] here we used that @xmath206 is identically one on the support of @xmath207 . by lemma [ lemma : sp1 ] , one now has @xmath208 \rvert } \leq c_{m , p}(a ) ( 1+{\lvert \theta \rvert})^m ( { \lvert \theta \rvert}^\mu + { \lvert k \rvert})^{-p},\ ] ] where @xmath209 is a seminorm on @xmath25 .
by using @xmath210 one can make the integral convergent and assure by choosing @xmath72 large enough .
[ ex : delta+ ] we consider the two - point function @xmath211 of a free massive scalar relativistic field , where one has @xmath212 and @xmath213 with @xmath214 here , we use the notation @xmath215 , with @xmath216 . note that @xmath40 is not homogeneous . in @xcite this problem
is circumvented by using @xmath217 as phase function and multiplying the symbol with the function @xmath218 .
it is then no longer a symbol ( as it is not smooth in @xmath219 ) , so one has to allow for so - called asymptotic symbols .
furthermore , one has to show that the multiplication with such a term does not spoil the fall - off properties , in particular that differentiation w.r.t .
@xmath44 lowers the order . in the present approach
, this is not necessary .
@xmath40 is a phase function in the sense of definition [ def : phase ] and therefore , by theorem [ thm : osc ] , it defines an oscillatory integral for every symbol @xmath162 . in order to find the wave front set ,
we compute @xmath220 it is easy to see that its modulus is bounded from below by a positive constant unless @xmath221 or @xmath222 and @xmath223 .
thus , we have @xmath224 furthermore , @xmath225 for large @xmath44 , this behaves as @xmath226 where the remainder term @xmath227 scales as @xmath228 .
thus , only in the directions @xmath229 the bound on the angle of @xmath8 and @xmath230 can not be fulfilled .
hence , we obtain the well - known result . in that convention ,
the sign of the zeroth component in the cotangent bundle has to be reversed . ]
@xmath231 a variant of this example is obtained by considering phase functions of the form @xmath232 with @xmath14 a positive function that is a symbol of order @xmath233 .
such expressions occur for example in quantum field theory on the moyal plane with hyperbolic signature and signify a distortion of the dispersion relations , cf .
the above trick to put @xmath234 into the symbol still works , but then the symbol will be of type @xmath235 , where @xmath236 is original type of the symbol .
it is straightforward to check that still defines a phase function of order 1 in the sense defined here , and that its stationary phase is as above .
if the function @xmath14 in is a symbol of order @xmath237 , of corresponds to solutions of the hyperbolic wave equation @xmath238 .
the modification suggested here means that the underlying pde is no longer hyperbolic . ]
then the shift of @xmath234 to the symbol is not possible , as this would no longer give a symbol of type @xmath239 .
thus , a treatment of in the context of phase functions that are homogeneous of order 1 is not possible .
however , one can still interpret as a phase function of order @xmath240 and easily computes @xmath241 [ ex : ncqft ] in quantum field theory on the moyal plane of even dimension @xmath242 with euclidean signature , one frequently finds phase funtions of the form @xmath243 here @xmath244 is some real antisymmetric @xmath245 matrix of maximal rank @xmath242 .
the above is clearly a symbol of order 2 , and we have @xmath246 as @xmath244 has rank @xmath242 , condition [ enum : phase ] of definition [ def : phase ] is fulfilled . from the above
it follows that @xmath247 and thus also @xmath248 , so that the resulting distributions are smooth .
we note that up to now such integrals could only be treated in the so - called adiabatic limit @xcite .
but then one loses the information about the singular behaviour in position space , contrary to the present case , where the wave front set is known completely .
[ ex : ncqftb ] in quantum field theory on the moyal plane with hyperbolic signature , one frequently finds phase functions of the form @xmath249 where @xmath250 with @xmath251 as in and @xmath244 as in example [ ex : ncqft ] .
the above is a symbol of order 2 , but it is not a phase function as defined here , as can most easily be seen in the case @xmath252
. then with @xmath253 one obtains @xmath254 if the signs of @xmath255 and @xmath256 coincide , then the above derivatives tend to a constant as a function of @xmath44 , so that condition [ enum : phase ] of definition [ def : phase ] is not fulfilled .
the rigourous treatment of such integrals is an open problem , which we plan to address in future work .
it is a pleasure to thank dorothea bahns for helpful discussions and her detailed comments on the manuscript .
i would also like to thank ingo witt for valuable comments .
this work was supported by the german research foundation ( deutsche forschungsgemeinschaft ( dfg ) ) through the institutional strategy of the university of gttingen .
d. bahns , s. doplicher , k. fredenhagen and g. piacitelli , _ field theory on noncommutative spacetimes : quasiplanar wick products _ , phys .
d * 71 * , 025022 ( 2005 ) . c. dscher and j. zahn , _ dispersion relations in the noncommutative @xmath258 and wess - zumino model in the yang - feldman formalism _ , annales henri poincare * 10 * , 35 ( 2009 ) .
r. wulkenhaar , _ field theories on deformed spaces _ , j. geom .
phys . *
56 * , 108 ( 2006 ) . c. dscher , _ yang - feldman formalism on noncommutative minkowski space _ , ph.d .
thesis , hamburg ( 2006 ) . | a generalized notion of oscillatory integrals that allows for inhomogeneous phase functions of arbitrary positive order is introduced . the wave front set of the resulting distributions
is characterized in a way that generalizes the well - known result for phase functions that are homogeneous of order one . |
a significant theory regarding inequalities and exponential convexity for real valued functions , has been developed @xcite . the intention to generalize such concepts for the @xmath0-semigroup of operators ,
is motivated from @xcite .
+ in the present article , we shall derive a jessen s type inequality and the corresponding adjoint - inequality , for some @xmath0-semigroup and the adjoint - semigroup , respectively .
+ + the notion of banach lattice was introduced to get a common abstract setting , within which one could talk about the ordering of elements .
therefore , the phenomena related to positivity can be generalized .
it had mostly been studied in various types to spaces of real - valued functions , e.g. the space @xmath1 of continuous functions over a compact topological space @xmath2 , the lebesque space @xmath3 or even more generally the space @xmath4 constructed over measure space @xmath5 for @xmath6 .
we shall use without further explanation the terms , order relation ( ordering ) , ordered set , supremum , infimum .
+ firstly , we shall go through the definition of vector lattice .
any ( real ) vector space @xmath7 with an ordering satisfying ; @xmath8 : : : @xmath9 implies @xmath10 for all @xmath11 @xmath12 : : : @xmath13 implies @xmath14 for al @xmath15 and @xmath16 is called an _ ordered vector space_. the axiom @xmath8 , expresses the translation invariance and therefore implies that the ordering of an ordered vector space @xmath7 is completely determined by the positive part @xmath17 of @xmath7 . in other words , @xmath9 if and only if @xmath18 .
moreover , the other property @xmath12 , reveals that the positive part of v is a convex set and a cone with vertex @xmath19 ( mostly called the _ positive cone _ of v ) . * an ordered vector space @xmath7 is called a _ vector lattice _
, if any two elements @xmath20 have a supremum , which is denoted by @xmath21 and an infimum denoted by @xmath22 .
+ it is trivially understood that the existence of supremum of any two elements in an ordered vector space implies the existence of supremum of finite number of elements in @xmath7 .
furthermore , @xmath23 implies @xmath24 , so the existence of finite infima therefore implied . *
few important quantities are defined as follows @xmath25 * some compatibility axiom is required , between norm and order .
this is given in the following short way : @xmath26 the norm defined on a vector lattice is called a lattice norm .
now , we are in a position , to define a banach lattice in a formal way .
a _ banach lattice _ is a banach space @xmath7 endowed with an ordering @xmath27 , such that @xmath28 is a vector lattice with a lattice norm defined on it . +
a banach lattice transforms to _ banach lattice algebra _ , provided @xmath29 implies @xmath30 .
@xmath31 + a linear mapping @xmath32 from an ordered banach space @xmath7 into itself is _ positive _ ( denoted by : @xmath33 ) if @xmath34 , for all @xmath35 .
the set of all positive linear mappings forms a convex cone in the space @xmath36 of all linear mappings from @xmath7 into itself , defining the natural ordering of @xmath36 .
the absolute value of @xmath32 , if it exists , is given by @xmath37 thus @xmath38 is positive if and only if @xmath39 holds for any @xmath15 . *
[ @xcite , p.249 ] * a bounded linear operator @xmath32 on a banach lattice v is a positive contraction if and only if @xmath40 for all @xmath15 .
@xmath31 + an operator @xmath41 on @xmath7 satisfies the positive minimum principle if for all @xmath42 , @xmath43 @xmath44 a ( one parameter ) @xmath0-semigroup ( or strongly continuous semigroup ) of operators on a banach space @xmath7 is a family @xmath45 such that ( i ) : : @xmath46 for all @xmath47 .
( ii ) : : z(0)=i , the identity operator on v. ( iii ) : : for each fixed @xmath15 , @xmath48(with respect to the norm on v ) as @xmath49 . where @xmath50 denotes the space of all bounded linear operators defined on a banach space v. the ( infinitesimal ) generator of @xmath51 is the densely defined closed linear operator @xmath52 such that @xmath53 @xmath54 where , for @xmath55 , @xmath56f}{t}\,\,\,\,(f\in v).\ ] ] @xmath31 + let @xmath51 be the strongly continuous positive semigroup , defined on a banach lattice v. the positivity of the semigroup is equivalent to @xmath57 where for positive contraction semigroups @xmath51 , defined on a banach lattice v we have
; @xmath58 the literature presented in @xcite , guarantees the existence of the strongly continuous positive semigroups and positive contraction semigroups on banach lattice v , with some conditions imposed on the generator .
the very important amongst them is , that it must always satisfy ( [ pmp ] ) .
+ a banach algebra @xmath59 , with the multiplicative identity element @xmath60 , is called the _
unital banach algebra_. we shall call the strongly continuous semigroup @xmath51 defined on @xmath59 , a _ normalized semigroup _
, whenever it satisfies @xmath61 the notion of normalized semigroup is inspired from normalized functionals @xcite .
the theory presented in next section , is defined on such semigroups of positive linear operators defined on a banach lattice @xmath7 .
in 1931 , jessen @xcite gave the generalization of the jensen s inequality for a convex function and positive linear functionals .
see ( @xcite , pp-47 ) .
we shall prove this inequality for a normalized positive @xmath0-semigroup and convex operator , defined on a banach lattice .
+ throughout the present section , @xmath7 will always denote a unital banach lattice algebra , endowed with an ordering @xmath27 .
let @xmath62 be a nonempty open convex subset of @xmath7 .
an operator @xmath63 is convex if it satisfies @xmath64 whenever @xmath65 and @xmath66 .
@xmath31 + let @xmath67 denotes the set of all differentiable convex functions @xmath68 . * ( jessen s type inequality ) * let @xmath51 be the positive @xmath0-semigroup on @xmath7 such that it satisfies ( [ nsg ] ) .
for an operator @xmath69 and @xmath70 ; @xmath71 * proof : * since @xmath68 is convex and differentiable , by considering an operator - analogue of [ theorem a , pp-98,@xcite ] , we have for any @xmath72 , there is a fixed vector @xmath73 such that @xmath74 using the property ( [ nsg ] ) along with the linearity and positivity of operators in a semigroup , we obtain @xmath75 in this inequality , set @xmath76 and the assertion ( [ jti ] ) follows .
@xmath31 + the existence of an identity element and condition ( [ nsg ] ) , imposed in hypothesis of the above theorem is necessary .
we shall elaborate the said , by following examples .
let @xmath77 , @xmath51 be the left shift semigroup defined on x and @xmath78 taking the mirroring along @xmath79-axis .
the identity function does not contain a compact support and therefore is not in @xmath59 . if we now take a bell - shaped curve like @xmath80 , @xmath81 .
then f is positive , @xmath82 , and @xmath83 has maximum at @xmath84 , and it is between 0 and 1 elsewhere . on the other hand , @xmath85 has a maximum at @xmath86 and it is immediate that we can not compare the two functions in the usual ordering . see figure 1(a ) .
let @xmath87 , and @xmath88 .
the rotation semigroup @xmath51 is defined as , @xmath89 , @xmath90 . the identity element @xmath91 , s.t . for all @xmath92 , @xmath93 . then @xmath94 . or
we can say that any complex number @xmath95 is mapped to @xmath96 .
@xmath97 satisfies ( [ nsg ] ) , only when @xmath98 is a multiple of @xmath99 .
let @xmath100 , then @xmath101 , hence @xmath102 . on the other hand , @xmath103 , and @xmath104 .
hence , the equality holds in ( [ jti ] ) when @xmath98 is a multiple of @xmath99 , but the two sides are not comparable in general . it can easily be verified that @xmath105 is a subgroup of @xmath106 , as @xmath107 . therefore @xmath108 is a normalized semigroup .
see figure 1(b ) .
+ @xmath31 +
in previous section , a jessen s type inequality has been derived , for a normalized positive @xmath0-semigroup @xmath51 .
this gives us the motivation towards knowing the behaviour of its corresponding adjoint semigroup @xmath109 on @xmath110 .
as the theory for dual spaces gets more complicated , we do not expect to have the analogous results .
it may ask for a detail introduction towards a part of the dual space @xmath110 , for which an adjoint of jessen s type inequality makes sense . given two banach spaces @xmath59 and @xmath111 and a bounded linear operator @xmath112 , recall that the adjoint @xmath113 is defined by @xmath114 for a strongly continuous positive semigroup @xmath51 on a banach space @xmath59 , by defining @xmath115 for every @xmath98 , we get a corresponding adjoint semigroup @xmath109 on the dual space @xmath116 . in @xcite
, it is obtained that , the adjoint semigroup @xmath109 fails in general to be strongly continuous .
the investigation @xcite , shows that @xmath109 acts in a strongly continuous way on ; @xmath117 this is the maximal such subspace on @xmath116 .
the space @xmath118 was introduced by philips in 1955 , and latter has been studied extensively by various authors . at the present moment , we do not necessarily require the strong continuity of the adjoint semigroup @xmath109 on @xmath116 .
+ if @xmath59 is an ordered vector space , we say that a functional @xmath119 on @xmath59 is positive if @xmath120 , for each @xmath121 . by the linearity of @xmath119 ,
this is equivalent to @xmath119 being order preserving .
i.e. @xmath122 implies @xmath123 .
the set @xmath124 of all positive linear functionals on @xmath59 , is a cone in @xmath116 .
+ we are mainly interested in the study of the space @xmath110 , where in our case @xmath7 is a banach lattice algebra .
let us consider the regular ordering among the elements of @xmath110 .
i.e. @xmath125 , whenever @xmath126 , for each @xmath127 .
+ consider the convex operator ( [ co ] ) . in case of the equality ,
@xmath128 is simply a linear operator and the adjoint @xmath128 can be defined as above . but how can it be defined in other case ?
this question has already been answered .
+ in @xcite , some kind of adjoint has been associated to a nonlinear operator @xmath128 .
in fact , this is possible for lipschitz continuous operators only .
consider the banach space @xmath129 of all lipschitz continuous operators @xmath130 satisfying @xmath131 , equipped with the norm @xmath132_{lip}= sup_{x_1\neq x_2}\frac { \|f(x_1)-f(x_2)\|}{\|x_1-x_2\|},\quad x_1,x_2\in x.\ ] ] where @xmath133 is the identity .
it is easy to see that the space @xmath134 of all bounded linear operators from x to y is a closed subspace of @xmath129 .
in particular , we set @xmath135 and call @xmath136 the pseudo - dual space of @xmath59 ; this space contains the usual dual space @xmath116 as closed subspace .
+ for @xmath137 , the pseudo - adjoint @xmath138 of @xmath128 is defined by ; @xmath139 this is of course a straightforward generalization of ( [ co ] ) ; in fact , for linear operators @xmath140 we have @xmath141 .
i.e. the restriction of the pseudo - adjoint to the dual space is the classical adjoint .
+ for the sake of convenience , we shall denote the adjoint of the operator @xmath128 by @xmath142 , throughout the present section .
either it s a classical adjoint or the pseudo - adjoint ( depending upon the operator @xmath128 ) .
+ similarly , the considered dual space of the vector lattice algebra @xmath7 will be denoted by @xmath110 , which can be the intersection of the pseudo - dual and classical dual spaces in case of a nonlinear convex operator .
let @xmath128 be the convex operator on a banach space @xmath59 , then the adjoint operator @xmath142 on the dual space @xmath116 is also convex . * * proof:**for @xmath121 and @xmath143@xmath144 where @xmath145 . by putting @xmath146 , for @xmath147 and using the convexity of the operator @xmath128 we finally get @xmath148 hence ,
@xmath142 is convex on @xmath116 . * ( adjoint - jessen s inequality ) * let @xmath109 be the adjoint semigroup on @xmath110 such that the original semigroup @xmath51 , the operator @xmath78 and the space @xmath7 are same as in theorem ( 1 ) .
for a convex operator @xmath149 and @xmath150 @xmath151 * proof : * for @xmath15 and @xmath150 , consider @xmath152,f ) & = & ( z^\ast(t)f^\ast,\phi(f ) ) \\ & = & ( f^\ast , z(t)(\phi f ) ) \\ & \geq & ( f^\ast,\phi(z(t)f ) ) \\ & = & ( \phi^\ast ( f^\ast),z(t)f ) \\ & = & ( z^\ast(t)[\phi^\ast f^\ast],f ) \end{aligned}\ ] ] therefore , the assertion ( [ ajti ] ) is satisfied .
in this section we shall define the exponential convexity of an operator .
moreover , few complex structures , involving the operators from a semigroup , will be proved to be exponentially convex .
let @xmath7 be a banach lattice endowed with ordering @xmath27 .
an operator @xmath153 is exponentially convex if it is continuous and for all @xmath154 @xmath155 where @xmath156 such that @xmath157 , @xmath158 . * * proof:**@xmath163 + take any @xmath156 and @xmath164 , @xmath162 . since the interval @xmath165 is convex , the midpoints , @xmath166 .
now set @xmath167 , for @xmath162 .
then we have , @xmath168 , for all @xmath158 . therefore ,
for all @xmath154 , we can apply @xmath169 to get , @xmath170 @xmath171 + let @xmath172 , such that @xmath173 , for @xmath158 .
define @xmath174 , so that @xmath175 .
therefore , for all @xmath154 , we can apply @xmath176 to get , @xmath177 @xmath31 + let @xmath159 be an exponentially convex operator .
writing down the fact for @xmath178 , in ( 10 ) , we get that @xmath179 , for @xmath180 and @xmath15 . for @xmath181
, we have @xmath182 hence , for @xmath183 and @xmath184 , we have @xmath185 i.e. @xmath153 , does indeed satisfy the condition of convexity .
@xmath31 + for @xmath186 , let us assume that @xmath63 is continuously differentiable on @xmath62 .
i.e. the mapping @xmath187 , is continuous .
moreover @xmath188 , will be a continuous linear transformation from @xmath7 to @xmath189 .
a bilinear transformation @xmath190 defined on @xmath191 is symmetric if @xmath192 for all @xmath20 .
such a transformation is * positive definite [ nonnegative definite ] * , if for every nonzero @xmath15 , @xmath193 [ @xmath194 .
then , @xmath188 is symmetric wherever it exists .
see [ @xcite , pp-69 ] .
[ condif ] let @xmath128 be continuously differentiable and suppose that second derivative exists throughout an open convex set @xmath186 .
then @xmath128 is convex on @xmath62 if and only if @xmath188 is nonnegative definite for each @xmath195 . and
if @xmath188 is positive definite on @xmath62 , then @xmath128 is strictly convex .
let @xmath7 be a unital banach algebra . for @xmath15 , a family of operators @xmath199
is defined as @xmath200 then @xmath201 .
whenever , @xmath35 , @xmath202 , therefore by theorem 3 , the mapping @xmath203 is convex .
[ t4 ] let @xmath51 be the positive @xmath0-semigroup , defined on a unital banach lattice algebra @xmath7 , such that it satisfies ( [ nsg ] )
. let @xmath15 , such that @xmath204 , for @xmath205 , @xmath206 , if @xmath207 and @xmath208 , if @xmath209 .
let us define @xmath210 then ( i ) : : for every @xmath154 and for every @xmath211 , @xmath212 , @xmath213_{i , j=1}^n\geq 0.\ ] ] ( ii ) : : if the mapping @xmath214 is continuous on @xmath215 , then it is exponentially convex on @xmath215 . *
* proof:**consider the operator @xmath216 for @xmath217 , @xmath218 and @xmath219 where @xmath220 .
then @xmath221 so , @xmath222 is a convex operator .
therefore by applying ( [ jti ] ) we get @xmath223 and the assertion ( [ a ] ) follows . assuming the continuity and using the * proposition 1 * we have also exponential convexity of the operator @xmath214 .
@xmath31 + * proof : * similar to the proof of theorem ( [ t4 ] ) .
@xmath31 + + * competing interests * + the authors declare that they have no competing interests . + * author s contribution * + all authors contributed equally and significantly in writing this paper .
all authors read and approved the final manuscript . + * acknowledgement * + authors of this paper are grateful to prof .
andrs btkai for his generous help in construction of examples . 20 , _ convexity , subadditivity and generalized jensen s inequality _ ,
vol 4 , no .
2 , 183194 . 2013 . , _ exponential convexity , positive semi - definite matrices and fundamental inequalities _ , j. math .
, vol 4 , no 2 , 171 - 189 .
2010 . , _ nonlinear spectral theory _
, walter de gruyter .
new york , 2004 .
, _ inequalities _ , springer - verlag , berlin , 1961 . , _ interpolation and approximation _ , dover , new york , 1975 .
, _ ber die entwicklung realen funktionen in reihen mittelst der methode der kleinsten quadrate _ ,
j. reine angewendte math . , no 94 , 4173 . 1883 .
, _ cauchy type means on one - parameter @xmath0-group of operators _ , j. math .
vol 9 , no .
2 , 631639 . 2015 .
, _ bemaerkinger om konvekse funktioner og uligheder imellem middelvaerdier i _ , muf .
tidsskrifr , 17 - 28 .
1931 . , _ konano dimenzionnalni vektorski prostori i primjene _ , tehnike knjige , zagreb , 1990 . , _ classical and new inequalities in analysis _ , kluwer academic publishers , the netherlands , 1993 . ,
_ one - parameter semigroups of positive operators _ , lect .
notes in math .
1184 , springer - verlag , 1986 . , _ convex functions , partial orderings , and statistical applications _ , academic press , inc .
new york , 1992 .
, _ the adjoint semi - group _ , pac . j. math .
no 5 , 269 - 283 , 1955 . , _ convex functions _ , academic press , new york and london , 1973 .
, _ the exponential and logarithmic functions on commutative banach algebras _ , int .
journal of math .
analysis , vol .
42 , 2079 - 2088 . 2010 . ,
_ the adjoint of a positive semigroup _ , compositio mathematica , no 90 , 99 - 118 , 1994 | in this paper the jessen s type inequality for normalized positive @xmath0-semigroups is obtained .
an adjoint of jessen s type inequality has also been derived for the corresponding adjoint - semigroup , which does not give the analogous results but the behavior is still interesting .
moreover , it is followed by some results regarding positive definiteness and exponential convexity of complex structures involving operators from a semigroup .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
the old - new interest in the problem of soliton - soliton intertaction and soliton molecules has been increasingly accumulating particularly over the past few years .
this is mainly motivated by the application of optical solitons as data carriers in optical fibers @xcite and the realization of matter - wave solitons in bose - einstein condensates @xcite .
one major problem limiting the high - bit rate data transfer in optical fibers is the soliton - soliton interaction . on the one hand ,
soliton - soliton interaction is considered as a problem since it may destroy information coded by solitons sequences . on the other hand
, it is part of the problem s solution , since the interaction between solitons leads to the formation of stable soliton molecules which can be used as data carriers with larger `` alphabet '' @xcite .
the interaction force between solitons was first studied by karpman and solovev using perturbation analysis @xcite , gordon who used the exact two solitons solution @xcite , and anderson and lisak who employed a variational approach @xcite .
it was shown that the force of interaction decays exponentially with the separation between the solitons and depends on the phase difference between them such that in - phase solitons attract and out - of - phase solitons repel .
this feature was demonstrated experimentally in matter - wave solitons of attractive bose - einstein condensates @xcite where a variational approach accounted for this repulsion and showed that , in spite of the attractive interatomic interaction , the phase difference between neighboring solitons indeed causes their repulsion @xcite . for shorter separations between the solitons ,
malomed @xcite used a perturbation approach to show that stationary solutions in the form of bound states of two solitons are possible .
however , detailed numerical analysis showed that such bound states are unstable @xcite .
stable bound states were then discovered by akhmediev @xcite and a mechanism of creating robust three - dimensional soliton molecules was suggested by crasovan _
recently , soliton molecules were realized experimentally by stratmann _
_ in dispersion - managed optical fibers @xcite and their phase structure was also measured @xcite .
perurbative analysis was used to account theoretically for the binding mechanism and the molecule s main features @xcite .
quantization of the binding energy was also predicted numerically by komarov _
_ @xcite . in refs.@xcite ,
a hamiltonian is constructed to describe the interaction dynamics of solitons .
the mechanism by which the relative phase between the solitons leads to their force of interaction , and hence the binding mechanism , is understood only qualitatively as follows . for in - phase ( out - of - phase ) solitons , constructive ( destructive ) interference takes place in the overlap region resulting in enhancement ( reduction ) in the intensity . as a result
, the attractive intensity - dependent nonlinear interaction causes the solitons to attract ( repel ) @xcite .
a more quantitative description is given in refs .
@xcite . in view of its above - mentioned importance from the applications and fundamental physics point of views
, we address here the problems of soliton - soliton interaction and soliton molecule formation using the exact two solitons solution .
this approach has been long pioneered by gordon @xcite where he used the exact two solitons solution of the homogeneous nonlinear schr@xmath0dinger equation to derive a formula for the force of interaction between two solitons , namely @xmath1 where @xmath2 is the solitons separation and @xmath3 is their phase difference .
this formula was derived in the limit of large solitons separation and for small difference in the center - of - mass speeds and intensities , which limits its validity to slow collisions
. with appropriately constructed hamiltonian , wu _ et al . _
have derived , essentially , a similar formula that gives the force between two identical solitons and reliefs the condition on slow collisions @xcite . here , we present a more comprehensive treatment where we derive the force between two solitons for arbitrary solitons intensities , center - of - mass speeds , and separation .
we also generalize gordon s formula to inhomogeneous cases corresponding to matter - wave bright solitons in attractive bose - einstein condensates with time - dependent parabolic potentials @xcite and to optical solitons in graded - index waveguide amplifiers @xcite .
many interesting situations can thus be investigated .
this includes the various soliton - soliton collision regimes with arbitrary relative speeds , intensities , and phases .
most importantly , soliton - soliton interaction at short solitons separations will now be accounted for more quantitatively than before .
specifically , soliton molecule formation is clearly shown to arise from the time - dependence of the relative phase which plays the role of the restoring force . in this case , the force between the two solitons is shown to be composed of a part oscillating between attractive and repulsive , which arises from the relative phase , and an attractive part that arises from the nonlinear interaction . the time - dependence of the relative phase results in a natural oscillation of the molecule s bond length around an equilibrium value .
the various features of the soliton molecule , including its equilibrium _ bond length _ , _ spring constant _ , _ frequency _ and _ amplitude _ of oscillation , and _ effective mass _ , will be derived in terms of the fundamental parameters of the solitons , namely their intensities and the nonlinear interaction strength .
the two solitons solution is derived here using the inverse scattering method @xcite .
although the two solitons solution of the homogeneous nonlinear schr@xmath0dinger equation is readily known @xcite , here we not only generalize this solution to inhomogeneous cases , but also present it in a new form that facilitates its analysis . the solution will be given in terms of the four fundamental parameters of each soliton , namely the initial amplitude , center - of - mass position and speed , and phase .
the main features of the solution will be shown clearly such as the contribution of the nonlinear interaction to the actual separation and phase difference between solitons where it turns that the separation between the two solitons increases with logarithm of the difference between the amplitudes of the two solitons .
furthermore , the general statement that _ a state of two equal solitons with zero relative speed and finite separation does not exist as a stationary state for the homogeneous nonlinear schr@xmath0dinger equation _ , will be transparently and rigorously proved .
stability of soliton molecules is an important issue since , in real systems , perturbations caused by various sources such as losses , raman scattering , higher order dispersion , and scattering from local impurities , tend to destroy the molecules . to investigate the stability of the soliton molecules described by our formalism , we have considered three situations .
first , we studied the reflection of the molecule from a hard wall and a softer one .
while for the hard wall the molecule preserves its molecular structure after reflection , it generally breaks up for the softer ones due to energy losses at the interface .
secondly , the scattering of the molecule by a potential barrier was also investigated .
we show that the molecular structure is maintained only for some specific heights of the barrier .
this suggests a quantization in the binding energy as predicted by komarov _
the oscillation period of the reflected molecule is noticed to be smaller than for the incident one .
in addition , the outcome of scattering depends on the phase of the molecule s oscillation at the interface of the barrier .
for instance , a dramatic change in the scattering outcome takes place if the _ coalescence _ point of the molecule lies exactly at the interface .
in such a case , the otherwise totally transmitting molecule will now split into reflecting and tunneling solitons .
thirdly , we have considered the collision between a single soliton with a stationary soliton molecule .
the effects of different initial speeds , amplitudes , and phases of the scatterer soliton were studied .
it turns out that for slower collisions , it is easier for the scatterer soliton to break up the soliton molecule , while for fast collisions the scatterer soliton expels and then replaces one of the solitons in the molecule .
the phase of the scatterer soliton plays also a crucial rule in preserving or breaking the bond of the molecule , which can be used as _ key _ tool to code or uncode data in the molecule .
the rest of the paper is organized as follows . in section [ exact_sec ] ,
we use the inverse - scattering method to derive the two solitons solution of the inhomogeneous nonlinear schr@xmath0dinger equation and present the solution in the above - mentioned appealing form .
the main features of the solution will be discussed in subsection [ mainsec ] .
the center - of - mass positions and relative phases will be derived in subsections [ com_sec ] and [ phases_sec ] , respectively .
the force between solitons will be derived in section [ force_sec ] where gordon s formula will be extracted as a special case in subsection [ gordon_sec ] and our more general formula will be derived in subsection [ our_sec ] . in subsection
[ num_sec ] , we compare our formula with the numerical calculation . in section [ molsec ] , we show the possibility of forming soliton molecules , derive their main features in subsection [ formation_sec ] , and investigate their stability in subsection [ stability_sec ] .
we end in section [ conc_sec ] with a summary of results and conclusions .
the details of the derivation of the two solitons solution and the center - of - mass positions are relegated to appendices a and b , respectively .
matter - wave solitons of trapped bose - einstein condensates and optical solitons in optical fibers can be both described by the dimensionless gross - pitaevskii equation @xmath4 , time to @xmath5 , and the wavefunction @xmath6 to @xmath7 , where @xmath8 and @xmath9 are the characteristic frequencies of the quasi one - dimensional ( @xmath10 ) trapping potential in the axial and radial directions , respectively . in these units
, the strength of the interatomic interaction will be given by the ratio @xmath11 , where @xmath12 is the @xmath13-wave scattering length .
for the case of optical solitons , the function @xmath6 represents the beam envelope , @xmath14 is the propagation distance , @xmath15 is the radial direction , and the intensity - dependent term represents the kerr nonlinearity . in this case
, scaling is in terms of the characteristic parameters of the fiber , as for instance in ref .
the specific form of the prefactors of the inhomogeneous and nonlinear terms guarantees the integrability of this equation @xcite . for the special case of @xmath16
, the homogeneous case is retrieved .
other interesting special cases have also been considered @xcite . as outlined in appendix [ appa ] , we use the darboux transformation method to derive the two solitons solution of this gross - pitaevskii equation , which can be put in the form @xmath17\nonumber\\ & + & \sqrt{\frac{n_2 \alpha_{22}(t)}{2}}e^{i ( \phi_{01}+\phi_{02}+\phi_2(x , t)+ \tan^{-1}({\alpha_2\over\alpha_1 } ) + \tan^{-1}({\alpha_4\over\alpha_3}))}\nonumber\\ & \times&{\rm{sech}}\left[\alpha_{22}(t)(x - x_{\rm cm2}(t ) ) + \frac{1}{2}\log\left(\frac{\alpha_1 ^ 2+\alpha_2 ^ 2}{\alpha_3 ^ 2 + \alpha_4 ^ 2}\right)\right]\label{exactsol},\end{aligned}\ ] ] where + + @xmath18 , @xmath19 , + + @xmath20 , + + @xmath21 , + + @xmath22 , + + @xmath23 , + + @xmath24 , + + @xmath25 , + + @xmath26 , + + @xmath27 , + + @xmath28 , + + @xmath29 , + + @xmath30 , + + @xmath31 , + + @xmath32 , + + @xmath33 , + + @xmath34 , + + @xmath35 , + + @xmath36 .
+ + the solution is put in this suggestive form to facilitate its analysis .
the first sech part corresponds to the exact single soliton solution with center - of - mass position @xmath37 , width @xmath38 , phase @xmath39 , and normalization @xmath40 .
hence , @xmath41 and @xmath42 correspond to the initial center - of - mass position and speed , respectively .
the second sech term contains the same features in addition to a shift in both the center - of - mass position and phase .
it should be noted , however , that @xmath43 , @xmath44 , @xmath45 , and @xmath46 correspond to the center - of - mass position and speed , normalization , and phase of the single noninteracting solitons . due to the interaction between solitons
, these four characteristic quantities may not correspond exactly to the values of the same physical quantities as they did for the single soliton solution .
for instance , @xmath37 will not correspond to the center - of - mass of one of the solitons .
instead , the soliton may be shifted from that position due to the interaction with the other soliton . in the following , we present a detailed analysis of the locations and phases of the two solitons .
inspection shows that there are two main regimes for the two solitons solution , namely the regime of resolved solitons and the regime of overlapping solitons . in the former case ,
the center - of - mass concept is well defined and analysis of the relative dynamics becomes feasible .
the solitons are considered resolved as long as the two main peaks are not overlapping , which means that partial overlap may occur in this regime .
the analysis in this section assumes the resolved solitons regime . in the resolved solitons regime ,
the argument of the second sech term of eq .
( [ exactsol ] ) , namely @xmath47 $ ] , simplifies to a function with three roots .
the fact that the sech function is peaked at the roots of its argument , leads to that the second sech term corresponds to three `` solitons '' .
this is shown in fig [ fig1 ] .
we denote these solitons as the `` left '' , `` central '' , and `` right '' solitons with peak locations at @xmath48 , @xmath49 , and @xmath50 , respectively .
we notice that the central soliton is located at the position of the soliton of the first sech term , namely near @xmath37 .
further inspection shows that the two solitons at this location are out of phase and interfere destructively such that they do not appear in the total profile .
therefore , the two solitons that our solution of eq .
( [ exactsol ] ) describes are in fact the left and right solitons arising from the second sech term .
this is different from what one aught to conclude from the form of the exact solution , namely that the first sech term corresponds to one soliton and the second sech term corresponds to the other soliton .
in general , the center - of - mass locations of the left and right solitons , @xmath48 and @xmath50 , do not match @xmath37 and @xmath51 .
typically , @xmath48 will be shifted to the left of @xmath37 , @xmath50 will be shifted to the right of @xmath51 , while @xmath52 remains near @xmath37 .
the amount of shift depends mainly on the normalization difference @xmath53 and the relative speed @xmath54 , as will be shown in the next subsection .
an interesting general and exact result is that , for the homogeneous case @xmath16 , a state of two equal solitons , @xmath55 , with zero relative speed , @xmath56 , and finite separation , does not exist as an exact solution of the gross - pitaevskii equation .
this can be proven by substituting @xmath55 and @xmath56 in @xmath57 to find that the right and left solitons migrate to @xmath58 and @xmath59 , respectively , while the center - of - mass of the central soliton matches exactly @xmath37 .
furthermore , we show in section [ phases_sec ] that the phase difference between this central soliton and that of the first sech term equals , in this case , @xmath60 ; guaranteeing their destructive interference .
thus , the three solitons disappear in such a special case and @xmath6 becomes the trivial solution . in view of the above , a need arises to derive formulae for the center - of - mass positions , @xmath48 and @xmath50 in terms of the solitons parameters , which will then be used to derive the force between the two solitons . in this section
, we derive formulae for the three roots of @xmath57 which correspond to the locations of the left , central , and right solitons , @xmath48 , @xmath52 , and @xmath50 , respectively . to facilitate the derivation ,
we define @xmath61}$ ] , @xmath62}$ ] , @xmath63 , @xmath64 , @xmath65 , and @xmath66 .
the equation @xmath67 is a third - order polynomial in both @xmath68 and @xmath69 . in principle
, this equation can be solved algebraically for @xmath68 or @xmath69 . however , extracting @xmath15 from the resulting three roots will not be possible analytically for general @xmath70 . alternatively , and as can be seen from fig .
[ fig1 ] , we can exploit the simple linear behavior of @xmath57 near its roots .
assuming , without loss of generality , @xmath71 , and noting that in the resolved solitons regime the solitons separation @xmath72 is large , we argue in appendix [ appb1 ] that @xmath73 for all @xmath15 , @xmath74 for @xmath75 , @xmath76 for @xmath77 , and @xmath78 for @xmath79 . based on this
, the center - of - mass position @xmath50 can be derived from a taylor expansion of @xmath57 in powers of large @xmath68 and @xmath69 .
expanding @xmath57 in powers of large @xmath68 only and leaving @xmath69 arbitrary , accounts for @xmath52 and @xmath48 simultaneously . keeping terms up to first order of @xmath80 , we show in appendix [ appb1 ] that the position of the right soliton is given by @xmath81 where @xmath82 and the position of the the left soliton is given by @xmath83 where @xmath84 , @xmath85 , and @xmath86 and @xmath87 are given in appendix [ appb1 ] .
the second and third terms on the right hand side of eqs .
( [ xreq2 ] ) and ( [ xleq2 ] ) account for the shift in the center - of - mass position with respect to the single soliton ones .
the third terms are much smaller than the second ones since they decay exponentially with the solitons distance @xmath88 . in the limit @xmath89 and @xmath90 ,
both @xmath91 and @xmath92 take the form @xmath93}$ ] .
thus , it is obvious that for @xmath94 , @xmath95 and @xmath96 .
this agrees with our earlier result that two equal solitons with zero relative speed and finite separation , do not exist as an exact solution . in this section
, we calculate the phases of the left , central , and right solitons in reference to the phase of the soliton of the first sech part in the two solitons solution . for simplicity , the special case of @xmath97 and @xmath98 will be assumed .
the phase difference between the left , central , and right solitons on one hand and the soliton of the first sech term on the other hand is generally given by @xmath99 which by observing that for all @xmath15 @xmath100 reduces to @xmath101 this expression gives the phases of the right soliton @xmath102 , the central soliton @xmath103 , and the left soliton @xmath104 , for @xmath75 , @xmath52 , and @xmath48 , respectively .
to calculate these phases we express the parameters @xmath105 in terms of @xmath68 and @xmath69 , as follows @xmath106 @xmath107 @xmath108 @xmath109 in general , @xmath73 for all @xmath15 , but @xmath74 only for @xmath110 , and @xmath111 for @xmath112 , as shown in appendix [ appb1 ] .
+ + * i. phase of the right soliton @xmath102 : * + + in this case @xmath113 which leads to @xmath114 and thus @xmath115 , @xmath116 , @xmath117 , @xmath118 .
therefore + + + @xmath119 + + + and @xmath120 , which gives @xmath121 where @xmath122 is a function that depends on the ratio @xmath123 for nonzero @xmath124 and @xmath125 .
+ + * ii .
phase of the central soliton @xmath103 : * + + in this case @xmath126 which gives @xmath127 , and hence @xmath128 , and @xmath129 . finally , we get @xmath130 the last result shows that , for @xmath131 , i.e. , two equal solitons with zero relative speed , the central soliton and the soliton of the first sech term of the two solitons solution are out of phase and therefore interfere destructively .
+ + + * iii .
phase of the left soliton @xmath104 : * + + in this case @xmath132 which results in @xmath78 and @xmath133 , @xmath134 , @xmath135 , @xmath136 .
therefore , @xmath128 and @xmath137 , which gives @xmath138 the phase difference between the left and right solitons is thus given by @xmath139 noting that @xmath140 for @xmath141 and small but finite @xmath125 , and @xmath142 for @xmath143 and small but finite @xmath124 , we finally conclude that the phase difference between the two solitons is given by @xmath144 the above results are verified in fig .
[ fig2 ] , where we plot @xmath145 , @xmath102 , @xmath104 , and @xmath103 versus @xmath15 .
the agreement between our estimated values and the exact curve is evident . in this section
, we use the results of the previous two subsections to derive the force between the left and right solitons .
the force is proportional to the acceleration of the solitons separation @xmath146\,e^{-\frac{1}{4 } g_0 { n_s } { x_d}(t ) e^{\gamma ( t)}}\label{deltaeq},\end{aligned}\ ] ] where @xmath147 and the coefficients @xmath148 are given in appendix [ appb2 ] .
the first two terms on the right hand side of eq .
( [ deltaeq ] ) are the dominant ones since they correspond to the noninteracting solitons separation , @xmath149 , and their logarithmic shifts @xmath150 arising from the interaction between solitons .
the time - dependence of @xmath151 originates from @xmath152 , @xmath153 , and @xmath149 .
the acceleration can thus be derived @xmath154 \ , e^{-\frac{1}{4 } g_0 n_s { \delta}(t ) e^{\gamma ( t)}+3 \gamma ( t)}\label{force2},\ ] ] where the coefficients @xmath155 are given in appendix [ appb2 ] . in the last equation
, we have used eq .
( [ deltaeq ] ) with only the first two terms of its right hand side to substitute for @xmath149 in terms of @xmath2 in the exponential factor . the first term on the right hand side of eq .
( [ force2 ] ) corresponds to the force due to the external potential which vanishes for the homogeneous case .
the rest of the terms correspond to the force of interaction between the solitons .
the interaction force depends , as expected , on the phase difference of the two solitons and decays exponentially with their separation .
it should be noted that this equation is a generalization of gordon s formula @xcite in two aspects .
first , it is derived for a time - dependent inhomogeneous medium .
secondly we have , essentially , no restriction on the difference between the two solitons amplitudes and speeds ; apart from some extreme cases which were mentioned in appendix [ appb1 ] and will be discussed further below . for the homogeneous case , @xmath16 , and in the limits @xmath89 and @xmath90 , the acceleration formula , eq . ( [ force2 ] ) , simplifies considerably .
an apparent inconsistency occurs when switching the order of these two limits , namely @xmath156 while @xmath157 which differ by an overall minus sign .
the conflict is resolved by invoking eq .
( [ phasediff ] ) where it is shown that in the first case : @xmath158 , while in the second case : @xmath159 . therefore , the two approaches agree on the following result @xmath160 this is essentially gordon s result , eq .
( [ gordoneq ] ) , since in his derivation gordon took @xmath161 and soliton amplitude @xmath162 , @xmath163 , as can be seen in eqs.(1 ) and ( 6 ) of ref .
@xcite . substituting @xmath161 and @xmath164 in the last equation ,
it becomes identical to eq .
( [ gordoneq ] ) .
it should be mentioned here that eq .
( [ gordoneq ] ) was also derived in ref .
@xcite using a perturbation analysis based on the inverse scattering method , and in ref .
@xcite using a variational calculation . for nonzero @xmath152 and in the limits @xmath89 and @xmath90
, the acceleration formula takes the form @xmath165 this is a generalization to gordon s formula for the inhomogeneous case as modeled by eq .
( [ gp ] ) . depending on the specific form of @xmath152 , the two force terms , namely the external ( first term ) and the interaction ( second term ) , can be repulsive , attractive , or oscillatory .
in addition , the phase difference @xmath3 also depends on @xmath152 .
it is established in the homogeneous case , as will also be shown in section [ molsec ] , that the time - dependence of the phase difference is responsible for binding the two solitons in the soliton molecule .
here , in addition to the possibility of forming soliton molecules , the dependence of @xmath3 on @xmath152 allows for controlling the parameters of the molecule such as its equilibrium bond length , period , and spring constant .
this and possibly other interesting phenomena will be left for future investigation . to obtain an estimate of the accuracy of the general acceleration formula , eq .
( [ force2 ] ) , we calculate numerically the acceleration from the exact two solitons solution , eq .
( [ exactsol ] ) , and compare the two results .
the distance between the two solitons of the function @xmath166 is determined using a numerical algorithm that employs our formulae for @xmath48 and @xmath50 given by eqs .
( [ xreq2 ] ) and ( [ xleq2 ] ) to calculate seed values .
the distance is then differentiated numerically twice at @xmath167 . in fig .
[ fig3 ] , we compare the two results .
good agreement is obtained for @xmath168 .
the analytical solution diverges at @xmath169 .
the value at which divergence takes place is set by the specific choice of parameters in fig .
as pointed out in section [ com_sec ] , the divergence occurs due to the merging of the central and left solitons .
this artifact divergency can be remedied by associating the location of the local maximum of @xmath57 to @xmath48 once this maximum has reached the @xmath15-axis from above .
restricting our study to the region where agreement is obtained , we interestingly notice that the acceleration is oscillating between positive and negative values .
this means that the force between the solitons is oscillating between attractive and repulsive .
the possibility of having attractive forces for finite @xmath124 is particularly interesting ; for two solitons with nonzero relative positive speed , i.e. the solitons are initially diverging from each other , the force between them is attractive .
this suggests that , if the force remains attractive for sufficient time , the two solitons will slow down and eventually converge at some point .
if true , this should occur at small distances since the force decays exponentially with distance and when the two solitons are allowed to diverge even for a short while , the force might be weakened such that the two solitons can not return back . to be able to judge on such a possibility , we need to know what happens to the acceleration at later times . to that end , we calculate numerically the acceleration in terms of @xmath124 and @xmath14 .
the result is plotted in fig .
[ fig4 ] , where it is clear that the acceleration indeed decays with time for all @xmath124 .
this leads to that any nontrivial effect of the oscillating force is most likely to take place at short solitons separations .
this is what we find in the next section where the possibility of forming stable soliton molecules is pointed out .
we have shown in the previous section that , as a result of the solitons time - dependent relative phase , the force of interaction between solitons is oscillating between repulsive and attractive .
since the force decays exponentially with the solitons separations , this oscillation will have a tangible effect only when the two solitons are close to each other . in this section ,
we investigate the force of interaction between solitons for short solitons separation .
in such a special case , eq . ( [ deltaeq ] ) takes a simple form that accounts for the solitons separation in terms of their relative phase . using this formula
, we show that the solitons will be bound to oscillate around some equilibrium distance where the phase plays the role of the restoring force .
comparison with exact numerical calculations shows that this formula is accurate for almost the full range of the solitons separation , except at the coalescence point ( if any ) . in subsection [ formation_sec ]
, we discuss the main features of the resulting soliton molecules , and in subsection [ stability_sec ] , we investigate numerically their stability in different scattering regimes . to focus on the role of relative phase , we simplify the analysis by restricting our treatment to the homogeneous case , @xmath16 , and zero relative speed , @xmath170 .
we also set @xmath171 so that any separation between the solitons to be as small as possible which in this case arises only from the logarithmic shifts ( @xmath172 in eq .
( [ deltaeq ] ) ) . in this case , the solitons separation @xmath151 , given by eq .
( [ deltaeq ] ) , simplifies in the limit @xmath173 to @xmath174-\frac{4}{g_0 n_s}\log \left(\frac{g_0 n_d^2}{n_s}\right)\label{delmol},\ ] ] and the acceleration is given by @xmath175 it is noted that for @xmath176 or @xmath177 , gordon s formula is retrieved , but here with an explicit time - dependence of the phase , @xmath178 .
this acceleration formula deviates considerably from gordon s formula for @xmath179 .
specifically , for @xmath180 , @xmath151 diverges to @xmath59 at @xmath181 , which indicates that the two solitons coalesce .
this is confirmed below by examining the exact solution at this condition .
we note here that an approximate expression for the solitons separation was also derived in refs .
in addition , our predicted molecule s oscillation frequency ( see eq .
( [ freqeq ] ) below ) agrees with these references . to verify this feature ,
we calculate numerically the distance between the two solitons directly from the exact solution , eq .
( [ exactsol ] ) , for different values of @xmath182 . for @xmath183 , the density plot in fig .
[ fig5]a , shows a soliton molecule of two clearly resolved solitons with a separation oscillating around some nonzero equilibrium distance . approaching the solitons coalescence point with @xmath184 , the density plot in fig .
[ fig6]a , shows the two solitons approaching each other more than the previous case .
furthermore , this figure shows a slight bounce back by one of the solitons in the region of collision .
approaching further the coalescence condition with @xmath185 , we indeed observe in fig .
[ fig7]a that the two solitons merge almost completely . for a more quantitative comparison ,
we calculate numerically the center - of - mass trajectories of the two solitons .
we show the trajectory curves in the density plots of figs .
[ fig5]a-[fig7]a . in figs .
[ fig5]b-[fig7]b , we plot the solitons separation obtained from formula ( [ delmol ] ) and the numerical trajectories obtained from the exact solution .
it is clear from these figures that this formula agrees well with the exact soliton separation except near the collision region . in fig .
[ fig5]b , the two solitons remain away from each other during the collision , and therefore good agreement is obtained with the exact result even in the collision region . in fig .
[ fig6]b the two solitons approach each other further such that formula ( [ delmol ] ) does not account for the above - mentioned slight bounce of one of the solitons . in fig .
[ fig7]b , agreement with the exact solution in the collision region is qualitative .
we found that at the condition @xmath180 and for @xmath173 , the analytic curve overlaps with that of the exact solution ; apart from the horizontal segments where formula ( [ delmol ] ) diverges to @xmath59 .
further insight is obtained by plotting the density profile of the soliton molecule at some specific times , as shown in figs .
[ fig5]c-[fig7]c . in fig .
[ fig5]c , we observe that the initial amplitude imbalance is never removed during the dynamics . instead , it becomes maximum when the two solitons are closest to each other .
in addition , we notice that the oscillation amplitude of the larger soliton around its equilibrium position is larger .
figure [ fig6]c shows clearly the soliton bounce which takes place in the time interval @xmath186 to @xmath187 . in these subfigures
we plot two vertical dashed lines that indicate the position of the solitons at the closest approach .
it is clear that after first closest approach at @xmath186 , the right soliton bounces back with a maximum displacement at @xmath188 . in fig .
[ fig7]c , it is shown that , although the two solitons coalesce , two small symmetric @xmath189 appear .
a detailed examination of these wings shows that they are the remnants of the two solitons after they coalesce and they both bounce back in the collision region similar to the case of fig .
[ fig6 ] .
it is also instructive to show the dynamics of the phase profile during the molecule s oscillation .
this is shown with the contour plots of fig .
[ fig8 ] which correspond to the molecules of figs .
[ fig5]-[fig7 ] . in fig .
[ fig8]a , which corresponds to fig .
[ fig5 ] , the two solitons start initially in phase . by time
the phase of the right soliton , which is the one with higher intensity amplitude and larger oscillation displacement , starts to exceed that of the left soliton . at the point of closest approach ,
the phase difference is exactly @xmath60 .
after that point , the two solitons diverge again , the phase difference starts to decrease , and the cycle is repeated .
similar behavior is seen in fig .
however , in fig .
[ fig8]c , where the two solitons coalesce for a considerable amount of time , the phase difference during the coalescence time is zero .
it is thus not completely understood why , in this case , the two solitons still repel each other and eventually split .
having established the existence of the soliton molecule from the exact two solitons solution and derived a formula that describes its bond length , here we use this formula to examine more closely the propertie of the soliton molecule and its mechanism of binding .
it is clear from eq .
( [ delppmol ] ) that the sinusoidal time - dependence of the solitons relative phase leads to a force of interaction that oscillates between attractive and repulsive and hence allowing for soliton molecule formation .
further details of the mechanism of binding will be uncovered by expressing the acceleration , @xmath190 in terms of @xmath2 by substituting for @xmath191 from eq .
( [ delmol ] ) into eq .
( [ delppmol ] ) , to get @xmath192 this shows that the interaction force between the two solitons is the resultant of an attractive part and a repulsive part .
the equilibrium bond length , defined by @xmath193 , is given by @xmath194 in consistence with our previous result , the equilibrium bond length diverges as @xmath195 . solving the last equation for @xmath196 and then substituting in eq .
( [ delpp2mol ] ) , @xmath197 simplifies to @xmath198.\ ] ] for small amplitude oscillations , @xmath199 , the last equation gives @xmath200 the restoring force ( @xmath201 ) originates from the phase - dependent terms , @xmath202 .
this appealing form of the force of interaction shows that the force between the solitons is of hooke s law type with a spring constant @xmath203 where @xmath204 is the _ bare mass _ of the molecule .
expressed in terms of @xmath205 and @xmath125 , the spring constant takes the form @xmath206 which shows that @xmath207 for @xmath143 ; corresponding to a soliton molecule of infinite bond length .
furthermore , @xmath208 diverges for @xmath180 , which signifies soliton coalescence , as we have pointed out in the previous subsection .
since the frequency of the soliton molecule is given by @xmath209 and the spring constant is given in eq .
( [ keq ] ) , the effective mass @xmath210 will be given by @xmath211 which again diverges at the soliton coalescence condition , @xmath180 .
having determined the main properties of the soliton molecule , we can now return back to eq .
( [ delmol ] ) to express @xmath151 as @xmath212 where @xmath213 is the initial value of @xmath151 , which is given by the solitons parameters through @xmath214 the amplitude of the oscillation @xmath215 is thus given by @xmath216 which gives an elastic potential energy @xmath217 it should be noted here that this is equal to the mechanical energy since the initial speed vanishes , @xmath218 .
the fact that the potential energy diverges at the coalescence condition @xmath180 is a gain an artifact of the calculation , but it at least indicates that the bond is tighter than cases where @xmath176 or @xmath177 . here , we investigate the stability of the soliton molecule against break up in the following three collision regimes : @xmath219 reflection by a hard wall , @xmath220 crossing a finite potential barrier , and @xmath221 collision with a single soliton . to that end
we solve the gross - pitaevskii equation , eq .
( [ gp ] ) , numerically . as an initial state ,
we use , for cases @xmath219 and @xmath220 the two solitons solution , eq .
( [ exactsol ] ) , which represents the soliton molecule . for case @xmath221
, we use the superposition of the exact single soliton , eq .
( [ psi1_exact ] ) , with the two solitons solution . before starting the discussion of results ,
we point out that in figs .
[ fig9]-[fig14 ] , we present the results of this section using spacio - temporal density plots .
since the solitons are too thin compared to the spacial range that we consider , a density plot with full spacial and time ranges will not show a clear solitons peak density or center - of - mass path , as fig .
[ fig9]c shows . to solve this problem , we restricted the density plotting to a finite range of @xmath166 , namely between 0.025 and 0.15 corresponding to the upper part of the solitons peaks .
this results in an easier tracking of both the solitons peak density and center - of - mass path , as shown in fig .
[ fig9]a , b , d and the rest of subsequent figures . for reflection from a hard wall we solve eq .
( [ gp ] ) with a potential step of the form @xmath222 where @xmath223 and @xmath224 are the hight and location of the potential wall , respectively .
the result of reflection from this hard wall , with @xmath225 , is shown in fig .
the soliton molecule preserves its molecular structure but with different characteristics .
the solitons in the reflected molecule do not coalesce as in the incident molecule .
in other words , the equilibrium bond length becomes larger .
the density plot shows that initially the two solitons are of comparable intensities .
after reflection , the brighter color of the left soliton and darker color of the right soliton indicate that the left soliton acquires higher intensity on the expense of the right soliton .
we also notice that the left soliton performs two reflections from the potential interface .
after the first reflection , it collides with the right soliton and then collides with the potential interface for the second time .
the picture becomes different when the hight of the wall is reduced to @xmath226 , as shown in fig .
the soliton molecule breaks up after reflection .
this is due to loss of energy at the interface of the potential .
part of the soliton molecule transmits as a nonsolitonic pulse that broadens and decays in intensity by time . by plotting @xmath166 in fig .
[ fig9]c with its full range , we can see the nonsolitonic part as the left- and right - going two red ejections corresponding to the transmitted and reflected nonsolitonic pulses , respectively . in fig .
[ fig9]d , we combine fig .
[ fig9]b and fig .
[ fig9]c to show the locations of the nonsolitonic ejections with respect to the solitons centers . in the case of reflection from a hard wall ,
the nonsolitonic ejections are essentially not present which results in the stability of the molecular structure . for reflection from a potential barrier we solve eq .
( [ gp ] ) with the potential @xmath227 where @xmath223 , @xmath228 , and @xmath224 , are the height , width , and location of the right side of the barrier , respectively . in fig .
[ fig10 ] , we show the many different possibilities that result when the height of the barrier is changed . the free evolution case with @xmath229
is shown as a reference plot .
the full reflection case is shown for @xmath225 , which is similar to the previous case of reflection from a hard wall . reducing the height of the barrier to @xmath230
, we notice that the soliton molecule breaks up after reflection . as pointed above
, this is due to the nonsolitonic ejections taking place at the interfaces of the potential . reducing the height of the barrier to @xmath231 , a sign of solitons recombining appears in the form of a soliton molecule of a short lifetime . at @xmath232 a stable molecule
is remarkably formed with a considerably shorter period than for the incident molecule .
we have confirmed numerically that this molecule remains stable for much longer time provided that the soliton molecule remains sufficiently far from the boundaries of the spacial grid .
this unique structure remains for some small domain around @xmath232 , but is lost for @xmath233 , where the molecule breaks for long evolution times . decreasing the height of the barrier to @xmath234 , the molecule breaks at the interface and splits into a reflected and transmitted solitons . for @xmath226 , the molecule breaks at the interface , but both solitons
transmit through the barrier . for @xmath235 ,
the transmitted solitons show a sign of recombining again but with shorter period than for the free evolution case and larger than in the case of @xmath232 . motivated by the fact that at the coalescence point the intensity of the molecule is considerably higher than at other instants , we expect to find different scattering dynamics when the soliton molecule meets the interface of the potential at different phases of its periodic oscillation . in fig .
[ fig11 ] , this is investigated by fixing the height of the potential barrier at @xmath232 and changing the initial launching position of the molecule .
starting at @xmath236 , the molecule breaks after reflection .
a shortly - lived molecule is obtained at @xmath237 , and a stable molecule is found for @xmath238 , which corresponds to the @xmath232 case of fig .
[ fig10 ] . at @xmath239 ,
the soliton splits at the interface into transmitted and reflected solitons .
the coalescence point is , in this case , located at the interface .
transmission takes place due to the high intensity of the soliton at the coalescence point . at @xmath240 ,
the two solitons still split as in the previous subfigure but with a weak transmitted soliton intensity less than 0.025 and hence will not be shown in our density plots which are restricted to the intensities between 0.025 and 0.15 , as pointed out previously . for @xmath241 ,
the coalescence point takes place before the molecule reaches the interface and both solitons reflect but the molecule breaks up . for @xmath242
the two reflected solitons start to recombine forming a stable molecule at @xmath243 . at @xmath244
the reflected molecule starts to break up since the second coalescence point becomes close to the interface .
thus , the conclusion from this figure is that the soliton molecule is more vulnerable to break up when it meets the interface at the coalescence point .
equivalently , soliton molecules with larger equilibrium bond length , such that coalescence does not occur , will be more stable against breakup post reflection from barriers . finally , we present in figs .
[ fig12]-[fig14 ] the results of scattering of a soliton molecule by a single soliton described by eq .
( [ newsol1 ] ) with normalization @xmath245 , center - of - mass position and speed @xmath246 and @xmath247 , respectively .
the effects of the phase , speed , and amplitude of the injected soliton are investigated separately . in fig .
[ fig12]a , a soliton initially at @xmath248 is launched towards a stationary soliton molecule near @xmath249 . at the impact , the molecule brakes up
, its right soliton is ejected in the direction of the positive @xmath15-axis , and the left soliton combines with the scatterer soliton to form a new stationary molecule shifted by about a bond length to the left .
we point out here that for such an outcome to occur , it is essential that the amplitude of the scatterer soliton is nearly equal to that of the right soliton of the molecule .
otherwise , a different outcome , as that of fig . [ fig14 ] , will be obtained . in fig .
[ fig12]b , the same numerical experiment is repeated but with adding a @xmath60 to the phase of the scatterer soliton .
clearly , this phase addition prevents the formation of a new molecule resulting in three solitons diverging from each other . from applications point of view
, the phase of @xmath60 could be used as a `` key '' to `` unlock '' the molecule for the purpose of extracting stored data . in fig .
[ fig13 ] , we show the effect of the initial speed of the injected soliton .
in contrary to one s first judgment , the molecule preserves its structure for fast collisions , as in fig .
[ fig13]a , and breaks up for slower collisions , as in fig .
[ fig13]b . in fig .
[ fig14 ] , the injected soliton has an amplitude that is approximately two times larger than any of the two solitons of the molecule .
the injected soliton penetrates the molecule leaving it almost unchanged apart from a center - of - mass shift to the left .
we have used the inverse scattering method to derive the two solitons solution of a nonlinear schr@xmath0diner equation with a parabolic potential and cubic nonlinearity with time - dependent coefficients , as given by eq .
( [ gp ] ) . the solution was then simplified and put in a suggestive form in terms of the fundamental parameters of the two solitons , namely their amplitudes , center - of - mass positions and speeds , and their phases . in this form ,
two different regimes of the solution , namely the resolved solitons and overlapping solitons , were distinguished and the main features such as the solitons separation and relative phase were extracted . from the expression for the solitons separation
we find that for the homogeneous case and zero solitons relative speed , the solitons separation diverges logarithmically with the solitons amplitude difference such that , for equal solitons , the trivial solution is obtained .
the force of interaction was then derived , essentially , for arbitrary solitons parameters .
this resulted in generalizing gordon s formula @xcite to _ i _ ) the generalized inhomogeneous case considered here , _ ii _ ) arbitrary solitons relative speed and amplitudes , and _
iii _ ) short solitons separations ( compared to their width ) . with this formula ,
the possibility of forming soliton molecules emerged naturally , where the force at short distance was shown to be composed of an attractive part , resulting from the nonlinearity , and another part that oscillates between repulsive and attractive resulting from the time - dependent relative phase .
the main features of the soliton molecule , including its equilibrium bond length and bond spring constant , frequency and amplitude of oscillation , effective mass , and its elastic potential energy , where then calculated in terms of the solitons parameters .
it turns out that the amplitudes difference @xmath63 plays an important role in determining these quantities .
furthermore , we show that at the condition @xmath250 , the two solitons coalesce while away from this condition , the solitons approach each other but remain resolved . at this condition ,
the molecule s effective mass and spring constant have maximum values . in our expressions eqs .
( [ keq ] ) and ( [ masseq ] ) diverge because these formulae were derived assuming the solitons remain resolved . to have a sense of its stability we investigated numerically _
i _ ) the collision of the soliton molecule with a hard wall and softer one , _ ii _ ) scattering by a potential barrier , and _ iii _ ) collision with a single molecule .
the first case showed that while the molecular structure is preserved after reflection from a hard wall , it breaks when reflecting from a softer one .
reflection from a finite barrier showed that the molecular structure is preserved only for specific heights of the barrier . for an incident molecule with the coalescence condition satisfied , the molecule will be most vulnerable to break up when the coalescence point takes place at the interface of the barrier .
this is simply understood by the fact that the intensity of the soliton molecule is maximum when the two solitons coalesce such that tunneling becomes possible .
stability of the molecule was also investigated by scattering the molecule with a single soliton .
it turned that slower collisions tend to break up the molecule more easily than faster ones .
in addition , the outcome of the collision depends on the phase of the incoming soliton such that a scatterer soliton which is in - phase with the molecule will typically preserve its molecular structure , but for an out - of - phase soliton , the molecule breaks up .
the two solitons solution presented here and the analysis that shows how to extract the solitons separation and relative phase may constitute the basis for a more accurate and detailed investigation of the origin of the soliton - soliton force , especially for short separations and at coalescence
. the results of this paper will be hopefully of relevance to possible future applications of soliton molecules as data carriers or memories .
the author acknowledges helpful discussions with vladimir n. serkin .
the lax pair associated with the gross - pitaevskii equation , eq .
( [ gp ] ) , is obtained using our lax pair search method @xcite and reads @xmath251 @xmath252 where + + @xmath253 , @xmath254 , @xmath255 , @xmath256 , + + + @xmath257 , + + + @xmath258 , @xmath259 , and @xmath260 and @xmath261 are arbitrary constants . here , @xmath6 is the solution of eq .
( [ gp ] ) and @xmath262 is the _ auxiliary _ field . the _ compatibility condition
_ @xmath263 of the linear system , eqs .
( [ phix ] ) and ( [ phit ] ) , is equivalent to the gross - pitaevskii equation , eq .
( [ gp ] ) , and its complex conjugate . for a known _ seed _ solution , @xmath264 , of eq .
( [ gp ] ) the linear system will have the solution @xmath265 .
the darboux transformation is defined as @xmath266={\bf\phi}\cdot{\lambda}-{\bf\sigma}\,{\bf\phi}$ ] , where @xmath266 $ ] is the transformed field and @xmath267 .
requiring the linear system to be _ covariant _ under the darboux transformation imposes the transformation @xmath268={\bf
p}+{\bf j}\cdot{\bf\sigma}-{\bf\sigma}\cdot{\bf j}$ ] , where @xmath268 $ ] is the transformed @xmath269 .
this gives the new solution in terms of the seed solution as @xmath270 using the trivial solution @xmath271 as a seed , the darboux transformation generates the well - known sech - shaped single soliton solution @xcite @xmath272 where @xmath273 @xmath274 @xmath275 @xmath276 , @xmath277 , @xmath278 , and the constant @xmath279 corresponds to an arbitrary overall phase .
this solution corresponds to a soliton density profile that is localized at @xmath37 , moving with center - of - mass speed @xmath280 , and normalized to @xmath40 . using this single soliton solution as a seed , the darboux transformation generates a two solitons solution .
the solution of the linear system , eqs .
( [ phix ] ) and ( [ phit ] ) , can in this case be derived and simplified to the following form @xmath281 @xmath282 @xmath283 @xmath284 @xmath285 where , @xmath286 @xmath287 @xmath288 @xmath289 finally , the two solitons solution is obtained by substituting for @xmath290 and @xmath291 in eq .
( [ newsol1 ] ) , which upon substituting @xmath292 and further simplification can then be put in the form of eq .
( [ exactsol ] ) .
since @xmath68 and @xmath69 are functions of @xmath15 , we start by examining their values near the roots of @xmath57 . this task can be simplified by rewriting @xmath68 as @xmath293 . for @xmath294 , we get @xmath295 , and for @xmath296 , we have @xmath297 . for @xmath113 the first term in the exponent of @xmath68 is larger than the second one , provided that @xmath40 is not much larger than @xmath298 , as remarked at the end of this section , hence @xmath73 .
for @xmath132 the magnitude of the first term in the exponent of @xmath68 is smaller than the magnitude of the second one , which again leads to @xmath73 . for the region @xmath299 ,
the condition @xmath73 is always satisfied . in conclusion ,
@xmath73 for all @xmath15 , apart from situations with extreme values of @xmath300 .
the situation is simpler for @xmath69 : @xmath74 for @xmath113 , @xmath301 for @xmath302 , and @xmath303 for @xmath132 .
thus , to find @xmath50 , we expand @xmath57 in powers of large @xmath68 and @xmath69 and to find @xmath48 and @xmath52 , we expand @xmath57 in powers of large @xmath68 .
expanding @xmath57 in powers of @xmath68 and @xmath69 around @xmath58 up to first order in @xmath80 and @xmath304 , we find @xmath305 where @xmath306 the root of this equation gives the center - of - mass position of the right soliton .
the first two log terms equal @xmath307 .
the third log term is constant and diverges for @xmath94 .
the last term is small since it is proportional to @xmath80 , but it is needed because it contains the phase - dependent contributions . due to the combination of @xmath308 and @xmath68 , finding an algebraic expression for the root of this equation will not be possible .
instead , we ignore at first the @xmath80 term to obtain the dominant contribution which will then be used to find the @xmath80 contribution .
this gives @xmath309 substituting back in eq .
( [ q1 ] ) and solving for @xmath15 , we finally get @xmath310 where @xmath311 . for the left and central solitons , we expand @xmath57 in powers of @xmath68 ; leaving @xmath69 arbitrary . in this manner
, we account for the two roots , @xmath48 and @xmath52 , simultaneously .
expanding @xmath57 in powers of large @xmath68 , we get @xmath312 where @xmath313\nonumber\\&/&{\left(g_0 ^ 2 e^{2 \gamma_0 } ( { n_d}+{n_s } { y } ) ^2 + 4 ( { y } + 1)^2 { \eta_d}^2\right)}.\end{aligned}\ ] ] similar to the above procedure for @xmath50 , we solve first eq . ( [ q2 ] ) without the @xmath80 term , which gives @xmath314 - 8 { \eta_d}^2}{2 g_0 ^ 2 { n_s}^2 e^{2 \gamma_0}+8 { \eta_d}^2},\ ] ] where @xmath315 in the limit @xmath89 and @xmath90 , the solution @xmath87 approaches 0 , which corresponds to @xmath316 . the solution @xmath317 approaches 1 , which corresponds to @xmath318 .
thus , the solutions @xmath87 and @xmath317 correspond the left and central solitons , respectively .
since we will be interested only in the left and right solitons , we take for the rest of this section the @xmath87 . to find the contribution of the @xmath80 term
, we substitute for @xmath87 in the @xmath80 term of eq .
( [ q2 ] ) , and then solve for @xmath69 to get a corrected expression for @xmath87 @xmath319 -8 { \eta_d}^2}{2 g_0 ^ 2 { n_s}^2 e^{2 \gamma_0}+8 { \eta_d}^2},\ ] ] where @xmath320 , @xmath321 , @xmath322 , @xmath323 , and @xmath324 . expanding for large @xmath325 and then solving for @xmath15 , we finally get @xmath326 final remarks about the validity of the above derivation are in order .
the condition @xmath73 will be met in the region @xmath113 only for @xmath327 .
note that the numerator of the right hand side of this inequality is less than the denominator by at least @xmath328 . for @xmath132
, the condition @xmath73 will be met only for @xmath329 . here
, the numerator of the right hand side of this inequality is larger than the denominator by at least @xmath328 .
a more quantitative estimate for the ratio @xmath300 can be obtained using the above results for @xmath330 and @xmath331 .
furthermore , the above - derived formula for @xmath48 is limited to values of the parameters for which the quantities @xmath332 and @xmath333 are positive . at @xmath334 , the two roots @xmath48 and @xmath52
coincide . in fig .
[ fig1 ] , this corresponds to the maximum of the @xmath57-curve lying on the @xmath15-axis . for @xmath335 ,
the maximum of the curve is below the @xmath15-axis and the two roots become nonreal . noting that : + @xmath341 + @xmath342 + @xmath343 + @xmath344 + @xmath345 + @xmath346 , + + the acceleration , @xmath190 , can be calculated to take the form of eq .
( [ force2 ] ) with coefficients @xmath347 @xmath348 @xmath349 @xmath350 a. hasegawa and y. kodama , _ solitons in optical communications _
( oxford university press , new york , 1995 ) ; l.f . mollenauer and j.p .
gordon , _ solitons in optical fibers _
( acadamic press , boston , 2006 ) ; g.p .
agrawal , _ nonlinear fiber optics _ ( academic press , san diego , 2001 ) , 3rd ed . ; n.n .
akhmediev and a. ankiexicz , _ solitons : nonlinear pulses and beams _ ( chapman and hall , london , 1997 ) .
l. khaykovich , f. schreck , g. ferrari , t. bourde , j. cubizolles , l.d .
carr , y. castin , c. salomon , science * 296 * , 1290 ( 2002 ) .
m. stratmann , t. pagel , and f. mitschke , phys .
* 95 * , 143902 ( 2005 ) . ) , @xmath47 $ ] .
thick ( green ) curve : the soliton intensity @xmath351 .
light ( red ) curve : the soliton intensity with only the second sech term of eq .
( [ exactsol ] ) .
the parameters used are : @xmath16 , @xmath352 , @xmath353 , @xmath354 , @xmath355 , @xmath356 , @xmath357 , @xmath358 , and @xmath167.,width=566 ] ) .
dotted ( blue ) curve corresponds to the phase of the right soliton , @xmath102 , calculated from eq .
( [ phaser ] ) .
thick dashed curve corresponds to the phase of the central soliton , @xmath103 , calculated from eq .
( [ phasec ] ) .
dashed - dotted curve corresponds to the phase of the left soliton , @xmath104 , calculated from eq .
( [ phasel ] ) .
the solid curves correspond to the density profiles as in fig .
the parameters used are : @xmath16 , @xmath352 , @xmath353 , @xmath354 , @xmath355 , @xmath359 , @xmath360 , @xmath357 , @xmath358 , and @xmath167.,width=566 ] , versus @xmath124 .
thick ( blue ) curve is calculated from eq .
( [ force2 ] ) .
light ( red ) curve is calculated numerically from the exact two solitons solution , eq .
( [ exactsol ] ) .
the parameters used are : @xmath16 , @xmath352 , @xmath353 , @xmath354 , @xmath355 , @xmath359 , @xmath357 , @xmath358 , and @xmath167.,width=566 ] ) . *
( b ) * thick ( green ) curve is solitons separation calculated from the exact solution ( [ exactsol ] ) .
light ( red ) curve is the solitons separation calculated from formula ( [ delmol ] ) .
* ( c ) * density profile at some specific times .
the parameters used are : @xmath16 , @xmath361 , @xmath362 , @xmath363 , @xmath356 , @xmath364 , @xmath358.,width=566 ] -[fig7 ] .
subfigures * ( a ) * , * ( b ) * , and * ( c ) * , correspond to fig .
[ fig5 ] , fig .
[ fig6 ] , and fig .
[ fig7 ] , respectively .
the blue ( lower ) and green ( upper ) curves correspond to the center - of - mass trajectories of the left and right solitons , respectively.,width=566 ] ) .
the dashed vertical line represents the interface of the potential step .
the parameters used are : @xmath364 , @xmath366 , @xmath367 , @xmath368 , @xmath369 , @xmath370 , @xmath371 .
* ( a ) * @xmath225 . *
( b),(c),(d ) * @xmath226 . , | the problem of soliton - soliton force is revisited . from the exact two solitons solution of a nonautonomous gross - pitaevskii equation
, we derive a generalized formula for the mutual force between two solitons .
the force is given for arbitrary solitons amplitude difference , relative speed , phase , and separation .
the latter allows for the investigation of soliton molecules formation , dynamics , and stability .
we reveal the role of the time - dependent relative phase between the solitons in binding them in a soliton molecule .
we derive its equilibrium bond length , spring constant , frequency , effective mass , and binding energy of the molecule .
we investigate the molecule s stability against perturbations such as reflection from surfaces , scattering by barriers , and collisions with other solitons . |
since its discovery by kageyama _ et al_.@xcite , the spin dimer compound srcu@xmath4(bo@xmath5)@xmath4 has attracted much attention as a suitable material for frustrated spin systems in low dimension .
srcu@xmath4(bo@xmath5)@xmath4 exhibits various interesting phenomena , such as a quantum disordered ground state @xcite and a complex shape of magnetization curve@xcite , because of its unique crystal structure . in consideration of the structure , miyahara and
ueda suggested that the magnetic properties of the spin dimer compound srcu@xmath4(bo@xmath5)@xmath4 can be described as a spin-@xmath6 two - dimensional ( 2d ) orthogonal - dimer model @xcite , equivalent to the shastry - sutherland model on square lattice with some diagonal bonds @xcite .
the ground state of the shastry - sutherland model in dimer phase is exactly represented by a direct product of singlets .
the low - energy dispersions possess six - fold degeneracy and are almost flat reflecting that the triplet tends to localize on vertical or horizontal bonds .
recent experiments by esr @xcite and neutron inelastic scattering ( nis ) have observed splitting of degenerate dispersions of srcu@xmath4(bo@xmath5)@xmath4 , which can not be explained by the _ isotropic _
shastry - sutherland model .
hence c ' epas _ et al .
_ pointed out that the dzyaloshinski - moriya ( dm ) interaction @xcite must be added between vertical and horizontal dimers in the isotropic shastry - sutherland model in order to explain the splitting .
@xcite in this paper , as a simple model to clarify effects of the dm interaction to low - energy excitations in orthogonal - dimer systems , one - dimensional ( 1d ) orthogonal - dimer model with the dm interaction is studied by using the perturbation theory and the numerical exact - diagonalization method . in the absence of the dm interactions , properties of ground state , low - energy excitations , and magnetization processes of the 1d orthogonal dimer model
has been studied by several authors .
the hamiltonian of the 1d orthogonal - dimer model with the dm interaction is given by @xmath7 where @xmath8 here @xmath9 is the number of unit cells in the system , as shown by a broken rectangle in fig .
the unit cell includes two dimers along vertical and horizontal direction , which are designated by the index , @xmath10 and @xmath11 , respectively .
@xmath12 ( @xmath13 and @xmath14 ) denotes a spin-@xmath6 operator on @xmath15-spin in @xmath10-th dimer . @xmath16 and
@xmath17 severally indicate the exchange coupling in intra - dimer and in inter - dimer . due to the structure of system ,
the dm exchange interaction , @xmath18 , exists only on inter - dimer bonds and has only a component perpendicular to two kinds of dimer in the unit cell .
the periodic boundary condition is imposed to the system . , that is @xmath19 .
the unit cell includes a vertical and horizontal dimer .
the former dimers are at @xmath10-site and the latter at @xmath20-site.,width=283 ]
in this section , let us discuss the ground state and low - energy excitations of the 1d orthogonal dimer model with the dm interaction .
we can expect that the ground state is in the dimer phase in the limit of strong intra - dimer coupling ( @xmath21 ) , even when the dm interaction is switched on the isotropic system .
therefore , it is reasonable to treat the intra - dimer hamiltonian ( [ eq : intra ] ) as an unperturbated one and the others as perturbation .
the inter - dimer interaction @xmath17 creates two adjacent triplets from a pair of a singlet and triplet and vice versa , and besides shows scatterings between two triplets .
the dm interaction not only causes the former process but also creates or annihilates two adjacent singlets . therefore the dm interaction can play a crucial role in the ground state and the low - energy excitations in the dimer phase .
first , we discuss the ground - state energy of hamiltonian ( [ eq : hamiltonian ] ) . in the absence of the dm interaction , the ground state
is exactly represented by a direct product of singlets and its energy is given as @xmath22 .
on the other hands , the ground - state energy of total hamiltonian ( [ eq : hamiltonian ] ) is estimated as @xmath23 from the perturbation expansion up to the third order in @xmath24 and @xmath25 .
the result means that the ground state can not be exactly described by the direct product of singlets owing to the dm interaction .
next , we argue the low - energy excitations in the system .
since the ground state belongs to the dimer phase in the region of strong-@xmath16 , the lowest excited states will be well described by @xmath26 here , @xmath27 and @xmath28 are the total magnetization and the wave number , respectivery . @xmath29 and
@xmath30 in ket severally denote a singlet and a triplet with @xmath31 at @xmath10-site and , @xmath32 ( @xmath33 ) is defined as an operator to create a triplet propagating on vertical ( horizontal ) dimers . by using two states of eqs .
( [ eq : vfourier ] ) and ( [ eq : pfourier ] ) , the hamiltonian ( 1 ) is projected on following ( @xmath34)-matrix : @xmath35 where @xmath36,~ { \mbox{\boldmath $ v$}}_m(k)\equiv \left [ \begin{array}{c } t_{m , k}^{\rm ver } \\ t_{m , k}^{\rm hor } \\
\end{array } \right].\end{aligned}\ ] ] the eq .
( [ eq : hm ] ) for @xmath1 has no off - diagonal elements within perturbation up to the third order .
therefore the excitation energies for @xmath1 are given by @xmath37 in contrast to the 2d orthogonal dimer model , two excitation energies , @xmath38 and @xmath39 , split in the case of 1d system .
it is also interesting to note that the curvature of @xmath39 appears in the third ordered correction in eq .
( [ eq : excitede1 ] ) .
on the other hand , the projected hamiltonian with @xmath40 has an off - diagonal element .
the perturbation calculation up to the third order leads to the matrix : @xmath41 , \label{eq : apm1}\end{aligned}\ ] ] where @xmath42 by diagonalizing eq .
( [ eq : apm1 ] ) , the excitation energies with @xmath40 are obtained as @xmath43 the curvature of @xmath44 is dominant by the first ordered correction with regard to @xmath25 in the off - diagonal element @xmath45 .
the correction derives from the scattering between a singlet and a triplet with @xmath2 due to the dm interaction . subtracting the ground - state energy of eq .
( [ eq : grounde ] ) from excited - state energies of eq .
( [ eq : excitede0 ] ) , ( [ eq : excitede1 ] ) , and ( [ eq : excitede3 ] ) , the low - energy dispersions , @xmath46 , are estimated as @xmath47 figure 2 shows the low - energy dispersions for @xmath48 and @xmath49 .
the low energy spectra @xmath50 and @xmath51 are severally represented by the lower and upper solid lines , and then the upper and lower dotted lines denote @xmath52 and @xmath53 in eqs .
( [ eq : omega2 ] ) .
the full and open circles represent the low - energy spectra with @xmath1 and @xmath2 .
the perturbation theory is in agreement with the numerical diagonalization , as to low - energy excitations .
the dispersions with same @xmath31 are not degenerate , which happens even if the dm interaction is not taken account of .
therefore the inter - dime coupling @xmath17 is also important for splitting of the low - energy dispersions with same @xmath31 .
this is because the parity on the vertical dimer is conserved in the one - dimension system without the dm interaction .
the dm interaction not only splits into branches with @xmath1 and @xmath2 , but also makes triplets move more easily . and @xmath49 .
the solid and dotted curves are of @xmath1 and @xmath40 , respectively .
the full ( open ) circles indicates the excitation energies for @xmath1 ( @xmath2 ) calculated by numerical diagonalizations.,width=264 ]
we investigated the low - energy excitations in the 1d orthogonal - dimer model with the dm interaction using the perturbation theory and numerical exact - diagonalization method .
the dm interaction allows a triplet to propagate in singlet sea as seen in the 2d system @xcite , while the triplet is localized on vertical or horizontal dimer in the absence of the dm interaction .
this curvature effect happens in a two - dimensional orthogonal - dimer model , but the splitting of spectra reflects a stringent constraint for the motion of a triplet due to one dimensionality as well as the dm interaction .
we gratefully acknowledge helpful discussions and comments with tetsuro nikuni , masaaki nakamura , and takahiro yamamoto on several points in this work .
we would also like to thank k@xmath54ichir@xmath54 ide for valuable advice on numerical techniques .
9 h. kageyama , et al .
* 82 * ( 1999 ) 3168 .
k. onizuka , et al . : j. phys .
* 69 * ( 2000 ) 1016 .
s. miyahara and k. ueda : phys .
87 * ( 2001 ) 3701 .
b. s. shastry and b. sutherland : physica b * 108 * ( 1981 ) 1089 .
h. nojiri , et al . : j. phys .
* 68 * ( 1999 ) 2906 .
i. dzyaloshinski : j. phys .
solids * 4 * ( 1958 ) 241 ; + t. moriya : phys . rev .
* 120 * ( 1960 ) 91 .
o. cpas , et al .
: phys . rev
. lett . * 87 * ( 2001 ) 167205 .
a. koga , k. okunishi and n. kawakami : phys .
b. * 62 * ( 2000 ) 5558 .
n. b. ivanov and j. richter : phys .
a * 232 * ( 1997 ) 308 .
j. richter , n. b. ivanov and j. shulenburg : phys .
* 88 * ( 2002 ) 201601 .
s. miyahara and k. ueda : j. phys .
( _ suppl . _ ) b * 70 * ( 2001 ) 180 . | effects of the dzyaloshinski - moriya ( dm ) interaction on low - energy excitations in a one - dimensional orthogonal - dimer model are studied by using the perturbation expansions and the numerical diagonalization method . in the absence of the dm interaction ,
the triplet excitations show two flat spectra with three - fold degeneracy , which are labeled by magnetization @xmath0 . these spectra split into two branches with @xmath1 and with @xmath2 by switching - on of the dm interaction and besides the curvature appears in the triplet excitations with @xmath3 more strongly than those of @xmath1 . ,
, quantum spin ; low - energy excitation ; dzyaloshinski - moriya interaction ; |
the parton model , according to which hadrons consist of quarks antiquarks and gluons ( partons ) , bound together in different ways , has been very successful in reproducing experiment .
this provides a relatively explicit and transparent technique for the description of high energy particle interactions .
the distributions of partons inside hadrons are characterized by the structure functions satisfying the dokshitzer
gribov lipatov
altarelli parisi ( dglap ) equations @xcite or ones that are basically similar .
such a function is the probability density of finding a parton in a hadron carrying a fraction of the total hadron s momentum .
numerical solutions of the equations are in a remarkable agreement with experimental measurements , especially for the nucleon @xcite .
photons being involved in high energy interactions are also able to manifest hadronic structure .
one can intuitively comprehend this since the photon directly couples to quarks and therefore may split into quark
antiquark pairs .
the parton contributions in two - photon processes and some crucial peculiarities of the kinematic behavior of the photon structure function have been described by walsh and zerwas @xcite . the first work in studying quantum - chromodynamics corrections to the naive pointlike structure of the photon belongs to witten @xcite .
this problem was also studied in refs .
@xcite . introducing the evolution equations , similar to the dglap ones , for photons as well as the properties of the corresponding solutions
were under scrutiny , for instance , in a series of papers by glk , reya , grassie and vogt @xcite .
a formulation of high energy @xmath2 interactions taking into account the hadronic properties of the photon was proposed in ref .
@xcite . today
, we possess convincing experimental evidence for the fact that the real photon has non - trivial parton structure @xcite . on the other hand , the cosmic microwave background ( cmb ) photons may play an important role in the formation of cosmic rays ( cr ) .
one of the brightest representatives is the greisen zatsepin kuzmin ( gzk ) limit on the energy of cr @xcite .
for example , protons of energies of over about @xmath3 ev would be decelerated by interaction with the cmb photons , mostly due to resonant pion production , @xmath4 .
other interesting processes , the @xmath5 reactions and their possible astrophysical implications , were extensively discussed in the literature ( see , _
e.g. _ , refs .
@xcite and the references cited therein ) . in this context , to invoke the parton content of the real photon could be convenient for calculations of the probabilities of such processes . here , we attempt to show the example of @xmath0 boson production in the @xmath5 scattering which may have important consequences for astrophysics @xcite . studying this reaction
could also provide a test of the universality of the parton distribution functions of the photon .
let us first consider inclusive on - shell @xmath0 boson production in the reaction @xmath6 at the resonance region using the parton language
. we will view it from the center - of - mass ( cms ) frame of the @xmath7 system . here , for example , a substantial fraction of the cmb photons will be of energies of about @xmath8 , where @xmath9 is the neutrino energy in the laboratory frame defined as the frame in which the cmb is isotropic , @xmath10 is the cmb photon energy ( typical value @xmath11 ev @xcite ) .
this reaction is standardly factorized into two subprocesses : the emission of a positron by the photon and annihilation of the neutrino with the positron into @xmath0 ( see fig .
then the corresponding cross section may be written as here @xmath13 is the total cms energy squared ( @xmath14 ) , @xmath15 is the probability density function to find the positron in the photon carrying the fraction @xmath16 of the total photon s momentum , and @xmath17 is the cross section of the annihilation subprocess .
note that we explicitly write the @xmath13 dependence of the function instead of the more traditional @xmath18 one ( 4-momentum transfer squared ) since we deal with an @xmath13-channel subprocess . where @xmath20 is the mass of the @xmath0 boson
, @xmath21 is the partial width of the initial channel ( the partial width for the decay @xmath22 ) , and @xmath23 is the total decay width of @xmath0 . in the leading order one can find that @xcite to determine the function @xmath15 , we adopt the formalism given in ref .
it is fair to expect @xmath15 to satisfy , up to factors associated with the quark colors and fractional electric charges , the same evolution equation as the quark distributions in the photon do , provided the gluons are excluded and one takes into account only the electromagnetic interaction . then ,
in the leading order we write the following equation for the positron distribution @xcite : the dependence of the cross section on @xmath13 is displayed in fig .
2a in comparison with calculations of the closely related process @xmath32 carried out by seckel @xcite . here
@xmath33 gev , @xmath34 gev@xmath35 , @xmath36 @xcite .
one can see that the values given by eq .
( [ eq:6 ] ) are about two times higher than those of ref .
@xcite .
let us turn now to the charged current interaction of the neutrino with the quark content of the photon ( see fig .
the corresponding cross section can be obtained in the same way as it is done for neutrino proton scattering @xcite : where @xmath39 is the probability density to find a quark @xmath40 ( antiquark @xmath41 ) in the photon carrying the fraction @xmath16 of the total photon s momentum .
taking into account only the densities of the lightest quarks @xmath42 and @xmath43 from ref .
@xcite , we found that here @xmath45 and @xmath46 are the electric charge and mass of the quark @xmath40 respectively ( for antiquarks the equation is analogous ) .
note that eq .
( [ eq:9 ] ) is valid in the limit @xmath47 .
we set @xmath48 gev and @xmath49 .
the dependence of the cross section thus determined on @xmath13 in the range 20 gev@xmath50 @xmath51 1000 gev@xmath50 at some values of @xmath18 is shown in fig .
this reaction may have interesting astrophysical implications because the struck quark may fragment into hadrons .
the latter can be highly boosted and on decaying ( if unstable ) may produce particles with energies exceeding their gzk limit . if it occurs in the vicinity of the earth the decay products may reach us without significant energy loss , provided the incident quark momentum pointed in the direction of the earth .
a similar idea has been proposed , for example , in ref .
@xcite , when photons would appear beyond the gzk limit from decays of highly boosted @xmath52 , which , in turn , were the decay products of real @xmath53 bosons excited in @xmath54 annihilation ( the so - called `` @xmath55-burst '' mechanism ) .
but there are problems here , mainly associated with the origin of such high energy neutrinos , @xmath56 ( see , _
e.g. _ , ref .
@xcite ) . in our case ,
the minimal neutrino energy required to produce hadrons is smaller than the latter one by about a factor of 400 , and the corresponding cross section is also suppressed by a factor @xmath57 .
anyway , one may expect that these processes were important for high energy neutrino absorption in the early universe . throughout this paper we implicitly used the assumption that the parton distributions are process
independent , which has been experimentally justified for the nucleon .
for example , the functions phenomenologically derived from electron nucleon and neutrino nucleon deep inelastic scattering data are close to each other . using them
one can correctly predict the probabilities of inclusive production of @xmath58 pairs in @xmath59 collisions ( drell yan process ) @xcite .
we have discussed only the @xmath7 interactions . meanwhile ,
all the things we said above may be straightforwardly applied to the reactions involving the antineutrino . likewise ,
heavier charged leptons can be considered .
one may also include neutral current interactions in the neutrino
quark scattering .
other processes involving the cmb photons can be treated in similar way .
99 v. n. gribov , l. n. lipatov , sov .
j. nucl .
phys . * 15 * , 438 ( 1972 ) .
g. altarelli , g. parisi , nucl .
b * 126 * , 298 ( 1977 ) .
y. l. dokshitzer , sov .
jetp * 46 * , 641 ( 1977 ) .
zeus collaboration , z. phys .
c * 72 * , 399 ( 1996 ) .
t. f. walsh and p. zerwas , phys . lett .
b * 44 * , 195 ( 1973 ) .
e. witten , nucl .
b * 120 * , 189 ( 1977 ) .
r. j. dewitt _ et al .
_ , phys .
d * 19 * , 2046 ( 1979 ) .
w. a. bardeen , a. j. buras , phys .
d * 20 * , 166 ( 1979 ) .
d. w. duke and j. f. owens , phys .
d * 22 * , 2280 ( 1980 ) .
m. glk and e. reya , phys .
d * 28 * , 2749 ( 1983 ) .
m. glk , k. grassie , and e. reya , phys .
d * 30 * , 1447 ( 1984 ) .
m. glck , e. reya , and a. vogt , phys .
d * 45 * , ( 1992 ) 3986 . g. a. schuler and t. sjstrand , phys .
b * 300 * , 169 ( 1993 ) .
j. cvach , nucl .
b ( proc . suppl . )
* 79 * , 501 ( 1999 ) .
k. greisen , phys .
* 16 * , 748 ( 1966 ) .
g. t. zatsepin and v. a. kuzmin , sov .
jetp lett .
* 4 * , 78 ( 1966 ) .
d. seckel , phys .
lett . * 80 * , 900 ( 1998 ) .
a. abbasabadi _
et al . _ ,
d * 59 * , 013012 ( 1999 ) .
a. abada , j. matias , r. pittau , nucl . phys .
b * 543 * , 255 ( 1999 ) .
e. masso , f. rota , phys .
b * 488 * , 326 ( 2000 ) .
e. v. bugaev , int .
j. mod .
a * 20 * , 6909 ( 2005 ) .
m. haghighat , m. m. ettefaghi , m. zeinali , phys .
d * 73 * , 013007 ( 2006 ) . t. k. gaisser , f. halzen , t. stanev , phys .
* 258 * , 173 ( 1995 ) .
p. bhattacharjee , g. sigl , phys .
rept . * 327 * , 109 ( 2000 ) .
_ , _ particles and nuclei .
an introduction to the physical concepts _ , 5th edn .
( springer , berlin , heidelberg 2006 ) . l. n. lipatov , _ standard model of elementary particle interactions _ ( lectures given at saint petersburg state university , 2006 ) .
w. furmanski and r. petronzio , phys .
b * 97 * , 437 ( 1980 ) .
e. g. floratos , c. kounnas , and r. lacaze , nucl .
b * 192 * , 417 ( 1981 ) .
chen and p. zerwas , phys .
d * 12 * , 187 ( 1975 ) .
( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) . t. j. weiler , astropart .
phys . * 11 * , 303 ( 1999 ) . f. e. close , _ an introduction to quarks and partons _
( academic press , london 1979 ) . | we possess convincing experimental evidence for the fact that the real photon has non - trivial parton structure .
on the other hand , interactions of the cosmic microwave background photons with high energy particles propagating through the universe play an important role in astrophysics . in this context , to invoke the parton content could be convenient for calculations of the probabilities of different processes involving these photons . as an example , the cross section of inclusive resonant @xmath0 boson production in the reaction @xmath1 is calculated by using the parton language .
neutrino
photon deep inelastic scattering is considered . |
let @xmath5 be the quantum enveloping algebra associated to an affine algebra @xmath6 without derivation .
let @xmath7 be finite - dimensional @xmath5-modules .
suppose @xmath8 is irreducible and @xmath7 have crystal bases @xmath9 .
then it is known @xcite that there exists a unique map @xmath0 from @xmath10 to @xmath11 commuting with any crystal operators @xmath12 and @xmath13 .
there also exists an integer - valued function @xmath14 on @xmath10 , called energy function , satisfying a certain recurrence relation under the action of @xmath12 ( see ) .
combinatorial @xmath0-matrices or energy functions play an important role in the affine crystal theory . in the kyoto path model @xcite , that realizes the affine highest weight crystal in terms of a semi - infinite tensor product of perfect crystals ,
the energy function is an essential ingredient for the computation of the affine weight . in the box - ball system @xcite or its generalizations @xcite in the formulation of crystal bases ,
the time evolution of the system is defined by using the combinatorial @xmath0-matrix .
energy functions are also crucial in the calculation of conserved quantities . in @xcite
a new connection was revealed between the energy function and the kkr or kss bijection @xcite that gives a one - to - one correspondence between highest weight paths and rigged configurations .
recently , for all nonexceptional affine types , all kr crystals , crystal bases of kirillov reshetikhin ( kr ) modules ( if they exist ) , were shown to exist and their combinatorial structures were clarified @xcite .
hence , it is natural to consider the problem of obtaining a rule to calculate the combinatorial @xmath0-matrix and energy function . in this paper , for type @xmath2 we calculate the image of the combinatorial @xmath0-matrix for any classical highest weight element in the tensor product of kr crystals @xmath1 ( theorem [ th : main ] ) .
( note that the first upper index of the second component is 1 . )
we also obtain the value of the energy function for such elements .
although we get the rule only for highest weight elements , there is an advantage from the computational point of view , since it is always easy to calculate the action of crystal operators @xmath15 for @xmath16 not only by hand but also by computer . to identify highest weight elements in the image
@xmath4 the notion of @xmath3-diagrams , introduced in @xcite , is used effectively .
the paper is organized as follows . in section 2
we briefly review crystals and @xmath3-diagrams . in section 3
we recall the kr crystal @xmath17 for type @xmath18 and @xmath19 , and the notion of combinatorial @xmath0-matrix and energy function .
the condition for an element of @xmath1 or @xmath4 to be classically highest is also presented .
the main theorem is given in section 4 . in section 5
we prove a special case of the theorem , and reduction to this case is discussed in section 6 according to whether @xmath20 is odd or even .
mo was supported by grant jsps 20540016 .
the work of rs is supported by the core research for evolutional science and technology of japan science and technology agency .
let @xmath6 stand for a simple lie algebra or affine kac
moody lie algebra with index set @xmath21 and @xmath22 the corresponding quantized enveloping algebra .
axiomatically , a @xmath6-crystal is a nonempty set @xmath23 together with maps @xmath24 where @xmath25 is the weight lattice associated to @xmath6 .
the maps @xmath12 and @xmath13 are called kashiwara operators and @xmath26 is the weight function . to each crystal one
can associate a crystal graph with vertices in @xmath23 and an arrow colored @xmath27 from @xmath28 to @xmath29 if @xmath30 or equivalently @xmath31 . for @xmath32 and @xmath27 , let @xmath33 in this paper we only consider crystal bases coming from @xmath22-modules
. for a complete definition of crystal bases see for example @xcite .
let @xmath34 be crystals .
then @xmath35 can be endowed with the structure of crystal .
the actions of kashiwara operators and the value of the weight function are given by @xmath36 the multiple tensor product is defined inductively . in order to compute the action of @xmath15 on multiple tensor products , it is convenient to use the rule called signature rule " @xcite
. let @xmath37 be an element of the tensor product of crystals @xmath38 .
one wishes to find the indices @xmath39 such that @xmath40 to do it , we introduce ( @xmath41-)signature by @xmath42 we then reduce the signature by deleting the adjacent @xmath43 pair successively .
eventually we obtain a reduced signature of the following form .
@xmath44 then the action of @xmath12 ( resp .
@xmath13 ) corresponds to changing the rightmost @xmath45 to @xmath46 ( resp .
leftmost @xmath46 to @xmath45 ) . if there is no @xmath45 ( resp .
@xmath46 ) in the signature , then the action of @xmath12 ( resp .
@xmath13 ) should be set to @xmath47 .
the value of @xmath48 ( resp .
@xmath49 ) is given by the number of @xmath45 ( resp .
@xmath46 ) in the reduced signature . consider , for instance , an element @xmath50 of the 3 fold tensor product @xmath51 .
suppose @xmath52 .
then the signature and reduced one read @xmath53 thus we have @xmath54 we denote by @xmath55 the highest weight crystal of highest weight @xmath56 , where @xmath56 is a dominant integral weight .
let @xmath57 with @xmath27 be the fundamental weights associated to a simple lie algebra . in this paper , we consider the types of @xmath58 and @xmath59 . as usual , a dominant integral weight @xmath60 is identified with a partition or young diagram with columns of height @xmath61 for @xmath62 , except when @xmath63 is a spin weight , namely , @xmath64 for type @xmath65 and @xmath66 and @xmath64 for type @xmath59 . to represent elements of @xmath55 we use kashiwara nakashima ( kn ) tableaux , a generalization of semistandard young tableaux for type @xmath67
. for kn tableaux refer to @xcite .
( see also @xcite for a summary . )
contrary to the original one , we use the french notation where parts are drawn in increasing order from top to bottom . to calculate the actions of @xmath15 on a kn tableau
it is convenient to use so - called the japanese reading word of a tableau . for a kn tableau @xmath68 move from right to
left and on each column move from bottom to top . during this process
we read letters , thereby obtaining a word @xmath69 .
a letter can be identified with an element of @xmath70 , crystal of the vector representation .
hence @xmath69 can be viewed as an element of @xmath71 with @xmath72 being the number of nodes in @xmath68 or length of @xmath69 .
then the action of @xmath12 or @xmath13 is calculated by using the signature rule . we still need to remember the crystal graph of @xmath70 for type @xmath73 , but it is easy as described in @xcite .
let @xmath74 be @xmath58 or @xmath59 .
for a subset @xmath75 , we say that @xmath32 is @xmath76-highest if @xmath77 for all @xmath78 .
we set @xmath79 .
we describe @xmath76-highest elements in terms of a notion of @xmath3-diagram @xcite .
a @xmath3-diagram @xmath25 of shape @xmath80 is a sequence of partitions @xmath81 such that @xmath82 and @xmath83 are horizontal strips .
we depict this @xmath3-diagram by the skew tableau of shape @xmath80 in which the cells of @xmath83 are filled with the symbol @xmath46 and those of @xmath82 are filled with the symbol @xmath45 .
write @xmath84 and @xmath85 for the outer and inner shapes of the @xmath3-diagram @xmath25 . for type @xmath86
we have a further requirement : the outer shape @xmath56 contains columns of height at most @xmath87 , but the inner shape @xmath88 is not allowed to be of height @xmath87 ( hence there are no empty columns of height @xmath87 ) . as we have discussed we identify a young diagram with a weight .
@xcite [ p : branch ] let @xmath56 be an @xmath74 weight that does not contain spin weights .
then there is an isomorphism of @xmath89-crystals @xmath90 that is , the multiplicity of @xmath91 in @xmath92 , is the number of @xmath3-diagrams of shape @xmath80 .
there is a bijection @xmath93 from @xmath3-diagrams @xmath25 of shape @xmath80 to the set of @xmath76-highest elements @xmath28 of @xmath89-weight @xmath94 . for any columns of height @xmath87 containing @xmath46 , place a column @xmath95
otherwise , place @xmath96 in all positions in @xmath25 that contain a @xmath45 , and fill the remainder of all columns by strings of the form @xmath97 .
we move through the columns of @xmath28 from top to bottom , left to right .
each @xmath46 in @xmath25 ( starting with the leftmost moving to the right ignoring @xmath46 at height @xmath87 ) will alter @xmath28 as we move through the columns .
suppose the @xmath46 is at height @xmath98 in @xmath25 .
if one encounters a @xmath96 , replace @xmath96 by @xmath99 .
if one encounters a @xmath100 , replace the string @xmath97 by @xmath101 .
[ ex : pm - diag ] let us consider the following @xmath3-diagram .
@xmath102 to obtain @xmath103 we first draw the tableau @xmath104 reading from left there are @xmath46 s at height 4,3,2,2,1 .
each @xmath46 alter the above tableau as follows .
the 1st @xmath46 changes the first column as @xmath105 ( reading from bottom ) , the 2nd and 3rd change the second column as @xmath106 , the 4th changes the third column as @xmath107 and the 5th changes the fourth column as @xmath108 .
therefore , @xmath103 is given by @xmath109 for a word @xmath110 let @xmath111 .
we use this convention also for @xmath112 . note that the order in @xmath113 is reversed from @xmath114 .
next proposition shows how we get to the highest element from a @xmath3-diagram by applying @xmath12 s .
[ prop : to highest ] let @xmath25 be a @xmath3-diagram whose outer shape has depth @xmath20 .
suppose @xmath115 for @xmath65 , @xmath116 for @xmath86 , @xmath117 for @xmath59 .
let @xmath118 be the number of columns of the outer shape with height @xmath41 .
let @xmath119 ( resp .
@xmath120 ) be the number of @xmath45 ( resp .
@xmath46 ) at height @xmath41 . define a word @xmath114 by @xmath121 where @xmath122 where @xmath123 for @xmath65 , @xmath124 for the other cases , and @xmath125 for @xmath59 , @xmath126 for the other cases .
then @xmath127 is the hightest weight element with highest weight given by its outer shape .
moreover , at each step when we apply @xmath128 or @xmath129 , including @xmath130 , the action is maximal , namely , if we apply @xmath131 or @xmath132 , the outcome turns out @xmath133 .
suppose @xmath134 for type @xmath86 .
we first prove the claim when there is no @xmath46 in @xmath25 .
set @xmath135 .
then the japanese reading word of the tableau corresponding to @xmath25 is given by @xmath136 the 1-signature is just given by @xmath137 , where @xmath138 , and there is no need to reduce . hence one can apply @xmath139 . calculating similarly for @xmath140 one always has a simple @xmath41-signature of the form @xmath141 , and we arrive at the highest weight element as desired .
next we consider the general case .
we prove by induction on @xmath72 , the number of @xmath46 . if @xmath142 , the claim is proven .
suppose @xmath143 and let @xmath98 be the height of the lowest @xmath46 in @xmath25 .
let @xmath144 be the same @xmath3-diagram as @xmath25 except that there are one less @xmath46 s at height @xmath98 .
compare the japanese reading word of the corresponding tableaux of @xmath25 and @xmath144 .
the difference is : @xmath145 apart from this difference in two words , there are subwords of the form @xmath146 or letters @xmath96 on the left and subwords of the form @xmath147 or @xmath148 for some @xmath149 on the right .
let us calculate the 1-signatures of both words .
they are @xmath150 for @xmath25 and @xmath151 for @xmath144 .
( there are no @xmath43 pairs . ) after applying @xmath152 on both @xmath25 and @xmath144 , the 2-signatures also turn out of the form @xmath153 for @xmath25 and @xmath154 for @xmath144 .
the difference is that there is @xmath155 or @xmath99 in @xmath25 but @xmath156 or @xmath157 in @xmath144 .
similar situations continue until we apply @xmath158 , and after applying @xmath159 , the two results coincide .
hence we should have the desired result .
the proof in the case of @xmath160 for type @xmath86 is almost the same .
the only difference is that we first treat the case when there is no @xmath46 in @xmath25 except at height @xmath87 , since there is no empty column of height @xmath87 .
hence we omit the proof . for a @xmath3-diagram given in example [ ex : pm - diag ] set @xmath161 .
then , according to the previous proposition @xmath127 is a highest weight element . later in this paper
we will need to apply @xmath162 to a @xmath3-diagram @xmath25 .
since @xmath163 is no longer @xmath76-highest , we have to use a pair of @xmath3-diagrams @xmath164 to consider @xmath165-highest elements .
namely , @xmath25 represents a @xmath76-highest element and @xmath166 represents a @xmath165-highest element in the @xmath89-component whose highest weight vector correponds to @xmath25 . under this bijection
we identify a @xmath165-highest element @xmath28 with a pair of @xmath3-diagram @xmath164 . to describe the action of @xmath162 on
@xmath164 perform the following algorithm : 1 .
successively run through all @xmath46 in @xmath166 from left to right and , if possible , pair it with the leftmost yet unpaired @xmath46 in @xmath25 weakly to the left of it .
2 . successively
run through all @xmath45 in @xmath166 from left to right and , if possible , pair it with the rightmost yet unpaired @xmath45 in @xmath25 weakly to the left .
3 . successively run through all yet unpaired @xmath46 in @xmath166 from left to right and , if possible , pair it with the leftmost yet unpaired @xmath45 in @xmath166 .
* lemma 5.1 ) [ prop : e1 action ] if there is an unpaired @xmath46 in @xmath166 , @xmath162 moves the rightmost unpaired @xmath46 in @xmath166 to @xmath25 .
else , if there is an unpaired @xmath45 in @xmath25 , @xmath162 moves the leftmost unpaired @xmath45 in @xmath25 to @xmath166 .
else @xmath162 annihilates @xmath164 .
let @xmath6 be an affine lie algebra of type @xmath167 , @xmath168 , or @xmath169 with the underlying finite - dimensional simple lie algebra @xmath170 of type @xmath171 , or @xmath86 , respectively .
we label the vertices of the corresponding dynkin diagram according to @xcite , so the index set of @xmath6 ( resp .
@xmath170 ) is @xmath172 ( resp .
@xmath173 ) . in this section
we review kr crystals @xmath17 of type @xmath6 given in @xcite for @xmath174 and @xmath175 for @xmath167 , @xmath176 for @xmath168 and @xmath177 for @xmath169 . as an @xmath74-crystal , @xmath17
is given by @xmath178 here @xmath55 is the @xmath74-crystal of highest weight @xmath56 and the sum runs over all dominant weights @xmath56 that can be obtained from @xmath179 by the removal of vertical dominoes , where @xmath57 are the @xmath41-th fundamental weights of @xmath74 . in order to define the actions of @xmath180 and @xmath181 we first consider an automorphism @xmath182 on the kr crystal @xmath17 .
the dynkin diagrams of type @xmath183 , and @xmath169 all have an automorphism interchanging nodes @xmath133 and @xmath184 .
@xmath182 corresponds to this dynkin diagram automorphism . by
construction @xmath182 commutes with @xmath185 and @xmath186 for @xmath187 .
hence it suffices to define @xmath182 on @xmath76-highest elements .
because of the bijection @xmath188 from @xmath3-diagrams to @xmath76-highest elements as described in section [ subsec : pm diag ] , it suffices to define the corresponding map @xmath189 on @xmath3-diagrams .
let @xmath25 be a @xmath3-diagram of shape @xmath80 .
let @xmath190 be the number of columns of height @xmath41 in @xmath94 for all @xmath191 with @xmath192 .
if @xmath193 , then in @xmath25 , above each column of @xmath94 of height @xmath41 , there must be a @xmath46 or a @xmath45 .
interchange the number of such @xmath46 and @xmath45 symbols . if @xmath194 , then in @xmath25 , above each column of @xmath94 of height @xmath41 , either there is no sign or a @xmath195 pair .
suppose there are @xmath196 @xmath195 pairs above the columns of height @xmath41 .
change this to @xmath197 @xmath195 pairs .
the result is @xmath198 , which has the same inner shape @xmath94 as @xmath25 but a possibly different outer shape .
let @xmath199 and @xmath200 be such that @xmath201 is a @xmath76-highest element .
then , @xmath202 is given by @xmath203 where @xmath204 .
the affine crystal operators @xmath205 and @xmath206 are then defined as @xmath207 if @xmath208 , the structure of the kr crystal turns out simple .
an crystal element of @xmath209 can be identified with one - row kn tableau of length @xmath210 with letters from @xmath211 ( @xmath212 ) ( and @xmath133 for @xmath168 ) . denoting the number of letters @xmath211 or @xmath133 by @xmath213 or @xmath214
, we have the so - called coordinate representation of @xmath209 @xcite . @xmath215
the action of @xmath15 for @xmath216 can be calculated as we explained in the last paragraph of section [ subsec : crystals ] .
the action of @xmath217 is given by @xmath218 we list the values of @xmath219 below .
@xmath220 where @xmath221 .
let us now consider a tensor product of kr crystals @xmath222 .
it is known @xcite that there exists a unique bijection @xmath0 , called combinatorial @xmath0-matrix , commuting with kashiwara operators @xmath15 for any @xmath223 .
since @xmath0 preserves the weight , @xmath224 should be sent to @xmath225 by @xmath0 , where @xmath226 ( resp .
@xmath227 ) is the @xmath228-highest elements of @xmath229 ( resp .
@xmath230 ) in @xmath17 ( resp .
@xmath231 ) . for the other elements the image
is uniquely determined , since @xmath222 is known to be connected @xcite .
next we explain the energy function @xmath14 .
let @xmath232 correspond to @xmath233 by @xmath0 .
suppose @xmath234 . applying @xmath12 on both sides of @xmath235 ,
we are led to consider the following four cases : @xmath236 then the function @xmath14 is uniquely determined , up to adding a constant , by @xmath237 although it is not obvious that such a function exists , it is shown to exist @xcite .
next we investigate conditions for an element of @xmath1 or @xmath4 to be @xmath228-highest .
recall the following fundamental fact : @xmath238 in particular , if @xmath239 is @xmath228-highest , then @xmath28 has to be @xmath228-highest . [ prop : ht cond 1 ] let @xmath240 be a dominant integral weight that appears in as highest weight
. by abuse of notation let @xmath241 also stand for the highest kn tableau of weight @xmath241 .
let @xmath242 be an element of @xmath209 represented by coordinates .
then , an element @xmath243 of @xmath1 is @xmath228-highest , if and only if @xmath244 for some @xmath241 as above and the following conditions for @xmath242 are satisfied .
* @xmath245 if @xmath246 , * @xmath247 if @xmath246 or @xmath248 , * @xmath249 if @xmath250 and @xmath251 , * @xmath252 if @xmath253 and @xmath251 . in the case of @xmath160
where @xmath254 , @xmath255 appearing in ( iii ) should be understood as @xmath133 .
apply for @xmath216 and use the formula for @xmath256 in . in
what follows , for a @xmath3-diagram @xmath25 we use the following notation .
let @xmath257 be one of @xmath258 ( @xmath259 stands for emptiness ) .
we denote by @xmath260 the number of columns of the outer shape of @xmath25 of height @xmath41 that contain @xmath257 .
( 30,10 ) ( 0,0)(1,0)30 ( 5,6)(1,0)20 ( 0,0)(5,0)4 ( 5,6)(5.3,6.5)(6.7,6.8 ) ( 8.3,6.8)(9.7,6.5)(10,6 ) ( 5,6)(0,1)2 ( 25,6)(0,-1)2 ( 25.3,3.9)(0.3,-0.1)16 ( 4.7,8.1)(-0.3,0.1)16 ( 7.1,6.8)@xmath261 ( 12.1,6.8)@xmath262 ( 17.1,6.8)@xmath263 ( 22.1,6.8)@xmath264 ( 5,3.7)(0,1)2.3 ( 5,2.3)(0,-1)2.3 ( 4.8,2.7)@xmath41 ( 10,5.2)@xmath46 ( 14.2,5.2)@xmath46 ( 15,5.2)@xmath45 ( 19.2,5.2)@xmath45 ( 20,5.2)@xmath45 ( 24.2,5.2)@xmath45 ( 20,4.3)@xmath46 ( 24.2,4.3)@xmath46 ( 0,0)(5,0)3 ( 10.95,5.42)(0.3,0)11 ( 20.95,4.52)(0.3,0)11 [ prop : ht cond 2 ] an element @xmath239 of @xmath4 is @xmath228-highest , if and only if @xmath265 and @xmath29 is a @xmath76-highest element whose corresponding @xmath3-diagram @xmath25 satisfies @xmath266 apply for @xmath216 and use proposition [ prop : e1 action ] to calculate @xmath267 of the @xmath3-diagram @xmath25 .
we consider the combinatorial @xmath0-matrix @xmath268 let @xmath269 be @xmath228-highest and @xmath270 .
then , @xmath271 is also @xmath228-highest , and from propositions [ prop : ht cond 1 ] and [ prop : ht cond 2 ] @xmath244 for some dominant integral weight @xmath241 , @xmath272 and there exists a @xmath3-diagram @xmath25 such that @xmath273
. thus we have @xmath274 for an element of @xmath209 we use both the coordinate representation and the japanese reading word of the corresponding one - row tableau .
let @xmath260 ( @xmath275 ) be data corresponding to @xmath25 as in the previous section .
then our main result is : [ th : main ] with the notations above we have the following formulas . @xmath276 here @xmath277 , and @xmath278 .
we also note @xmath279 ( mod 2 ) .
moreover , the value of the energy function is given by @xmath280 if we normalize @xmath14 in such a way as @xmath281 . here
@xmath282 is the first part of the partition corresponding to the weight of @xmath283 . solving the formulas for @xmath260 with respect to @xmath241 and @xmath242
, we obtain the coordinates of the image @xmath283 of the inverse of @xmath0 for an element @xmath284 of @xmath4 are given by @xmath285 here we should understand @xmath286 . in what follows in this section we prove theorem [ th : main ] by assuming technical propositions in later sections .
we give a proof only for type @xmath169 , since the difference from the other cases is very small as we have seen in proposition [ prop : to highest ] .
suppose we need to apply @xmath113 with such a word @xmath114 as @xmath287 for type @xmath169 ( see e.g. ) .
then for type @xmath168 we replace it with @xmath288 and for type @xmath167 @xmath289 consider first the case when @xmath20 is odd .
suppose @xmath290 and @xmath25 are related as in the statement of the theorem .
we are to show @xmath291 by proposition [ reduction_odd3 ] showing is reduced to the case where @xmath242 is of the form @xmath292 . applying this proposition again to this case
, it is then reduced to the case where @xmath293 , since there is no @xmath100 or @xmath96 in @xmath242 of the previous case .
hence proposition [ prop_special ] completes the proof of .
( notice that when @xmath294 for any odd @xmath41 and the other @xmath260 are all zero . ) using these propositions we can calculate @xmath14 as @xmath295 since @xmath296 .
the case when @xmath20 is even can be proven similarly by using propositions [ reduction_even3 ] and [ prop_special ] .
let @xmath297 be a dominant integral weight whose corresponding young diagram is depicted as follows .
( 25,11.5)(-1,0 ) ( -2.3,5)@xmath298 ( 0,0)(1,0)25 ( 0,0)(0,1)10 ( 0,10)(1,0)5 ( 5,10)(0,-1)2 ( 5,8)(1,0)5 ( 10,8)(0,-1)1 ( 10.3,7)(0.5,-0.1)20 ( 20,4)(0,1)1 ( 25,4)(-1,0)5 ( 25,0)(0,1)4 ( 2.3,4.8)@xmath299 ( 2.5,6)(0,1)4 ( 2.5,4)(0,-1)4 ( 7.3,3.8)@xmath300 ( 7.5,5)(0,1)3 ( 7.5,3)(0,-1)3 ( 22.3,1.8)@xmath301 ( 22.5,3)(0,1)1 ( 22.5,1)(0,-1)1 ( 0,10)(1,10.6)(1.8,10.7 ) ( 5,10)(4,10.6)(3.2,10.7 ) ( 2.1,10.6)@xmath302 ( 5,8)(6,8.6)(6.8,8.7 ) ( 10,8)(9,8.6)(8.2,8.7 ) ( 7.1,8.6)@xmath303 ( 20,4)(21,4.6)(21.8,4.7 ) ( 25,4)(24,4.6)(23.2,4.7 ) ( 22.0,4.6)@xmath304 here one can assume @xmath305 and @xmath306 .
we also assume that @xmath307 then the claim of this section is the following special version of the main theorem : [ prop_special ] let @xmath241 be as above
. then we have @xmath308 the condition ( [ s : eq : sumc_i = k ] ) for @xmath309 is necessary .
for example , in type @xmath310 , the image of the combinatorial @xmath0-matrix and the value of the energy function for @xmath311 are @xmath312 and @xmath313 .
we divide the proof of this proposition into three parts .
let us define two words that will be used in the proof .
@xmath314 where underbraces are introduced to show a unit of repetitions , and @xmath315 are defined as follows .
set @xmath316 , then @xmath317 where @xmath318 and for @xmath319 , @xmath320 @xmath321 is defined by @xmath322 and @xmath323 where for @xmath324 , @xmath325 and @xmath326 is @xmath327 finally , @xmath328 in the process of proof , we use a dominant integral weight @xmath329 given by @xmath330 .
we assume that @xmath304 is even .
the proof for odd @xmath304 is similar . during the proof
, we often identify a kn tableau with its japanese reading word .
the goal of this subsection is the following lemma : [ lem_special1 ] @xmath331 this is a direct consequence of the following three sublemmas .
@xmath332 [ lem : special1 ] @xmath333 to begin with we apply @xmath180 on @xmath334 for maximal times .
define a word @xmath335 by @xmath336 then the @xmath76-highest element of @xmath334 is @xmath337 by the map @xmath338 we get the corresponding @xmath3-diagram and @xmath189 acts on it as follows : ( 31.2,8 ) ( 0,0)(18,0)2 ( 0,0)(1,0)13.2 ( 0,0)(0,1)8 ( 0,8)(1,0)3.3 ( 3.3,8)(0,-1)2 ( 3.3,6)(1,0)3.3 ( 6.6,6)(0,-1)2 ( 6.6,4)(1,0)3.3 ( 13.2,2)(0,-1)2 ( 9.9,4)(0,-1)2 ( 9.9,2)(1,0)3.3 ( 27.9,4)(1,0)3.3 ( 31.2,4)(0,-1)2 ( 0.15,7.2)@xmath339 ( 3.45,5.2)@xmath339 ( 6.75,3.2)@xmath339 ( 14,3)@xmath340 ( 18.15,7.2)@xmath341 ( 21.35,5.2)@xmath341 ( 24.65,3.2)@xmath341 ( 28.1,3.2)@xmath341 ( 28.1,2.2)@xmath339 note that there are @xmath304 @xmath46 s at height @xmath342 of the right @xmath3-diagram .
assume that @xmath304 satisfies @xmath343 .
then @xmath344 is @xmath345 and @xmath346 is @xmath347 from this expression , we get @xmath348 . applying @xmath349
we get @xmath350 to convert the action of @xmath351 into that of @xmath352 , we need to define the words @xmath353 as follows .
@xmath354 where the subwords @xmath355 are @xmath356 define @xmath357 and @xmath358 where the subwords @xmath359 are @xmath360 then , starting from @xmath361 , one calculates @xmath362 here @xmath363 means @xmath364 . with @xmath338 ,
this corresponds to the following @xmath3-diagram and @xmath189 acts on it as follows : ( 25,8 ) ( 0,0)(15,0)2 ( 0,0)(1,0)9.9 ( 0,0)(0,1)7 ( 0,7)(1,0)3.3 ( 3.3,7)(0,-1)2 ( 3.3,5)(1,0)3.3 ( 6.6,5)(0,-1)2 ( 6.6,3)(1,0)3.3 ( 9.9,3)(0,-1)3 ( 0.1,6.3)@xmath339 ( 3.4,4.3)@xmath339 ( 6.7,2.3)@xmath339 ( 10.9,3)@xmath340 ( 15.1,6.3)@xmath341 ( 18.4,4.3)@xmath341 ( 21.7,2.3)@xmath341 thus , starting from @xmath365 , we calculate @xmath366 where the final formula gives @xmath367 . from @xmath368 and @xmath369 , we see that the 0-signature of @xmath370 is @xmath371
. therefore we get @xmath372 which gives the desired expression .
[ lem : special2 ] starting from @xmath373 , we have @xmath374 where the final expression is equal to @xmath375 . since @xmath376
, we have finished the proof of lemma [ lem_special1 ] .
the goal of this subsection is the following lemma : [ lem_special2 ] @xmath377 to begin with , we have @xmath378 here , we need to divide the calculation into two cases .
+ _ case 1 : _ if @xmath379 , we have @xmath380 _ case 2 : _ if @xmath381 , we have @xmath382 to begin with we remark that in this case we have @xmath383 assume that @xmath210 satisfies @xmath384 .
then @xmath385 define a word @xmath386 by @xmath387 .
then we have @xmath388 by the map @xmath338 , @xmath389 corresponds to the following @xmath3-diagram , @xmath189 acts on it as follows : ( 35,10 ) ( 0,1)(20,0)2 ( 0,0)(1,0)15 ( 0,0)(0,1)8 ( 0,8)(1,0)3.3 ( 3.3,8)(0,-1)2 ( 3.3,6)(1,0)3.3 ( 11.6,4)(1,0)3.4 ( 15,4)(0,-1)4 ( 7.7,-1)@xmath390 ( 13.2,-1)@xmath210 ( 6.6,-1)(0,0.4)18(0,1)0.2 ( 11.6,-1)(0,0.4)13(0,1)0.2 ( 15,-1)(0,0.4)5(0,1)0.2 ( 6.6,7)(0,-1)2 ( 6.6,5)(1,0)5 ( 26.6,7)(1,0)5 ( 31.6,7)(0,-1)2 ( 0.15,8.3)@xmath339 ( 3.45,6.3)@xmath339 ( 11.85,4.3)@xmath339 ( 15.8,4)@xmath391 ( 20.1,8.3)@xmath341 ( 23.4,6.3)@xmath341 ( 26.95,5.3)@xmath392 ( 26.95,6.3)@xmath393 ( 31.8,4.3)@xmath341 there are @xmath394 @xmath46 s at height @xmath342 in the right @xmath3-diagram .
assume that @xmath394 satisfies @xmath395 .
we also assume that this @xmath41 satisfies @xmath396 for the sake of simplicity .
then @xmath397 is @xmath398 and @xmath399 is @xmath400 from this expression , we have @xmath401 . applying @xmath402 we get @xmath403 note that the length of the string @xmath404 is @xmath301 whereas that of the string @xmath405 is @xmath406 . in order to convert the action of @xmath351 into that of @xmath352 , we define the word @xmath407 as follows .
@xmath408 , @xmath409 where subwords are defined by @xmath410 @xmath411 , @xmath412 where subwords are defined by @xmath413 and @xmath414 .
computation of @xmath415 proceeds as follows : @xmath416 by @xmath338 , this corresponds to the following @xmath3-diagram , and @xmath189 acts on it as follows : ( 35,11 ) ( 0,1)(20,0)2 ( 0,0)(1,0)15 ( 0,0)(0,1)9 ( 0,9)(1,0)3.3 ( 3.3,9)(0,-1)2 ( 3.3,7)(1,0)3.3 ( 6.6,7)(0,-1)2 ( 6.6,5)(1,0)5 ( 15,3)(0,-1)3 ( 7.7,-1)@xmath390 ( 13.1,-1)@xmath210 ( 6.6,-1)(0,0.4)15(0,1)0.2 ( 11.6,-1)(0,0.4)15(0,1)0.2 ( 15,-1)(0,0.4)5(0,1)0.2 ( 11.6,6)(0,-1)2 ( 11.6,4)(1,0)3.4 ( 35,6)(0,-1)2 ( 31.6,6)(1,0)3.4 ( 0.2,9.3)@xmath339 ( 3.5,7.3)@xmath339 ( 7.0,5.3)@xmath392 ( 15.8,4)@xmath391 ( 20.1,9.3)@xmath341 ( 23.5,7.3)@xmath341 ( 27.0,5.3)@xmath393 ( 31.9,5.3)@xmath341 ( 31.9,4.3)@xmath339 let us assume that @xmath417
. then the right @xmath3-diagram corresponds to the expression ( [ eq : w_3 ] ) with @xmath418 and @xmath419 in ( [ eq : w_3 ] ) being replaced with @xmath420 and @xmath421 .
application of @xmath422 is similar to that of @xmath423 on ( [ eq : w_3 ] ) and we obtain @xmath424 as @xmath425 the remaining computation of @xmath426 is almost the same as the computation of @xmath427 given in the final part of the proof of lemma [ lem : special1 ] .
the only difference in @xmath428 is caused by the fact that letters @xmath184 and @xmath429 appear @xmath210 times in @xmath424 .
as for @xmath430 , the beginning two steps @xmath431 gives @xmath432 by comparing this with @xmath433 , we see that the rest of the computation of @xmath430 is almost the same as that given in lemma [ lem : special2 ] .
this completes the proof for case 1 .
note that in this case we have @xmath434 .
action of @xmath352 is obtained by formally setting @xmath435 in case 1 .
therefore we have @xmath436 where @xmath437 is given in ( [ eq : special3 ] ) . when we further apply @xmath430 on this formula
, we realize that there are extra exponents originating from @xmath438 in the first tensor component of the right hand side .
these extra contributions coincide with the exponents @xmath439 in @xmath440 .
we have completed the proof of lemma [ lem_special2 ] .
now we can prove proposition [ prop_special ] .
we prove the first relation by descending induction on @xmath301 .
if @xmath441 ( this is the maximal possible value ) , we have @xmath442 .
in this case we see @xmath443 by weight consideration .
the induction proceeds by using lemmas [ lem_special1 ] and [ lem_special2 ] . as for the energy function
, we have to look carefully the action of @xmath180 in lemmas [ lem_special1 ] and [ lem_special2 ] . if @xmath180 acts on the second component of the tensor product , we write @xmath0 , and @xmath444 on the first component .
we summarize actions of @xmath180 to get @xmath375 and @xmath445 in two lemmas as follows ( proceeds from left to right ) : @xmath446 the diagram is drawn in the case of @xmath447 . including the other inequality case
, we see that we have exactly the same number of @xmath448 and @xmath449 cases ( see ) .
therefore we have @xmath450 . using the same induction as above we obtain @xmath451 .
this completes the proof of proposition [ prop_special ] .
let @xmath453 be @xmath228-highest . recall that we defined @xmath454 by @xmath240 .
( readers are warned that it is not the multiplicity of @xmath41 in the corresponding partition @xmath241 but its conjugate @xmath455 . )
note that @xmath456 unless @xmath457 and @xmath41 is odd .
we also know that the coordinates other than @xmath458 are all @xmath133 by proposition [ prop : ht cond 1 ] .
let us define a word @xmath459 by @xmath460 where the exponents are defined as follows . for @xmath461 , @xmath462 define @xmath463
. then @xmath464 . for @xmath465
, @xmath466 set @xmath467 and define other @xmath468 by @xmath469 for @xmath470 .
note that @xmath471 .
set @xmath472 , @xmath473 and define other @xmath474 by @xmath475 for @xmath476 .
since @xmath482 is the @xmath3-diagram of outer shape @xmath241 such that all the columns have @xmath46 as symbol , we see @xmath483 .
thus one has @xmath484 .
we have @xmath485 . to convert the result into that for @xmath486 we define a word @xmath487 as follows : @xmath488 where @xmath489 , @xmath490 , @xmath491
. then we have @xmath492 applying @xmath493 further , we obtain @xmath494 finally , applying @xmath495 we obtain the desired relation .
let us consider the operation of @xmath502 in @xmath503 .
since @xmath504 , @xmath505 acts on the second component at most @xmath506 times and the rest goes to the first .
the 2-signature of the second component of ( [ tochu6_2 ] ) is @xmath507 .
from the highest condition for @xmath508 we have @xmath509 , thus @xmath505 acts on @xmath510 only . we can continue similarly and obtain the desired result .
finally , we consider the action of @xmath511 .
the 1-signature of eq.@xmath501 is @xmath512 . by applying @xmath513 ,
we get @xmath514 the 2-signature of the above element is @xmath515 . from the highest condition for @xmath508 we have @xmath516 ,
thus @xmath505 does not act on @xmath517 .
therefore @xmath518 acts on the first component and obtain @xmath519 we can continue the computation and arrive at proposition [ reduction_odd ] . in this subsection ,
let @xmath25 and @xmath144 be the @xmath3-diagrams .
as before , corresponding to @xmath25 and @xmath144 , we use the parametrization @xmath260 and @xmath521 ( @xmath522 ) respectively . note that by definition @xmath523 . define a word @xmath524 by @xmath525 here
the exponents for @xmath526 are @xmath527 the exponents for @xmath528 are @xmath529 we use proposition [ prop : e1 action ] .
schematically , the pair of @xmath3-diagrams corresponding to @xmath198 looks as follows : @xmath541{0,0,0,0.2 } \put(2,9){\rule{20pt}{20pt } } \put(6,8){\rule{20pt}{20pt } } \put(13,5){\rule{20pt}{20pt } } \put(17,4){\rule{20pt}{20pt } } \put(21,3){\rule{20pt}{20pt } } \put(28,0){\rule{20pt}{20pt } } \color{black } \thicklines \put(0,0){\line(1,0){36 } } \put(0,0){\line(0,1){11 } } \put(0,11){\line(1,0){8 } } \put(8,11){\line(0,-1){2 } } \put(8,9){\line(1,0){2 } } \multiput(10,9)(0.3,-0.1){10}{\circle*{0.1 } } \put(13,8){\line(1,0){2 } } \put(15,8){\line(0,-1){2 } } \put(15,6){\line(1,0){8 } } \put(23,6){\line(0,-1){2 } } \put(23,4){\line(1,0){2 } } \multiput(25,4)(0.3,-0.1){10}{\circle*{0.1 } } \put(28,3){\line(1,0){2 } } \put(30,3){\line(0,-1){2 } } \put(30,1){\line(1,0){6 } } \put(36,1){\line(0,-1){1 } } \thinlines \put(2,11){\line(0,-1){1 } } \put(2,10){\line(1,0){4 } } \put(6,10){\line(0,-1){1 } } \put(6,9){\line(1,0){2 } } \put(13,6){\line(1,0){2 } } \put(17,6){\line(0,-1){1 } } \put(17,5){\line(1,0){4 } } \put(21,5){\line(0,-1){1 } } \put(21,4){\line(1,0){2 } } \put(28,1){\line(1,0){2 } } \put(32,1){\line(0,-1){1 } } \put(0.2,10.3){$+$ } \put(1.2,10.3){$+$ } \put(2.2,10.3){$+$ } \put(3.2,10.3){$+$ } \put(4.2,10.3){$-$ } \put(5.2,10.3){$-$ } \put(6.2,10.3){$-$ } \put(7.1,10.3){$-$ } \put(2.2,9.3){$+$ } \put(3.2,9.3){$+$ } \put(4.2,9.3){$+$ } \put(5.2,9.3){$+$ } \put(6.2,9.3){$+$ } \put(7.1,9.3){$+$ } \put(6.2,8.3){$+$ } \put(7.1,8.3){$+$ } \put(8.2,8.3){$+$ } \put(9.2,8.3){$+$ } \put(13.2,7.3){$-$ } \put(14.1,7.3){$-$ } \put(13.2,6.3){$+$ } \put(14.1,6.3){$+$ } \put(13.2,5.3){$+$ } \put(14.1,5.3){$+$ } \put(15.2,5.3){$+$ } \put(16.2,5.3){$+$ } \put(17.2,5.3){$+$ } \put(18.2,5.3){$+$ } \put(19.2,5.3){$-$ } \put(20.2,5.3){$-$ } \put(21.2,5.3){$-$ } \put(22.2,5.3){$-$ } \put(17.2,4.3){$+$ } \put(18.2,4.3){$+$ } \put(19.2,4.3){$+$ } \put(20.2,4.3){$+$ } \put(21.2,4.3){$+$ } \put(22.2,4.3){$+$ } \put(21.2,3.3){$+$ } \put(22.2,3.3){$+$ } \put(23.2,3.3){$+$ } \put(24.2,3.3){$+$ } \put(28.2,2.3){$-$ } \put(29.2,2.3){$-$ } \put(28.2,1.3){$+$ } \put(29.2,1.3){$+$ } \put(28.2,0.3){$+$ } \put(29.2,0.3){$+$ } \put(30.2,0.3){$+$ } \put(31.2,0.3){$+$ } \put(32.2,0.3){$+$ } \put(33.2,0.3){$+$ } \put(34.2,0.3){$-$ } \put(35.2,0.3){$-$ } \put(0.5,11.6){$p_r^\cdot$ } \put(2.5,11.6){$p_r^-$ } \put(4.5,11.6){$p_r^+$ } \put(6.0,11.6){$p_{r-2}^\cdot$ } \put(8.5,9.6){$p_{r}^\mp$ } \put(13.5,8.6){$p_i^\cdot$ } \put(15.2,6.6){$p_{i+2}^\mp$ } \put(17.5,6.6){$p_i^-$ } \put(19.5,6.6){$p_i^+$ } \put(21.1,6.6){$p_{i-2}^\cdot$ } \put(23.5,4.6){$p_i^\mp$ } \put(28.6,3.6){$p_1^\cdot$ } \put(30.5,1.6){$p_3^\mp$ } \put(32.6,1.6){$p_1 ^ -$ } \put(34.6,1.6){$p_1^+$ } \end{picture } \nonumber\end{aligned}\ ] ] here the thick lines represent outer shape of @xmath198 and the thin lines represent the inner @xmath3-diagram .
( since we are to consider the @xmath162 action , we need such a pair of @xmath3-diagrams . )
the numbers @xmath260 represent the numbers of columns which have the same pattern of @xmath46 and @xmath45 indicated below @xmath260 . according to proposition [ prop : e1 action ] , we make pairs of two @xmath46 symbols which we indicate by gray squares in the diagram
. then we see that we can apply @xmath162 up to @xmath542 times , which gives the value for @xmath543 .
the pair of @xmath3-diagrams corresponding to @xmath544 looks as follows : @xmath545 note that the numbers of columns of height 1 have changed from @xmath546 , @xmath547 , @xmath548 to @xmath548 , @xmath546 , @xmath547 . in order to compute @xmath549
, we usually make @xmath544 into @xmath550-highest by applying suitable @xmath113 , apply @xmath189 and then apply @xmath551 ( see ) .
however , since @xmath189 commutes with the action of @xmath12 @xmath552 , we can apply @xmath189 on the pair of @xmath3-diagrams directly .
namely , @xmath189 changes the outer @xmath3-diagram only .
the pair of @xmath3-diagrams corresponding to @xmath553 looks as follows : @xmath554 note that the outer shape has also been changed at @xmath548 .
we use proposition [ prop : to highest ] .
the quantities @xmath118 , @xmath119 and @xmath120 there should be used for the corresponding numbers of the inner @xmath3-diagram of ( [ pair+-diagram ] ) .
since we are considering the inner @xmath3-diagram , we have to understand the word @xmath114 there as follows : @xmath556 and the formula for @xmath557 and @xmath558 are the same in terms of @xmath118 , @xmath119 and @xmath120 . then
, @xmath559 and differences @xmath560 are @xmath561 and differences @xmath562 are @xmath563 we see that the word @xmath114 computed here coincides with @xmath526 .
again , we use proposition [ prop : to highest ] . in this case , the quantities @xmath118 , @xmath119 and @xmath120 there mean those for the outer @xmath3-diagram of ( [ pair+-diagram ] ) .
let us compute the word @xmath565 there in the case of our ( [ pair+-diagram ] ) . to begin with , @xmath566 is @xmath567 and differences @xmath560 are @xmath568 and differences @xmath562 are @xmath569 we see that the word @xmath114 computed here coincides with @xmath528 except for @xmath570 which does not appear in @xmath528 .
let @xmath571 be the @xmath228-highest weight element whose outer shape coincides with ( [ pair+-diagram ] ) .
then the above lemma shows that @xmath572 .
since there are exactly @xmath573 columns of height 1 in @xmath571 , we see that the content of columns of height 1 in the tableau @xmath574 are all 2 and that the other columns are the same as @xmath571 . from the shape of ( [ pair+-diagram ] )
we see that @xmath574 coincides with @xmath144 given in proposition [ reduction_odd2 ] .
to summarize , we have @xmath575 , hence we complete the proof of proposition [ reduction_odd2 ] .
\(i ) we have @xmath586 where we have used @xmath587 in the final line .
thus @xmath588 .
+ ( ii ) to begin with let us show @xmath589 .
we compute @xmath590 thus @xmath591 , which shows the coincidence of the first letters of @xmath592 and @xmath593 . as for the other @xmath594 and @xmath595 , note that @xmath596 when @xmath41 is odd , we see @xmath597 . when @xmath41 is even , we have @xmath598 and thus we have @xmath597 , i.e. , @xmath599 for all @xmath41 .
we have @xmath600 , i.e. , @xmath601 .
similarly , we have @xmath602 . as for other @xmath603 and @xmath604
, we have @xmath605 thus we have @xmath606 for all @xmath41 and obtain @xmath589
. similarly we can show @xmath607 .
we compute @xmath608 thus @xmath609 , i.e. , the coincidence of the first letters of @xmath592 and @xmath593 .
next , we have @xmath610 . on the other hand
, we have @xmath467 and @xmath611 , thus @xmath612 , i.e. , @xmath613 .
similarly , we can recursively show @xmath614 for all @xmath41 , @xmath615 , @xmath616 for all @xmath41 .
so we have @xmath607 , and therefore we get the final result @xmath617 . + ( iii ) the 0-signature of @xmath508 is @xmath618 for some @xmath619 and that of @xmath579 is @xmath620 for some @xmath621 . here
we divide into two cases .
let us first assume @xmath622 .
then the actions of @xmath180 on two tensor products look as follows ( proceed from left to right ) : @xmath623 thus we have @xmath624 ( rr ) pairs and @xmath210 ( ll ) pairs .
therefore we have @xmath625 which gives the desired relation .
next assume @xmath626 .
then we have @xmath627 ( rr ) pairs and @xmath628 ( ll ) pairs and again we obtain @xmath625 .
let @xmath453 be @xmath228-highest and @xmath240 .
@xmath456 unless @xmath457 and @xmath41 is even .
we also know that the coordinates other than @xmath629 are @xmath133 by proposition [ prop : ht cond 1 ] .
let us set @xmath630 throughout this subsection .
let us define a word @xmath631 by @xmath632 where the exponents are defined as follows .
for @xmath633 , @xmath634 define @xmath635 .
then @xmath464 . for @xmath636 , @xmath637 set @xmath638 and define other @xmath468 by @xmath639 for @xmath640 .
note that @xmath471 .
set @xmath472 and define other @xmath474 by @xmath641 for @xmath642 . in this subsection ,
let @xmath25 and @xmath144 be the @xmath3-diagrams .
as before , corresponding to @xmath25 and @xmath144 , we use the parametrization @xmath260 and @xmath521 ( @xmath522 ) respectively . define a word @xmath649 by @xmath650 here
the exponents for @xmath651 are @xmath652 the exponents for @xmath653 are @xmath654 s. v. kerov , a. n. kirillov , n. yu .
reshetikhin , _ combinatorics , the bethe ansatz and representations of the symmetric group _ , zap.nauchn .
( lomi ) * 155 * ( 1986 ) 5064 ( english translation : j. sov . math .
* 41 * ( 1988 ) 916924 ) . | we calculate the image of the combinatorial @xmath0-matrix for any classical highest weight element in the tensor product of kirillov reshetikhin crystals @xmath1 of type @xmath2 .
the notion of @xmath3-diagrams is effectively used for the identification of classical highest weight elements in @xmath4 . |
the recognised convention for representing the overall galaxy population is the luminosity distribution or space density of galaxies ( see review by binggeli , sandage & tammann 1988 and more recently @xcite ) .
this distribution is typically derived from a magnitude limited redshift survey which is corrected for malmquist bias ( e.g. , efstathiou , ellis & peterson 1988 ) and fitted with the three parameter schechter function @xcite .
generally , the schechter function is found to be a formally good representation , although some surveys ( e.g. , @xcite ; @xcite ) have hinted at an upturn at the faintest limits , where the statistics become poor ( typically @xmath25 mag , see also @xcite ) .
the process of fitting a luminosity function ( lf ) reduces the galaxy population to three crucial numbers : the characteristic luminosity , @xmath10 ( or @xmath26 ) , the normalisation , @xmath27 , and the faint - end slope , @xmath28 . since the pioneering work of @xcite numerous measurements of these three crucial parameters have been made ( see @xcite ) . among the most recent are those derived from the two - degree field galaxy redshift survey ( 2dfgrs ; @xcite ) and the sloan digital sky survey ( sdss1&2 ; @xcite ; 2003a ) .
however , although individual surveys yield schechter function parameters to high accuracy , a comparison between independently published values shows excessively large variations , indicative of strong systematic errors ( see @xcite ; @xcite ; @xcite ) . @xcite
discuss the impact of surface brightness selection effects on measurements of the lf ( see also @xcite ; @xcite ) and find that such effects _ may _ provide an explanation for the variation seen .
in @xcite we compared a number of recent @xmath24-band schechter function values and used the precision number counts of the millennium galaxy catalogue with a uniform isophotal detection limit of @xmath29 mag arcsec@xmath1 in the @xmath24-band to revise the normalisation parameters .
this resulted in a reasonably consistent global schechter function with : @xmath30 mag , @xmath31 mpc@xmath7 and @xmath32 ( taken from @xcite ) . despite the apparent convergence brought about by @xcite there are two strong reasons to now go beyond the monovariate schechter function : \(1 ) numerical and semi - analytic simulations are sufficiently developed that it is now possible to predict not just the luminosity distribution but also the size distribution ( see e.g. @xcite ; buchalter , jimenez & kamionkowski 2001 ; @xcite ) and hence the _ bivariate _ distribution of luminosity and size ( or luminosity and surface brightness , see e.g. @xcite ; @xcite ; @xcite ; @xcite ) .
this bivariate brightness distribution ( bbd ) can be fitted by a 6-parameter chooniewski function @xcite essentially a schechter function with a luminosity dependent gaussian distribution in surface brightness ( see e.g. @xcite ; @xcite ) or compared directly without confinement by any functional form .
either way , a multivariate distribution provides a more stringent constraint than the monovariate schechter function ( see also @xcite ) ; \(2 ) @xcite essentially circumvent the surface brightness issues raised by @xcite by allowing the normalisation of the lfs to accommodate the missing flux from low surface brightness galaxies .
this approximation succeeds because the galaxies that dominate the bright galaxy number counts are all around the @xmath26 value , where the surface brightness distribution appears well defined ( see @xcite , @xcite ) .
we also know from studies of giant ellipticals ( @xcite ; @xcite ) , disk galaxies ( @xcite ) , dwarf galaxies in the local group ( @xcite ) , rich clusters ( @xcite ; @xcite ) and from early work on the global field bbd ( @xcite ; @xcite ; @xcite ; @xcite ) that there exists a luminosity surface brightness relation ( see also @xcite for the near - ir relation and a discussion of its variation with wavelength ) .
together these imply that the impact of surface brightness selection effects will be a function of luminosity . therefore recovering the full luminosity distribution
can only be done by simultaneously considering luminosity _ and _ surface brightness
a conclusion reached many times over the past half - century ( @xcite ; @xcite ; @xcite ) .
an additional motivation is that given the size of contemporary datasets , it becomes logical to start to explore multivariate distributions to quantify trends such as the luminosity - surface brightness relation @xcite or the colour - luminosity relation ( e.g. , @xcite ; see also @xcite ) . here
we reconstruct the bivariate luminosity - surface brightness distribution with careful attention to the impact of surface brightness and size selection effects . in section [ data ]
we describe the imaging and redshift data . in section [ cosmoke ]
we describe the conversion of our catalogue from apparent to absolute properties , the various selection boundaries and our methodology for deriving the joint luminosity - surface brightness distribution .
we apply it to our data in section [ mgcbbd ] and compare our results to previous estimates of the bbd ( @xcite ; @xcite ) , the lf ( @xcite ; @xcite ; @xcite ; @xcite ) and the surface brightness or size distribution ( @xcite ; @xcite ) .
we summarise our key results in section [ conclusions ] . throughout
we assume @xmath33 as defined in section [ cosmology ] .
the millennium galaxy catalogue ( mgc ) is a deep ( @xmath34 mag arcsec@xmath1 ) , wide - field ( @xmath35 deg@xmath36 ) @xmath24-band imaging survey obtained with the wide field camera on the 2.5-m isaac newton telescope .
the survey region is a long , 35-arcmin wide strip along the equator , covering from @xmath37 to @xmath38 ( j2000 ) , and is fully contained within the regions of both the 2dfgrs and sdss data release 1 ( dr1 ; @xcite ) .
in @xcite we gave a detailed description of the observations , reduction , object detection , galactic extinction correction and classification ( using sextractor ; @xcite ) and presented precision galaxy number counts over the range @xmath39 mag . all non - stellar sources to @xmath40 mag were visually inspected and if an object was found to be incorrectly deblended or if the object s parameters were obviously wrong , the object was re - extracted by manually changing the sextractor extraction parameters until a satisfactory result was achieved . in addition , all low - quality regions in the survey ( e.g. near ccd defects ) were carefully masked out ( see @xcite and @xcite ) .
we demonstrated that the internal photometric accuracy of the mgc is @xmath41 mag and that the astrometric accuracy is @xmath42 arcsec .
in @xcite we then compared the photometric accuracy , completeness and contamination of the mgc to the 2dfgrs and sdss - edr and dr1 datasets .
being deeper and of higher resolution , we found that the mgc is both more accurate and more complete than either of these surveys .
the mgc is currently the highest quality and most complete representation of the nearby galaxy population .
[ mgcexamples ] shows images of three galaxies from the digital sky survey ( left ) , the sdss ( centre ) and the mgc ( right ) , illustrating the superior resolution and greater depth of the mgc data .
the mgc - bright catalogue of @xcite was bounded at the bright end by @xmath43 mag , where stars begin to flood in our best seeing fields .
we now extend this catalogue brightwards to @xmath44 mag by carefully examining all non - stellar objects in the range @xmath45 mag , repairing objects that had been over - deblended by sextractor and checking completeness with ned at @xmath46 mag , where sextractor
optimised for faint galaxy detection and photometry may occasionally miss very large objects .
a second change to mgc - bright with respect to @xcite is the deletion of some exclusion regions .
liske et al .
had placed exclusion regions around very bright objects ( @xmath47 mag ) to avoid spurious halo detections and problems with background estimation in their vicinity .
@xmath48 such regions had been erroneously placed around galaxies fainter than @xmath49 mag and these have been deleted , resulting in an increase of the good - quality survey area from @xmath50 to @xmath35 deg@xmath36 .
further minor changes were introduced to mgc - bright , each affecting @xmath51tens of objects : different deblending decisions ; spectroscopic identification of galaxies previously misclassified as stars ( compact emission line galaxies ) and vice versa ( stars which appear extended because of almost perfect alignment with small , faint background galaxies ) ; and identification of asteroids previously misclassified as galaxies ( discovered by comparison with sdss - dr1 imaging data ) .
the final mgc - bright sample now comprises @xmath2 galaxies . by design the mgc survey region
is fully contained within the 2dfgrs and sdss survey regions .
hence we first turned to these surveys in order to obtain redshifts for mgc - bright .
in addition we obtained redshifts from the 2qz @xcite , ned ( excluding 2dfgrs and sdss objects ) and the smaller surveys of francis , nelson & cutri ( 2004 ) ( pf ) and @xcite ( lsbg ) . in turn each survey , @xmath52 , was matched against mgc galaxies and stars with @xmath53 mag .
this magnitude limit and the following matching procedure were chosen in order to be able to account for every @xmath52 object and hence to verify the completeness of the mgc . objects with multiple redshift measurements .
the distribution has an rms dispersion of @xmath54 km s@xmath55 .
the solid line is a gaussian with the same median and rms as the data . ]
first we found all pairs of @xmath52 and mgc objects where each is identified as the other s nearest neighbour in that catalogue .
a pair with separation @xmath56 smaller than some @xmath57 is classified as a match unless one or more other mgc objects have also identified the @xmath52 object as their nearest neighbour in @xmath52 with a separation @xmath58 ( where @xmath56 refers to the separation of the closest pair ) .
these @xmath52 objects as well as objects with @xmath59 and those which have not been identified as a nearest neighbour by any mgc objects are classified as initially unmatched .
the value of @xmath60 is chosen as the separation where the invariably occurring tail in the distribution of separations begins to dominate . the above definition of initially unmatched objects is designed to conservatively flag all possible mismatches .
almost all mismatches fall into one of three categories , which are a priori difficult to distinguish from one another but for which one needs to proceed differently : ( i ) unusually large positional disagreement ; ( ii ) two ( or more ) objects in one catalogue where the other catalogue only has one object , because of different deblending choices or different resolution ; ( iii ) one object in @xmath52 but none in mgc because of inaccurate photometry in either @xmath52 or mgc , or because the @xmath52 object is an asteroid or a spurious noise detection , or because the object is genuinely missing from the mgc .
all initially unmatched objects are visually inspected in order to reliably identify those objects that were in fact correctly matched .
the few cases where the mismatch was due to poor object extraction in the mgc were fixed .
each redshift of a matched object is assigned a redshift quality , @xmath61 , which has the same meaning as in the 2dfgrs @xcite : @xmath62 means no redshift could be measured ; @xmath63 means possible but doubtful redshift ; @xmath64 means probable redshift ; and @xmath65 or @xmath66 means good or very good redshift . @xmath62 and @xmath67 are considered failures and we will use only redshifts with @xmath68 in this paper .
in addition to the 2dfgrs this system is also already used by pf .
we translate the various quality parameters of the other surveys to our system in the following manner : sdss - dr1 redshifts are assigned @xmath65 unless the sdss zstatus or zwarning flags indicate a serious problem ; for the 2qz we use @xmath69 ; and ned and lsbg redshifts are assigned @xmath64 . in total the public data provided good quality redshifts for @xmath70 mgc galaxies . in order to complete the redshifts for the full mgc - bright sample we conducted our own redshift survey ( which we label mgcz ) , targeting those galaxies which did not already have publicly available redshifts .
these observation were mainly carried out with the two degree field ( 2df ) facility on the anglo australian telescope .
the data were collected , reduced and redshifted in an identical manner to that of the 2dfgrs ( see @xcite for full details ) . in particular
we again used the same redshift quality parameter , @xmath61 , as the 2dfgrs ( see above ) .
the redshift incompleteness of the 2df data showed a clear bias against low - surface brightness galaxies .
hence we undertook additional single - object long - slit observations , targeting low - surface brightness galaxies ( @xmath71 mag arcsec@xmath1 , where @xmath72 is the apparent effective surface brightness within the half - light radius , see section [ mgcsb ] ) as well as gaps in our 2df coverage , with the double beam spectrograph ( dbs ) on the australian national university s @xmath73-m , the low resolution spectrograph ( lrs ) on the @xmath74-m telescopio nazionale galileo on la palma , the eso multi - mode instrument ( emmi ) on the new technology telescope in la silla and the gemini multiple object spectrograph ( gmos , long - slit nod & shuffle mode ) on gemini north . in all cases we observed between @xmath75 and @xmath76 s through a @xmath77 to @xmath67-arcsec wide slit , aligning the slit with the major axis of the object to collect as much flux as possible .
all the data were reduced in a similar manner using standard iraf procedures .
the resulting spectra are of similar resolution as our 2df data , but of higher quality : they increased our redshift completeness at @xmath78 mag and @xmath71 mag arcsec@xmath1 from @xmath79 to @xmath80 per cent .
using all available data and mgcz data qthe overall redshift completeness of mgc - bright is @xmath81 per cent . in fig .
[ completeness ] ( third row from top ) we show the final incompleteness as a function of apparent effective surface brightness and apparent magnitude , ( @xmath82 ) and ( @xmath83 ) model colours . in table
[ alldata ] we list the numbers of redshifts contributed by each of the surveys . for @xmath84 galaxies we have more than one @xmath85 redshift measurement , totaling @xmath86 redshifts .
these duplications are mainly due to overlaps between the 2dfgrs and sdss - dr1 , but @xmath87 of these galaxies have at least one mgcz redshift . for these objects
we pick a ` best ' redshift by sorting first by @xmath61 , then by signal - to - noise ratio ( where available ) and then by the order of the surveys in table [ alldata ] . fig .
[ multiplezs ] shows the distribution of differences between the multiple redshifts which has an rms of @xmath54 km s@xmath55 .
to calculate the mgc bbd we first need to adopt a cosmological framework , appropriate @xmath5-corrections [ @xmath88 and an evolutionary model [ @xmath89 .
it is of course possible to determine a lf while simultaneously solving for one or more of these quantities
see e.g. @xcite , who derive a lf while simultaneously solving for both the number - density evolution and the luminosity evolution . however , our concern with this approach is that it can lead to erroneous results _ if _ the incompleteness is significant and biased ( as indicated in @xcite and fig .
[ completeness ] ) . in this case
the evolutionary functions would be fitting not only evolution but also the uncorrected selection biases . instead , our philosophy is to adopt reasonable assumptions for the evolution and concentrate on a strategy to account for selection bias .
-corrections as defined in section [ kcorrections ] .
the solid curves show the bright and faint magnitude limits assuming the mean @xmath5-correction and a universal evolution of @xmath90 .
the dashed lines again show the magnitude limits but now assuming the bluest and reddest @xmath5-corrections used in the mgc .
the upper and lower redshift limits are marked as vertical lines and the @xmath91 value from norberg et al .
( 2002 ) is shown as a horizontal dotted line .
note that the lower redshift limit defines the minimum absolute magnitude of @xmath92 mag .
the locations of known abell clusters are indicated and clearly show up as redshift over - densities . ]
we use @xmath93 and @xmath94 broadly confirmed by the wmap mission combined with the 2dfgrs power spectrum but adopt @xmath95 km s@xmath55 mpc@xmath96 for ease of comparison with previous results .
we also define maximum and minimum redshift limits .
the maximum limit is chosen as the point at which red @xmath26-galaxies fall below our imposed @xmath40 mag limit ( this occurs at @xmath97 , see fig .
[ mz ] , excluding @xmath98 galaxies ) .
the lower limit is determined by the point at which the redshifts are significantly influenced by the local velocity field . from @xcite
we find that the pairwise velocity dispersion , as measured from the 2dfgrs , is @xmath99 km s@xmath55 . hence to ensure minimal impact from peculiar velocities ( i.e. , @xmath100 ) we must set @xmath101 ( excluding 44 galaxies ) . fig .
[ mz ] shows the absolute magnitude versus redshift distribution for our data along with the upper and lower redshift limits ( vertical lines ) and the typical value of @xmath91 ( @xcite ; dotted line ) .
note that the lower @xmath102 limit , combined with the faint apparent magnitude limit ( @xmath40 mag ) ultimately defines how far down the luminosity distribution one can probe ( @xmath103 mag ) .
one can only _ reliably _ push fainter by extending the faint apparent magnitude limit of the survey and not by covering a wider area ( unless direct distances are measured or the detailed local peculiar velocity field is modeled ) . in total
there are @xmath104 galaxies within these redshift limits .
-band @xmath5-correction . ]
@xmath5-corrections , or bandpass corrections , are galaxy specific ultimately galaxy component specific and given the known variation in galaxy types , a mean @xmath5-correction ( as adopted in @xcite and in most previous studies ) may be overly simplistic ( see also @xcite who reach a similar conclusion ) . here
we identify a best fit synthetic spectrum from the combined broad - band mgc and sdss - dr1 colours for each individual galaxy and then directly measure the @xmath5-correction from this optimal template .
we use the extinction - corrected mgc kron ( @xmath105 kron radii ) and extinction - corrected sdss - dr1 petrosian magnitudes for this purpose .
the petrosian magnitudes are corrected to the ab system using @xmath106 ; @xmath107 and @xmath108 as recommended on the sdss - dr2 website
. the @xmath109 magnitudes are adjusted for the known zeropoint offset of @xmath110 mag between the mgc and sdss - dr1 @xcite and converted from vega to ab .
this conversion was found to be @xmath111 for the mgc system .
-probability distribution from our @xmath112 fitting algorithm ( histogram ) as compared to the expected distribution ( for five degrees of freedom , solid line ) .
note that because we re - analyse galaxies with probabilities less than @xmath66 per cent no values with @xmath113 occur . ]
we now wish to derive the optimal spectral template match to these six filters in the ab system .
we elect to use the spectrum library of @xcite which includes ellipticals and early and late - type spirals , with ages ranging from @xmath114 gyr to @xmath115 gyr ( @xmath116 synthetic templates in all ) . to determine the best template we calculate the ab magnitudes as if observed through the @xmath109 and sdss @xmath117 filters for a single template and iteratively rescale the amplitude of the spectral template to minimise @xmath118 : @xmath119 ^ 2 $ ] where the errors are given by : @xmath120 and @xmath121 mag .
these error values were derived as follows .
firstly , using just the sdss filters , we determined the @xmath122-distribution for the optimal spectral fits based on the errors listed on the sdss website ( @xmath123 [ red leak ] , @xmath124 ) .
we then scaled the @xmath125 errors until the recovered @xmath126-distribution matched the expected distribution for four degrees of freedom .
we then repeated this process now including the @xmath109 filter ( see fig .
[ prob ] ) .
these final errors are in good agreement with the study of @xcite . the @xmath118-minimisation was repeated for each of the @xmath116 spectral templates and the smallest of the @xmath116 @xmath126 values used to identify the most appropriate template . finally if the @xmath126-probability for the optimal template was less than @xmath66 per cent we repeated the process , first rejecting the @xmath109 filter and then the @xmath127 and @xmath102 filters in turn ( hence the truncation at @xmath128 on fig .
[ prob ] ) .
magnitudes fainter than the quoted @xmath129 per cent completeness limit were not used ( @xmath130 , @xmath131 , @xmath132 , @xmath133 , @xmath134 mag , see @xcite ) .
[ kcorrexamples ] shows two example galaxies with the best - fit spectral templates overlaid on the measured magnitudes .
[ prob ] shows the resulting @xmath122 distribution ( histogram ) and the expected @xmath118-distribution ( for five degrees of freedom , solid line ) .
the mgc @xmath5-corrections trace out the 27 tracks of the synthetic spectra of poggianti ( 1997 ) .
the filled circles indicate the mean of our data and the vertical dashed lines show the adopted redshift limits used in the bbd analysis . ] shown in fig .
[ kcorr ] are the resulting @xmath5-corrections for the mgc , which trace out the @xmath116 different spectral templates , as well as the mean @xmath5-correction in @xmath135 redshift intervals ( large dots ) . finally , in the few cases ( @xmath136 galaxies in total ) where a galaxy did not have an sdss - dr1 match we assigned the mean @xmath5-correction value at the galaxy s redshift .
we adopt a global evolution of the form @xmath137 with an initial value of @xmath138 ( i.e. , @xmath139 , see @xcite ) .
given the redshift limits this implies an evolutionary correction in the range @xmath140 mag .
this is a passive luminosity evolution model with no merging over this redshift range ( @xmath141 ) . in section [ evolmstar ]
we solve for @xmath142 and find @xmath143 ( see fig .
[ beta ] ) .
more complex evolutionary scenarioes will be considered in future papers where we sub - divide the galaxy population by type and component .
however we do note that recent studies ( e.g. , @xcite ) including our own ( @xcite ) indicate a very low low - z merger rate based on counts of dynamically close pairs .
it is also worth noting that the step - wise maximum likelihood method used assumes no number density evolution .
we adopt the effective surface brightness , i.e. the average surface brightness within the half - light radius , as our structural measurement .
the effective surface brightness is chosen as it makes no assumption of the radial flux profile and can be measured empirically from the data for any galaxy shape . hence it is a direct measure reflecting the global compactness regardless of whether the galaxy is smooth or lumpy .
the half - light radius is the semi - major axis of the ellipse containing half the flux of sextractor s best magnitude .
the ellipse s centre , ellipticity and position angle are taken from sextractor . for compact objects
the half - light radius is affected by the seeing . through simulations of both exponential and de vaucouleurs profile galaxies using iraf artdata ( see appendix [ appsim ] for details )
we find that we can correct the observed half - light radius for this blurring and recover its true value , @xmath144 , to an accuracy of @xmath145 per cent ( see fig .
[ selection ] ) by : @xmath146 where @xmath147 is the observed half - light radius and @xmath148 is the fwhm of the seeing profile , all in arcsec .
work is currently underway to determine 2d bulge - disk decompositions from the mgc data which will be presented in a future paper .
however at this stage it is unclear whether the effective radius or independent bulge and disk parameters are the more fundamental . here
, we define our absolute effective surface brightness , @xmath149 , as : @xmath150 - 10 \log ( 1+z ) - k(z ) - e(z),\ ] ] we note that @xmath144 is defined along the semi - major axis and hence the above prescription assumes galaxies are optically thin , i.e. , derived parameters are independent of inclination .
the role of dust will be explored in a future paper .
like all surveys our catalogue will suffer from some level of incompleteness in both the imaging and spectroscopic surveys .
as our data are the deepest available over this region of sky we can not empirically quantify our imaging incompleteness ( see @xcite for an assessment of the incompleteness of the supercosmos sky survey , 2dfgrs and sdss - edr / dr1 relative to the mgc ) .
we can , however , define the selection limits at any redshift , following @xcite , and use this information later ( section [ swml ] ) to quantify the selection boundary in our final space density distribution . and the solid line shows the fit .
the low surface brightness and maximum size limits are defined as limits of reliable photometry by requiring errors in the photometry of less than @xmath151 mag and @xmath151 mag arcsec@xmath1 .
the dashed line is the fundamental detection limit . ]
we begin by noting the mgc s apparent survey limits : a bright and faint apparent magnitude limit of @xmath152 mag and @xmath153 mag , a limiting apparent effective surface brightness of @xmath154 mag arcsec@xmath1 , a minimum size of @xmath155 arcsec ( where @xmath148 is the seeing fwhm ) and a maximum size of @xmath156 arcsec .
the last three limits refer to the seeing - corrected quantities ( see section [ mgcsb ] ) and were derived from simulations of @xmath157-inclined optically thin exponential disks using iraf s artdata package ( see appendix [ appsim ] for details ) .
it is important to note that the surface brightness and maximum size limits do _ not _ define the _ absolute detection _ limits but the boundary of _ reliable photometry _ ( where ` reliable ' is arbitrarily defined as @xmath151 mag and @xmath151 mag arcsec@xmath1 accuracy , see fig .
[ selection ] ) .
[ maxhlr ] shows these limits in the apparent magnitude - apparent effective surface brightness plane , along with the galaxy distribution with ( dots ) and without ( triangles ) redshifts .
-corrections ) and the full ( @xmath141 ) mgc distribution ( dots ) .
the figure illustrates that the selection boundaries vary dramatically with redshift ( bottom to top ) such that each redshift slice , within the survey limitations , only samples a restricted region of the luminosity - surface brightness plane .
note that some points may lie outside the box if their @xmath5-corrections are significantly different from the galaxy for which the selection lines have been drawn .
figure degraded see http://www.eso.org/@xmath51jliske/mgc/ for full pdf copy of this paper . ] for each galaxy @xmath158 with known redshift , @xmath159 , we can now determine the parameter space in absolute magnitude , @xmath160 , and absolute effective surface brightness , @xmath149 , over which this galaxy could have been observed , given the survey limits .
the appropriate limits are ( following @xcite ) : @xmath161 note that @xmath162 is defined by both the minimum effective surface brightness for reliable detection , @xmath163 , and the maximum reliable size limit , @xmath164 .
the maximum reliable size limit is imposed by the smoothing box size when calculating the background sky levels .
these limits are introduced in more detail in @xcite .
one can also envisage a sixth limit ( also @xmath165 ) due to the dynamic range of the ccd detector . in our case none of the galaxies have their central regions flooded and so this limit can be ignored .
[ limfig ] shows the selection limits for three galaxies at redshifts @xmath166 ( lower panel ) , @xmath167 ( middle ) and @xmath168 ( upper ) along with the location of the galaxy ( large solid circle ) , galaxies with similar redshifts ( open squares ) and the full ( @xmath169 ) mgc distribution ( dots ) .
the selection boundary consistently forms a five - sided figure which glides through the full distribution as the redshift increases . because of the individual @xmath5-corrections each galaxy has its own unique selection window .
it is worth noting that large luminous galaxies are undetectable , or at least photometrically unreliable , at very low redshift .
hence the methods of @xcite which use @xmath170-style corrections , _ may _ underestimate this population because they implicitly assume selection boundaries only cut into the distribution at some upper redshift limit .
@xcite accounted for this effect in their analysis . for each galaxy with known redshift
we now define an observable window function , @xmath171 , as : @xmath172 the functions @xmath173 will be used in section [ swml ] in the construction of the bbd .
we now construct the mgc bbd using a bivariate brightness step - wise maximum likelihood ( swml ) estimator similar to that described by @xcite .
this is an extension of the standard swml method defined by @xcite in the sense that the sample is now divided into bins of both absolute magnitude _ and _ absolute surface brightness as opposed to just absolute magnitude alone .
note that @xcite formulated their method in terms of absolute magnitude and a physical diameter .
however , the mgc is primarily magnitude and surface brightness limited with a fixed isophotal detection limit of @xmath29 mag arcsec@xmath1 . for this reason
we elect to work in @xmath160-@xmath149 space rather than @xmath160-@xmath174 space .
essentially , for @xmath175 objects the volume - corrected relative bbd in @xmath176 absolute magnitude bins and @xmath177 surface brightness bins with widths @xmath178 and @xmath179 can be evaluated by ( cf .
@xcite ) : @xmath180},\ ] ] where the relative space density of galaxies , @xmath181 , is given in terms of the space density of the previous iteration , @xmath182 , the weighting function , @xmath183 , and the visibility function , @xmath184 .
the weighting function , @xmath183 , corrects for the incompleteness in our redshift survey .
this incompleteness is demonstrably not random ( see fig . [ completeness ] ) and we are preferentially missing faint dim galaxies with extreme colours . when constructing lfs this bias is often ignored ( e.g. , @xcite ) or considered a function of magnitude only ( e.g. , @xcite ) . here
, we account for luminosity and surface brightness dependent incompleteness and define @xmath183 as : @xmath185 where @xmath186 is the total number of galaxies lying in the same apparent magnitude - apparent surface brightness bin as galaxy @xmath158 and @xmath187 is the number of galaxies with known redshifts ( i.e. , those with @xmath85 , see section [ data ] ) in the same bin .
hence each galaxy is weighted by the inverse of the redshift completeness in the galaxy s apparent magnitude - surface brightness bin .
the visibility function @xmath184 is given by : @xmath188 where @xmath189 is defined in section [ sellim ] . hence @xmath184 is the fraction of the @xmath190-@xmath191 bin which lies inside the observable window of galaxy @xmath158 .
[ matrices ] shows example @xmath183 and @xmath184 matrices for @xmath192 mgc16504 ( upper and lower left respectively ) as well as the summed @xmath193 and @xmath194 matrices ( upper right and lower right respectively ) .
@xmath195 is essentially the observed galaxy distribution corrected for incompleteness and @xmath196 is analogous to the visibility surface of visibility theory ( phillipps , davies & disney 1990 ) , although not precisely as it does not account for the variation of physical cross - section with redshift or for isophotal corrections .
the steep sides of the @xmath196 surface ( fig .
[ matrices ] ) highlights the narrow range over which surface brightness selection effects go from negligible to extreme and hence the importance of understanding a survey s selection limits .
we now apply the above methodology to our mgc sample . to recap , we extract galaxies in the range @xmath197 mag , determine individual @xmath5-corrections , determine seeing - corrected surface brightness estimates , attach an incompleteness weight to each galaxy , reject galaxies which lie outside of their designated reliably observable regions ( @xmath198 galaxies too large , @xmath199 too compact and @xmath200 too dim ) , generate weighting and visibility matrices for each galaxy and then iteratively apply equation ( [ phijk ] ) until a stable set of @xmath201-values is found . fig .
[ bbd ] shows the result of this procedure after renormalisation ( described below ) .
it is worth noting that the majority of the rejected galaxies are extremely compact and have entered our sample only through spectroscopy of apparently stellar objects ( 2qz , sdss qso survey , ned and mgcz ) .
mpc@xmath7 mag@xmath55 ( mag arcsec@xmath1@xmath202.,width=377,height=302 ] the swml method , by its construction , loses all information regarding the absolute normalisation ( cf .
@xcite ) . here
the normalisation issues are compounded by the consideration of selection bias .
we overcome this by calculating the number of galaxies in the absolute magnitude and absolute surface brightness interval @xmath203 mag and @xmath204 mag arcsec@xmath1 .
this parameter space is observable for any @xmath112 over the redshift range , @xmath205 ( equivalent to a co - moving volume of @xmath206 mpc@xmath207 for the @xmath208 deg@xmath36 field - of - view ) . within these redshift , absolute magnitude and surface brightness
ranges there are @xmath209 galaxies with a summed weight , due to incompleteness , of @xmath210 .
the @xmath211 values are now rescaled to reproduce the correct space density for these galaxies ( i.e. , @xmath212 mpc@xmath7 ) .
the error includes both possion and large scale structure considerations and was derived from mock 2df ngp catalogues ( @xcite ; http://star-www.dur.ac.uk/@xmath51cole/mocks/main.html ) extracted from the virgo consortium hubble volume .
volumes equivalent in shape to the mgc normalisation volume were extracted and the standard deviation of @xmath213 mag galaxies determined .
the final error , inclusive of large scale structure , is roughly twice that of the poisson error alone .
[ bbdcont ] shows the final normalised bbd on a logarithmic scale .
the thick dotted line shows the region within which the statistical errors are @xmath136 per cent or less .
the thick solid line shows the effective detection limit defined as the bbd region sampled by at least @xmath214 galaxies ( equivalent to a volume limit of @xmath215 mpc@xmath216 ) .
the detection limit bounds the bbd at the low surface brightness and faint limits and clearly starts to impinge on the distribution from around @xmath217 mag ( coincidentally the usually adopted boundary between the dwarf and giant galaxy populations ) .
this confirms that even the mgc , the deepest local imaging survey to date , remains incomplete for both high ( @xmath218 mag ) and low surface brightness galaxies . to sample these regions requires both deeper ( @xmath219 mag arcsec@xmath1 ) _ and _ higher resolution imaging data ( @xmath220 arcsec ) over a comparable or larger area ( @xmath221 deg@xmath36 ) with high spectroscopic completeness .
following @xcite and @xcite we attempted to fit the mgc bbd with a chooniewski function @xcite , given by : @xmath222 ^ 2\right\rbrace\end{aligned}\ ] ] the parameters @xmath27 , @xmath10 and @xmath28 are the conventional schechter function parameters @xcite .
the three additional parameters , @xmath223 , @xmath224 and @xmath142 are the characteristic surface brightness , the gaussian width of the surface brightness distribution and the luminosity - surface brightness relation respectively . this function has been derived analytically by @xcite , @xcite and @xcite . note that the additional normalisation factor , @xmath225 , ensures that @xmath226 and @xmath28 are directly equivalent to the more familiar schechter function parameters .
the fit is achieved using the downward simplex method @xcite and the 1-@xmath227 errors are derived from a one - dimensional @xmath118-minimisation across each parameter ( @xmath228 ) . for the @xmath229 degrees of freedom ( @xmath230 data values minus @xmath231 model parameters )
we find an unreduced @xmath126 of @xmath232 .
this suggests that the chooniewksi function is a poor fit to the data ( see fig .
[ bbdfit ] ) .
the main reason appears to be the intrinsic assumption of a constant @xmath224 , whereas the distribution shown in fig .
[ bbd ] clearly broadens , a feature also observed by @xcite in their sdss data .
we discuss this further in section [ sbd ] .
chooniewski parameters and associated errors are tabulated in the upper half of table [ lfdata ] along with other published data including our previous measurement based on the 2dfgrs @xcite .
mpc@xmath7 mag@xmath55 ( mag arcsec@xmath1@xmath202 .
the greyscale image and the thick solid line are the same as those in fig .
[ bbdcont].,width=377,height=302 ] of particular interest over the past few decades has been the galaxy luminosity distribution .
this can be derived from the bbd by integrating over the surface brightness axis , ( the associated errors are the root sum squares of the individual errors for each bin ) .
[ lf ] ( left panel ) shows the result : @xmath233 mpc@xmath7 , @xmath234 mag and @xmath9 .
note that the error in the normalisation parameter , @xmath27 , is the combination ( in quadrature ) of the error from the fitting algorithm combined with the error from the number of galaxies within the normalisation volume . for completeness
we also show the implied lf from our chooniewski function fit ( dotted line ) .
however recall the poor @xmath126 value from this fit which highlighted the fit s inability to model the dwarf population correctly ( mainly due to the broadening of the surface brightness distribution ) .
the right panel of fig.[lf ] shows the formal 1,2,3-@xmath227 error contours for the schechter function fit to the collapsed mgc bbd luminosity distribution .
contours of the schechter function fit . ] fig .
[ lfcomp ] compares our lf based on the bbd analysis of the deeper , higher quality , higher completeness mgc data to a number of other recent lf estimates ( tabulated in the lower half of table [ lfdata ] ) .
note that these schechter functions are shown with their original @xmath27-values as opposed to the @xcite corrected values .
we find our lf recovers a comparable @xmath10 value to most previous surveys , but a slightly higher normalisation ( as one might expect from a bbd - style analysis which recovers more light from the low surface brightness population than a non - bbd style analysis ) , and an @xmath28 value consistent with the median of these other surveys . in comparison to the esp
lf ( @xcite ) we find consistent schechter function parameters within the quoted errors of the two surveys . in comparison to the 2dfgrs lf ( @xcite ) we find consistent @xmath10 and @xmath27 values but a flatter ( move positive ) @xmath28 .
this discrepancy in @xmath28 is formally significant at the @xmath235-level , however we do note that the 2dfgrs has undergone a major overhaul of its photometry since this result ( see discussion in @xcite ) .
the final 2dfgrs post - recalibration lf has yet to be determined .
our result are consistent with the @xmath10 value only for sdss1 and inconsistent with all three schechter function values for sdss2 .
a detailed discussion of this discrepancy follows in the next section . and @xmath28 together with the mgc @xmath236 , @xmath67 and @xmath237-@xmath227 contours ( right panel ) .
] -@xmath227 error ellipses for the schechter function fits along with the data points from the right panel of fig .
[ lfcomp ] . ] the most noticeable outliers in fig .
[ lfcomp ] are the two sdss results , sdss1 @xcite and sdss2 @xcite .
the discrepancy in the normalisation of sdss1 has been comprehensively discussed by both @xcite and @xcite and ascribed to the lack of an evolutionary model in the sdss1 analysis , coupled with their method of normalisation .
an over - density in the sdss commissioning data region also contributes .
the sdss1 normalisation has been revised by @xcite : @xmath238 ) .
once corrected , the sdss1 lf agrees well with the mgc ( 2@xmath227-discrepancy in @xmath28 ) , the 2dfgrs and the eso slice project ( esp ) estimates .
the more recent sdss2 result has a significantly fainter @xmath10 to @xmath109 is given by @xcite as @xmath239 . ] and a much flatter faint - end slope than any of the other surveys .
@xcite argued that the difference in the @xmath240-band lfs of sdss1 and sdss2 was simply due to the inclusion of luminosity evolution in the latter and that the luminosity densities measured by the 2dfgrs and sdss2 agreed to within the errors once the differing amounts of evolution had been taken into account .
however , in the @xmath241-band the discrepancy in @xmath28 between sdss1 and sdss2 is more than twice that in the @xmath240-band .
furthermore , even though the luminosity evolution derived in the sdss2 fitting process is very strong ( @xmath242 mag per unit redshift , where @xmath243 ) it can not explain away the difference between the mgc and sdss2 values .
our evolutionary correction given in section [ evolution ] corresponds to @xmath244 .
if we replace this with the sdss2 evolution , @xmath245 , and re - derive our collapsed mgc lf we find @xmath246 mag and @xmath247 ( or @xmath248 mag and @xmath249 if we restrict the lf fitting to the equivalent sdss2 limit of @xmath250 mag ) .
hence some tension remains between the sdss2 and mgc lfs irrespective of evolution .
the same argument holds when comparing sdss2 to the 2dfgrs since the latter adopted an evolution equivalent to @xmath251 . to test whether the surface brightness selection limits might be responsible , fig .
[ lfcomp2 ] shows a re - analysis of the mgc data , following the same procedures as before , but now imposing progressively shallower surface brightness limits ( as indicated ) .
the left panel again shows the lfs and the right panel the @xmath236-@xmath227 error ellipses of the schechter function fits .
it is quite apparent that shallower surveys progressively miss a greater fraction of the dwarf population(s ) and ultimately giants recovering both a fainter @xmath10 and a more positive @xmath28 .
in fact , fig.[lfcomp2 ] based on mgc data mirrors the expectation from the simulations of @xcite remarkably well ( see their fig . 5 ) . on fig .
[ lfcomp2 ] the lf with an imposed selection limit of @xmath252 mag arcsec@xmath1 appears to match the sdss2 lf closely and initially suggests that surface brightness selection might explain the sdss2 result .
( note that sdss1 implemented a bbd - style analysis ) .
however fig .
[ completeness ] ( left panels ) do not show a significant difference in the low surface brightness spectroscopic completeness between the 2dfgrs and sdss surveys suggesting limits closer to @xmath253 mag arcsec@xmath1 .
we note that the sdss2 surface brightness @xmath254 per cent completeness limit ( including consideration of deblending and flux loss issues , etc . )
is quoted as @xmath255 \approx 23.4 $ ] mag arcsec@xmath1 ( @xcite , their fig .
3 ) , where @xmath256 $ ] indicates that the sdss surface brightness measures are derived from ` circular ' half - light radii .
conversely the mgc ( and most contemporary datasets ) use major axis half - light radii .
these are derived by growing the optimal ellipse until it encloses @xmath254 per cent of the flux .
the use of a circular aperture will always result in a lower half - light radius than an elliptical aperture ( except for purely face - on systems ) , and hence in a higher effective surface brightness .
the two effective surface brightnesses are related by : @xmath257 = \mu^e[{\rm circular}]-2.5\log \left(\frac{b}{a } \right),\ ] ] where @xmath258 is the axis ratio . for a random distribution of infinitely thin disks
the expected median axis ratio @xmath259 is @xmath260 . in reality
the galaxy distribution is more diverse consisting of spheroids , bulges and thick disk components / systems .
circular half - light radii measurements are also somewhat resolution dependent , introducing both seeing and redshift dependencies into the conversion .
empirically we find for our mgc data : @xmath261 , implying a correction of @xmath262 mag arcsec@xmath1 .
although we can not directly compare our mgc elliptical half - light radii to the circular s ' ersic half - light radii used by sdss2 , we can compare them to the standard sdss - dr1 circular petrosian half - light radii .
we find @xmath263 , independently confirming the correction of @xmath262 mag arcsec@xmath1 .
hence the quoted sdss2 @xmath254 per cent completeness limit of @xmath255=23.4 $ ] mag arcsec@xmath1 equates to @xmath264=23.77 $ ] mag arcsec@xmath1 .
the colour correction is more complex as it depends strongly on galaxy type and luminosity .
the low luminosity population is the bluest ( @xmath265 mag ) and hence surface brightness incompleteness will preferentially affect the faint end of the lf .
evidence for such a colour completeness bias can be seen in the observed sdss incompleteness within the mgc region .
this is shown as a function of @xmath266 or @xmath267 versus @xmath72 in the middle row ( centre and right panels ) of fig .
[ completeness ] . as the sdss spectroscopic survey is @xmath240-band selected it should not be surprising that the @xmath109 completeness is a more complex surface in magnitude , surface brightness and colour . in the worst case scenario we find a 50 per cent completeness limit for the bluest systems of @xmath268 mag arcsec@xmath1 which is consistent with fig .
[ completeness ] ( middle left ) , but more importantly significantly fainter than the surface brightness limit implied by fig .
[ lfcomp2 ] .
that a colour selection is obvious in both the @xmath266 and @xmath267 sdss completeness maps of fig .
[ completeness ] argues that the colour selection bias is not a feature of the @xmath127 band data but more a general trend across all filters .
adopting an appropriate @xmath269 selection criterion is clearly non - trivial , however a simple way forward is to construct a lf from the available sdss2 redshifts within the mgc region ( see table .
1 ) . this will have the sdss selection function inherently built in .
we adopt @xmath270 mag , @xmath271 mag arcsec@xmath1 , the sdss2 @xmath241-band evolutionary correction and apply a conventional monovariate step - wise maximum likelihood @xcite .
we now recover @xmath272 mag and @xmath273 ( or @xmath274 mag and @xmath275 for @xmath276 mag ) and consistent at the @xmath277-@xmath227 level with sdss2 .
we therefore conclude that the discrepancy between the mgc and sdss2 lfs is complex , but predominantly due to the colour bias within the sdss ( cf .
[ completeness ] middle right ) .
this highlights the additional complexity in measuring lfs for filters other than the spectroscopic selection filter .
finally we note the recent preprint by @xcite in which an attempt is made to sample the very low luminosity regime from their new york university value added catalogue @xcite .
this incorporates a strategy for correcting the surface brightness incompleteness within the sdss2 dataset resulting in a significantly steeper faint - end slope ( @xmath278 to @xmath279 in @xmath240 ) .
the premise is that the surface brightness distribution of the dwarf population can be extrapolated from that of the giant population .
this results in correction factors for the space density of dwarf systems of up to @xmath280 ( see @xcite , their fig .
6 ) . in a similar spirit of producing a ` best guess ' lf at low luminosities fig.[guesstimate ] shows the mgc lf with the selection boundaries and lower redshift bounds removed .
this undoubtedly represents an underestimate of the true space density at low luminosities ( @xmath281 mag ) , but in so far as a comparison can be made it is consistent with the sdss result @xcite , in the sense that the sdss data lie above the mgc lower limits .
our lf is also consistent with a compilation of @xmath282 galaxies within the local sphere ( @xmath283 mpc ) taken from @xcite .
apart from indicating the amount of work that is yet to be done , fig.[guesstimate ] does suggest that the low redshift global luminosity function must steepen beyond @xmath284 at some intermediate magnitude ( @xmath285 mag ) .
however it would seem unlikely that it could steepen enough for the dwarf systems to make a significant contribution to the total local luminosity density ( as indicated by the dashed line ) .
mpc ) data taken from the catalogue of @xcite .
the schechter function from fig .
[ lfcomp ] is shown as the solid line and its extrapolation as the dotted line .
the dashed line shows the contour of equal contribution to the overall luminosity density . ] as we found earlier the chooniewski function adopted by previous studies ( @xcite ; @xcite ) provides a poor fit to the joint luminosity - surface brightness distribution .
this is mainly due to the apparent broadening of the surface brightness distribution towards fainter luminosities .
[ sbcomp ] highlights this by showing the surface brightness distribution at progressively fainter absolute magnitude intervals along with the selection limits taken from fig .
[ bbdcont ] .
we attempt to fit each of these distributions ( fig .
[ sbcomp ] , solid lines ) with a gaussian function defined as : @xmath286,\ ] ] where , @xmath223 is the characteristic ( peak ) surface brightness and @xmath224 is the dispersion ( analogous to those values in the chooniewski function ) .
the errors reflect @xmath287-statistics from the original observed number of galaxies contributing to each histogram element . for elements with zero values ( i.e. , no galaxies detected )
we adopt the error appropriate for a detection of one galaxy so that the significance of these ` null ' detections can be used to help constrain the fits .
the number of degrees of freedom ( listed in the final column of table [ gfdata ] ) is the number of data values plus any zero values either side of the distribution within the selection boundaries minus the number of fitting parameters ( three ) .
[ sbgauss](a ) shows the @xmath223 ( dark grey line ) and 1-@xmath224 values ( shaded region ) as a function of absolute magnitude .
these results are also tabulated in table [ gfdata ] .
[ sbgauss](a ) shows a trend of an initially invariant luminosity - surface brightness relation until @xmath288 mag , and then a steady decline to the limits of detection at @xmath289 mag .
the dispersion of the surface brightness distributions grows steadily broader as the luminosity decreases .
the changes in both the luminosity - surface brightness relation and its dispersion explain the poor fit of the chooniewski function which can not accommodate such features . at this stage
it is also worth noting that the gaussian fits return poor @xmath126 values at intermediate luminosities with even a hint of bimodality ( at @xmath290 mag ) .
this may imply that the low luminosity distribution consists of two overlapping populations : rotationally supported disks ( lower surface brightnesses ) and pressure supported spheroids ( higher surface brightnesses ) .
if borne out by more detailed structural studies then one might expect the rotationally supported disk systems to exhibit both a steeper @xmath160-@xmath291 relation and a narrower surface brightness distribution than the overall population .
three luminosity - surface brightness trends ( @xmath292-@xmath293 ) are well known ( see table [ lfdata ] ) : the kormendy relation for spheroids ( @xcite ) ; freeman s law for disks @xcite ; and the virgo / local group dwarf relation ( @xcite ; @xcite ) .
more recently @xcite identified a relation for late - type disks similar to that seen for dwarf systems in virgo and the local group .
the general trends for these established relations are overlain on fig .
[ sbgauss](a ) .
one can speculate that the combination of these trends might well lead to the overall shape of the global @xmath292-@xmath293-distribution shown on fig .
[ sbgauss](a ) , as the spheroid - to - disk ratio typically declines with decreasing luminosity .
hence the global trend initially follows a kormendy relation , switching to a more constant freeman law and finally into a declining dwarf relation . in comparison to the @xcite sample for late - type disks we find a shallower luminosity - surface brightness relation and a broader surface brightness distribution
this may have two reasons : firstly , the djl sample is @xmath294-band selected and our use of a global filter conversion may introduce this error ; secondly , the djl sample is far more local and both inclination and dust - corrected .
an obvious extension to the mgc analysis would be to also implement these corrections .
obviously , detailed structural analysis is required before a meaningful comparison to these ad hoc local studies can be made .
however , we can highlight two key questions : ( i ) are there two distinct disk populations , i.e. a giant disk population adhering to freeman s law and a dwarf - disk population which does not , or just a single dwarf - disk relation ? ( ii ) is the dwarf surface brightness distribution bimodal reflecting the two dominant classes of dwarf galaxies ( dwarf ellipticals and dwarf irregulars ) ?
there are three significant studies of the global @xmath292-@xmath293 distribution ( see table [ lfdata ] ) : a 2dfgrs study by @xcite , an sdss study by @xcite and a local galaxy sdss study by @xcite . in comparison to @xcite
we find that the mgc data follow a similar @xmath292-@xmath293 relation but show significantly broader surface brightness distributions .
[ sbgauss](b ) shows the comparison where we have taken the data from table c2 of @xcite and fitted gaussian distributions in a similar manner as for the mgc data .
the obvious discrepancy is in the @xmath224 values and this reflects a number of significant improvements .
firstly , the deeper mgc isophote coupled with the improved redshift completeness of mgcz ( see fig . [ completeness ] ) enables the inclusion of lower surface brightness systems . as these preferentially occur at lower luminosity the effect acts to broaden the surface brightness distribution at fainter luminosities .
secondly , the 2dfgrs sample was not seeing - corrected and hence the sizes of the high surface brightness galaxies were overestimated .
thirdly , the cross et al.analysis derived the surface brightness measurements under the assumption that all galaxies were perfect exponential disks . these factors conspire to push the 2dfgrs bbd into a narrower distribution .
a fourth factor , hard to qualitatively assess , is the major revisions to the 2dfgrs photometry which have taken place since @xcite ( see @xcite for further details on the 2dfgrs recalibration history ) . in fig .
[ sbgauss](c ) we overlay the recent sdss results by @xcite .
the shen et al .
study is @xmath295-band selected with an imposed surface brightness cut at @xmath296=23 $ ] mag arcsec@xmath1 .
shen et al .
divide their sample into a number of sub - groupings and we elect to show the data from their fig .
13 which separates galaxies according to concentration index , @xmath297 ( see @xcite for details ) .
this divides the population into galaxies earlier or later then s0/a ( with an @xmath298 per cent success rate ) . to convert the shen et al .
@xmath240-band results to @xmath109 we derive the early and late - type @xmath299 colour in each of the @xmath240-band absolute magnitude bins specified by shen et al . to do this
we assume our eyeball e / s0 type corresponds to ` early ' and our eyeball sabcd / irr to ` late ' ( a detailed account of the eyeball morphologies and morphological luminosity functions is presented in driver et al .
, in preparation ) .
hence each data point is adjusted independently .
in addition we must also correct the sdss effective surface brightness measurements from circular to major axis measurements as discussed in section [ sdssdiff ] , i.e. @xmath300 = \mu^e[{\rm circular}]+0.37 $ ] .
[ sbgauss](c ) shows the comparison between the shen et al .
data and the mgc and we see a uniform offset of @xmath301 mag arcsec@xmath1 . despite extensive experimentation with various data subsets and selection boundaries
we have been unable to reproduce this result from the mgc data .
[ sbgauss](d ) shows a comparison between the recent very low redshift sdss @xmath240 results @xcite and the mgc .
we implement the same colour and surface brightness corrections as for the shen et al . data .
the ridge line of the very local sdss data shows good agreement with the mgc data , albeit with a narrower distribution ( i.e. , lower @xmath227 values ) .
however this appears to be a further manifestation of the sdss use of circular apertures .
for example , if we derive the surface brightness distribution around the @xmath10-point for the mgc data using circular and major axis derived surface brightness measures we find a peak offset of @xmath302 ( slightly larger than our estimated median offset of @xmath262 ) with @xmath303 and @xmath304 .
this latter value is now consistent with the recent sdss estimate ( @xcite ) .
note that the dashed lines on fig .
[ sbgauss ] indicated the sdss extrapolated data used to correct the faint - end of the luminosity function . to the mgc limit
these extrapolated distributions appear to match our data well providing some vindication for the extrapolation process ( see @xcite , section [ llfe ] and fig . [ guesstimate ] ) .
the above two comparison between the mgc and the @xcite and @xcite studies must imply an inconsistency between the two sdss results .
this is perhaps more obvious if one bypasses the mgc data and directly compare fig . 14 of shen et al with fig . 5 of blanton et al .
the peaks of the late - type / low sersic index sb distributions ( both measured in @xmath240 ) are clearly distinct . given the consistency with the 2dfgrs and the traditional structural studies discussed earlier we conclude that the problem most likely lies with the shen et al . analysis .
if we disregard the shen et al study , we can conclude that the ridge line of the effective surface brightness distribution is well constrained and consistent between the mgc , 2dfgrs and the most recent sdss results .
however the mgc distribution is significantly broader than either the 2dfgrs or sdss results which is due to improvements in the analysis ( in comparison to the 2dfgrs ) and the use of elliptical as opposed to circular aperture photometry ( in comparison to the sdss ) . in section [ evolution ]
we described our adopted evolutionary parameter , @xmath142 , where the luminosity evolution follows the form : @xmath305 $ ] . in section [ sdssdiff ]
we saw that this evolution differs from that adopted by the 2dfgrs and sdss2 and discussed the possible impact .
[ beta ] ( upper panels ) highlights this dependency by showing the derived values for @xmath10 and @xmath28 for various values of @xmath142 . taking this one step
further we can attempt to constrain @xmath142 directly by adopting the additional constraint that the recovered @xmath10 parameter should be invariant to redshift .
we divide our sample into two redshift intervals : @xmath306 and @xmath307 ( with median redshifts of @xmath308 and @xmath309 respectively ) and redetermine the collapsed lf distributions using our bbd analysis for regular intervals of @xmath142 .
we hold @xmath28 fixed to the value derived from the full redshift range for each @xmath142 ( i.e. , the value indicated in the middle panel of fig . [ beta ] ) and
solve for @xmath10 and @xmath27 .
it is necessary to hold @xmath28 fixed because the high-@xmath102 sample has too few low - luminosity galaxies to adequately constrain it .
[ beta ] ( main panel ) shows the resulting @xmath310 versus @xmath142 and we find the weak constraint : @xmath311 . while this is not particularly stringent , it is consistent with our adopted value of @xmath312 .
this constraint on beta is marginally inconsistent ( @xmath313 ) with the evolution found by @xcite for the @xmath241-band ( @xmath314 ) , but consistent with the evolution adopted by the 2dfgrs ( @xmath315 ) .
perhaps more importantly this range of @xmath142 implies a potential systematic uncertainty on the quoted @xmath10 ( @xmath28 ) values of @xmath316 ( @xmath317 ) . as noted at the start of section [ cosmoke ] , it is also equally plausible that this error does not actually reflect evolution per se but could instead be interpreted as a redshift dependent photometric error or similar . from the variation of @xmath10 between low and high redshift samples .
upper panels : the @xmath10 and @xmath28 values for the entire dataset for different beta values . ]
in this paper we have attempted to recover the space density and surface brightness distributions of galaxies , paying careful attention to surface brightness selection biases .
we have used the millennium galaxy catalogue which , although smaller in size , has significantly higher resolution , greater depth and higher redshift completeness than either the 2dfgrs or sdss datasets , hence probing to fainter intrinsic luminosities . in comparison to most earlier studies this work includes the following enhancements .
\1 . imaging data which probes @xmath277@xmath67 mag arcsec@xmath1 deeper than the 2dfgrs and sdss surveys
. \2 . exceptionally high spectroscopic completeness ( @xmath4 per cent ) .
individual @xmath5-corrections derived from @xmath318 fits to spectral templates .
seeing - corrected half - light radius measurements without assumption of profile shape .
modeling of the mgc surface brightness detection and reliability limits which includes the effects of the analysis software .
\6 . a weighting system which accounts for apparent magnitude and apparent surface brightness dependent redshift incompleteness .
\7 . a joint luminosity - surface brightness step - wise maximum likelihood method which incorporates the selection boundaries .
having incorporated these additional improvements to the classical measurement of galaxy luminosity functions , we generally find close agreement with previous results . in particular , our collapsed luminosity distribution is marginally higher in normalisation and marginally flatter than the 2dfgrs @xcite , esp @xcite and sdss1 ( @xcite ; after renormalisation by @xcite ) results .
it is significantly brighter and steeper than the more recent sdss2 luminosity function @xcite .
we infer that this discrepancy arises because of a colour - selection bias in the sdss .
the mgc survey extends to lower luminosity than these previous surveys , providing more leverage and hence more reliability at the faint end .
our final schechter function parameters are : @xmath319 mpc@xmath7 , @xmath320 mag and @xmath9 with a @xmath321 of @xmath322 indicating a respectable fit .
overall we find that surface brightness selection effects do not play a significant role at bright luminosities .
however , the combination of the luminosity - surface brightness relation and the broadening of the surface brightness distribution is likely to lead to the loss of both low surface brightness and compact dwarf systems in contemporary surveys hence the variation in @xmath28 ( cf .
[ lfcomp ] & [ lfcomp2 ] and discussion in section [ spacedensity ] ) . at the very faint end
we provide lower limits which suggest that the luminosity function must turn up at some point faintwards of @xmath323 mag . the final mgc schechter function parameters , which include compensation for the effects of selection bias , imply a luminosity density of @xmath325 mpc@xmath7 ( adopting @xmath326 mag and @xmath327 , cf .
this is fully consistent with the mgc revised values for the 2dfgrs , esp and sdss1 surveys ( see table 3 of @xcite ) .
it is also worth noting that this value is also consistent with the original 2dfgrs ( @xcite ) and esp ( @xcite ) values .
the more recent sdss2 result ( @xcite ) finds @xmath328 mpc@xmath7 and @xmath329 mpc@xmath7 .
hence while the @xmath12 value is consistent the @xmath13 value is not ( @xmath330 ) .
the sdss @xmath13 value is derived by adopting the evolutionary parameters derived by @xcite .
if we use our evolution instead ( @xmath138 , see section [ evolution ] ) the sdss @xmath13 luminosity density becomes @xmath331 mpc@xmath7 which although closer is also inconsistent with our result . from section [ sdssdiff ]
we concluded that a colour bias may have resulted in an underestimate of the sdss @xmath241-band lf and hence to an underestimate of the sdss luminosity densities .
based on the consistency between the mgc , 2dfgrs and esp results we can also find that while the more sophisticated bbd analysis does recover some additional flux from the low surface brightness population , it does not constitute a significant change ( consistent with the findings of @xcite ) .
our recovered bivariate brightness distribution exhibits a clear ` ridge ' of data well separated from the selection boundaries down to @xmath332 mag at which point both very compact and highly extended dwarf systems will be missed .
this indicates that the recovered schechter function is generally robust to surface brightness selection effects for luminous galaxies but that the dwarf population(s ) remain elusive .
we attempted to fit a chooniewski function @xcite , essentially a schechter function with a gaussian distribution in surface brightness , to the overall distribution .
it introduces three additional parameters : a characteristic surface brightness ; a gaussian dispersion ; and a luminosity surface brightness relation .
we find that this function fails to provide a satisfactory fit due to the broadening of the surface brightness distribution with decreasing luminosity and a distinct change in slope of the luminosity - surface brightness relation at @xmath333 mag .
we find that the surface brightness distribution of galaxies exhibits interesting behaviour , with a well bounded gaussian - like distribution at @xmath26 , qualitatively consistent with the combination of the kormendy relation for spheroids and a @xmath334 broader freeman s law for bright disks . at @xmath26
we find that the surface brightness distribution is well described by a gaussian with parameters : @xmath335 mpc@xmath7 , @xmath336 mag arcsec@xmath1 , @xmath337 mag arcsec@xmath1 with a @xmath321 of @xmath338 . at fainter luminosities the normalisation increases ,
the characteristic surface brightness becomes fainter and the dispersion broadens until the distribution appears almost flat ( or possibly bimodal ) .
the schechter function fit to the gaussian normalisations ( column 4 of table 3 ) provides a basic completeness correction and yields consistent schechter function parameters to those quoted in section [ lfresult ] , i.e. @xmath339 mpc@xmath7 , @xmath340 mag and @xmath341 with a @xmath321 of @xmath342 . in comparison to the 2dfgrs study of @xcite
we find a consistent luminosity - surface brightness relation but with a significantly broader dispersion in surface brightness .
this is predominantly due to the implementation of seeing corrected size , direct ( i.e. , profile - independent ) half - light radius measurements and improved completeness for extended systems .
a comparison to the sdss results of @xcite and @xcite are complex because of the sdss standard of circular aperture photometry .
however we find that our results can be reconciled with those of @xcite but not with those of @xcite .
we also note that the sdss use of circular apertures also leads to significantly narrower surface brightness distributions . various studies ( @xcite ; @xcite ; @xcite ; @xcite ;
@xcite ) relate the gaussian dispersion of the surface brightness distribution , or equivalently of the @xmath174 is the half - light radius measured in kpc .
it is related to surface brightness by @xmath343 and hence @xmath344 . ]
distribution , to the dimensionless spin parameter of the dark matter halo , @xmath345 @xcite .
the basic concept is that the baryons in the disk reflect the angular momentum state ( rotational or dispersive ) of the dark matter halo .
while the luminosity relates to the mass , the disk scale - length relates to the specific angular momentum via the spin parameter .
numerical simulations generally predict values for @xmath21 ( @xcite ; see also @xcite ) .
this is most consistent with our observations at fainter luminosities .
however , at brighter luminosities we find a significantly narrower dispersion @xmath346 .
this may imply that brighter , more mature galaxies have evolved via mechanisms currently beyond the standard numerical simulations ( i.e. , the baryons have decoupled from the dark matter halo and evolved away from the halo properties ) .
alternatively it may argue for significant non - directional baryon infall onto the halo , muddying any coupling between the primordial baryons and the halo @xcite .
further possibilities might also include the presence of bulges in high luminosity systems whose properties depend more critically on a central supermassive black hole @xcite . one intriguing result of the simulations is the lack of variation of the surface brightness or size distribution with mass ( @xcite .
however , the more detailed studies which actually follow individual halos _ do _ imply a greater variation in @xmath347 for low - mass systems .
for example , @xcite trace the history of three halos through the merging process .
they clearly show that @xmath347 initially varies significantly as the halos grow but that it becomes more resilient to mergers once the halos reach sufficient mass .
ultimately all three halos converge towards a common value at high mass .
this seems plausible as one would expect smaller systems to be more affected by mergers than more massive systems .
the surface brightness distribution of galaxies has been hotly debated for over thirty years since the advent of freeman s law and the conjecture that it is wholly a selection bias @xcite . while in many minds
the debate is resolved through deep hi studies ( e.g.@xcite ) and ly-@xmath28 absorption line studies ( @xcite ; @xcite ) , others remain loyal to disney s conjecture ( e.g. oneil , andreon & culliandre 2003 ) and still argue for a flat surface brightness distribution at all luminosities ( @xcite ; @xcite ; @xcite ) implying a significant amount of hidden mass . at least in oneil
et al.s case this stance derives from neglecting the clear luminosity - surface brightness relation seen in their data .
the global surface brightness distribution may indeed be flat but the low surface brightness systems are predominantly dwarfs , providing little enhancement to the global luminosity , baryon or mass densities .
we have attempted to address this debate through careful consideration of _ most _ selection issues .
we probe to depths where the distribution is clearly bounded and where we robustly demonstrate we could reliably recover giant low surface brightness galaxies were they to exist in significant numbers .
they do not .
while we address the selection bias in the redshift survey as a function of apparent magnitude and apparent surface brightness other potential selection effects remain unexplored .
for example , the redshift completeness may well be a function of spectral type , in which case the assumption that all galaxies of similar apparent magnitude and surface brightness are drawn from the same redshift distribution will not be correct . in particular , inert low surface brightness galaxies which may have insufficient features to readily yield a redshift and may be missed entirely .
we also do not at this stage decompose galaxies into bulge and disk components .
this is clearly required by both the observations ( which show two , or possibly three , distinct luminosity - surface brightness relations for spheroid and disk components ) as well as by theory and simulations which argue for distinct formation mechanisms for these components ( merger and accretion respectively ; e.g. pierani , mohayaee & de freitas pacheco 2004 ) .
we hence intend to extend our analysis in three ways : @xmath214 per cent spectroscopic completeness ; bulge - disk decompositions ; and extension to longer wavelengths and in particular the near - ir which promises smoother , dust - free profiles .
we thank shiyin shen , eric bell and michael blanton for useful discussions and exchange of data , steve phillipps , alister graham and stefan andreon for comments on the early drafts and the referee for a number of detailed scientific and stylistic improvements .
the millennium galaxy catalogue consists of imaging data from the isaac newton telescope and spectroscopic data from the anglo australian telescope , the anu 2.3 m , the eso new technology telescope , the telescopio nazionale galileo , and the gemini telescope .
the survey has been supported through grants from the particle physics and astronomy research council ( uk ) and the australian research council ( aus ) .
the data and data products are publicly available from http://www.eso.org/@xmath51jliske/mgc/ or on request from j. liske or s.p . driver .
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 .
funding for the sloan digital sky survey ( sdss ) has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , princeton university , the united states naval observatory , and the university of washington . | we recover the joint and individual space density and surface brightness distribution(s ) of galaxies from the millennium galaxy catalogue .
the mgc is a local survey spanning 30.9 deg@xmath0 and probing approximately one two mag arcsec@xmath1 deeper than either the two - degree field galaxy redshift survey ( 2dfgrs ) or the sloan digital sky survey ( sdss ) .
the mgc contains @xmath2 galaxies to @xmath3 mag with @xmath4 per cent spectroscopic completeness . for each galaxy
we derive individual @xmath5-corrections and seeing - corrected sizes .
we implement a joint luminosity - surface brightness step - wise maximum likelihood method to recover the bivariate brightness distribution ( bbd ) inclusive of most selection effects . integrating the bbd over surface brightness
we recover the following schechter function parameters : @xmath6 mpc@xmath7 , @xmath8 mag and @xmath9 .
compared to the 2dfgrs @xcite we find a consistent @xmath10 value but a slightly flatter faint - end slope and a higher normalisation , resulting in a final luminosity density @xmath11 mpc@xmath7 marginally higher than , but consistent with , the earlier 2dfgrs ( @xcite ) , esp ( @xcite ) , and sdss @xmath12 @xcite results .
the mgc is inconsistent with the sdss @xmath13 result ( @xmath14 ) if one adopts the derived sdss evolution ( @xcite ) .
the mgc surface brightness distribution is a well bounded gaussian at the @xmath10 point with @xmath15 mpc@xmath7 , @xmath16 mag arcsec@xmath1 and @xmath17 .
the characteristic surface brightness for luminous systems is invariant to @xmath18 mag faintwards of which it moves to lower surface brightness .
the surface brightness distribution also broadens ( @xmath19 ) towards lower luminosities .
the luminosity dependence of @xmath20 provides a new constraint for both the theoretical development ( dalcanton , spergel & summers 1997 ; mo , mao & white 1998 ) and numerical simulations ( e.g. , @xcite ) which typically predict a mass - independent @xmath21 ( see @xcite and @xcite ) . higher resolution ( fwhm @xmath22 arcsec ) and
deeper ( @xmath23 mag arcsec@xmath1 in the @xmath24-band ) observations of the local universe are now essential to probe to lower luminosity and lower surface brightness levels .
[ firstpage ] galaxies : general galaxies : fundamental parameters galaxies : luminosity function , mass function galaxies : formation galaxies : evolution astronomical data bases : catalogues |
the current picture of galaxies is that they are composed of baryons ( stars , gas , dust ) and non baryonic dark matter . while observationally the distribution of the baryons can be studied , it is difficult to probe how the dark matter is distributed compared to the baryons . traditionally , the study of galaxy dynamics has been the strongest proof of the existence of dark matter in galaxies ( e.g. , @xcite ; @xcite ; @xcite ) . but given the inherent degeneracies in the inversion of dynamical data to obtain density profiles , it is hard to measure how the dark matter is distributed .
another approach is to use gravitational lens systems . for strongly lensed quasars ( qsos ) , the geometry and photometric properties of the lens system
depend on the projected mass inside the lensed qso images and therefore can be used to constrain the mass distribution . for the particular case of spiral galaxies , generally composed of a bulge , a disk and a dark matter halo
, the rotational velocity curves can not disentangle the relative contributions of the different mass components in the inner ( luminous ) parts of the galaxy . on the other hand ,
gravitational lenses offer the possibility of doing so , especially in the case of edge - on spirals .
there are only five spiral gravitational lenses known so far ( b0218 + 357 : @xcite ; pks 1830 - 211 : @xcite ; q2237 + 0305 : @xcite ; b1600 + 434 : @xcite ; pmn j2004 - 1349 : @xcite ) . only one of them is seen edge - on , b1600 + 434 . however , none of them is ideal for studying the relative mass contribution of the different components . here
, we report the discovery of an edge - on spiral lens galaxy as part of the cyder survey , which may be the best spiral lens system to separate the relative mass contributions of its constituents . the caln - yale deep extragalactic research ( cyder )
survey ( @xcite ; @xcite , hereafter t05 ) is an optical and near - infrared imaging and spectroscopic program carried out in archived , moderately deep _
chandra _ fields .
cxocy j220132.8 - 320144 ( hereafter cy 2201 - 3201 ) is one of the faint x - ray sources detected by cyder in its d1 field .
optical follow - up of this field seemed to suggest that cy 2201 - 3201 was an edge - on spiral with a very bright nucleus .
however , optical spectroscopy and imaging of this source in good seeing conditions revealed its true nature . cy 2201 - 3201 is a lensing system , showing two images of a distant @xmath1 quasar being lensed by an edge - on spiral galaxy at @xmath0 .
the paper is structured as follows . in
2 we present in chronological order the observational data gathered to characterize this system . in
3 we analyze the observational constraints obtained , and discuss them in 4 .
finally we draw our conclusions in 5 . throughout
, we assume @xmath4 , @xmath5 and @xmath6 km s@xmath3 mpc@xmath3 .
cy 2201 - 32 was observed by the @xmath7 x - ray observatory with the acis - i instrument for 50.16 ks in 2000 july as part of the observation of the hcg90 field ( pi bothum ) .
we retrieved this image from the archive and analyzed it using standard techniques with the ciao package ( @xcite , t05 ) .
cy 2201 - 32 was detected in the soft band ( 0.5 - 2.0 kev ) with 11.6 @xmath8 3.9 counts and undetected in the hard band ( 2.0 - 8.0 kev ) .
we computed an x - ray flux of @xmath9 erg @xmath10 s@xmath3 and x - ray luminosity of @xmath11 erg s@xmath3 in the soft band ( t05 ) .
as part of the cyder survey we imaged the cyder d1 field ( t05 ) with the cerro tololo inter - american observatory ( ctio ) 4 m telescope using the mosaic ii camera in 2001 august .
we took images for a total of 6600 and 1800s in the @xmath12 and @xmath13 bands , reaching limiting magnitudes of @xmath14 and @xmath15 and effective seeings of 1.0 and 0.9 , respectively .
cy 2201 - 3201 appears in these ctio images as an edge - on spiral with a very bright nucleus / bulge with a total measured magnitude of @xmath16 and @xmath17 . on 2003
october 31 , we took spectra of the x - ray optical counterparts in the cyder d1 field with the ut4 vlt fors2 instrument .
we used the 300v grism which gives a resolution of @xmath18 ( 10.5 ) for our 1 slits .
cy2201 - 3201 was observed in one of our masks .
given the multi - object purpose of the masks , all slits were oriented north - south .
we took five exposures of 1800 s in seeing condition of 0.50-0.75 and bright / grey sky conditions .
we reduced the spectra using standard iraf routines .
surprisingly , for this source we found two spectra of a qso at a redshift of z=3.90 in our best seeing spectra ( see fig .
[ fig1 ] ) .
the two qso spectra were blended in our worse seeing exposures .
our 30 s mask acquisition image taken with a seeing of 0.45 confirmed the existence of two point sources close to the center of the edge - on spiral galaxy , and therefore confirmed the lensing nature of the system .
we observed the cy 2201 - 3201 system with the magellan clay telescope using the magic instrument on 2004 september 8 .
magic has a pixel scale of 0.0691 pixel@xmath3 .
we took a series of 300 s exposures in three filters ( sdss @xmath19 , @xmath20 and @xmath21 ) for a total of 1800 , 2400 , and 1800 s in the @xmath19 , @xmath20 and @xmath21 filters respectively .
we reduced the images using standard procedures in iraf .
the resulting combined images have effective seeings of 0.68 , 0.65 and 0.62 in @xmath19 , @xmath20 , and @xmath21 , respectively .
figure [ fig2 ] shows the combined @xmath21-band image .
the system configuration is composed of an edge - on spiral with two images of the qso at each side of the disk .
this is one of the expected configurations produced by an edge - on spiral galaxy lens @xcite , except that it is missing a third image that should be located close to the disk . to obtain tight constraints in the lens modeling
it is important to know as accurately as possible the position and the fluxes of the observed components . in our images
the seeing is only slightly smaller than the qso images separation .
therefore the overlap between the qso images and the galaxy light distribution makes it hard to separate their relative fluxes and determine their positions .
we have thus used a monte carlo markov chain method ( mcmc ; @xcite ; @xcite , @xcite ) to try to determine the positions and fluxes ( and their errors ) of the three components as accurately as possible .
we model the system with three components , two point sources ( qso_a and qso_b ) and a spiral galaxy . each point source
is modeled with three parameters : the @xmath22 and @xmath23 of its center and its flux .
the spiral galaxy is modeled as an elliptical exponential with six parameters : @xmath22 and @xmath23 of the galaxy center , flux , axis ratio , position angle of the major axis and effective radius .
our model thus has 12 parameters . in order to obtain the best values we generate images with varying values of these 12 parameters and
convolve them with the point spread function as given by bright nearby stars .
we then compare the resulting model image with the observed one and compute a @xmath24 goodness of fit .
we run a mcmc to obtain the best values for 11 of the 12 parameters .
we fix one parameter , the position angle of the major axis , because the large elongation of the galaxy allows a robust direct determination .
table [ tab1 ] presents the parameters used to constrain the lens model which result from a combination of the parameters found in the three images ( @xmath19 , @xmath20 , and @xmath21 ) .
the errors on the fluxes have been artificially increased ( see below ) .
lccc[b ] parameter & qso a & qso b & galaxy + @xmath25 ( arcsec ) & @xmath26 & @xmath27 & @xmath28 + @xmath29 ( arcsec ) & @xmath26 & @xmath30 & @xmath31 + relative flux & @xmath32 & @xmath33 & ... + axis ratio ( b / a ) & ... & ... & @xmath34 + position angle ( degrees ) & ... & ... & @xmath35 + @xmath36_e ( arcsec ) & ... & ... & @xmath37 + @xmath38 ( km s@xmath3 ) & ... & ... & @xmath39 + table [ tab2 ] gives the measured magnitudes of the two qso images and the lens galaxy .
on 2004 september 7 , 8 and 9 , we took spectra of the lens galaxy using the magellan clay boller & chivens spectrograph ( b&c ) .
we use a 1 wide slit aligned along the major axis of the lens galaxy .
we use two grisms : 600 and 1200 lines mm@xmath3 , giving dispersions of 1.6 and 0.8 pixel@xmath3 , respectively .
we observed the lens for 30 minutes in 0.8 conditions and 3 hr in 1.2 conditions with the lower resolution setting and for 3.5 hrs in 0.7 conditions with the higher resolution set - up .
unfortunately , the slit position moved with respect to the object during the observations and the effective on - source exposure is shorter .
we reduced the images using standard procedures in iraf .
figure [ fig3 ] shows the two dimensional low - resolution spectrum at the wavelength of the observed [ oii ] doublet .
we measure the rotational velocity of the disk using the [ oii ] doublet .
we manage to detect signal out to a radius of 3.8 equivalent to 2.7 the effective radius where the rotational velocity appears to have already flattened .
in fact for an exponential mass distribution the maximum of the rotational velocity is at approximately twice the effective radius @xcite . in order to obtain a good estimate of the rotational velocity we use an mcmc .
we model the two - dimensional spectrum assuming an exponential distribution for the galaxy light and an arc - tangent function for the rotational velocity curve .
we generate two - dimensional spectra with the expression @xmath40 where @xmath41 is the normalization , @xmath42 and @xmath43 are the pixel coordinates of the center of the galaxy in the two dimensional spectrum , @xmath44 is the scale radius of the exponential profile , @xmath45 is the `` turnover '' rotational velocity radius and @xmath46 is the asymptotic rotational velocity .
we then convolve the model image with a gaussian of the same width of the seeing in the spatial direction and with a gaussian of the same width of the spectral resolution in the spectral direction .
we generate four model images : one for the low - resolution data taken on september 7 , one for the low resolution data taken on september 9 , one for the high resolution data taken on september 7 and one for the high resolution data taken on september 8 .
we compare the model images to the observed ones and compute a global @xmath24 of goodness of fit .
we run an mcmc with two free parameters , @xmath41 and @xmath46 , to obtain the asymptotic rotational velocity that best fit the observations .
we fix the rest of the parameters ( @xmath42 , @xmath43 @xmath44 and @xmath45 ) to the values directly measured from the two dimensional spectra ( @xmath42 and @xmath43 ) or the magic images ( @xmath44 ) .
we fix @xmath45 to the same value in pixels of @xmath44 .
we obtain a best value for the asymptotic rotational velocity of @xmath47 km s@xmath3 . with the arc - tangent rotational velocity model used , we compute a rotational velocity at 2.7 times the scale radius of @xmath48 km s@xmath3 .
most of the signal in the fit comes from data interior to this radius and therefore we will use the value at this radius to compare to the lens models that best fit the qso image positions and fluxes . lccc filter & qso a & qso b & galaxy + @xmath19 & 25.06@xmath49 & 25.22@xmath50 & 22.11@xmath51 + @xmath20 & 23.09@xmath52 & 23.20@xmath53 & 21.06@xmath51 + @xmath21xs & 22.62@xmath49 & 22.74@xmath51 & 20.61@xmath54 + @xmath55 & 18.75@xmath56 & 18.96@xmath57 & 17.18@xmath58
+ in addition we measure the galaxy and qso redshifts to be @xmath59 and @xmath60 , respectively .
we observed cy2201 - 3201 with the eso new technology telescope ( ntt ) on 2004 october 30 in service mode .
we took an exposure of 4185 s in the @xmath55 band . the reduced combined image provided by eso had an effective seeing of 0.75 .
we calibrated it with standard stars .
we use the same mcmc method used for the magellan images to calculate the positions and fluxes of the three modeled components in this @xmath55 image .
the calculated positions are much more uncertain than the ones calculated with the magellan images .
the fluxes nevertheless help us constrain the spectral energy distributions of the qso and lens galaxy ( see table [ tab2 ] ) .
we model the cy 2201 - 3201 system using the positions and fluxes of the two qso images together with the parameters measured for the lens galaxy .
we use the gravlens softwareckeeton / gravlens/ ] @xcite developed by c. keeton .
the models used are described in @xcite .
table [ tab1 ] summarizes the observational constraints used for the modeling .
note that we have artificially increased the error in the measured fluxes to allow for micro - lensing and/or qso variability effects .
we would like to test the hypothesis that mass is distributed similarly to light , and therefore start modeling the system with one ellipsoidal mass component forced to have the same ellipticity as the one observed in the light .
we then add a spherical component meant to represent the galaxy halo , which we then allow to be slightly elongated . we also add external shear to see the role it plays in the overall modeling . in this section
we discuss only the constraints enforced by the observed qso positions and fluxes .
we ignore the constraints imposed by the third and sometimes fourth images predicted but not seen and the predicted circular velocity of the model .
we discuss those constraints in the next section . for simplicity , we start with a singular isothermal ellipsoid ( sie ) mass model which gives a flat rotation curve . we follow @xcite and use his `` alpha '' model .
the projected mass surface density is given by : @xmath61 , where @xmath62^{1/2}$ ] .
@xmath63 is related to the axis ratio @xmath64 by : @xmath65 .
we fix the position and ellipticity of the galaxy and only allow the normalization to vary to get the best fit to the observed qso positions and fluxes .
figure [ fig4 ] shows the basic configuration of the best fit .
three images of the qso are produced at one side of the rotation axis of the galaxy .
this is a typical spiral galaxy lens configuration @xcite , with the two brighter images at each side of the disk and the fainter image closer to the disk .
figure [ fig4 ] shows that the position of the qso in the source plane ( denoted by the plus sign , @xmath66 ) is very close and outside of the radial caustic ( @xmath67 ) .
if the qso were closer to the galaxy center and inside this caustic then five images would be produced ( with the central one highly demagnified ) .
next , we try an elliptical navarro - frenk - white ( nfw ) model @xcite
. the spherical mass density in the nfw model is given by @xmath68 , where @xmath69 .
the projected surface mass density is @xmath70 , where @xmath71 is given by equation ( 48 ) in @xcite .
see also @xcite ; @xcite ; @xcite ; @xcite ; @xcite .
keeton s gravlens code defines an elliptical model from this spherical one replacing the polar radius @xmath20 in the projected surface mass density expression by the ellipse coordinate @xmath72^{1/2}$ ] .
this model produces the same `` disk '' configuration as the previous one .
we fit the qso positions and fluxes allowing only the normalization and characteristic radius of the model ( @xmath73 ) to vary .
we find that a large range of values produce acceptable fits .
the values of the characteristic radius and normalization that produce acceptable fits are degenerate .
the trends in this degeneracy are that , when all other parameters are fixed , the larger the characteristic radius , the lower the normalization and the smaller the radial caustic .
that is , the central concentration of the model dictates the size of the radial caustic .
figure [ fig4 ] shows a typical configuration of this model .
finally , we model cy2201 - 3201 with an exponential mass component @xcite . the projected surface mass density of this model is given by @xmath74 where @xmath75 is the same as above , @xmath64 is the axis ratio and @xmath76 is the scale radius of the exponential distribution .
we fix the scale radius , the position angle and the axis ratio to the values obtained with the light distribution ( table [ tab1 ] ) and solve for the normalization that best reproduces the position and fluxes .
the model shows the disk configuration as before and reproduces the position and fluxes of the two qso images fairly well ( see fig . [ fig4 ] ) .
given that the three models produce acceptable fits , one would like to find another constraint to differentiate between them .
comparing the three models , we find that the position and relative magnification of the third image depend on the concentration of the mass models . for the most concentrated model ( sie ) ,
the position of the third image is further away from the galaxy center and the relative magnification compared to the brightest qso image is the lowest ( 30% ) . for the least concentrated model ( exponential ) ,
the position is closer to the galaxy center and the relative magnification is the largest ( 59% ) .
we start with a simple two component model : one sie , as in the previous section , to which we add a single isothermal sphere ( sis ) .
the sis model uses the alpha model of @xcite ( the same as the sie , see above ) in which @xmath77 , the axis ratio @xmath78 , and @xmath79^{1/2}$ ] .
we vary the normalizations of the sis and the sie models , fixing all other parameters , to find the best fitting model for the observed data .
a wide range of normalizations in the models produce acceptable fits to the position and fluxes .
all acceptable fits produce five images of the qso ( the central one strongly de - magnified ) .
only models in which the rotational velocity of the sis is approximately 50% larger than the rotational velocity of the sie do not give acceptable fits .
figure [ fig5 ] shows an example of an acceptable fit in which both components have similar rotational velocities .
best fits are produced for models with larger sie contribution .
in general when the contribution from the spherical component ( sis ) gets larger in the fits , then the following trends are observed : the area enclosed by the radial caustic increases and the area enclosed by the tangential caustic decreases ; the magnification of the qso images gets larger and the relative magnification of the fourth brightest image compared to the brightest gets larger .
on the other hand , the relative magnification of the third image compared to the brightest is almost independent ( staying around 30% ) of the ratio of the sie and sis contributions .
the fourth and fifth images disappear at almost the point where the sis contribution is negligible and the model is dominated by the sie component .
we next model the system with an expectedly more realistic mass distribution with two components : an exponential distribution ( as in 3.1 ) to account for the disk and a spherical nfw ( same as above but with the axis ratio @xmath78 ) representing the halo .
we search for models that best fit the data , varying the two normalizations and the characteristic scale @xmath73 of the nfw model , and fixing all other parameters . again
, acceptable fits to the positions and fluxes are produced for a wide range of model parameters .
the only models that are rejected are those in which the spherical nfw component dominates .
this occurs when the rotational velocity of the nfw is equal to or 25% larger ( depending on the concentration parameter of the nfw ) than the rotational velocity of the exponential component at 2.7 disk scale radii .
these rejected models produce five images , with the fifth central image strongly demagnified .
the only exception is when the combination of the two mass components is such that the qso needs to be close to the caustic to fit the observed position and fluxes . in this case
the fifth image moves away from the center towards the fourth image in the image plane .
the trends with respect to the relative strengths of the exponential and nfw components are the following : the larger the nfw spherical component , the worse the fit to the positions and the larger the magnification of the qso images .
the relative magnification of the third brightest image compared to the brightest image goes from @xmath8050% when the nfw dominates to @xmath8059% when the exponential dominates . if we model the halos with some ellipticity , the results present general trends that are between the previous one mass component and two mass components cases discussed .
the most significant difference in the case of an exponential plus elliptical nfw model is the magnification of the resulting images .
the total magnification of the qso images is in general lower and the relative magnification of the third image compared to the brightest becomes larger than in the previous cases . in the previous section we restricted ourselves to modeling the system taking into account the positions and fluxes of the two observed qso images as constraints .
we have also measured the rotational velocity of the lens galaxy and can use it as an additional constraint . for each model
we compute the rotational velocity predicted at 2.7 @xmath44 , where @xmath44 is the scale radius of the exponential distribution of the galaxy light .
this is the value of the radius for which we can measure the rotational velocity with our spectroscopic data ( see 2.5 ) .
all rotational velocity comparisons are performed at this radius , neglecting the rotational velocity dependence with radius .
for the one mass component models , the best fitting sie model to the qso image positions and fluxes predicts a lower rotational velocity than observed , but consistent within the 1 @xmath81 error .
the exponential and elliptical nfw models predict larger rotational velocities than observed , which are consistent only at the 2 @xmath81 level .
if we include the rotational velocity in the overall minimization of the two component models , we find that for the sie+sis model the best fit occurs when the rotational velocities of the sis is approximately 20% lower than the rotational velocity of the sie .
unacceptable models happen when the rotational velocity of the sis is @xmath80 5% larger than the rotational velocity of the sie .
for the exponential+nfw case , we do not find acceptable models that satisfy the three constraints : positions and fluxes of the qso images and rotational velocity of the galaxy .
the general trend is that the larger the relative contribution of the spherical component ( halo ) , the lower the rotational velocity predicted to fit the positions and fluxes .
cy 2201 - 3201 is an edge - on galaxy lens .
such systems offer the possibility of decomposing the relative mass contributions of the disk , bulge and halo . in our optical images
cy 2201 - 3201 appears as a bulgeless edge - on spiral galaxy , producing two images of a background @xmath1 qso .
we have therefore modeled the system with one ( disk ) and two ( disk+halo ) mass components , neglecting the bulge . all viable models explored
produce either three or four visible images . however , our optical images only show two images of the qso ( fig [ fig2 ] ) .
we observe the two images that are farther away from the disk but miss the third ( and in some configurations the fourth ) image located very close to the disk .
we have investigated whether dust extinction could be responsible for this missing image .
the colors of the qso images a and b ( table [ tab2 ] ) are in fact redder than expected for a typical qso at that redshift .
although low in signal - to - noise ratioand covering a small wavelength range , the qso spectrum does not reveal indication of strong intrinsic absorption in the qso itself .
one is then led to conclude that the redder colors are due to dust extinction between the qso and the observer .
the most probable source of attenuation comes from the disk itself .
we have computed the amount of dust extinction at the galaxy redshift ( @xmath82 ) necessary to explain the observed colors assuming a mean qso spectrum at this redshift , no intrinsic absorption at the qso itself and the value of the local galactic extinction @xmath83 @xcite .
we find that the qso image a requires an extinction by the galaxy disk of @xmath84 and the image b an extinction of @xmath85 .
images a and b are seen at a projected distance of 0.35 and 0.47 ( 1.7 and 2.2 @xmath86 kpc ) of the disk plane respectively .
the inferred values of the extinction by the disk are then likely taking into account thee projected distances to the disk .
our @xmath21 band image is the one with the highest signal - to - noise detection of the qso images .
we have inserted a third image of the qso in our model images at the position given by our lens models and verify through our mcmc method the maximum flux it could have without being significantly detected .
if the ratio of expected flux to maximum observed flux are completely due to absorption by dust at the lens galaxy , then there are at least 3 mag of extinction more at the position of the third image than at the brightest . taking into account the possible extinction at the brightest image , the third image would suffer @xmath87 of extinction going through the lens galaxy . in our models of the lens ,
we have first explored the assumption that the mass distribution has the same ellipticity as the light . we have tried three one - mass component models fixing the galaxy center , axis ratio and position angle to those observed in the optical .
we have fit an sie , an exponential and an nfw model .
all three models reproduce the two qso images and predict a third image that is unobserved .
the predicted rotational velocities are consistent with the observed value at the 1 @xmath81 level for the sie model and at the 2 @xmath81 level for the exponential and elliptical nfw models .
we have also tried other methods to estimate the galaxy mass .
we have fit the observed galaxy photometry with the pegase synthesis evolutionary models @xcite .
we obtain an absolute magnitude of @xmath88 , adding an evolutionary correction . ] which is a factor 3 fainter than l@xmath89 @xcite .
the stellar mass - to - light ratio(@xmath90/@xmath91 ) of the best fitting model is @xmath92 , so its stellar mass would be @xmath93 @xmath94 @xmath95 . the expected rotational velocity for such a mass at 2.7 times the scale radius is @xmath96 km s@xmath3 ( depending on the mass model ) . the stellar mass by itself is thus insufficient to produce the observed ( or predicted ) rotational velocity .
if we assume the local tully - fisher relation @xcite neglecting evolution and the measured absolute magnitude , we obtain a value for the expected rotational velocity of @xmath97 km s@xmath3 , which is consistent with the measured rotational velocity .
we have also studied more realistic models .
the galaxy does not appear to have a bulge in the optical images , and therefore we have modelled the system with two mass components : one for the disk and one for the halo .
we have tried an sie+sis model and an exponential+nfw model .
these models reproduce the positions and fluxes of the two observed qso images . however , the sie+sis model predicts four visible images and the best fitting exponential+nfw predicts three visible images .
none of these predicted additional images are seen in our images ( see above ) .
we measure a rotational velocity of @xmath2 km s@xmath3 at 2.7 disk scale radius for the lens galaxy .
the sie+sis model predicts values of the rotational velocity consistent with this value for certain combinations of the relative contributions of the sie and sis components ( see 3.3 ) .
however , the exponential+nfw model predicts higher values of the rotational velocity if we fit the observed position and fluxes of the qso images .
our magellan spectroscopic data were obtained on three different nights with two different settings . as explained in 2.5 , we have four sets of same night / grating data .
the individual fits to the rotational velocity at 2.7@xmath44 for each set of data are @xmath98 , @xmath99,@xmath100 , and @xmath101 ( low - resolution september 7 , low - resolution september 9 , high - resolution september 7 and high - resolution september 8 , respectively ) .
the relative dispersion of these values may hint at a possible underestimation of the errors .
it this were the case , the range of allowed @xmath46 values would be larger and the exponential+nfw model prediction would still be viable .
we have also explored other effects that can affect our modeling .
our system is likely to be influenced by some external shear which will contribute to the image separation but not to the rotational velocity of the lens galaxy .
in fact , cy2201 - 3201 lies 7 away from the hickson compact group hcg 90 .
we have computed what would be the external shear produced by the group assuming it is modeled with a sie with the same velocity dispersion as measured from the galaxy members .
hcg 90 is relatively small and the external shear induced in the cy 2201 - 3201 system is negligible for our purposes .
apart from hcg 90 , cy 2201 - 3201 appears to be isolated and not in any group , cluster , or obvious large - scale structure .
in fact , as part of the cyder survey we have obtained spectra of several sources in the field and , with the limited spectroscopic data we have , not found any sign of a massive structure . another possible flaw in our modeling
could be that the galaxy center is miscalculated .
if the galaxy center is much closer to the qso images than the position we have measured then most of the discussed configurations would no longer apply and the system would display other image configurations , which can in fact place tighter constraints on the relative contribution of the halo and disk to the total mass budget in the central regions of the lens galaxy .
however , the seeing and pixel size of our images and the consistency of the galaxy center in our different filter images make us believe that the true center if different from the one measured should not be very far off .
we have presented the discovery and subsequent follow up observations of the cy2201 - 3201 system composed of an edge - on spiral at @xmath82 splitting a background @xmath102 qso into two observed images each at opposite sides of the disk .
we have modeled the system with one ( disk ) and two ( disk+halo ) mass components .
the most likely configuration is the `` disk''configuration with three images of the qso one at each side of the disk and the fainter one approximately behind the disk .
there are also possible configurations that produce four or even five ( this one unlikely ) observable images .
however we only observe two images .
we have discussed the possibility that the third ( and fourth , if existent ) image is extincted by the disk .
we estimate that an @xmath87 at th epredicted position of the third image at the galaxy lens redshift is required to be consistent with our observations .
we have measured the rotational velocity of the lens galaxy to be @xmath103 km s@xmath3 at a radius 2.7 times the scale radius of the galaxy exponential light distribution .
if we use an sie+sis model to fit the qso image positions and fluxes and this value of the rotational velocity , we find that the contribution by the sie ( disk ) to the @xmath46 at this radius is required to be the same or larger than the contribution of the sis ( halo ) .
if we use an exponential+nfw model then we are unable to reproduce this value of the rotational velocity if we fit the positions and fluxes of the qso images .
we have speculated whether we have underestimated the error in our determination of the @xmath46 which would make the exponential+nfw model viable .
cy 2201 - 3201 is the best lensing spiral galaxy known to date that can be used to disentangle the contributions of its different mass components .
unfortunately our current follow up observations are not constraining enough to elucidate between different possible models .
more accurate image source positions , the discovery or not of the predicted third image and the measurement of its properties and a precise rotational velocity would make this system fulfill its potential .
cy 2201 - 3201 has been awarded _ hubble space telescope _ ( hst ) acs time in cycle 13 .
we have also been awarded more spectroscopic time at magellan .
we expect that the higher quality images and extra spectra will help us improve our modeling and place strong constraints on the relative contribution of the disk and halo mass components .
we thank c. keeton for making public his gravlens code and replying to our questions .
we thank paul schechter for his encouragement and helpful discussions .
we thank all observatory staff for their help during observations .
we thank the anonymous referee for his / her comments that have helped us improve the paper . f.j.c .
acknowledges support from the spanish ministerio de educacin y ciencia ( mec ) , project aya2005 - 09413-c02 - 01 with ec - feder funding and from the research project 2005sgr00728 from the generalitat de catalunya .
j.m . gratefully acknowledges support from the chilean centro de astrofsica fondap 15010003 .
f.j.c . and j.m .
acknowledge support from a `` convenio bilateral csic - universidad de chile '' .
e.g. is supported by the national science foundation under grant ast 02 - 01667 . | we present the cxocy j220132.8 - 320144 system , which is composed of an edge - on spiral galaxy at @xmath0 lensing a @xmath1 background quasar .
two images of the quasar are seen .
the geometry of the system is favorable to separate the relative mass contribution of the disk and halo in the inner parts of the galaxy .
we model the system with one elliptical mass component with the same ellipticity as the light distribution and manage to reproduce the quasar image positions and fluxes .
we also model the system with two mass components , disk and halo .
again , we manage to reproduce the quasar image positions and fluxes .
however , all models predict at least a third visible image close to the disk that is not seen in our images . we speculate that this is most likely due to extinction by the disk .
we also measure the rotational velocity of the galaxy at 2.7 disk scale radius to be @xmath2 km s@xmath3 from the [ oii ] emission lines . when adding the rotational velocity constraint to the models , we find that the contribution to the rotational velocity of the disk is likely to be equal to or larger than the contribution of the halo at this radius .
the detection of the third image and a more accurate measurement of the rotational velocity would help to set tighter constraints on the mass distribution of this edge - on spiral galaxy . |
the aim of this work is to present more evidence that excitation spectra of complex open - shell atoms , and probably any other atom at sufficient excitation energies , display clear quantum chaotic features .
this phenomenon is caused by strong mixing of many - electron excited states by the residual two - body coulomb interaction .
it manifests in particular in a gaussian statistics of the @xmath0 amplitudes for these states . since the time of bohr s hydrogen atom theory atoms were considered as perfectly regular dynamical systems .
as the classical theory of chaos evolved it became apparent that highly excited atomic states in the rydberg range could become chaotic if an external field is applied @xcite , as long as the underlying classical motion is chaotic .
on the other hand , it was also due to bohr that the notion of compound nuclei was introduced in physics .
the behaviour of these highly excited nuclear states is essentially quantum - mechanical .
nevertheless , they display a number of chaotic properties .
for example , the statistics of their energy spectra show certain universal features , and transition amplitudes involving compound states obey gaussian statistics @xcite . to describe these properties
it was suggested by wigner that the hamiltonian of a compound nucleus could be modeled by a random matrix , and different characteristics found by averaging over ensembles of such matrices ( see reviews @xcite ) .
the first insight into quantum chaotic properties of complex atoms was given by rosenzweig and porter @xcite who analyzed experimental spectra of some neutral atoms and showed that in heavy open - shell atoms the spectral statistics are similar to those of compound nuclei .
that analysis was later extended and refined in @xcite .
of course , the study of eigenvalues provides valuable information about the system . on the other hand
, the spectral statistics observed in heavy open - shell atoms are similar to those of the hydrogen atom in a strong magnetic field @xcite , or even a particle in a 2-dimensional classically ergodic billiard @xcite .
however , the eigenstates of these quantum systems must be completely different , and it is clear that the eigenvalue statistics can not really tell us much about the origin of chaotic behaviour , or indeed the structure of the chaotic eigenstates .
the first inquiry into the possibility of chaos in the eigenstates of complex atoms was done by b. chirikov @xcite .
he studied configuration compositions of eigenstates of the ce atom using data from the tables @xcite , and came to the conclusion that the ` eigenfunctions are random superpositions of some few basic states ' .
inspired by that work we conducted an extensive numerical study of the spectra and eigenstates of complex open - shell atoms , using the rare - earth atom of ce as an example @xcite .
this allowed us to investigate many - body quantum chaos in a real system .
we showed that atomic excited states are in fact similar to nuclear compound states and developed a statistical approach for analyzing their properties . unlike eigenvalues ,
the eigenfunctions are not observable directly . to probe the structure of the chaotic eigenstates one can look at the transition probabilities or matrix elements of some external perturbation coupling them to each other , or to regular , simple eigenstates ( like the ground state ) .
the matrix elements involving chaotic eigenstates must have gaussian statistics .
we showed that its main characteristics the mean squared value of the matrix element between the chaotic multiparticle states ( compound states ) , can be calculated in terms of statistical parameters of the eigenstates and single - particle amplitudes and occupation numbers of the orbitals present in the compound states @xcite . in this work
we have chosen the quantity most easily accessible experimentally the @xmath0 amplitudes .
it also gives us an opportunity to look for experimental signatures of chaos in the ce atom using the work by bisson _
et al _ @xcite , where over 200 line strengths were measured for transitions between a large number of levels within 3.5 ev of the ground state .
it should be mentioned that there are many other possible atomic systems to search for quantum chaos , e.g. , in doubly excited states and inner - shell excitation spectra of alkaline - earth atoms @xcite , or even multiply excited states of light atoms @xcite .
let us now recall briefly what chaotic many - electron atomic eigenstates are .
suppose one uses a basis of some single - electron orbitals ( e.g. , the hartree - fock ones ) to construct many - electron basis states @xmath3 .
the states @xmath4 can be taken as single - determinant states corresponding to certain configurations of a few valence electrons , or constructed from them through some coupling scheme to be of definite total angular momentum @xmath5 .
the true atomic eigenstates @xmath6 and eigenvalues @xmath7 are obtained by diagonalizing the hamiltonian matrix @xmath8 .
the coefficients @xmath9 describe mixing of the basis states by the residual coulomb interaction . in the multi - electron excitation range the number of basis states @xmath3 formed by distributing several electrons among a few open orbitals is large . many of these states are nearly degenerate and the mean spacing between the basis state energies @xmath10 is likely to be smaller than the typical value of the off - diagonal matrix element @xmath11 . in this situation
the basis states are strongly mixed together @xcite .
apart from a few lowest levels , each of the eigenstates is a superposition of a large number of basis states . of course ,
by a simple perturbation theory argument , the mixing must be weak for distant basis states ( large @xmath12 ) .
the strong mixing takes place within a certain energy range @xmath13 , where @xmath14 is the mean level spacing , @xmath15 , and @xmath16 is called the _ spreading width _ , since it characterizes the spread of the eigenstates to which a given basis state contributes noticeably .
one can estimate the number of _ principal components _
, i.e. those that contribute significantly to a given eigenstate ( [ a ] ) , as @xmath17 .
the coefficients @xmath9 corresponding to the principal components have typical values @xmath18 .
their statistics is close to that of independent random variables , and tends towards gaussian when the mixing is strong . in this case
even the single - electron orbital occupancies are far from integer and only the total angular momentum , parity and the energy itself remain good quantum numbers @xcite .
thus , we can talk about _ quantum chaos _ in the system .
this situation is similar to that in compound nuclei and the corresponding chaotic eigenstates can be called atomic compound states .
the model configuration interaction calculations performed for ce produced a value of @xmath19 ev , and demonstrated the existence of a dense spectrum of chaotic compound excited states with @xmath20 ( @xmath21 ev ) just few ev from the ground state @xcite .
consider two chaotic many - body states ( compound states , for short ) that are superpositions of large numbers of basis states , @xmath22 and @xmath23 .
if the expansion coefficients @xmath9 are random , the matrix element of some operator @xmath24 @xmath25 is a sum of a large number of almost uncorrelated random items @xcite .
therefore , one should expect that such matrix elements display gaussian statistics with zero mean .
hence , the probability distribution of the matrix elements between compound states can be characterized by their mean squared value alone .
if @xmath24 is a single - particle operator , e.g. , the electric dipole moment @xmath26 ( @xmath27 and @xmath28 are single - particle states ) , it is convenient to express its matrix elements in terms of the matrix elements of the density matrix operator @xmath29 , @xmath30 where @xmath31 . in @xcite and @xcite a statistical approach to calculation of mean squared matrix elements between compound states has been developed .
it is first based on the assumption that contributions from different single - particle transitions @xmath32 in the matrix element ( [ mel ] ) are uncorrelated . the mean squared value is then given by @xmath33 where averaging is done over a number of compound states around @xmath34 and/or @xmath35 .
the mean squared value of the density matrix operator @xmath36 is expressed in terms of the parameters of the compound states 1 and 2 ( i.e. their energies and spreading widths ) , and the average occupation numbers of the single - particle states @xmath27 and @xmath28 . in a spherically symmetric system where the states 1 and 2 are characterized by their total angular momenta @xmath37 and projections @xmath38 , the wigner - eckhart theorem applies , and it is convenient to deal with the reduced matrix elements @xmath39 independent of the projections @xmath38 . for example , the mean squared value of the zero - rank reduced density matrix operator ( @xmath40 then ) is obtained in the following two forms @xcite : @xmath41 where @xmath42 are the mean level spacings near the states 1 and 2 , @xmath43 and @xmath44 are the orbital occupation numbers , and @xmath45 is a `` finite - width @xmath46 function '' .
it depends on the spreading widths @xmath47 of the compound states and on the energy difference @xmath48 between the transition frequency for the compound many - electron states @xmath49 and the frequency @xmath50 of the single - particle transition between the orbitals @xmath51 and @xmath52 .
the function @xmath45 has a maximum at @xmath53 and describes the energy conservation for the compound states .
its width is determined by the spreading widths @xmath54 .
note that @xmath55 in eq .
( [ answ0 ] ) denote averaging of the occupation - number factors over the compound states 1 or 2 .
note also that the exact form of the function @xmath56 depends of the spreading of the compound states over the basis components , i.e. on the `` shapes '' of the eigenstates . in the simplest approximation
this spreading is described by the breit - wigner formula ( see numerical studies in @xcite ) and @xmath45 is also a breit - wigner profile @xmath57 to calculate the mean squared value of the @xmath0 amplitude we now need a formula for the reduced density matrix operator of the first rank .
starting from the definition @xcite @xmath58 for @xmath59 ( linear polarization along the quantization axis ) and assuming that transitions between different magnetic sublevels @xmath60 are uncorrelated we can derive a formula for the mean square of ( [ rank1 ] ) , and then use it to obtain the mean - squared @xmath0 amplitude , @xmath61 analogous to the lower formula in eq .
( [ answ0 ] ) , or an alternative form with @xmath62 and @xmath63 on the right - hand side .
the factor @xmath64 on the right hand side of eq .
( [ answe1 ] ) is due to the fact that there are three final - state momenta @xmath65 accessible from a given @xmath66 by means of a dipole transition . in deriving this expression
an additional assumption has been made that the occupancies of the @xmath67 and @xmath68 states are statistically independent , and the states with different @xmath60 within the same @xmath51 shell are equally populated .
this supposition influences only the `` emptiness '' factors @xmath69 , which are close to unity anyway when the number of single - electron states available is much greater than the number of active electrons .
the square of the reduced dipole matrix element @xmath70 is called the strength of the line @xmath71 , so eq . ( [ answe1 ] ) allows one to estimate _
mean line strengths _ for transitions involving compound states .
it is interesting to note that the statistical theory expression ( [ answe1 ] ) satisfies the dipole sum rule @xcite ( in atomic units ) , @xmath72 is the number of active valence electrons included in the configuration space of the problem . to obtain this result
one should replace summation over the final states @xmath73 with integration over @xmath74 , take into account that @xmath75 [ see eq .
( [ bw ] ) ] , neglect the `` emptiness '' factor @xmath76 and use @xmath77 , and rely on the single - particle sum rules for the orbitals @xmath78 occupied in the initial state @xmath34 , @xmath79 numerical results for the ce atom --------------------------------- cerium , @xmath80 , is the second of the lanthanide atoms .
its electronic structure consists of the xe - like @xmath81 core and four valence electrons .
the atomic ground state is described by the @xmath82 configuration with @xmath83 @xcite .
the origin of the extremely complex and dense excitation spectra of the rare - earth atoms is the existence of several open orbitals near the ground state , namely @xmath84 , @xmath85 , @xmath86 , and @xmath87 , or , in relativistic notation , @xmath88 , @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 , and @xmath94 .
these make a total of @xmath95 single - electron states . for ce with @xmath96 valence electrons
there are about @xmath97 possible many - electron states constructed of them .
if we allow for the two possible parities , about ten possible total angular momenta @xmath5 , and @xmath98 different projections ( another factor of ten ) , there will be still hundreds of energy levels within a given @xmath99 manifold .
in the present work we perform relativistic configuration interaction calculations in the hartree - fock - dirac basis analogous to those in @xcite . in that work we limited ourselves to just 7 nonrelativistic configurations constructed of the @xmath84 , @xmath85 , @xmath86 , and @xmath87 orbitals , for both odd and even states , which produced 260 and 276 states with @xmath83 and @xmath100 , respectively . to make the results more realistic
we have extended the configuration basis set by 9 odd and 23 even nonrelativistic configurations .
basically , the additional configurations were obtained by exciting one of the four electrons of an `` old '' configuration into the next orbital , e.g. , the lowest even @xmath101 configuration would produce @xmath102 , @xmath103 , @xmath104 , and @xmath105 configurations . to keep the size of the configuration space reasonable we included only the configurations with mean energies within about 10 ev from the ce ground state .
this increased the total number of @xmath106 and @xmath100 states to 862 and 1433 , respectively .
note that @xmath2 states have been chosen because these manifolds are among the most abundant . as a result
, the level density @xmath107 has increased greatly above 4 ev and become closer to that observed experimentally .
of course , to be meaningful the level density must be averaged over some small energy interval to obtain a smooth function rather than a set of spikes .
an alternative procedure is to look at the cumulative number of levels @xmath108 which we present in fig .
[ fdens ] for @xmath109 states .
each @xmath110 plot is a staircase of steps of the unit height occurring at successive excited state energies .
the level density can be easily estimated from the slope of the @xmath110 plot .
the experimental data for the 132 even levels with @xmath2 known from @xcite is shown by the solid - line staircase , and the energies are given with respect to either experimental , or calculated ground state energy .
they can be compared with the dashed line that shows @xmath110 for our earlier small - basis calculation @xcite ( 276 states ) , and the dotted line for the present calculation ( 1433 states ) .
the improvement is obvious , however , the agreement is not perfect .
we believe that the remaining disagreement is not due to some missing configurations in the ci calculation , but rather due to an overall `` softening '' of the spectra due to screening of the coulomb repulsion between the valence electrons by the electrons of the core @xcite . in the ci language
this effect is produced by the high - energy excitations of the valence electrons into the continuum together with the electron excitations from the core .
two typical features can be observed in the spectra of complex atoms @xcite .
the first clearly seen in fig .
[ fdens ] , is the rapid increase of the level density @xmath111 with energy @xcite .
its origin is purely combinatorial the larger the excitation energy , the greater the number of ways it can be distributed among a few single - particle excitations . in the independent - particle model
this dependence is described by the following exponent @xcite : @xmath112 where @xmath113 and @xmath114 are some constants , and @xmath115 is the ground state energy of the system .
this dependence also follows from the thermodynamic definition of the temperature , @xmath116\ } /de$ ] , combined with the estimates of the average number of excited fermi particles , @xmath117 , and that of the excitation energy , @xmath118 .
the experimental spectra of rare - earth atoms and their ions examined in @xcite are in agreement with eq .
( [ rho ] ) .
figure [ fdens ] shows that the calculated cumulative level number plot is fitted well by @xmath119 with @xmath120 ev@xmath121 , @xmath122 ev@xmath123 , and the `` ground state '' energy of the @xmath100 sequence @xmath115 shifted by 0.25 ev up from the true @xmath1 ground state of ce .
thus , eq . ( [ rho ] ) gives a good overall fit of the calculated level density below 6 ev .
the second feature typical for the spectra of complex many - body systems is level repulsion .
it is a basic quantum mechanics fact that two levels with identical quantum numbers can not be degenerate if they are coupled by a non - zero matrix elements they `` repel '' each other . in quantum chaotic systems this repulsion is characterized by the wigner level spacing distribution @xmath124 where @xmath125 is the nearest - neighbour level spacing normalized so that @xmath126 . equation ( [ wigner ] ) shows that the probability of finding small level spacings is indeed vanishingly small . as we pointed out in the introduction , spectral statistics do not tell much about the eigenstates of the system . however , eq .
( [ wigner ] ) is still a good test for some possible hidden quantum numbers , e.g. , the total spin or orbital momentum , which might characterize atomic eigenstates besides @xmath127 .
if these do exist , small level spacings ( `` degeneracies '' ) will be more abundant than predicted by eq .
( [ wigner ] ) .
these statistics were checked for many experimental @xcite and calculated @xcite complex atomic spectra , as well as for molecular vibronic spectra @xcite . as seen from fig .
[ fdens ] the level density changes significantly for the first 500 levels of the calculated spectrum . to analyze the distribution of the corresponding level spacings we use the analytical density fit @xmath128 to normalize the spacings :
@xmath129 their distribution shown on the inset in fig .
[ fdens ] is in reasonable agreement with the wigner formula .
the deviations are probably due to the long - range fluctuations of the level density , not accounted for by the simple exponential ( [ rho ] ) . in the previous calculation @xcite , where only the lowest orbitals of each symmetry were included
, we also observed the wigner distribution .
when orbitals with higher principal quantum numbers become involved ( as seen from fig .
[ fdens ] above 3.5 ev ) the spatial extent of the eigenstates increases .
this should cause a decrease of the residual coulomb interaction between the electrons . on the other hand ,
the level spacings also become smaller . as a result , the state mixing at these excitation energies remains strong , which is confirmed by the agreement with the wigner distribution , and the eigenstates are chaotic .
our estimate of the number of principal components @xmath130 shows that it becomes even greater as the energy increases , in accord with the estimate @xmath131 ( @xmath132 ev , and the mean level spacing @xmath133 ev at @xmath134 ev ) . in sec .
[ mael ] we explained that matrix elements involving chaotic compound states should have gaussian statistics , and the mean squared value of the matrix elements could be estimated in terms of some average characteristics of the compound states . in this section
we concentrate on the dipole matrix elements ( @xmath0 amplitudes ) @xmath135 between the 14 lowest states with @xmath136 and 80 consecutive @xmath100 states obtained numerically in our ci calculations of ce .
we have chosen this energy region to cover the range explored in the experiment @xcite , where absolute values were derived for 228 of the most intense lines of neutral ce between 10706 and 22184 cm@xmath121 .
of course , low - lying atomic states , e.g. , the ground state , have well - defined configuration composition and are not chaotic , hence , the @xmath0 amplitudes between them should not be distributed in any particular statistical way .
however , the matrix elements ( [ m21 ] ) will become random ( and close to gaussian ) as soon as at least one of the states involved , the initial or the final , moves into the compound - state energy range and becomes a superposition of many random components .
besides that , the mean squared value of the matrix element is expected to show some smooth secular variation with the energy of the states involved . for these reasons
we skip the first 20 states with @xmath138 and analyze the statistics of the @xmath139 @xmath0 amplitudes for the following 80 even states by grouping them in bunches of twenty : 2140 , 4160 , 6180 , and 81100 , which correspond to the mean excitation energies of 2.49 , 2.95 , 3.40 , and 3.70 ev above the atomic ground state ( the mean energy of the lowest 14 odd states is 0.68 ev ) . thus , each plate in fig .
[ fcomp ] shows the distribution of the 280 reduced dipole matrix elements together with their root - mean - square ( r.m.s . )
value . also shown in fig .
[ fcomp ] are the gaussian distributions @xmath140 , where the root - mean - square parameter @xmath141 has been adjusted to minimize @xmath142 around the center of the histogram .
the values of @xmath141 and @xmath142 are given in table [ comp ] .
two effects can be seen in fig .
[ fcomp ] .
first , the distributions of the matrix elements are indeed close to gaussian .
second , the width of the distributions ( the mean squared value of the matrix elements ) varies with the energy of the even states .
it is mostly this effect that is responsible for the visible discrepancies between the histograms and the gaussian fits . to eliminate it
we can use a running average procedure to normalize the amplitudes : @xmath143 where @xmath144 is the r.m.s .
value over the 14 odd states , calculated for every even state @xmath145 .
figure [ fnorm ] confirms that the 1120 normalized @xmath0 amplitudes for the 21100 even states are distributed according to the normal law .
the inset shows the dependence of the r.m.s .
@xmath0 amplitude @xmath144 on the energy of the even state @xmath7 .
fluctuations aside , it is in agreement with the r.m.s .
values calculated from the statistical theory , eq .
( [ answe1 ] ) , at the energies of the 30th , 50th , 70th and 90th even states .
the numerical values of the r.m.s .
@xmath0 amplitudes are listed in table [ comp ] .
note that we have chosen eq .
( [ answe1 ] ) with 1 standing for the odd states and 2 for the even ones . in our numerical example
we consider the dependence of the r.m.s .
@xmath0 amplitude on the energy of the even states , and keep the odd states the same .
therefore , as in eq .
( [ answe1 ] ) , we only need to know the average occupation numbers for the lowest 14 odd states , and the result depends on the final even state via its energy @xmath146 , mean level spacing @xmath147 and spreading width @xmath148 . as we saw in our previous calculations @xcite , the even states of ce with
@xmath2 become very much chaotic at excitation energies of just 2 ev , i.e. from the 20th level up .
also , as earlier in @xcite , we use average configuration energies rather than single - particle hartree - fock energies to determine the transition frequencies @xmath149 needed for calculation of @xmath150 in eq .
( [ answe1 ] ) .
the ground state of ce is described as @xmath82 , however , the dominant configuration among the 14 lowest odd states is @xmath151 , and we used it to calculate the transition energies .
for example , the energy of the @xmath152 transition @xmath153 was determined as the difference between the average energies of the @xmath154 and @xmath151 configurations .
physically , this corresponds to choosing a particular mean field close to that of the low - lying odd states of ce for calculation of the transition energies .
it should be mentioned though that the results obtained with the hartree - fock frequencies @xmath155 were not too different .
gaussian statistics of the dipole matrix elements result in the porter - thomas ( pt ) distribution of the line strengths @xmath156 @xmath157 where @xmath158 is the mean line strength .
divergence of this function at small @xmath159 means that if the @xmath0 amplitudes are gaussian , there should be many weak lines in the spectrum .
earlier evidence of the pt statistics of line strengths can be found in calculations of dipole excitations in complex atoms @xcite , and transitions between the vibronic levels in molecules measured in @xcite .
in@xcite absolute values of @xmath160 were obtained for 228 of the most intense observed lines between 10706 and 22184 cm@xmath121 in ce .
it is interesting to analyze these data to see whether they support our theoretical and numerical considerations .
the values of @xmath160 listed in @xcite are defined as @xmath161 , where @xmath162 is the @xmath0 transition rate from the upper level @xmath163 into the lower level @xmath145 @xcite .
we use the experimental values of @xmath160 , @xmath164 and transition frequencies @xmath165 to extract values of the line strengths @xmath166 in fig . [ port ] the probability distribution of the 228 experimental line strengths is shown .
compared to the expected pt formula ( [ pt ] ) , there is a clear lack of small line strengths .
nevertheless , the decreasing part of the histogram can be fitted well by a pt distribution with an additional normalization factor @xmath167 , @xmath168 shown in fig .
[ port ] by a solid line for @xmath169 and @xmath170 a.u .
that minimize @xmath142 for the 22 bins with @xmath171 a.u .
it would be tempting to say that the excellent agreement between the pt curve and the histogram is a confirmation of the gaussian statistics of the @xmath0 amplitudes in ce .
the value of @xmath167 would then indicate that about one half of all lines are missing in the experimental data .
however , the value of @xmath170 a.u . corresponds to the r.m.s .
@xmath0 amplitude of 1.8 , which is more than 2 times greater than our numerical results in fig .
[ fcomp ] and fig .
[ fnorm ] ( inset ) and in table [ comp ] . on the other hand ,
the experimentally observed 228 lines include transitions between levels with various total angular momenta between @xmath172 and 8 ( @xmath173 , of course ) whereas we have about 500 hundred lines with just @xmath174 in our calculation in the analogous energy range .
this means that in @xcite only the strongest 10% or less of all lines have in fact been measured .
the very suggestive agreement with the pt distribution in fig . [ port ] should then be considered as merely fortuitous .
it is worth noting that in experiment the lines are selected by their intensities proportional to @xmath160 , rather then strengths .
hence , even lines with large strengths can be omitted if their frequencies are small .
let us look at the simplest model of this effect and see how it influences the observed strengths distribution .
assume that transitions in a certain frequency range @xmath175 are studied , and different values within this interval are equally probable .
the observed intensities of the lines are proportional to @xmath176 .
if we assume that there is a minimal threshold intensity that can be registered , the original pt distribution of strengths would be modified as follows : @xmath177 ~,\quad s > s_0\cr } \end{aligned}\ ] ] where @xmath178 is the minimal strength that can be observed at @xmath179 , and @xmath167 is the normalization factor . as seen from fig .
[ port ] , eq .
( [ mpt ] ) also gives a very good fit of the experimental data with @xmath180 , @xmath181 and @xmath182 @xcite .
note , however , that the new value of the mean line strength is 1.5 times smaller than the one we had from the pure pt fit .
therefore , the assumptions used in our processing of the experimental data affect the estimates of the experimental r.m.s .
@xmath0 amplitudes , and we should not be too concerned about the apparent disagreement with our numerical calculations . besides that , extraction of absolute line strengths from the experimental data is not free from uncertainties estimated in @xcite at 1020% . for 30 transitions in ce the @xmath160 values were obtained more accurately from branching ratios and delayed photoionization measurements of lifetimes ( @xcite , table 2 ) . when we look at the statistics of the corresponding line strengths ( fig .
[ port ] , inset ) and compare it with the pt distribution ( [ pt ] ) , a value of @xmath183 is obtained , much smaller than the estimates of @xmath158 from the statistics of the 228 lines .
thus , it appears that to make firm conclusions about gaussian statistics of the @xmath0 amplitudes a much more thorough experimental survey is needed . on the other hand ,
even relative measurements of a large number of line strengths could be very valuable for examining these statistics @xcite .
in this work we have extended the configuration interaction approach of @xcite to calculate large numbers of eigenstates in ce . in agreement with our earlier studies
the energy level statistics indicate that the simple configurational basis states are strongly mixed together by the residual electron interaction , and the only good quantum numbers in the spectrum are parity and the total angular momentum . the total orbital momentum @xmath184 and spin @xmath159 are not conserved due to the spin - orbit interaction , whose effect is dynamically enhanced , just as that of any other perturbation in a chaotic many - body system @xcite .
the strong configuration mixing makes multielectron atomic eigenstates chaotic .
this in turn results in a gaussian statistics of the matrix elements for chaotic atomic eigenstates ( compound states ) .
this understanding is fully confirmed by our numerical calculations of the 1120 @xmath0 amplitudes between the 14 lowest @xmath83 states and 80 @xmath138 states above 2 ev .
it is important that the parameter of the gaussian , the r.m.s .
@xmath0 amplitude , varies slowly with the excitation energy .
this effect should be taken into account when analyzing the statistics of the matrix elements .
we also show that a statistical theory can be used to estimate mean squared matrix elements involving compound states .
it enables one to express the answer in terms of the single - particle matrix elements and occupation numbers , and parameters of the compound states , namely the number of principal components and the spreading width .
this approach has already been applied to calculation of matrix elements between compound states in nuclei @xcite .
it could be useful in various other many - body systems , e.g. , atomic clusters or quantum dots , where direct diagonalization of the hamiltonian matrix is not feasible because of a huge size of the hilbert space of the problem .
an attempt has been made to analyze existing experimental data for the line strengths in ce @xcite .
it appears that the statistics of the measured line strengths is compatible with the porter - thomas distribution , with allowance for the missing weak lines .
however , the discrepancy between the calculated r.m.s . @xmath0 amplitudes and those inferred from the experimental data does not allow us to say that the existence of quantum chaos in the ce eigenstates has been confirmed experimentally . to make this statement one would have to do a much more complete survey and statistical analysis of the line strengths in the ce spectrum .
on the other hand , this means that a comparison between the experimental and theoretical line strengths in ce is not yet possible , even at the level of their mean values .
theoretically , to calculate precisely the dipole matrix elements between particular levels in the compound - state energy range of complex atoms like ce looks a prohibitively difficult problem . experimentally ,
identification of specific lines in enormously complicated spectra is also a very difficult task .
however , we would like to suggest that extraction of mean characteristics from the experiment and comparison with the corresponding theoretical estimates is a meaningful way of exploring such complex systems . as a result
, one might hope to get a deeper insight into the existence of quantum chaos in many - body systems on the whole , and in complex open - shell atoms , in particular .
10 [email protected] , a selection of papers compiled and introduced by g. casati and b. v. chirikov ( university press , cambridge , 1995 ) , part 2 ; see also m. courtney and d. kleppner , phys . rev . a * 53 * , 178 ( 1996 ) and references therein .
a. bohr , and b. mottelson , _ nuclear structure _
( benjamin , new york , 1969 ) , vol . 1 . t. a. brody , j. flores , j. b. french , p. a. mello , a. pandey and s. s. m. wong , rev . mod
53 * , 385 ( 1981 ) .
t. guhr , a. mller - groeling , and h. a. weidenmller , phys .
( to be published ) , preprint cond - mat/9707301 .
n. rosenzweig and c. e. porter , phys . rev .
* 120 * 1698 ( 1960 ) .
s. camarda and p. d. georgopulos , phys .
lett . * 50 * , 492 ( 1983 ) .
d. delande and j. c. gay , phys .
lett . * 57 * , 2006 ( 1986 ) .
m. v. berry , ann .
phys . * 131 * 163 ( 1981 ) ; o. bohigas , m. j. giannoni , and c. schmit , phys .
* 52 * , 1 ( 1983 ) .
b. v. chirikov , phys . lett . *
108a * , 68 ( 1985 ) .
w. c. martin , r. zalubas and l. hagan , _ atomic energy levels - the rare - earth elements _ , natl . bur .
stand . ref .
( u.s . ) , nbs-60 , ( u.s . , gpo , washington , dc , 1978 ) .
v. v. flambaum , a. a. gribakina , g. f. gribakin , and m. g. kozlov , phys .
a * 50 * , 267 ( 1994 ) .
a. a. gribakina , v. v. flambaum and g. f. gribakin , phys .
e , * 52 * , 5667 ( 1995 ) . v. v. flambaum , a. a. gribakina and g. f. gribakin , phys .
a , * 54 * ( 1996 ) . v. v. flambaum and o. k. vorov , phys .
* 70 * , 4051 ( 1993 ) ; v. v. flambaum , in _ time reversal and parity violation in neutron reactions _ , edited by c. r. gould , j. d. bowman and yu .
p. popov ( world scientific , singapore , 1994 ) , p. 39 .
s. e. bisson , e. f. worden , j. g. conway , b. comaskey , j. a. d. stockdale , and f. nehring , j. opt .
8 * , 1545 ( 1991 ) .
j. p. connerade , m. a. baig , and m. sweeney , j. phys .
b * 23 * , 713 ( 1990 ) .
j. p. connerade , i. p. grant , p. marketos , and j. oberdisse , j. phys .
b * 28 * , 2539 ( 1995 ) .
j. p. connerade , j. phys .
b * 30 * , l31 ( 1997 ) . n. vaeck and n. j. kylstra , in xx icpeac , abstracts of contributed papers , edited by f. aumayr , g. betz , and h. p. winter ( unpublished ) , vol . 1 , th128 .
if the number of active electrons in the system is more than two , the hamiltonian matrix @xmath11 appears to be sparse , i.e. , there is a certain fraction of zero off - diagonal matrix elements in it ( only the basis states that differ by the positions of no more than two electrons can be coupled by the two - body coulomb interaction ) . in this case
one should compare the magnitude of @xmath11 with the mean `` distance '' @xmath185 between the basis states directly coupled by non - zero @xmath11 .
see , e.g. p. jacquod and d. l. shepelyansky , phys .
lett . * 79 * , 1837 ( 1997 ) ; b. georgeot and d. l. shepelyansky , _ ibid _ * 79 * , 4365 ( 1997 ) and references therein .
there are some specific correlations caused by the two - body character of the electron interaction , but they do not change the statistics of the matrix elements . however , they can influence the estimate of the mean square value of the matrix element , see v. v. flambaum , g. f. gribakin , and f. m. izrailev , phys .
e * 53 * , 5729 ( 1996 ) .
see , e.g. v. a. dzuba , v. v. flambaum , and m. g. kozlov , phys .
a , * 54 * , 3948 ( 1996 ) , and references therein . note that there are no rydberg series in our ci calculation for ce , and only the @xmath125 , @xmath186 , @xmath187 , and @xmath188 orbitals with the two lowest principal quantum numbers are considered .
of course , the inclusion of single - particle rydberg excitation series would give infinite density peaks at the positive ion thresholds .
however , the extended large - radius rydberg states are physically different from the compact compound states , and they decouple from the compound state spectra at high @xmath189 , a. a. gribakina and g. f. gribakin , j. phys .
b , * 29 * , l809 ( 1996 ) . th .
zimmermann , h. @xmath190 , and l. s. cederbaum , phys .
* 61 * , 3 ( 1988 ) .
r. karazija , _ sums of atomic quantities and mean characteristics of spectra _ ( mokslas , vilnius , 1991 ) . i. s. sobelman , _ atomic spectra and radiative transitions _ ( springer - verlag , berlin , 1992 ) .
the normalization integral @xmath191 for this set of values is equal to 0.902 , as this fit , just as the pure pt fit , evidently fails to describe the large number of line strengths greater than 15 a.u .
observed in @xcite . | using a configuration interaction approach we study statistics of the dipole matrix elements ( @xmath0 amplitudes ) between the 14 lower states with @xmath1 and 21st to 100th even states with @xmath2 in the ce atom ( 1120 lines ) .
we show that the distribution of the matrix elements is close to gaussian , although the width of the gaussian distribution , i.e. the root - mean - square matrix element , changes with the excitation energy .
the corresponding line strengths are distributed according to the porter - thomas law which describes statistics of transition strengths between chaotic states in compound nuclei .
we also show how to use a statistical theory to calculate mean squared values of the matrix elements or transition amplitudes between chaotic many - body states .
we draw some support for our conclusions from the analysis of the 228 experimental line strengths in ce [ j. opt .
. am . *
8 * , 1545 ( 1991 ) ] , although direct comparison with the calculations is impeded by incompleteness of the experimental data .
nevertheless , the statistics observed evidence that highly excited many - electron states in atoms are indeed chaotic . |
if the topology of space - time is @xmath5 then the only known exact solution representing a black object is the uniform black string with horizon topology @xmath6 @xcite .
though this solution exists for all values of the mass it is unstable below a critical value , @xmath7 , as shown by gregory and laflamme @xcite .
horowitz and maeda @xcite argued that a uniform black string can not change its topology into a black hole in finite affine time , making the possibility of such a transition questionable .
they suggested the possibility of a transition to a nonuniform black string .
gubser @xcite showed the existence of nonuniform black string solutions .
the non - uniform black string solution in 6 dimensions was investigated numerically @xcite for a range of the mass values above @xmath7 .
nonuniform black string configurations do not exist for masses below the gregory - laflamme point , in the region where the uniform black string solution is unstable @xcite .
a natural candidate for a black object in this mass range is a black hole .
unfortunately , no exact black hole solutions are known in a 5 ( or more ) dimensional space with a compactified dimension . still , on the basis of physical intuition , a black hole solution should exist for very small values of the mass .
when the radius of the horizon is much smaller than the size of the compactified dimension , i.e. when @xmath8 , the black hole should be unaware of the compactification . then a myers - perry @xcite solution should become asymtotically exact .
indeed , a numerical solution , extending the myers - perry solution to larger values of the mass was found by harmark and obers @xcite and kudoh and wiseman @xcite . using general arguments
kol @xcite suggested that the black hole branch and the nonuniform black string branch meet at a point when the black hole fills the compact dimension .
further arguments for such a transition were presented in @xcite .
recently , we have studied a related problem , namely the existence of black holes in randall - sundrum @xcite theories .
we used @xcite an approximation scheme based on the expansion of solutions in the ratio of the radius of the horizon of the black hole to the ads curvature . in this paper
we employ a similar strategy by expanding the metric and other relevant quantities in the ratio of the two natural lengths associated with a black hole configuration , ( i ) the five dimensional schwarzschild radius , @xmath1 , associated with the mass @xmath9 and ( ii ) the compactification length , @xmath2 , defined as the proper circumference of the compact dimension in the region far away from the mass .
the dimensionless ratio of these two quantities serves as an excellent expansion parameter . as the solution is an even function of @xmath1 and @xmath2 we use @xmath10 as our expansion parameter .
such an expansion for add black holes has recently been proposed and harmark @xcite and gorbonos and kol @xcite evaluated the leading term of the expansion . to find an unique solution we must fix boundary conditions , both at infinity and at the horizon of the black hole .
we find solutions in two different regions . in the asymptotic region , @xmath11 an ` asymptotic solution ' , and in the near horizon region , @xmath12 a ` near solution ' is found .
using the asymptotic solution we satisfy asymptotic boundary conditions but the boundary conditions at the horizon can not be satisfied .
the near solution suffers from the opposite problem , as it can not be used to satisfy asymptotic boundary conditions . to solve those problems we combine the two solutions .
we match the asymptotic solution with the near solution in the region @xmath13 where both are valid .
this method was used in @xcite to find small black holes in randall - sundrum scenario , and in @xcite to find small black holes on cylinders .
the work in this paper is parallel to and agrees with previous results @xcite , but here we calculate the metric up to fourth order in @xmath14 ( second order in @xmath4 ) .
the organization of this paper is as follows : in section [ sec : method ] we describe the general method of our calculations , and list the parameters that will be calculated from the metric . in sections
[ sec : zeroth]-[sec : second ] we present the detailed calculations up to second order . in section [ sec : summary ] we summarize our results for the metric , up to second order , and discuss the possible scenarios for black holes of increasing mass .
a reader not interested in technical details should read the next section , ( [ sec : method ] ) and then proceed directly to the summary and discussions ( [ sec : summary ] ) .
as we indicated in the introduction we will investigate black hole solutions of the einstein equation in 5 dimensional space when one of the dimensions is compactified .
recently , this problem has attracted the attention of several groups .
harmark and obers@xcite introduced the relative tension of black holes as an order parameter , wrote down a generalized smarr formula @xcite , investigated the phase diagram for black objects , and studied the solutions numerically .
gorbonos and kol @xcite proposed an analytic approximation scheme based on the expansion in the ratio of the radius of the horizon and the compactification length .
this method is similar to the expansion method we used @xcite to investigate black holes in the randall - sundrum @xcite scenario .
we will follow a similar path here and calculate further terms of the expansion . by necessity
, the perturbation method is applied in two overlapping regions . in the asymptotic region , @xmath11
, @xmath1 is the smallest scale .
therefore , an expansion in @xmath14 is actually an expansion in @xmath1 and in every order the metric is a general function of the dimensionless coordinates @xmath15 , @xmath16 we refer to this solution as the ` asymptotic solution . ' in the near horizon region , @xmath12 , @xmath17 is the smallest scale .
therefore , an expansion in @xmath14 is actually an expansion in @xmath17 . in every order ,
the metric is a general function of the dimensionless coordinates @xmath18 , @xmath19 we refer to this solution as the ` near solution . ' in each region , when we calculate the @xmath20-th order terms , einstein s equations is solved in terms of two functions : a gauge function and a wave function , which satisfies a linear , ( inhomogeneous ) , master equation .
the differential operator for the master equation in the asymptotic region is @xmath21 where @xmath22 is a coordinate in the compact dimension and @xmath23 is a radial coordinate in 3-dimensional space . in the near region
the operator is @xmath24 where @xmath25 is a radial coordinate and @xmath26 is an angular coordinate in 4-dimensional space .
the driving term in the master equation for the @xmath20th order contribution depends on the solution in lower orders .
it determines an inhomogeneous solution in the wave function .
the additional homogeneous solution introduces new parameters that should be fixed by boundary conditions . in empty space , or for the uniform black string configuration ,
the cylinder is invariant under translations in the compact direction .
when a point mass is introduced translation invariance is broken , but a @xmath27 reflection symmetry remains around the location of the mass in the compact direction .
so , if we endow the compact direction with the coordinate @xmath28 $ ] and put the mass at @xmath29 then the solution is symmetric about @xmath30 and periodic in @xmath22 with period @xmath2 .
asymptotically ( far away from the mass ) , we assume that the metric is minkowski .
the leading order correction to the minkowski metric is of order @xmath31 . such a term provides , among others , the four - dimensional newtonian potential .
we must impose constraints on the near solution , as well .
the horizon of a small black hole has @xmath32 topology .
we require that there is no black string attached to the horizon .
consequently , the kretchmann scalar must be regular on the @xmath22 axis everywhere , when @xmath33 .
furthermore , the surface gravity must be constant , otherwise the horizon is not regular @xcite . to match the asymptotic and the near solutions we must work in the intermediate region @xmath34 , where @xmath35 is a four - space - dimensional radial coordinate . in this region the functions
@xmath36 can be expanded in the coordinates .
owing to the @xmath37 symmetry the components depend only on @xmath35 and @xmath26 , the angle between the compact direction and the four - dimensional sub - manifold
. then one can use a double expansion in @xmath38 and @xmath39 to write the metric as @xmath40 the expansion includes only even powers due to the @xmath27-symmetry as will be explained later .
globally , our expansion parameter is @xmath41 . in the two regions
@xmath4 has a different meaning .
the asymptotic solution is constructed as an expansion in @xmath1 , such that @xmath42 , so one keeps @xmath20 constant and solves the equations for all values of @xmath43 .
for example ,
minkowski metric corresponds to the zeroth order ( @xmath44 , @xmath45 ) approximation to our problem , with appropriate coefficients , @xmath46 .
the first order , @xmath47 , @xmath45 , provides linearized gravity ( newtonian potential ) , etc .
the near solution is constructed as an expansion in @xmath17 , such that @xmath48 , so at fixed @xmath43 we solve for all values of @xmath20 . here , @xmath49 , with appropriately chosen coefficients , @xmath46 , corresponds to the myers and perry five dimensional black hole ( mps ) @xcite .
the matching is done in the intersecting points denoted by circles , as can be seen in fig.([gridfig ] ) .
for example , first order of the asymptotic solution intersects zeroth order of the near solution at @xmath50 .
we use this information to fix parameters in the asymptotic solution .
terms as points in the @xmath51 grid .
the horizontal ( vertical ) lines describe the asymptotic ( near ) solution in increasing order of @xmath4 .
the matching is done at the intersecting points denoted by circles . ] in the intermediate region the asymptotic solution is expanded in @xmath39 , while the near solution is expanded in @xmath38 .
the table shows the sequence of calculating terms in the asymptotic and near solutions .
terms connected by an arrow are identified by the double expansion .
a similar table has been presented in @xcite ] the calculation is carried out order by order .
both solutions ( asymptotic and near ) must be known up to order @xmath52 before calculating order @xmath53 . within order @xmath53
, one should calculate the asymptotic solution first and then the near solution .
the sequence of calculation of terms in the two solutions is depicted in fig.([tablefig ] ) .
two parameters can be measured in the asymptotic region , the adm mass and the relative tension around the compact dimension .
@xcite expanding the asymptotic metric around the minkowski metric we write @xmath54 , then the adm mass , @xmath55 , and the tension , @xmath56 , are defined as @xmath57dw~ , \label{defadmmass}\\ b&=&\int d^{4}x t^{ww}=-\frac{1}{4g_{5}}\lim_{r\rightarrow\infty}r^{2 } \int_{-l/2}^{l/2}\left[\frac{2}{r}h_{rr}+h_{tt , r}\right]dw~. \label{deftension}\end{aligned}\ ] ] it is useful to define the relative tension , @xmath58 . in @xcite
the pair @xmath59 is used to describe a phase diagram of all possible black objects .
the near solution provides the thermodynamic parameters of the horizon , i.e. entropy and temperature .
the entropy is defined as the three dimensional area of the horizon divided by @xmath60 .
the temperature is defined through the zeroth law of black holes thermodynamics as the surface gravity divided by @xmath61 . a smarr formula @xcite relates the entropy , temperature , mass , and relative tension as @xcite @xmath62
in the asymptotic region zeroth order in @xmath4 means that we neglect the mass . therefore , the configuration is asymptotically flat ( since the mass is confined to a small region ) .
so , the asymptotic form of the metric is minkowski on a cylinder @xmath63 where @xmath22 is the compact coordinate such that @xmath28 $ ] . in the near region
zeroth order in @xmath4 means that @xmath64 .
the mass point is in five dimensional asymptotic flat infinite manifold .
therefore , the near solution is given by the myers - perry @xcite solution ( mps ) , @xmath65 where @xmath1 is the five dimensional schwarzschild radius . in order to match the metrics ( [ a0 ] , [ n0 ] ) one needs to specify a coordinate transformation between @xmath66 and @xmath67 .
we choose to work with a specific coordinate transformation throughout the calculation of all orders .
( one can , however , choose to redefine the transformation in each order ) .
the coordinates @xmath68 , @xmath69 , and @xmath70 are shared by the coordinate systems .
furthermore , we want to preserve the periodic structure in @xmath22 so we choose the transformation [ transformation ] @xmath71~ , \label{z}\\ r&=&\rho\sin\psi~,\label{r}\\ \rho&=&\sqrt{r^{2}+v^{2}}=\sqrt{r^{2}+\frac{l^{2}}{\pi^{2}}\sin^{2}\frac{\pi w}{l } } ~,\label{rho}\\ \psi&=&\arctan\frac{r}{v}=\arctan\frac{\pi r}{l\sin(\pi w / l)}~,\label{psi}\end{aligned}\ ] ] one can verify that the metrics ( [ a0],[n0 ] ) agree up to order @xmath72 .
we show now that the expansion contains only terms with integer powers of @xmath73 .
it is sufficient to establish that the solution must be an even function of @xmath1 and of @xmath2 ( at fixed @xmath74 ) .
the solution is an even function of @xmath1 , because the driving term , the zeroth order near metric ( [ n0 ] ) depends only on @xmath75 . the zeroth order near metric , ( [ n0 ] ) , combined with ( [ transformation ] ) is also an even function of @xmath2 .
the conditions imposed on the metric also force it to be even in @xmath2
. the mass is located at @xmath76 , ( @xmath77 ) .
the configuration of a mass point on a cylinder has a @xmath27 symmetry about @xmath29 , ( @xmath78 ) , requiring that the terms of the metric depend on @xmath22 and @xmath2 through functions of the form @xmath79 .
in the asymptotic region , @xmath11 , we expand the metric as @xmath80 where @xmath81 is the flat metric ( [ a0 ] ) . the general form of the solution for @xmath82 appears in many places in the literature @xcite .
we re - derive it here using the matching method .
we expand the einstein equation to first order in @xmath4 and get a set of homogeneous linear equations for @xmath82 . using the coordinate transformation @xmath83
we choose the gauge @xmath84 .
next we solve @xmath85 for @xmath86 , and @xmath87 for @xmath88 , where @xmath89 is the einstein tensor .
the solution is given by [ h1 ] @xmath90 where the gauge function satisfies the symmetry properties @xmath91 and @xmath92 is a constant .
the function @xmath93 satisfies the equations @xmath94 the solution to eq.([htteq ] ) can be written as a fourier series @xmath95 the constants @xmath96 are determined by matching ( [ httmodes ] ) to the zeroth order near solution ( [ n0 ] ) .
then it follows that up to leading order in @xmath17 the two solutions must be identical , @xmath97 .
the constants @xmath96 are given by the fourier expansion of @xmath98 .
[ fourier1 ] @xmath99 using ( [ fourier1 ] ) we are able to sum eq.([httmodes ] ) as @xmath100 for @xmath101 the metric is given in terms of the ` zero mode , ' independent of @xmath22 , and can be summarized as [ h1asymp ] @xmath102~,\\ g_{rw}&=&\epsilon w_{,r}(r , w)~.\end{aligned}\ ] ] the constant @xmath92 , which appears in the metric ( [ h1asymp ] ) is related to the tension in the compact dimension @xmath103dw = -\frac{\alpha\epsilon l}{2g_{5}}~. \label{tension}\ ] ] the tension @xmath56 should be zero because we deal here with linearized gravity , with no interaction between the mass and its periodic images . consequently , we must set @xmath104 .
dimensional analysis also requires @xmath104 .
first , we can deduce from eqs.([h1 ] ) that @xmath105 .
the asymptotic solution is an expansion in the mass , so @xmath92 must be independent of the mass and proportional to @xmath2 , the only quantity of the correct dimension at our disposal .
as we have discussed earlier , function ( [ htt1final ] ) is even in @xmath2 .
if @xmath106 then the functions @xmath86 and @xmath88 , in ( [ h1 ] ) , would have a mixed symmetry under @xmath107 .
note that this symmetry argument can and will be used in higher ( @xmath20th ) orders of the expansion to eliminate terms of the form @xmath108 from @xmath109 and @xmath110 .
next , we turn to evaluate the conserved adm mass using eqs.([h1asymp ] ) @xmath111dw = \frac{3\pi\mu^{2}}{8g_{5}}~. \label{admmass}\ ] ] the effective four - dimensional newton constant is defined by @xmath112 . using eqs.([h1asymp ] , [ admmass ] )
we find that @xmath113 .
we should emphasize that the physically measurable quantities , the relative tension , the mass , and newton s constant , are independent of the gauge function @xmath114 .
as we mentioned earlier , in higher orders of the expansion the relative tension in the compact dimension is not zero .
however , we require that the adm mass is completely determined by the first order , such that higher orders do not change eq.([admmass ] ) . this way
, we make sure that the expansion in @xmath4 is also an expansion in @xmath55 and that thermodynamic properties of the black hole are well defined in each order
. we will return to this issue in section [ sec : second ] when we discuss the second order contributions .
the gauge function @xmath114 , which appears in ( [ h1 ] ) , can be partially determined by matching to the zeroth order of the near solution .
however , we prefer to determine the function completely by matching to the first order of the near solution , as well . in the near region
it is convenient to use the coordinates @xmath67 just like in eq.([n0 ] ) .
we use the following ansatz for the metric : @xmath115 the functions @xmath116 are expanded to first order in @xmath4 as [ n1 ] @xmath117 the first step is to use a gauge transformation of the form @xmath118 , @xmath119 , which leaves the last term of ( [ spherical ] ) unchanged to leading order , to choose a gauge where @xmath120 this gauge choice simplifies the equations for the rest of the functions to be determined .
next , one can solve the equation @xmath121 for @xmath122 .
the rest of the equations can be solved in terms of a single function , @xmath123 , obeying a second order partial differential equation .
the solution is [ n1solution ] @xmath124\nonumber\\ & & -\frac{2\rho^{2}h_{1,\rho}(\rho,\psi)}{\sin^{2}\psi\,\cos\psi } + \frac{\rho^{3}}{\rho^{2}-\mu^{2}}f_{1,\psi}(\rho,\psi ) -\rho^{2}\tan\psi f_{1,\rho}(\rho,\psi ) ~. \label{vn1}\end{aligned}\ ] ] the wave function @xmath125 satisfies the differential equation @xmath126 the solution of eq.([h1equation ] ) can be written as @xmath127h_{\nu}(\psi)~ , \label{h1nu}\\ h_{\nu}(\psi)&=&\frac{4}{\sqrt{2\pi}}\sin^{3/2}2\psi\left [ \pi\cos\alpha_{\nu}p^{3/2}_{\nu}(\cos2\psi ) -2\sin\alpha_{\nu}q^{3/2}_{\nu}(\cos2\psi)\right]\nonumber\\ & & = ( 2\nu-1)\cos[(2\nu+3)\psi-\alpha_{\nu } ] -(2\nu+3)\cos[(2\nu-1)\psi-\alpha_{\nu}]~ , \label{hpsi}\end{aligned}\ ] ] where @xmath128 and @xmath129 are associated legendre functions of the first and second kind .
the functions @xmath130 and @xmath131 will be fixed below using the constraints on function @xmath132 .
these constraints come from symmetries , from regularity requirements , and from boundary conditions .
@xmath27 symmetry about the direction @xmath78 ( @xmath29 ) implies the conditions [ z2brane ] @xmath133 it can be verified that condition ( [ h1,psi ] ) is automatically satisfied by ( [ h1nu ] ) . for a black hole configuration , unlike for a black string configuration , the mass is localized at the origin
. then the components of the metric , ( [ n1 ] ) , are finite at @xmath134 and arbitrary @xmath35 .
this will only be true if @xmath135 it can be verified that ( [ nostring ] ) implies @xmath136 .
then ( [ hpsi ] ) simplifies to @xmath137 where we have omit a possible overall sign .
as we mentioned earlier , the expansion in @xmath41 and the fact that the zeroth order of the near solution ( [ n0 ] ) depends only on @xmath75 imply that the metric is even in @xmath1 .
this means that the function @xmath132 should be even in @xmath138 . for large @xmath25 the legendre functions in ( [ h1nu ] ) behave as @xmath139 where @xmath140 are analytic functions . since we require that @xmath132 is even in @xmath25 , @xmath141 should be integer .
the integral in ( [ h1nu ] ) is reduced to a sum over integer values of @xmath141 . @xmath142\sin^{3}\psi + 2r\sqrt{r^{2}-1}\sum_{n=1}^{\infty } \left[a_{n } p^{1}_{n}(2r^{2}-1)+b_{n } q^{1}_{n}(2r^{2}-1)\right]\nonumber\\ & & \hspace{2in}\times\left[(2n-1)\sin(2n+3)\psi -(2n+3)\sin(2n-1)\psi\right]~. \label{h1sum}\end{aligned}\ ] ] the case @xmath44 requires special attention , since the @xmath143 .
the metric ( [ spherical ] ) is static .
therefore , the surface @xmath144 is a killing horizon .
the normal vector @xmath145 should be null on the horizon .
this implies that on the horizon @xmath146 , thus the horizon is located at constant @xmath147 . using metric ( [ n1 ] ) we expand @xmath148 in @xmath4 as @xmath149 .
the conditions @xmath150 and @xmath151 restrict the gauge function @xmath152 @xmath153 the surface gravity for metric ( [ spherical ] ) is defined as @xmath154 the surface gravity should be constant on the horizon , otherwise the horizon is singular @xcite .
when we expand the metric in @xmath4 , the surface gravity must be constant in every order of the expansion .
we use metric ( [ n1 ] ) and conditions ( [ f1horizon ] ) to evaluate the surface gravity in order @xmath4 @xmath155 the requirement of a constant surface gravity constrains the function @xmath132 . at @xmath156 the legendre functions in eq.([h1sum ] ) take the values @xmath157 , @xmath158 .
we use the representation ( [ h1sum ] ) to evaluate the surface gravity ( [ kappah1 ] ) @xmath159 the set @xmath160 is complete on the interval @xmath161 $ ] , therefore , the solution to eq.([chi1 ] ) is @xmath162 in other words , the sum in eq.([h1sum ] ) contains legendre functions of the first kind only .
the functions @xmath163 are polynomials of order @xmath164 in @xmath165 .
the remaining free parameters of the near solution , @xmath166 , must be determined from matching the two gauge functions , @xmath114 in the asymptotic solution ( [ a1 ] ) and @xmath167 in the near solution ( [ n1 ] ) .
we use ( [ transformation ] ) to transform the asymptotic solution to @xmath168 coordinates and then we expand it in @xmath17 .
the gauge function @xmath114 is also expanded as @xmath169 we keep the explicit factor @xmath170 in ( [ wn ] ) to insure that @xmath171 .
we start with zeroth order in @xmath17 .
we transform the asymptotic metric using ( [ transformation ] ) and expand it to zeroth order in @xmath17 .
[ asymprhopsi0 ] @xmath172~ , \label{vasymp0}\end{aligned}\ ] ] where the superscript @xmath173 stands for asymptotic and the functions are defined in ansatz ( [ spherical ] ) . matching ( [ asymprhopsi0 ] ) to zeroth order of the near solution ( [ n0 ] )
determines the function @xmath174 @xmath175 the asymptotic solution is [ asymprhopsi1 ] @xmath177~ , \label{aasymp1}\\ u^{a}&=&1+\frac{\pi^{2}}{l^{2}}\left[\rho^{2}\cos^{2}\psi\sin^{2}\psi -\frac{\mu^{2}\sin\psi\cos^{2}\psi(5\sin3\psi-37\sin\psi)}{12 } -2\mu^{2}\frac{\sin\psi}{\rho } w_{2,\psi}(\rho,\psi)\right]~ , \label{uasymp1}\\ v^{a}&= & -\frac{\pi^{2}}{l^{2}}\left[\rho^{3}\sin\psi\cos^{3}\psi + \mu^{2}\rho\sin\psi\cos^{3}\psi(\sin^{2}\psi+2 ) + \mu^{2}\rho\sin\psi w_{2,\rho}(\rho,\psi ) -\mu^{2}\cos\psi w_{2,\psi}(\rho,\psi)\right ] ~. \label{vasymp1}\end{aligned}\ ] ] the near solution ( [ n1 ] ) should be expanded to second order in @xmath1 , which appears in @xmath41 and in the rescaled coordinate @xmath138 .
the large @xmath25 behavior of the terms of @xmath132 , ( [ h1sum ] ) , is @xmath178 these expressions contribute by terms of order @xmath179 in metric ( [ spherical ] ) .
we deduce that @xmath180 terms contribute by negative orders in @xmath1 and must be eliminated
. therefore we impose @xmath181 if @xmath180 . for similar reasons ,
gauge function @xmath152 should also be a polynomial of order @xmath182 in @xmath35 .
thus , functions @xmath132 and @xmath152 are @xmath183\sin^{3}\psi~,\label{h10}\\ f_{1}&=&(\mu^{-2}\rho^{2}-1)f_{2}(\psi)-\zeta_{1}~,\label{f1asymp}\end{aligned}\ ] ] where we have already imposed @xmath184 , ( [ f1horizon ] ) .
the exact form of the periodic function @xmath185 will be fixed below .
then the near metric , expanded to second order in @xmath1 , is [ nearmu2 ] @xmath186~ , \label{anear2}\\ u^{n}&=&1-\frac{2\tan^{2}\psi}{l^{2}}\left[\rho^{2}[2a_{0}+f_{2}(\psi ) + \cot\psi f_{2}'(\psi)]-\mu^{2}[4a_{0}-6b_{0}+f_{2}(\psi)+\zeta_{2 } + \cot\psi f_{2}'(\psi)]\right]~ , \label{unear2}\\ v^{n}&= & \frac{\rho^{3}\tan\psi}{l^{2 } } \left[-4a_{0}-2f_{2}(\psi)+\cot\psi f_{2}'(\psi)\right ] ~. \label{vnear2}\end{aligned}\ ] ] comparing ( [ asymprhopsi1 ] ) and ( [ nearmu2 ] ) we find that @xmath187 ~.\label{w2}\end{aligned}\ ] ] at this point the near solution still contains two free parameters , @xmath188 and @xmath189 . the @xmath27-symmetry condition , ( [ f1brane ] ) , imposes one constraint on these parameters @xmath190 a constraint , determining @xmath189 , is derived from the first law of black hole thermodynamics .
it will be discussed in the next section .
the zeroth law of black hole thermodynamics states that the temperature of a black hole is @xmath191 where @xmath192 is the surface gravity , which is constant on the horizon . to calculate the temperature for the near solution we use eqs.([kappa ] , [ kappah1 ] , [ bn ] ) to find that @xmath193 the first law of black hole thermodynamics is @xmath194 , where the entropy is proportional to the area of the horizon @xmath195 we calculate the entropy for the near solution and find that @xmath196 according to the first law the temperature can be found as @xmath197 .
the mass appears in the entropy only through @xmath9 and @xmath41 , so if we combine the first law and the zeroth law we get @xmath198=2\pi\mu[1 + 5\zeta_{1}\epsilon]~. \label{t2}\ ] ] this fixes the last parameter . if we combine eqs.([b0],[t2 ] ) we find that @xmath199 to summarize , the near metric , to first order in @xmath4 , is [ near1final ] @xmath200~ , \label{b1final}\\ a^{n}&=&1+\frac{\pi^{2}\epsilon}{3}\left[(3\frac{\rho^{2}}{\mu^{2}}-1)\cos^{4}\psi-1 \right]~ , \label{a1final}\\ u^{n}&=&1+\pi^{2}\epsilon(\frac{\rho^{2}}{\mu^{2}}-1)\sin^{2}\psi\cos^{2}\psi~ , \label{u1final}\\ v^{n}&= & -\pi^{2}\epsilon\frac{\rho^{3}}{\mu^{2}}\sin\psi\cos^{3}\psi ~. \label{v1final}\end{aligned}\ ] ] the location of the horizon is at @xmath201 .
the entropy and the temperature are @xmath202 these expressions are in agreement with previous results @xcite .
the only freedom left in the first order metric is the @xmath203 terms of the gauge function in the asymptotic solution ( [ wn ] ) .
the @xmath44 and @xmath204 terms are fixed in the region @xmath205 by our matching procedure , as given in eqs.([w0],[w2 ] ) .
in addition , in the asymptotic region , @xmath101 , the metric should be minkowski , therefore , we require that @xmath206 . a form of the gauge function which is consistent with these conditions appears in appendix [ app : asymp2 ] .
prior to completing calculations in second order we describe our general procedure for calculating higher order contributions .
first we consider the asymptotic solution , which is fully determined ( up to gauge freedom ) by the lower order contributions . in the asymptotic region
we expand the metric in @xmath4 as @xmath207 assume that we know the solution up to order @xmath208 and we intend to obtain the @xmath20th order solution .
einstein s equation is linear in the @xmath209 but due to lower order contributions it is inhomogeneous .
the solution is similar to ( [ h1 ] ) but , in addition to the solution of the homogeneous equation includes extra terms corresponding to a particular solution ( the particular solution is denoted by @xmath210 ) , as follows [ hn ] @xmath211 where @xmath92 is a constant , and @xmath212 is a gauge function which is periodic and antisymmetric in @xmath22 .
the function @xmath213 is also periodic and symmetric in @xmath22 .
it satisfies the equation @xmath214 where @xmath215 is the deriving term , which depends on lower orders , @xmath216 , @xmath217 .
the homogeneous solution to eq.([httneq ] ) can be written as a fourier series .
consequently , we have @xmath218 the particular solution @xmath210 is completely determined by the lower orders of the asymptotic solution ( without any use of the near solution ) .
the free parameters are @xmath92 and the set @xmath219 .
these can be determined by matching to the lower orders of the near solution as follows .
take @xmath220 from the near solution up to order @xmath52 and expand it in @xmath1 .
take the term of order @xmath221 and find its fourier series just like in eq.([fourier1 ] ) .
compare that fourier series with the forier series of @xmath213 to lowest order in @xmath17 , and determine the constants @xmath219 .
just like in first order , we can show that @xmath222 .
solution ( [ hn ] ) should be even in @xmath2 .
the functions @xmath223 and @xmath213 are even functions of @xmath2 since these functions are determined by lower orders but for dimensional reasons @xmath224 is odd .
therefore , @xmath225 must vanish .
next , we have to make sure that the definition of the adm mass does not change .
so , we require that @xmath226dw=0~. \label{fixm}\ ] ] in general , the tension in the compact dimension , ( [ tension ] ) , does not vanish . as a result ,
the effective four - dimensional newton s constant , @xmath227 , acquires a correction of order @xmath53 .
this means that @xmath227 depends on the mass and the _ equivalence principle is violated .
_ at this point the asymptotic solution is determined up to order @xmath53 , except for the gauge function @xmath212 , which is ( partially ) determined by matching to the near solution to orders up to @xmath53 . near the horizon
it is convenient to use the metric in form ( [ spherical ] ) and expand it in @xmath4 as follows [ emetric ] @xmath228 we need the near solution up to order @xmath208 and the asymptotic solution up to order @xmath20 to calculate the @xmath20th order near solution .
again , the @xmath20th order einstein s equation is a linear inhomogeneous equation in @xmath229 .
the inhomogeneity depends on @xmath230 .
we solve the equations , in a way , similar to that for the first order correction .
the first step is to apply a gauge transformation of the form @xmath231 , @xmath232 to arrive at a gauge , in which @xmath233 next , one can solve the inhomogeneous equation @xmath121 for @xmath234 . the rest of the equations can be solved in terms of a single function , @xmath235 , obeying a second order partial differential equation , which is similar to ( [ h1equation ] ) , @xmath236 where @xmath237 depends on the lower order corrections .
the solution of eq.([hnequation ] ) can be found by the method of separation of variables , just like it has been done when we have calculated the first order correction ( [ h1nu ] ) .
the eigenfunctions of eq.([hnequation ] ) are the functions @xmath238 , which appear in eq.([hpsi ] ) .
the driving term , @xmath237 , can also be expanded in the set @xmath238 .
in addition , the boundary conditions at @xmath78 and at @xmath134 in @xmath20th order are the same as in first order , eqs.([h1,psi ] , [ nostring ] ) . and just like in first order the function @xmath239 should be even in @xmath35 .
so , the general solution of eq.([hnequation ] ) is @xmath240\sin^{3}\psi\nonumber\\ & & + \sum_{n=1}^{\infty } \left[2r\sqrt{r^{2}-1}\left(a_{n } p^{1}_{n}(2r^{2}-1)+b_{n } q^{1}_{n}(2r^{2}-1)\right)+p_{n}(r)\right]\nonumber\\ & & \hspace{2in}\times\left[(2n-1)\sin(2n+3)\psi -(2n+3)\sin(2n-1)\psi\right]~ , \label{hnsum}\end{aligned}\ ] ] where @xmath241 is a particular solution of the inhomogeneous equation ( [ hnequation ] ) .
the particular solutions are completely determined by the lower orders of the near solution ( without using the asymptotic solution ) .
the homogeneous part should be completely fixed by matching to the asymptotic solution .
the matching is done as follows : take the asymptotic solution up to order @xmath53 and transform it to the @xmath168 coordinates , using ( [ transformation ] ) .
expand the solution in @xmath17 , take the term of order @xmath242 and compare it to the near solution of order @xmath53 to fix the parameters @xmath243 , and the gauge functions @xmath244 and @xmath212 . at this point
the function @xmath212 is determined only up to order @xmath242 and one should fix it completely before proceeding to the next order . in the next section
we apply this method to calculate the second order contributions .
up to first order , the asymptotic solution is given by equations ( [ a0 ] ) , ( [ a1 ] ) , ( [ h1 ] ) , and ( [ htt1final ] ) .
the second order solution is described in the appendix [ app : asymp2 ] .
it contains an inhomogeneous part , @xmath245 , which is completely determined by the first order solution , and a homogeneous part , which includes the wave function @xmath246 , and the gauge function @xmath247 [ h2 ] @xmath248 the wave function @xmath249 satisfies the homogeneous part of eq.([httneq ] ) , and can be written in the form of eq.([httnmodes ] ) @xmath250 to fix the constants , @xmath219 , we follow the prescription that appears after eq.([httnmodes ] ) .
we take @xmath220 from the first order near solution , ( [ b1final ] ) , and expand it in @xmath1 , up to order @xmath251 , to get @xmath252 + \mu^{4}\frac{\pi^{2}(4\cos^{4}\psi-3)}{12l^{2}\rho^{2}}~. \label{gttn1}\ ] ] then we take @xmath220 from the second order asymptotic solution and transform it to the @xmath168 coordinates , using ( [ transformation ] ) . after expanding it in @xmath17 , up to order @xmath253
, we obtain @xmath254 + \mu^{4}\left[\frac{2\cos2\psi+1}{4\rho^{4 } } -\frac{\pi^{2}(3\cos6\psi+124\cos^{4}\psi+21\cos2\psi+4 ) } { 96l^{2}\rho^{2}}\right]+\epsilon^{2}h_{tt}^{(h)}~. \label{gtta2}\ ] ] we find @xmath249 from @xmath255 and calculate its fourier coefficients @xmath256\nonumber\\ & & = \frac{\pi^{3}l\sinh\frac{2\pi r}{l}\left(-\cosh\frac{4\pi r}{l } + 16\cosh\frac{2\pi r}{l}\cos\frac{2\pi w}{l}+5\cos\frac{4\pi w}{l}-20\right ) } { 24r\left(\cosh\frac{2\pi r}{l}-\cos\frac{2\pi w}{l}\right)^{3}}~. \label{htth}\end{aligned}\ ] ] the asymptotic ( @xmath101 ) form of the metric is [ h2asymp ] @xmath257 the contribution to the adm mass vanishes indeed , @xmath258dw
= \frac{\pi^{3}l^{2}}{4}-2r^{2}\left.w_{,r}^{(2)}\right|_{w =- l/2}^{w = l/2}=0 ~ , \label{deltam2}\ ] ] where the last equality follows from boundary condition ( [ w2l2 ] ) .
the tension ( [ tension ] ) is now @xmath259dw = -m\frac{\pi^{2}\epsilon}{6}~. \label{tension2}\ ] ] following section [ subsec : nearn ] , we find that the second order solution is [ emetric2 ] @xmath260~ , \label{b2}\\ a_{2}&=&\frac{4h_{2,\psi}(\rho,\psi)}{\sin^{2}\psi\,\cos\psi } + 2f_{2}(\rho,\psi)+2\rho f_{2,\rho}(\rho,\psi)-\frac{\pi^{4}\cos^{2}\psi}{864\mu^{4 } } \left[\mu^{2}(105\rho^{2}+11\mu^{2})\cos6\psi\right.\nonumber\\ & & \left.+(-64\rho^{4}+138\mu^{2}\rho^{2}+58\mu^{4})\cos4\psi + ( 1472\rho^{4}+1383\mu^{2}\rho^{2}+237\mu^{4})\cos2\psi -768\rho^{4}+774\mu^{2}\rho^{2}-2\mu^{4}\right]~ , \label{a2}\\ u_{2}&=&-2 \frac{6\rho h_{2}(\rho,\psi ) -(2\rho^{2}-\mu^{2})h_{2,\rho}(\rho,\psi)}{\rho\sin\psi\,\cos^{2}\psi } -2\tan^{2}\psi f_{2}(\rho,\psi)-2\tan\psi f_{2,\psi}(\rho,\psi)~ , \label{u2}\\ v_{2}&= & \frac{\rho(2\rho^{2}-\mu^{2})}{(\rho^{2}-\mu^{2})\sin^{2}\psi\,\cos\psi } \left[3h_{2}(\rho\psi)-\tan\psi h_{2,\psi}(\rho\psi)\right ] -\frac{2\rho^{2}h_{2,\rho}(\rho\psi)}{\sin^{2}\psi\,\cos\psi}\nonumber\\ & & + \frac{\rho^{3}}{\rho^{2}-\mu^{2}}f_{2,\psi}(\rho,\psi ) -\rho^{2}\tan\psi f_{2,\rho}(\rho,\psi ) + \frac{\pi^{4}\rho(2\rho^{2}-\mu^{2})\sin2\psi}{864\mu^{4}(\rho^{2}-\mu^{2})}\left [ \mu^{2}(15\rho^{2}+\mu^{2})\cos6\psi\right.\nonumber\\ & & \left .
+ ( -8\rho^{4}+29\mu^{2}\rho^{2}+4\mu^{4})\cos4\psi + ( 184\rho^{4}+185\mu^{2}\rho^{2}+11\mu^{4})\cos2\psi -96\rho^{4}+99\mu^{2}\rho^{2}-16\mu^{4}\right]~. \label{v2}\end{aligned}\ ] ] the function @xmath261 satisfies the differential equation ( [ hnequation ] ) with @xmath262 where the functions @xmath263 are given in eq.([hpsisin ] ) . the solution of eq.([hnequation ] ) with inhomogeneity ( [ s2 ] ) can be written in form eq.([hnsum ] ) with @xmath264 .
for each of the non vanishing @xmath265 one can add combinations of the homogeneous solution which includes the legendre functions as appear in eq.([hnsum ] ) .
these should be chosen such that @xmath265 is regular at @xmath266 , and is a polynomial in @xmath35 of the lowest order possible ( in principle , the homogeneous solution is of order @xmath267 ) .
in fact , the particular solution for eq.([hnequation ] ) with inhomogeneity ( [ s2 ] ) includes polynomials of order @xmath268 in @xmath35 , only , [ h2p ] @xmath269 in addition , to avoid negative powers of @xmath1 , the homogeneous part of @xmath261 and the gauge function @xmath270 should be polynomials of order @xmath268 in @xmath271 .
@xmath272\sin^{3}\psi + 8a_{1}r^{2}(r^{2}-1)(1+\cos^{2}\psi)\sin^{3}\psi~,\label{h2}\\ f_{2}(\rho,\psi)&\equiv & ( r^{2}-1)^{2}f_{4}(\psi ) + ( r^{2}-1)f_{2}(\psi)+f_{0}(\psi)~.\label{f2}\end{aligned}\ ] ] we now turn to imposing boundary conditions .
the horizon is located at a constant @xmath147 .
we set @xmath273 and solve @xmath150 for @xmath274 .
we find @xmath275~.\label{f0}\ ] ] the surface gravity is constant ( due to the fact that the legendre functions of the second kind are not included in ( [ h2 ] ) ) , @xmath276~.\label{kappa2}\ ] ] we match the near solution , ( [ emetric2 ] ) , to the asymptotic solution , ( [ asymp2 ] ) , expanded to order @xmath277 . comparing the expressions for @xmath220 , we find that @xmath278~,\label{2f4}\\ f_{2}(\psi)&=&\frac{\pi^{4}}{17280}\left[15\cos8\psi+188\cos6\psi + 1248\cos4\psi+3156\cos2\psi+2993\right]~.\label{2f2}\end{aligned}\ ] ] matching the other components , as well , we find the constants @xmath279 the gauge function , @xmath280 , is @xmath281~. \label{w2lexpand}\end{aligned}\ ] ] the metric ( [ emetric2 ] ) should be symmetric about the plane @xmath78 . as a result
we find that @xmath282 we calculate the entropy of the black hole @xmath283~. \label{entropy2}\ ] ] we compare the zeroth law @xmath284 and the first law @xmath197 to find that @xmath285 to summarize , the location of the horizon , the entropy , and the temperature are @xmath286~. \label{rhoh2}\\ s&=&\frac{\pi^{2}\mu^{3}}{2g_{5}}\left[1+\frac{\pi^{2}\epsilon}{8 } + \frac{\pi^{4}\epsilon^{2}}{384}\right]~. \label{s2final}\\ t&=&\frac{1}{2\pi\mu}\left[1-\frac{5\pi^{2}\epsilon}{24 } + \frac{43\pi^{4}\epsilon^{2}}{1152}\right]~. \label{t2final}\end{aligned}\ ] ] the near metric is given by [ near4final ] @xmath287 + \frac{\pi^{4}\epsilon^{2}}{5760\mu^{2}}\left[1570\mu^{2}+17\rho^{2 } -4(14\rho^{2}-395\mu^{2})\cos2\psi\right.\right.\nonumber\\ & & \left.\left.+(52\rho^{2}+580\mu^{2})\cos4\psi -4(2\rho^{2}-25\mu^{2})\cos6\psi -(\rho^{2}-2\mu^{2})\cos8\psi\right]\frac{}{}\right)~ , \label{b4final}\\ a^{n}&=&1+\frac{\pi^{2}\epsilon}{3\mu^{2}}\left[(3\rho^{2}-\mu^{2})\cos^{4}\psi-1 \right]+\frac{\pi^{4}\epsilon^{2}}{5760\mu^{4}}\left[1800\rho^{4}+1324\mu^{2}\rho^{2 } -307\mu^{4 } + 4(675\rho^{4}+422\mu^{2}\rho^{2}-325\mu^{4})\cos2\psi\right.\nonumber\\ & & \left .
+ 8(135\rho^{4}+88\mu^{2}\rho^{2}-55\mu^{4})\cos4\psi + 4(45\rho^{4}+26\mu^{2}\rho^{2}-15\mu^{4})\cos6\psi + 5\mu^{2}(4\rho^{2}-\mu^{2})\cos8\psi\right]~ , \label{a4final}\\ u^{n}&=&1+\frac{\pi^{2}\epsilon}{\mu^{2}}(\rho^{2}-\mu^{2})\sin^{2}\psi\cos^{2}\psi -\frac{\pi^{4}\epsilon^{2}\sin^{2}\psi}{1440\mu^{4 } } \left[-540\rho^{4}-970\mu^{2}\rho^{2}+385\mu^{4}\right.\nonumber\\ & & \left . -5(144\rho^{4}+229\mu^{2}\rho^{2}-385\mu^{4})\cos2\psi
-5(\rho^{2}-\mu^{2})(36\rho^{2}+82\mu^{2})\cos4\psi -55\mu^{2}(\rho^{2}-\mu^{2})\cos6\psi\right]~ , \label{u4final}\\ v^{n}&= & -\frac{\pi^{2}\epsilon\rho^{3}\sin\psi\cos^{3}\psi}{\mu^{2 } } -\frac{\pi^{4}\epsilon^{2}\rho^{3}\sin2\psi}{2880\mu^{4 } } \left[990\mu^{2}+540\rho^{2 } + ( 720\rho^{2}+1207\mu^{2})\cos2\psi\right.\nonumber\\&+&\left.2(90\rho^{2}+89\mu^{2})\cos4\psi + 25\mu^{2}\cos6\psi\right ] ~. \label{v4final}\end{aligned}\ ] ]
in this paper we have calculated the metric of a small black hole in a five dimensional cylinder .
the metric is found perturbatively , with an expansion parameter @xmath288 .
@xmath55 is the physical adm mass and @xmath2 is the circumference of the compact dimension , both measured at infinity . we calculate the metric , up to second order in @xmath4 ( fourth order in the ratio @xmath14 ) . the asymptotic solution is given in eqs.([a1],[h1],[htt1final ] ) and in the appendix [ app : asymp2 ] .
the near solution is provided in eqs .
( [ spherical],[near4final ] ) .
we obtained the following expressions for the entropy and the temperature of the black hole @xmath289~ , \label{sfinal}\\ t&=&\frac{1}{2\pi\mu}\left[1-\frac{5\pi^{2}\epsilon}{24 } + \frac{43\pi^{4}\epsilon^{2}}{1152}\right]~. \label{tfinal}\end{aligned}\ ] ] these , and other results , presented below , were previously obtained in first order of @xmath4 by gorbonos and kol @xcite and by harmark @xcite . using the smarr formula we find the relative binding energy @xmath290 we have chosen the coordinate system such that the horizon is independent of the angle , @xmath26 , between the three dimensional space and the @xmath22 axis . as the scale of the five dimensional radial coordinate , @xmath35 , is fixed by the term of the metric , @xmath291
, the radius of the horizon is uniquely determined .
the horizon is located at @xmath147 as in eq.([spherical ] ) .
we find that @xmath292~. \label{rhohfinal}\ ] ]
the question of transition from black hole to black string , as a function of the mass parameter , has been discussed at length in the literature @xcite , @xcite .
one scenario proposed was the transition at the intersection of the black hole and nonuniform black string lines in the @xmath293 ( mass - relative tension ) phase diagram .
we use the second order results for the small black hole to extrapolate to such a transition .
we study three aspects of the transition ; the mass - relative tension phase diagram , comparison of entropies , and the change of topology ( using two different approaches ) .
a uniform black string is described by the metric @xmath294 the relative tension is constant , @xmath295 , and the entropy is @xmath296 the entropy and relative tension of the non - uniform black string were have not been found numerically yet.@xcite .
such a configuration exists for masses larger than the gregory - laflamme mass , which is , in terms of our variables , @xmath297 . in fig.([phasediagram ] ) we draw the @xmath298 phase diagram for the uniform black string , the non - uniform black string ( just the leading order term near the gl transition ) , and the black hole branches .
the black hole turns into a nonuniform black string beyond the transition point , @xmath299 .
we did not include recent numerical results @xcite , which , due to numerical difficulties , have large errors @xcite in fig.([entropies ] ) we draw the entropies of the black objects in units of @xmath300 . comparing ( [ sfinal ] ) and ( [ bsentropy ] )
we find that the enropies of the uniform black string and the black hole are equal for @xmath301 .
below that value the black hole has higher entropy , and above that value the black string has higher entropy .
this simple check suggests that the black hole will be unstable for @xmath302 .
note again that our second order expansion in @xmath4 shows good convergence . using the first order expansion
only the intersection of entropies would occur at @xmath303 .
a transition between a black hole and a black string requires a topology change of the horizon .
this change is likely to happen when the horizon of the black hole fills the compact dimension and touch itself . at this point
a small pinch of the horizon will turn it into a non - uniform black string . in the notation we use , eq.([transformation ] ) , this will happen when @xmath304 the first order formula would give @xmath305 .
again , we see a fairly good convergence , making our result reasonably reliable .
we will see a further indication for this point in another , independent , calculation of the critical value of @xmath4 .
the apparent agreement of @xmath306 with @xmath307 ( equality of the entropies of the uniform black string and the black hole ) is somewhat problematic .
we see no mathematical reason for such a agreement , though it would be interesting if these points also coincided .
note that for geometric reasons the black hole branch should cease to exist above the critical value of @xmath4 so the stable state of the system is the uniform black string , which has a higher entropy than the nonuniform black string . in terms of the coordinates we chose the horizon
is located at constant @xmath35 . however , this does not mean it is spherical .
compactification breaks the @xmath308 symmetry .
the horizon can be an oblate ( elongated perpendicular to the compact direction ) or a prolate ( elongated parallel to the compact dimension ) ellipsoid .
there is no generally accepted method for distinguishing between the oblate and the prolate configurations .
some authors @xcite define the eccentricity of the black hole as @xmath309 where @xmath310 is the maximal area of a cross section of the horizon parallel to the compact dimension , and @xmath311 is the maximal area of a cross section of the horizon perpendicular to the compact dimension . @xmath312
a prolate ( oblate ) horizon has positive ( negative ) eccentricity .
we use the near solution ( [ emetric2 ] ) to find that @xmath313 .
we see that the horizon is prolate .
another possible measure of the eccentricity is given by the intrinsic ricci scalar of the horizon , which is calculated in terms of the three metric @xmath314 ~.\label{horizonmetric}\ ] ] using the near solution , the intrinsic curvature of ( [ horizonmetric ] ) is @xmath315~.\label{intrinsicr}\ ] ] it is maximal at @xmath134 , which indicates that the horizon is indeed prolate .
one might speculate that a prolate horizon will tend to grow in the periodic direction and turn into a nonuniform black string .
however , the circumference of the compact direction grows as well @xcite and the transition into a black string is not obvious . in @xcite
it is suggested to measure the inter polar distance , which is the proper distance between the poles of the prolate horizon along the compact dimension . in principle , to calculate this distance we need to break the integral defining this distance into two parts . in the first part we should use the near solution and in the other we should use the asymptotic solution . in the near horizon region
one should integrate along @xmath134 from @xmath147 to some arbitrary point @xmath316 . in the asymptotic region
we should integrate along @xmath317 from @xmath318 to @xmath319 @xmath320 in @xcite it was found that the zeroth order approximation is @xmath321 , which means that the circumference grows enough to make room for the black hole .
if the mass ( @xmath4 ) value , at which the black hole fills the compact dimension , is small enough then it is possible that the integral over the near solution is sufficient in ( [ lpoles ] ) , or , in other words , we can choose @xmath322 .
then we obtain @xmath323 first of all , we see that the distance between the poles of the horizon is a decreasing function of @xmath4 .
if we solve the equation @xmath324 we obtain @xmath325 , which is in a rough agreement with values we obtained for the critical value , @xmath326 , above . taking the first order term only we would obtain @xmath327 . clearly , the convergence of @xmath328 is not as good as those of @xmath326 and @xmath306 .
one reason for the poorer convergence is that for very small @xmath4 ( [ lpoles2 ] ) must fail , as a significant contribution coming from the asymptotic solution was omitted .
it should work , however in the region where @xmath329 .
fortunately , in this region the @xmath4-expansion is still fairly reliable and the rough agreement of @xmath330 is encouraging .
this work is supported in part by the u.s .
department of energy grant no .
de - fg02 - 84er40153 .
we thank richard gass and cenalo vaz for fruitful discussions .
we also thank barak kol , hideaki kudoh , evgeny sorkin , and toby wiseman for their comments , which helped us improving this work .
the second order asymptotic solution is given by eqs.([h2 ] ) [ h2anh ] @xmath332-\frac{r}{8}\left[h\,h_{,r } + 4w\,h_{,rw}\right]+(w_{,r})^{2}-r\,k_{,w } ~,\\ \bar{h}_{ww}^{(2)}&=&\frac{r^{2}}{16}\left[(h_{,r})^{2 } + ( h_{,w})^{2}\right]-\frac{r}{2}\left[2w_{,w}h_{,r } + w\,h_{,rw}\right]+\frac{3}{16}h^{2}+(w_{,w})^{2}-r\,k_{,w } ~,\\ \bar{h}_{rw}^{(2)}&=&k-\frac{r}{16}\left[h\,h_{,w}+8w_{,r}h_{,r}\right ] + w_{,w}w_{,r}~,\end{aligned}\ ] ] where @xmath333 , @xmath334 , and @xmath335 , @xmath336~ , \label{w1final}\\ k&=&\frac{\pi l^{3}}{16r^{3}}\coth\frac{2\pi r}{l } \left[1-\frac{2\pi r}{l}\coth\frac{2\pi r}{l}\right ] \arctan\left(\tan\frac{\pi w}{l}\coth\frac{\pi r}{l}\right ) + \frac{\pi l^{3}\sin\frac{2\pi w}{l } \left[1-\frac{2\pi r}{l}\coth\frac{2\pi r}{l}\right ] } { 32r^{3}\left[\cosh\frac{2\pi r}{l}-\cos\frac{2\pi w}{l}\right]}~.\end{aligned}\ ] ] the function @xmath337 was chosen such that @xmath338 .
it satisfies eqs .
( [ w0],[w2 ] ) , and it has a structure similar to @xmath339 .
@xmath340 is an odd function of @xmath22 , but @xmath341 .
however , @xmath27-symmetry implies that @xmath342 .
this means that the gauge function , @xmath280 , has nontrivial boundary conditions at @xmath343 @xmath344~. \label{w2l2}\ ] ] to match the asymptotic and the near solution we have to transform the asymptotic solution to the @xmath168 coordinates , using ( [ transformation ] ) , and expand to order @xmath277 .
we also expand the gauge function @xmath280 as @xmath345~. \label{w2near}\ ] ] the asymptotic metric in the @xmath168 coordinates , expanded to order @xmath277 is [ asymp2 ] @xmath346~ , \label{uasymp2}\\ v^{a}&= & -\frac{\pi^{2}}{l^{2}}\left[\rho^{3}\sin\psi\cos^{3}\psi + \mu^{2}\rho\sin\psi\cos^{3}\psi(\sin^{2}\psi+2 ) + \mu^{2}\rho\sin\psi w_{2,\rho}(\rho,\psi ) -\mu^{2}\cos\psi w_{2,\psi}(\rho,\psi)\right ] ~. \label{vasymp2}\end{aligned}\ ] ] t. harmark and n.a .
obers , jhep 0205,032 ( 2002 ) , e - print archive : hep - th/0204047 ; t. harmark , phys.rev.d69:104015,2004 e - print archive : hep - th/0310259 t. harmark and n.a .
obers , e - print archive : hep - th/0309230 h. kudoh and t. wiseman , prog.theor.phys.111:475-507,2004 e - print archive : hep - th/0310104 ; e - print archive : hep - th/0409111 ; b. kol , e - print archive : hep - th/0206220 ; b. kol and t. wiseman , e - print archive : hep - th/02304070 ; e. sorkin , b. kol , t. piran _ phys.rev _
.d 69:064032 ( 2004 ) e - print archive : hep - th/0310096 ; d. karasik , c. sahabandu , p. suranyi , and l.c.r .
wijewardhana , _ phys . rev . _ d70:064007 ( 2004 ) , e - print archive : gr - qc/0404015 t. harmark , phys.rev.d69:104015,2004 e - print archive : hep - th/0310259 ; d. gorbonos and b. kol , jhep 0406:053,200 , e - print archive : hep - th/0406002 l. smarr , j.w .
york _ phys .
_ d17 , 2529 ( 1978 ) i. rcz and r. m. wald , _ class .
_ * 9*:2643 - 2656 ( 1992 ) . | we study black hole solutions in @xmath0 space , using an expansion to fourth order in the ratio of the radius of the horizon , @xmath1 , and the circumference of the compact dimension , @xmath2 .
a study of geometric and thermodynamic properties indicates that the black hole fills the space in the compact dimension at @xmath3 . at the same value of @xmath4 the entropies of the uniform black string and of the black hole are approximately equal |
the physics of quantum dots continuously attracted a lot of experimental and theoretical interest .
@xcite within the assumption that the thouless energy ( @xmath0 ) is much larger than mean single - particle level spacing ( @xmath1 ) , @xmath2 , an effective zero - dimensional hamiltonian has been derived .
@xcite in this so - called universal hamiltonian the electron - electron interaction that involves a set of matrix elements in the single - particle basis is reduced to just three parameters : the charging energy ( @xmath3 ) , the ferromagnetic exchange ( @xmath4 ) and the interaction in the cooper channel .
the single particle energies are random quantities with wigner - dyson statistics .
thus the universal hamiltonian provides a convenient framework for the theoretical description of quantum dots .
the charging energy ( typically @xmath5 ) restricts the probability of real electron tunneling through a quantum dot at low temperatures @xmath6 .
@xcite this phenomenon of the coulomb blockade leads to suppression of the tunneling density of states in quantum dots at low temperatures @xcite .
it was also understood that a small enough exchange interaction @xmath7 is important for a quantitative description of the experiments on low temperature ( @xmath8 ) transport through quantum dots fabricated in a two - dimensional electron gas .
@xcite for a quantum dot of size @xmath9 ( @xmath10 stands for the fermi wave length ) the exchange interaction can be estimated by bulk value of the fermi - liquid interaction parameter ( @xmath11 ) : @xmath12 .
as it is well - known , strong enough exchange interaction in bulk materials leads to a stoner instability at @xmath13 and a corresponding quantum phase transition between a paramagnet and a ferromagnet . in quantum dots
it is possible to realize an interesting situation in which the ground state has a finite total spin .
@xcite in the case of the equidistant single - particle spectrum it occurs for @xmath14 . as @xmath15 increases towards @xmath1 , the total spin in the ground state increases and at @xmath16 all electrons in a quantum dot become spin polarized .
this phenomenon of mesoscopic stoner instability is specific to finite size systems and disappears in the thermodynamic limit @xmath17 . due to the entanglement of the charge and spin degrees of freedom in the universal hamiltonian
, the mesoscopic stoner instability affects the electron transport through a quantum dot .
for example , it leads to an additional nonmonotonicity of the energy dependence of the tunneling density of states @xcite and to the enhancement of the shot noise .
@xcite the cooper channel interaction in the description within the universal hamiltonian framework is responsible for superconducting correlations in quantum dots .
@xcite we shall assume throughout the paper that the cooper channel interaction is repulsive and , therefore , omit it .
@xcite we also neglect corrections to the universal hamiltonian due to the fluctuations in the matrix elements of the electron - electron interaction .
@xcite they are small in the regime @xmath18 but lead to interesting physics beyond the universal hamiltonian .
@xcite in the presence of a spin - orbit coupling the description of a quantum dot in the framework of the universal hamiltonian breaks down . even for a weak spin - orbit coupling ( large spin - orbit length , @xmath19 ) fluctuations of the matrix elements of the electron - electron interaction can not be neglected in spite of the condition @xmath18 .
@xcite for a quantum dot in a two - dimensional electron gas the orbital degrees of freedom are coupled to in - plane components of the spin .
then in the regime @xmath20 the low energy description is again possible in terms of the universal hamiltonian but with the ising exchange interaction ( @xmath21 ) .
@xcite in this case mesoscopic stoner instability is absent for the equidistant single - particle spectrum . @xcite as a consequence , the tunneling density of states is almost independent of @xmath22 while the longitudinal spin susceptibility @xmath23 is independent of @xmath24 as in a clean fermi liquid .
@xcite the experiments on tunneling spectra in nanometer - scale ferromagnetic nanoparticles revealed the presence of an exchange interaction with significant anisotropy .
@xcite the simplest model which allows to explain the main features of experimentally measured excitation spectra of ferromagnetic nanoparticles resembles the universal hamiltonian with uniaxial anisotropy in exchange interaction .
@xcite such modification of exchange interaction can arise due to shape , surface , or bulk magnetocrystalline anisotropy .
in addition , in the presence of spin - orbit scattering the anisotropic part of the exchange interaction can experience large mesoscopic fluctuations .
@xcite the alternative reason for appearance of anisotropy in the exchange interaction in quantum dots is the presence of ferromagnetic leads .
@xcite the universal hamiltonian with an anisotropic exchange interaction ( albeit it is not microscopically justified ) is interesting on its own as the simplest model interpolating between the cases of the heisenberg and ising exchange interactions . since in the latter case
there is no mesoscopic stoner instability for the equidistant single - particle spectrum , it is interesting to understand how it disappears as the exchange develops anisotropy .
does the spin of the ground state vanish continuously or discontinuously as the anisotropy increases ?
for the ising exchange interaction transverse dynamical spin susceptibility @xmath25 is nontrivial .
its imaginary part is odd in frequency with maxima and minima at @xmath26 , respectively .
@xcite in the case of the heisenberg exchange @xmath27 reduces to a delta - function . but how does this reduction occur with decrease in anisotropy ? in low dimensions @xmath28 interaction and disorder can induce a transition between paramagnetic and ferromagnetic phases at a finite temperature @xmath24 .
@xcite in @xmath29 the stoner instability can be promoted by disorder and occurs at smaller values of exchange interaction @xcite . in the universal hamiltonian
the disorder remains in randomness of the single - particle levels .
as it is known , @xcite level fluctuations affect the temperature dependence of the average static spin susceptibility @xmath23 in the case of the heisenberg exchange . in the case of the ising exchange
the role of disorder is even more dramatic . for the equidistant single - particle levels @xmath23
is temperature independent . due to level fluctuations the average spin susceptibility acquires a curie type @xmath24-dependence dominating at low enough @xmath24 and for @xmath30 .
@xcite in this regime of strong ( with respect to the small distance @xmath31 to the average position of the stoner instability at @xmath32 ) level fluctuations a quantum dot is in the paramagnetic phase on average but it can be fully spin - polarized for a particular realization of the single - particle levels .
these fully spin - polarized realizations should affect the tails of the distribution functions for @xmath23 and dynamical transverse spin susceptibility @xmath25 , but how exactly ?
can it be possible that at zero temperature the level fluctuations shift the position of the stoner instability from its average position , @xmath32 , and lead to the existence of a finite temperature transition between the paramagnetic and the ferromagnetic phases in quantum dots ?
of course , the very same questions can be asked for the case of the heisenberg exchange . in this paper
we address these questions within the universal hamiltonian framework extended to the case of exchange interaction with uniaxial anisotropy .
we compute the temperature and magnetic field dependence of the static longitudinal spin susceptibility @xmath23 for equidistant single - particle spectrum . except the case of the ising exchange
it always has a non - zero temperature - dependent contribution of curie type ( @xmath33 ) or of @xmath34 type .
this indicates that destruction of the mesoscopic stoner instability by uniaxial anisotropy is not abrupt . for equidistant single - particle levels we also compute the transverse spin susceptibility .
it always has a maximum and a minimum whose positions tend to zero frequency with decrease of anisotropy .
we show that at low temperatures and for @xmath30 the statistical properties of the longitudinal spin susceptibility ( both for the ising and heisenberg exchanges ) are determined by the statistics of the extrema of a certain gaussian process with a drift .
this random process resembles locally a fractional brownian motion with the hurst exponent @xmath35 where @xmath36 .
we recall that the fractional brownian motion with the hurst exponent @xmath37 is the gaussian process @xmath38 with zero mean @xmath39 and the two - point correlation function @xmath40 ^ 2 } = |t - t^\prime|^{2h}$ ] .
we rigorously prove that in the case of ising ( heisenberg ) exchange all moments of static longitudinal spin susceptibility @xmath23 are finite for @xmath41 ( @xmath42 ) . for the ising exchange
we argue also that all moments of dynamic transverse spin susceptibility @xmath25 do not diverge for @xmath41 .
we estimate the tail of the complementary cumulative distribution function for @xmath23 for both ising and heisenberg exchange interactions . we demonstrate that the average static longitudinal spin susceptibility @xmath23 has nonmonotonous dependence on magnetic field in the case of ising exchange .
our results mean that the level fluctuations do not shift the stoner instability from its average position and do not induce a finite temperature transition between the paramagnetic and the ferromagnetic phases .
the outline of the paper is as follows . in sec .
[ sec : formalism ] we introduce the model hamiltonian , derive exact analytical expressions for the corresponding grand canonical partition function and longitudinal static spin susceptibility . in sec .
[ sec : eqspectrum ] we analyze the temperature and magnetic field dependence of longitudinal static spin susceptibility in the case of equidistant single - particle spectrum and anisotropic exchange interaction . in sec .
[ sec : lf : ic ] we present a detailed analysis of the effect of level fluctuations on the longitudinal static spin susceptibility for the cases of ising and heisenberg exchange interactions . in sec .
[ sec : trss ] we compute the transverse dynamical spin susceptibility in the case of equidistant single - particle spectrum and anisotropic exchange interaction and analyze the effect of level fluctuations in the case of ising exchange interaction .
we conclude the paper with summary of the main results and discussion of how our predictions can be experimentally verified ( sec .
[ sec : dc ] ) .
some of the results were published in a brief form in ref .
we consider the following hamiltonian with direct coulomb and anisotropic exchange interactions : @xmath43 the noninteracting hamiltonian , @xmath44 is given as usual in terms of the single - particle creation ( @xmath45 ) and annihilation ( @xmath46 ) operators .
it involves the spin - dependent ( @xmath47 ) single - particle energy levels @xmath48 . in what follows
, we assume that they depend on applied magnetic field @xmath49 via the zeeman splitting , @xmath50 . here
@xmath51 and @xmath52 stand for the land g - factor and the bohr magneton , respectively .
the charging interaction part of the hamiltonian , @xmath53 describes the direct coulomb interaction in a quantum dot in the zero - dimensional approximation , @xmath54 .
here @xmath55 denotes the particle number operator , and @xmath56 is the background charge .
the term @xmath57 represents the anisotropic exchange interaction within the qd .
the total spin operator @xmath58 is defined in terms of the standard pauli matrices @xmath59 . in the case of isotropic heisenberg exchange , @xmath60
, the hamiltonian reduces to the universal hamiltonian which describes a quantum dot in the limit @xmath2 .
@xcite in this case the single - particle levels @xmath61 are random .
their statistics ( in the absence of magnetic field , @xmath62 ) is described by the orthogonal wigner - dyson ensemble .
the hamiltonian with the ising exchange , @xmath63 , and @xmath62 can be used for description of lateral quantum dots with spin - orbit coupling .
@xcite in this case the statistics of @xmath61 is described by the unitary wigner - dyson ensemble .
the grand canonical partition function for the hamiltonian is defined as @xmath64 ( @xmath65 denotes the chemical potential ) .
it can be found by using the following trick .
let us separate @xmath66 into the heisenberg and ising parts : @xmath67 then the time evolution operator in the imaginary time can be rewritten as @xmath68 where @xmath69 .
the exponent in the second line of eq .
indicates that the grand canonical partition function for the hamiltonian can be found in two steps . at first
, one can use well - known results for the partition function for the case of isotropic exchange and effective magnetic field @xmath70 .
@xcite secondly , one needs to integrate over the effective magnetic field @xmath71 with the kernel given in the first line of eq . .
thus we obtain the following exact result for the grand canonical partition function of hamiltonian : @xmath72 here @xmath73 .
the integers @xmath74 and @xmath75 represent the number of spin - up and spin - down electrons , respectively .
the total number of electrons is @xmath76 , and @xmath77 .
we note that for a configuration with given @xmath74 and @xmath75 electrons the total spin equals @xmath78 .
the integers @xmath79 denote @xmath80 projection of the total spin @xmath81 .
the factors @xmath82 and @xmath83 are canonical partition functions for @xmath74 and @xmath75 noninteracting spinless electrons , respectively .
the canonical partition function takes into account the contributions from the single - particle energies and is given by darwin - fowler integral : @xmath84 for the heisenberg exchange interaction , @xmath60 our result coincides with the result known in the literature . @xcite in the case of purely ising exchange interaction , @xmath63 , our result agrees with the result obtained in ref . [ ] .
we note that the result can be also derived directly from hamiltonian with the help of wei - norman - kolokolov transformation ( see appendix [ wnk - der ] ) . in order to analyze the exact result for the grand canonical partition function
, it will be convenient to use the following completely equivalent integral representation : @xmath85 the grand canonical partition function for non - interacting spinless electrons is defined in a standard way @xmath86 the variables @xmath87 and @xmath88 have the meaning of the zero - frequency matsubara components of an electric potential and a magnetic field which can be used to decouple the direct coulomb @xcite and exchange interaction @xcite terms , respectively .
the general expressions and for the grand partition function @xmath89 allow us to extract the results for the longitudinal spin susceptibility : @xmath90 it is worthwhile to mention that in zero magnetic field one can use the equivalent formula @xmath91 to simplify calculations .
as it is well - known,@xcite at @xmath92 ( the regime we are interested in ) we can perform integration over @xmath87 in eq . in the saddle - point approximation .
then the grand canonical partition function is factorized into two multipliers : @xmath93 where @xmath94 describes the effect of charging energy . here
@xmath95 is the solution of the saddle - point equation @xmath96 and @xmath97 stands for the thermodynamic density of states at the fermi level .
we note that in the regime @xmath6 ( which we are interested in ) one can approximate @xmath95 by @xmath98 .
the term @xmath99 } \notag \\ & \hspace{1cm}\times \frac{\sinh ( h ) \sinh \bigl ( ( b+\eta \mathcal{b})h / j_\perp\bigr ) } { \sinh \bigl ( \beta(b+\eta \mathcal{b})/2\bigr ) } e^{-\frac{\beta\mathcal{b}^2}{4 |j_z - j_\perp| } ] } \notag \\ & \hspace{1 cm } \times \prod_{\sigma}e^{\beta \omega_0(\tilde\mu)-\beta \omega_0(\tilde\mu+h \sigma/\beta)}\label{zs1}\end{aligned}\ ] ] describes the contribution due to exchange interaction .
the function @xmath100 = \int\limits_{-\infty}^\infty de\ , \nu_0(e)\ , \notag \\ \times \ln \left [ 1+\frac{\sinh^2(h/2)}{\cosh^2(e/2t)}\right ] \end{gathered}\ ] ] that appears in eq .
depends on a particular realization of the single - particle spectrum via the single - particle density of states @xmath101 .
provided @xmath102 , we can write @xmath103 = \frac{h^2}{\beta \delta}-v(h ) , \label{zzeq2}\ ] ] where @xmath104 .\label{vh_def}\ ] ] here @xmath105 stands for the deviation of the single - particle density of states @xmath106 from its average ( over realizations of the single - particle spectrum ) value : @xmath107 .
the charging energy contribution @xmath108 is independent of the magnetic field and therefore does not affect the spin susceptibility .
we note that the normalization is such that @xmath109 for @xmath110 .
in what follows we will discuss @xmath111 only .
we start our analysis from the case of the equidistant single - particle spectrum , i.e. we completely neglect the effect of level fluctuations ( we set @xmath112 in eq . to zero ) .
we discuss the role of level fluctuations in sec .
[ sec : lf : ic ] . using the integral representation
and we can perform integration over @xmath88 and find @xmath114 here @xmath115 is the energy scale specific for the anisotropic problem that interpolates between @xmath116 ( for @xmath117 ) and @xmath118 ( for @xmath63 ) . the function @xmath119 is defined as follows @xmath120 using eq .
, the zero field longitudinal spin susceptibility can be written as @xmath121 at high temperatures @xmath122 , the result for the zero - field longitudinal static spin susceptibility can be simplified ( cf .
. then we obtain @xmath123 away from the isotropic case ( @xmath117 ) a set of temperature intervals with different temperature behavior of the longitudinal spin susceptibility exists .
below we use the asymptotic result from appendix [ app : f1:zzf1 ] . at temperatures @xmath124 , we find @xmath125 for the temperature range @xmath126 , we obtain @xmath127 if the temperature is within the interval @xmath128 , the zero field longitudinal static spin susceptibility becomes @xmath129 finally , for the lowest temperature range @xmath130 we find ( cf . eq . )
@xmath131 we mention that @xmath23 consists of two contributions ( see eqs . - and ): the one which resembles the fermi - liquid result for spin susceptibility , @xmath132 , and the other which is of curie type , @xmath133 .
such behavior is illustrated in fig .
[ fig : plot_chi_zz ] where the dependence of longitudinal spin susceptibility on temperature and @xmath22 at a fixed ratio @xmath134 is shown .
we emphasize that longitudinal spin susceptibility diverges at @xmath32 regardless of the value of @xmath135 .
to understand the origin of such interesting behavior of the zero field longitudinal spin susceptibility it is useful to rewrite eq . in terms
a series form again : @xmath136 here we used the following result @xmath137 which is valid provided the following conditions hold : @xmath138 and @xmath139 ( see appendix [ app : znzn ] ) . $ ] to the fermi - liquid - like result on dimensionless magnetic field and inverse temperature , @xmath140 and @xmath141 .
we chose @xmath142 and @xmath143.,width=264 ] in the case of large temperatures @xmath144 our results - imply the fermi - liquid behavior of @xmath145 . in this temperature range
all terms except the first one with @xmath146 in the sum over @xmath147 in eq .
cancel each other
. then we find @xmath148 and , consequently , @xmath149 $ ] .
this result implies that the average value of @xmath150 is of the order of @xmath151 \gg 1 $ ] regardless of @xmath135 . at the same time the average value of the squared total spin @xmath152
is of the order of @xmath151 + 1/[\beta(\delta - j_\perp)]$ ] .
therefore , at @xmath153 the total spin strongly fluctuates in all three directions so that @xmath154 whereas for @xmath155 the total spin fluctuates along the @xmath80 axis only so that @xmath156 .
we mention the unusual ( inverse square - root ) temperature dependence of the longitudinal spin susceptibility in eq . . however , the result is valid in a temperature range that exists only if @xmath157 .
then the restrictions for the temperature become @xmath158 .
therefore we can use the arguments from the previous paragraph . in order to explain the @xmath159 dependence of @xmath23
, one needs to perform the perturbation expansion in @xmath160 for eq . .
at low temperatures @xmath161 our results - and imply a curie type longitudinal spin susceptibility . in this case the second term in brackets in the right hand side of eq . can be neglected .
the sum over @xmath147 can be estimated by the integral which is dominated by @xmath162 .
then we find @xmath163 this estimate yields the typical value of @xmath164 $ ] and , thus , the curie type behavior of the longitudinal spin susceptibility : @xmath165 ^ 2 $ ] .
therefore , at relatively low temperatures @xmath166 the configuration with a non - zero total spin @xmath167 $ ] gives the main contribution to the thermodynamic quantities . }
{ d ( \delta - j_z)}$ ] on @xmath168 and @xmath169 for @xmath170 . in the left upper region
the curie - like behavior dominates the fermi - liquid - like result . in the right lower region the curie - like correction to the fermi - liquid - like result , @xmath171\propto \frac{1}{(\delta - j_z)^2}$ ] is small .
red region corresponds to the intermediate regime in which there is a correction @xmath172\propto \frac{1}{t^{1/2}(\delta - j_z)^{3/2}}$ ] to the fermi - liquid result due to transverse degrees of freedom.,width=264 ] for small magnetic fields @xmath173 the longitudinal spin susceptibility @xmath174 can be well approximated by the zero field result . for larger magnetic fields
@xmath175 there are two regions of temperature with different behavior . in the range of temperatures @xmath176
the longitudinal spin susceptibility becomes linear in temperature ( cf .
): @xmath177 at higher temperatures @xmath178 the temperature dependence of the longitudinal spin susceptibility saturates : @xmath179 in the limit of large magnetic fields the ground state energy for the configuration with the total spin projection @xmath180 is equal to @xmath181 .
thus the projection of the total spin in the ground state is @xmath182 $ ] .
it allows us to estimate the longitudinal spin susceptibility as @xmath183 $ ] in agreement with eq . .
using the integral representation and , we integrate over @xmath88 and obtain @xmath185 here @xmath186 stands for the jacobi theta function . since @xmath187 , the jacobi theta function @xmath188 becomes equal to unity .
then for @xmath189 we find @xmath190 where @xmath191 at temperatures @xmath192 , with the help of eq .
we obtain that the longitudinal spin susceptibility is given by eq . .
in the temperature range @xmath193 , the behavior of @xmath23 is described by eq . .
in the case of an easy plane anisotropy the interplay between fermi - liquid and curie - like temperature dependencies of the longitudinal spin susceptibility can be explained in exactly the same way as it was done for the case of an easy axis anisotropy .
the longitudinal static spin susceptibility is almost insensitive to the presence of a small magnetic field @xmath194 . in the opposite case @xmath195
, one can neglect @xmath196 in the @xmath197 s arguments in eq . .
then at @xmath195 we find @xmath198 the result implies that for magnetic fields in the range @xmath199 the longitudinal spin susceptibility is described by eq .
whereas for @xmath200 , @xmath23 is given by eq . .
as it was explained above , the hamiltonian describes a quantum dot in the zero - dimensional limit for the ising and heisenberg exchange interactions only .
therefore , it is reasonable to study the effect of level fluctuations on the results obtained above for @xmath63 and @xmath60 .
we start with the case of ising exchange . to simplify the general result in the case of the ising exchange ,
it is convenient to make a change of variable @xmath201 , to take the limit @xmath202 , and then to integrate over @xmath71 . thus we find @xmath203 where @xmath204 the information on fluctuations of single - particle levels is encoded in the even random function @xmath112 via the density of states ( see eq . ) .
we remind that the single - particle density of states @xmath205(e ) has non - gaussian statistics .
@xcite however , for @xmath206 the function @xmath112 is a gaussian random variable with zero mean value .
@xcite the two - point correlation function of @xmath207 can be written as follows ( see appendix [ app : lh ] ) : @xmath208 . \label{corrvv}\end{gathered}\ ] ] here @xmath209 is the euler digamma function and @xmath210 is the euler mascheroni constant . in the case of the ising exchange , the parameter @xmath211 in eq .
is equal to @xmath212 since the energy levels @xmath61 in the hamiltonian are described by the unitary wigner - dyson ensemble ( class a ) .
@xcite the asymptotics of @xmath213 are as follows : @xcite @xmath214 according to eq . , the average longitudinal spin susceptibility @xmath215 is determined by the quantity @xmath216 .
although @xmath112 is a gaussian random variable , exact evaluation of @xmath216 for arbitrary values of @xmath217 and @xmath218 is a complicated problem .
we start from the perturbation theory in the correlation function @xmath219 . expanding expression for @xmath220 to the second order in @xmath207 and performing the averaging of @xmath221 with the help of eq . , we find @xmath222 . \label{averlnxi1}\end{aligned}\ ] ] there exist four regions of different behavior of @xmath216 .
they are shown in fig .
[ figureregions ] .
it is convenient to introduce the renormalized exchange @xmath223 . in the region
i , @xmath224 , the arguments of @xmath220 satisfy the condition @xmath225 .
the latter allows one to use the asymptotics of @xmath213 for @xmath226 ( see eq . )
. then we find @xmath227 .\label{eq : xi : regi}\ ] ] hence we obtain the following result for the average longitudinal spin susceptibility at temperatures @xmath228 : @xmath229 .\quad \label{eq : chizz : regi}\end{aligned}\ ] ] in the region i the corrections to the longitudinal spin susceptibility are always small and , therefore , the perturbation theory is well justified .
we present a more transparent way for derivation of eq . .
at first , one can substitute @xmath230 for @xmath231 in the expression ( with @xmath63 ) for the equidistant spectrum .
secondly , we expand @xmath23 to the second order in the deviation @xmath232 . finally , one can perform averaging with the help of the relation @xcite @xmath233 and obtain the result ( with @xmath189 ) . in the region
ii , @xmath234 , one can perform an expansion in @xmath235 in the right hand side of eq . since the condition @xmath236 holds .
however , the argument of @xmath237 is typically large and we need to use its asymptotics for @xmath238 ( see eq . ) .
then we obtain @xmath239 therefore , the average longitudinal spin susceptibility in the region ii ( @xmath240 is as follows : @xmath241 at zero magnetic field we check that the contribution of the second order in @xmath237 to @xmath242 is of order of @xmath243 ( see appendix [ app:2dorder ] ) . therefore the perturbation theory in the two - point correlation function of @xmath207 is justified for @xmath244 only . in this regime
the variance of @xmath23 is small @xmath245/\overline{\chi}_{zz}^2
\sim \bar{j_z}/(\pi^2 \bm{\beta } t ) \ll 1 $ ] ( see appendix [ app:2dorder ] ) .
therefore , at @xmath244 one can expect the normal distribution of @xmath23
. finally , in the region iii , @xmath246 , the typical value of @xmath247 contributing to the integral in the right hand side of eq .
can be not only of the order unity but also of the order of @xmath248 . in the latter case ,
since @xmath249 one needs to use the asymptotics of @xmath213 for @xmath238 ( see eq . )
. then we find @xmath250 we thus obtain the average longitudinal spin susceptibility in the region iii ( @xmath251 ) : @xmath252 for magnetic fields @xmath253 the effect of level fluctuations is suppressed and the perturbation theory is justified . at @xmath254
the result agrees with the result whereas at @xmath255 the corrections due to level fluctuations in and become of the same order .
due to fluctuations for the case of ising exchange in the plane of dimensionless magnetic field and inverse temperature , @xmath140 and @xmath256 .
note that in our analysis we assume @xmath92.,width=226 ] results and imply non - monotonous behavior of the average longitudinal spin susceptibility with magnetic field @xmath257 in the temperature range @xmath258 ( see fig .
[ fig : chi_zz_on_b ] ) . the susceptibility @xmath259 as a function of @xmath257 has a minimum at @xmath260 . in the region of strong fluctuations
@xmath261 we expect similar behavior of the average longitudinal spin susceptibility .
although the result is derived for @xmath92 , for @xmath262 it can be obtained from the following zero - temperature arguments .
the difference in the ground state energies for the state with projections @xmath263 and @xmath180 of the total spin can be estimated as @xmath264 here @xmath265 is the fluctuation of the energy window in which there are @xmath266 levels on average .
it can be expressed as @xmath267 where @xmath268 is the fluctuation of the number of single - particle levels in the strip with @xmath266 levels in average . from the random matrix theory
it is well known that @xcite @xmath269 .
\label{eq : crit2}\ ] ] comparing the energies of the ground states with total spin projections @xmath263 and @xmath180 , we find from eq . that @xmath270 .
\label{eq : crits3}\ ] ] hence the average longitudinal spin susceptibility can be estimated as @xmath271 , \ ] ] where @xmath272 .
using eq .
, we reproduce the result .
$ ] due to fluctuations on @xmath140 ( see eqs . and ) .
the temperature @xmath273$].,width=226 ] the average longitudinal spin susceptibility is mostly affected by the level fluctuations in the region ii ( @xmath274 ) .
the perturbative result loses its validity at @xmath275 . such a regime is realized in the close vicinity of the stoner instability @xmath276 . in this case of strong fluctuations it is useful to know the distribution function of @xmath23 rather than the average value . in the range of temperatures @xmath277 , the integral in the right hand side of eq .
is dominated by large values of @xmath278 .
then , using the asymptotic expression , one can check that for @xmath279 the two - point correlation function is homogeneous of degree two : @xcite @xmath280 with the help of eq .
, at zero magnetic field @xmath189 and for @xmath281 , eqs . and can be simplified to @xmath282 we remind that the normalization is such that @xmath283 at @xmath284 .
according to eq . for @xmath285 ,
the grand canonical partition function increases as @xmath22 increases .
hence it follows that @xmath286 . according to eq .
, the statistics of the zero field longitudinal spin susceptibility is determined by the single parameter @xmath287^{1/2}$ ] .
the gaussian random process @xmath288 has zero mean and is even in @xmath88 , @xmath289 .
its two - point correlation function reads @xmath290 hence we find that @xmath291 ^ 2 } = - 2 u^2 \ln |u| + o(u^2)=o(u^{2h } ) \label{eq : local}\ ] ] for any @xmath292 .
thus the trajectories of @xmath288 are continuous and its increments are strongly positively correlated ( see fig .
[ fig : process ] ) .
in fact the process @xmath288 is in many aspects close to the ballistic one @xmath293 with @xmath294 being a gaussian random variable ( recall that @xmath295 is the unique process with @xmath296 ) .
the process @xmath288 has arisen before in a seemingly unrelated context .
@xcite ; dashed lines @xmath297 are guides for the eye.,width=226 ] we are interested in the complementary cumulative distribution function @xmath298 , i.e. the probability that @xmath299 exceeds @xmath300 : @xmath301 .
it has the following properties : @xmath302 , @xmath303 and @xmath298 is monotonously decreasing as @xmath300 increases .
the average moments of @xmath299 can be conveniently written as @xmath304^k}= k \int_{0}^{\infty}dw w^{k-1 } \mathcal{p}(w)$ ] .
although a closed analytical expression for the complementary cumulative distribution function is not known , we bound @xmath298 from above to prove that all moments of @xmath299 ( and consequently all moments of @xmath23 ) are finite for @xmath41 . at first
, we split the gaussian weight @xmath305 in the integral in the right hand side of eq . and obtain ( @xmath306 is an arbitrary splitting parameter ) @xmath307 the inequality allows us to reduce the problem of finding an upper bound for @xmath298 to the statistics of the maxima of the gaussian process @xmath308 which locally resembles a fractional brownian motion with a drift .
indeed , from eq .
we find @xmath309 to give an upper bound for the probability @xmath310 we employ the slepian s inequality .
@xcite let us consider an auxiliary gaussian process @xmath311 where @xmath312 is the standard brownian motion ( @xmath313 ; the hurst exponent @xmath314 ) .
for any interval @xmath315 the sample paths @xmath316 and @xmath317 are bounded .
the following relations hold : @xmath318 ^ 2 } \geqslant \overline{[y_\gamma(h_1)-y_\gamma(h_2)]^2 } .\end{gathered}\ ] ] the first two equalities are trivially satisfied while the last inequality follows from an easily verifiable inequality @xmath319
^ 2}\leqslant 8 r \ln 2 $ ] for @xmath320 .
then the processes @xmath321 and @xmath322 satisfy the slepian s inequality : @xmath323 for all real @xmath324 . using a well - known result for the brownian motion with a linear drift ( see e.g. , ref . [ ] ) @xmath325 we find the following upper bound for the complementary cumulative distribution function : @xmath326 \right\ } .
\label{eq : tail}\ ] ] from eq .
it follows that for @xmath327 all moments of @xmath299 ( and hence all moments of @xmath23 ) are finite for @xmath41 .
therefore even in the presence of the strong level fluctuations the stoner instability occurs at @xmath32 only . for @xmath41 and for temperatures @xmath92
the quantum dot is in the paramagnetic state .
for @xmath328 the saddle - point approximation in eq . becomes exact and the statistics of @xmath299 reduces to the statistics of maxima of the process @xmath329 . as it can be seen from rescaling of @xmath88 , the probability that the maximum of @xmath330 exceeds @xmath324 equals the probability that the maximum of @xmath331 defined on @xmath332 exceeds @xmath333 . from the results of hsler and piterbarg@xcite it follows that the large-@xmath324 tail of @xmath334 is determined by a small vicinity of the point @xmath335 where the variance of @xmath336 attains its maximum @xmath337 .
furthermore , should we have a finite limit @xmath338 ^ 2 } } { k^2(s - t ) } > 0 \label{eq : kdef}\ ] ] for some function @xmath339 regularly varying at @xmath116 with index @xmath340 , the precise asymptotics would read @xmath341 , \quad w / z^2 \gg 1 .
\label{eq : probtail}\end{gathered}\ ] ] here @xmath342 stands for the functional inverse of @xmath339 . in our case
. translates into @xmath343 which is regularly varying with index @xmath344 [ recall that a function @xmath345 is regular varying at @xmath346 with index @xmath347 if @xmath348 for any @xmath349 .
the result of ref .
[ ] is therefore not directly applicable , but we believe this to be a technicality . in analogy with a similar situation for fractional brownian motion , we expect the asymptotics to hold with only the @xmath300-independent factor @xmath350 modified .
note that the exponential part can be tracked to be the tail of a normal distribution with variance @xmath337 taken at @xmath351 , and that it had been correctly reproduced by our initial estimate .
therefore we find with logarithmic accuracy that the tail of the complementary cumulative distribution function is given by ( @xmath352 \delta j_z/[t ( \delta - j_z)]$ ] ) @xmath353 this result is valid in the temperature range @xmath327 and is consistent with the upper bound . to illustrate the result we approximate the gaussian process @xmath288 by a degenerate one @xmath354 , where @xmath294 is the gaussian random variable with zero mean @xmath355 and variance @xmath356 . substituting the process @xmath357 for @xmath288 into the right hand side of eq .
, we estimate the partition function as @xmath358 $ ] .
the large values of @xmath111 correspond to large negative values of @xmath294 such that @xmath359 .
therefore , the tail of distribution of @xmath299 is simple exponential .
hence we find that for @xmath328 the tail of the complementary cumulative distribution function @xmath298 is given by eq . without the logarithm in the pre - exponent .
as shown in fig .
[ fig : ccdf ] the overall behavior of @xmath298 for @xmath328 is well enough approximated by the complementary cumulative distribution function for the degenerate process @xmath357 .
also we mention that the behavior of @xmath298 for @xmath328 is very different from its behavior at @xmath360 . for the later
, @xmath298 is given by the complementary cumulative distribution function of the normal distribution ( see fig . [
fig : ccdf ] ) .
equation implies that the average moments of @xmath299 scale as @xmath361 for @xmath328 .
hence for @xmath261 the @xmath362-th moment of the spin susceptibility is given by @xmath363^k , \quad k=1 , 2 , \dots .
\label{eq : chi : kth}\ ] ] the result can be obtained from the saddle - point analysis of the integral in the right hand side of eq .
, i.e. , in essence , by larkin - imry - ma type arguments . @xcite the scaling of the average spin susceptibility ( eq . with @xmath364 ) was proposed in ref .
[ ] using arguments of larkin - imry - ma type .
on @xmath365 at @xmath366 computed numerically for @xmath367 ( @xmath368 ) ( upper solid curve ) and @xmath369 ( @xmath370 ) ( lower solid curve ) .
the black dotted curve is the complementary cumulative distribution function for the normal distribution with mean and variance as one can find from the lowest order perturbation theory in @xmath207 for @xmath366 and @xmath367 ( cf .
eqs . and ) .
the red dashed curve is the complementary cumulative distribution function of the degenerate process @xmath357 for @xmath366 and @xmath369 .
inset : comparison of the tail of @xmath298 computed numerically for @xmath369 ( @xmath370 ) and asymptotic result .,width=226 ] for the case of the isotropic exchange , @xmath371 , the integration over @xmath71 in eq .
becomes trivial .
then for @xmath92 we obtain @xcite @xmath372 where @xmath373 since in the absence of magnetic field @xmath89 grows with increase of @xmath15 ( see eq . ) , one can check that for the heisenberg exchange @xmath286 .
the detailed results of the perturbative expansion in @xmath207 for the longitudinal spin susceptibility can be found in ref .
[ ] . similarly to the case of the ising exchange ,
the effect of fluctuations is important at @xmath189 and @xmath374 $ ] . in this range of parameters the typical value of @xmath278 in the integral in the right hand side of eq .
is large , @xmath375 where @xmath376 . then for @xmath189 eq
can be rewritten as @xmath377 where @xmath378^{1/2}$ ] . here
@xmath379 which corresponds to the orthogonal wigner - dyson ensemble .
the complementary cumulative distribution function @xmath380 can be estimated in a similar way as in the previous section .
writing @xmath381 \notag \\
\times \max\limits_{h\geqslant 0 } \bigl \{e^{-h^2-(z/\sqrt{\gamma})v(h)}\bigr \ } , \label{eq : split : iso}\end{gathered}\ ] ] with arbitrary splitting parameter @xmath382 ( @xmath306 ) , we obtain the following upper bound : @xmath383 \biggr \ } , \label{eq : tail : iso}\end{aligned}\ ] ] where @xmath384\}$ ] .
this upper bound implies that all moments of @xmath299 ( and of @xmath23 ) are finite for @xmath42 . at @xmath385
the integral in the right hand side of eq . can be evaluated in the saddle - point approximation , reducing the statistics of @xmath299 to the statistics of maxima of the process @xmath386 . then using as in the previous section the results of hsler and piterbarg@xcite
, we find the tail of the complementary cumulative distribution function at @xmath387 $ ] is given by @xmath388 ( see eq . ) .
we note that for this tail the drift term @xmath389 in the process @xmath330 is not important .
the typical value of @xmath88 contributing to the integral in the right hand side of eq . is @xmath390 .
then for @xmath328 we find , with logarithmic accuracy , @xmath391 .
hence for @xmath392 the average @xmath362-th moment of the longitudinal spin susceptibility can be estimated as @xmath393 where @xmath394 $ ] is the spin susceptibility in the absence of level fluctuations .
we note that for @xmath392 the scaling of the average longitudinal spin susceptibility similar to eq . with @xmath364
was derived in ref .
the transverse spin susceptibility is defined as follows ( see , e.g. , ref . [ ] ) @xmath395e^{-\beta h}\bigr ) , \label{eq : chi_perp_def}\ ] ] where @xmath396 . since
, in contrast with @xmath397 , the operators @xmath398 , @xmath399 of the total spin do not commute with the hamiltonian @xmath37 ( for @xmath400 ) , the transverse spin susceptibility can acquire non - trivial frequency dependence . in order to find the dynamic transverse spin susceptibility we use the heisenberg equations of motion for the spin operators : @xmath401 $ ] .
since the operator @xmath180 commutes with the hamiltonian , it has no dynamics , @xmath402 . for the other components of the total
spin we find @xmath403 using expressions , we integrate over time in eq . and obtain the following operator expression for the transverse spin susceptibility : @xmath404-\hat{s}_z\bigr ) e^{-\beta h } } { \omega+b+(j_\perp - j_z)(2\hat{s}_z+\sigma)+i0^+ } .
\label{eq : chi_perp_2}\end{aligned}\ ] ] since operators @xmath397 and @xmath405 commute with @xmath37 , one easily evaluates the trace in eq . with the help of eq . .
thus we derive the exact result for the dynamic transverse spin susceptibility : @xmath406-l}{\omega+b+(j_\perp - j_z)(2l+\sigma)+i0^+ } .
\label{eqzb2}\end{aligned}\ ] ] in what follows we will be interested in the imaginary part of @xmath25 .
the real part can be restored from the kramers - kronig relations .
using eq . , the imaginary part of the dynamic transverse spin susceptibility can be written as @xmath407 here we introduce the fourier transform of the partition function @xmath408 in the complex magnetic field @xmath409 : @xmath410 as it follows from eq . , the imaginary part of the transverse spin susceptibility obeys the sum rule @xmath411 where the magnetization @xmath412 . since at @xmath189 the function @xmath413 is even , the imaginary part of the zero - field transverse spin susceptibility is odd in frequency : @xmath414 , so the sum rule is trivially satisfied .
we mention that in the case of an isotropic exchange , @xmath117 , eq . becomes trivial , @xmath415 . in this case
the behavior of the transverse spin susceptibility is fully determined by the behavior of the magnetization @xmath416 .
therefore , in what follows we shall not discuss the transverse spin susceptibility for the isotropic exchange .
at first we consider the case of an equidistant single - particle spectrum and , therefore , neglect effects related to the level fluctuations . as it was discussed in sec .
[ sec : longsusc ] , for @xmath138 the partition function can be factorized in accordance with eq . .
since the factor @xmath108 does not depend on the magnetic field , it does not influence the results for @xmath25 and we omit it below in this section .
it implies that @xmath111 , @xmath417 , and @xmath418 should be substituted for @xmath89 , @xmath413 , and @xmath408 in eqs . - , respectively . using eq . for the equidistant single - particle spectrum we can rewrite @xmath417 in the following way : @xmath419 next ,
performing integration over @xmath420 , we obtain the following result : @xmath421 . \label{im_general}\end{aligned}\ ] ] in the case @xmath422 , applying the euler - maclaurin formula to estimate the sums over @xmath423 , we find @xmath424 in the opposite case @xmath425 , the term with @xmath426 in the right hand side of eq .
provides the main contribution
. then we obtain @xmath427 we note that for @xmath63 , both expressions and coincide and are valid , in fact , for arbitrary @xmath428 . according to eq .
, @xmath429 is represented as the sum of delta - peaks . since their positions are independent of the realization of single - particle levels , the delta - peaks survive averaging of @xmath429 over level fluctuations .
therefore , in order to discuss the frequency dependence of the transverse spin susceptibility in a form of a smooth curve , we assume some natural broadening @xmath430 for these delta - peaks . then the sum over @xmath428 in eq . can be replaced by an integral and we obtain @xmath431 where @xmath432 $ ] . in the limit of large frequencies or large magnetic fields , @xmath433 ,
the imaginary part of the transverse spin susceptibility is exponentially small : @xmath434 .
\label{eq : imagchi : tail}\end{gathered}\ ] ] in the absence of magnetic field , @xmath189 , @xmath435 is an odd function of the frequency @xmath436 .
for @xmath437 the imaginary part of the transverse spin susceptibility has linear behavior : @xmath438\end{aligned}\ ] ] where @xmath439 of the traverse spin susceptibility on @xmath134 for @xmath440.,width=226 ] the slope of @xmath429 at @xmath441 has different behaviors for @xmath442 and for @xmath443 . in the interval
@xmath444 the slope grows monotonously with the increase of @xmath135 and diverges at @xmath60 . in the range @xmath445 the slope
has a minimum ( see fig .
[ fig : imchislope ] ) .
the imaginary part of the zero field transverse spin susceptibility has two extrema ( a minimum at a negative frequency and a maximum at a positive frequency ) . in the case @xmath446 and @xmath447 the positions of the extrema can be estimated as @xmath448 .
\label{eq : ext : gen}\end{gathered}\ ] ] the behavior of @xmath25 as a function of frequency is shown in fig .
[ fig : dssplots ] . in the presence of a magnetic field @xmath429
is shifted along the frequency axis and becomes asymmetric ( see fig .
[ fig : dssplots ] ) .
on @xmath436 for @xmath449 and several values of @xmath135 : @xmath450 ( red solid line ) , @xmath451 ( blue dashed line ) and @xmath63(green dotted line ) .
the curves shrink to @xmath441 as one moves closer to the isotropic case.,width=226 ] it is worthwhile to discuss the case of the ising exchange ( @xmath63 ) in more detail . in the regime of small frequencies and magnetic fields , @xmath452 ,
the imaginary part of the dynamic spin susceptibility reads @xmath453 ^ 2}{4j_z^2(\delta - j_z)}\right \ } .
\label{eqnising2s}\end{aligned}\ ] ] although @xmath27 is asymmetric in the presence of magnetic field , it still vanishes at zero frequency , @xmath454 . in the opposite limit @xmath455 , from eq .
we find @xmath456 ^ 2}{4j_z^2(\delta - j_z)}\right \ } .
\label{eqnising2l}\end{aligned}\ ] ] in the case @xmath189 , our results and coincide with the small and large frequency asymptotics of the result obtained in ref .
the presence of a magnetic field leads to a shift of the extrema of the imaginary part of the dynamic spin susceptibility according to @xmath457 the above results for the dynamic spin susceptibility have been obtained without taking the level fluctuations into account .
below we consider how the level fluctuations affect the dynamic spin susceptibility in the case of the ising exchange . as we shall demonstrate , the effect of level fluctuations on @xmath27 is small in most cases .
since the effect of level fluctuations is suppressed by the magnetic field , below we consider only the case @xmath189 .
we start from a generalization of eq . to an arbitrary spectrum ( see appendix [ app : znzn ] ) : @xmath458 with the help of eqs .
, , and we rewrite the imaginary part of the dynamic spin susceptibility as follows : @xmath459 \biggl |_{n=(\sigma j_z-\omega)/(2j_z ) } .\end{gathered}\ ] ] here the random function @xmath460 is equal to unity in the absence of level fluctuations ( for @xmath461 ) .
expanding the right hand side of eq . to the second order in @xmath207 we find @xmath462 in the high temperature regime , @xmath463 , and for @xmath464 , all three integrals in the right hand side of eq .
are of the same order . using the asymptotic expression for the function @xmath213 at @xmath465 , we obtain the following result for the imaginary part of the average dynamic spin susceptibility at low frequencies @xmath466 and high temperatures @xmath463 : @xmath467 .
\label{eq : ising : fluc1}\end{aligned}\ ] ] here @xmath468 is given by eq . with @xmath189 .
we mention that eq . can be obtained from eq .
if one substitutes @xmath230 for @xmath231 and performs averaging with the help of eq . .
in the regime of low frequencies and high temperatures the effect of level fluctuations is small . in the case of high frequencies and high temperatures , @xmath469 , the first and second lines in the right hand side of eq .
provide the main contribution . then with the help of the asymptotic expression for @xmath213 at @xmath470 we find that for @xmath471 the imaginary part of the average dynamic spin susceptibility can be written as @xmath472 here @xmath468 is given by eq . with @xmath189 .
we note that the result is valid provided @xmath473 ^ 2\ll \pi^2\bm{\beta}$ ] so that the perturbation theory in @xmath207 is justified .
we emphasize that although the result eq .
is valid at high temperatures @xmath474 , it can not be obtained from eq . by a substitution of @xmath230 for @xmath231 and averaging with the help of eq . .
in the case of low temperatures @xmath475 , the @xmath423-independent contributions in the right hand side of eq .
vanish in the leading order .
using the asymptotic result for @xmath213 at @xmath470 ( see eq . )
we obtain @xmath476 hence we find the following result for the imaginary part of the average dynamical spin susceptibility at low frequencies , @xmath477 : @xmath478 here @xmath468 is given by eq . with @xmath189 . in the temperature range @xmath477 the effect of level fluctuations
is suppressed by an additional small factor @xmath479 .
thus we expect that the perturbation theory is valid even for @xmath480 $ ] .
in the high frequency regime , @xmath481 , and at low temperatures @xmath475 we obtain from eq . the following result for the average dynamical spin susceptibility : @xmath482 here @xmath468 is given by eq . for @xmath189 .
the perturbation theory is justified for @xmath483 ^ 2 , \delta j_z/[t(\delta - j_z ) ] \right \ } \ll \pi^2\bm{\beta}$ ] .
we remind that the maximum of @xmath468 is close to the frequency @xmath484 . then , as it follows from eq . , the fluctuations yield an enhancement of the maximal value of the average dynamical spin susceptibility of the relative order @xmath485 $ ] . due to fluctuations
there is a small shift of the maximum towards zero frequency , @xmath486 . since @xmath487 , we can bound the function @xmath488 from above as @xmath489 therefore @xmath488 remains finite for @xmath41 .
thus , in spite of the level fluctuations , the stoner instability in @xmath27 emerges only at @xmath32 . according eq .
, averaging over level fluctuations keeps @xmath27 finite .
however , the form of the curve can be changed drastically in the regime of strong fluctuations . to estimate @xmath27 at @xmath490 $ ] we substitute the degenerate process @xmath357 for @xmath112 into eq . .
then a straightforward calculation yields @xmath491\ ] ] for @xmath492 .
we recall that @xmath493 $ ] .
this result implies that @xmath494 has a minimum and a maximum at frequencies @xmath495 due to strong fluctuations of the single - particle levels the frequency of the extremum shifts towards higher frequencies ( in comparison with the corresponding result without fluctuations ) and becomes temperature independent .
the fluctuations do not affect considerably the values of @xmath494 at the extrema .
therefore the slope at @xmath441 becomes smaller , @xmath496 .
in this paper we have addressed the spin fluctuations and dynamics in quantum dots and ferromagnetic nanoparticles . within the framework of the model hamiltonian which is an extension of the universal hamiltonian to the case of uniaxial anisotropic exchange interaction , we have derived exact analytic expressions for the static longitudinal and dynamic transverse spin susceptibilities for arbitrary single - particle spectrum . for the equidistant single - particle levels we analyzed the temperature and magnetic field dependence of @xmath23 .
for @xmath497 the zero - field longitudinal spin susceptibility has temperature dependence of type @xmath33 ( curie - like ) or @xmath34 .
this indicates that the destruction of the mesoscopic stoner instability by uniaxial anisotropy is not abrupt .
the magnetic field suppresses the temperature dependence of @xmath23 making spins aligned along the field . for the case of the ising exchange interaction we study the effect of single - particle level fluctuations on @xmath23 in detail
the temperature dependence of @xmath23 appears only due to level fluctuations .
we showed that at low temperatures and for @xmath30 ( where fluctuations are strong ) the statistical properties of the longitudinal spin susceptibility are determined by the statistics of the extrema of a gaussian process with a drift .
this random process resembles locally a fractional brownian motion .
we rigorously prove that in this regime of strong fluctuations all moments of zero - field static longitudinal spin susceptibility @xmath23 are finite for @xmath41 and temperatures @xmath498 .
this means that the stoner instability is not shifted by the level fluctuations away from its average position at @xmath499 .
also , our results imply that randomness in the single - particle levels does not lead to a transition at finite @xmath500 between a paramagnetic and a ferromagnetic phase .
we expect that these conclusions hold also for temperatures @xmath501 . however , we can not argue it within our approach ; a separate ( perhaps numerical ) analysis is needed .
we found that the magnetic field suppresses the effect of level fluctuations on the average longitudinal spin susceptibility .
interestingly , the dependence of @xmath215 on @xmath257 is non - monotonous with a minimum .
we extended the analysis of the effect of strong level fluctuations to the case of heisenberg exchange .
we demonstrated that in this case the very same conclusions as for the ising exchange hold . for equidistant single - particles levels we computed the temperature and magnetic field dependence of the imaginary part of the transverse spin susceptibility @xmath27 .
we found that it always has a maximum and a minimum whose positions tend to zero frequency with the decrease of anisotropy .
the height of the maximum and the depth of the minimum increase with the decrease of anisotropy . for the ising exchange we took into account the effect of single - particle level fluctuations on @xmath27 .
we argued that all moments of the dynamic transverse spin susceptibility @xmath25 do not diverge for @xmath41 .
we found that at @xmath502 the positions of the extrema of @xmath503 have a @xmath504-type dependence at high temperatures and become independent of @xmath24 at low temperatures ( in the regime of strong level fluctuations ) .
interestingly , the level fluctuations do not change the minimal and maximal values of @xmath503 significantly .
our results , in principle , can be checked in quantum dots and nanoparticles made of materials close to the stoner instability , such as co impurities in a pd or pt host , fe or mn dissolved in various transition - metal alloys , ni impurities in a pd host , and co in fe grains , as well as nearly ferromagnetic rare - earth materials .
@xcite however , to test our most interesting results on spin susceptibility in the regime of strong level fluctuations one needs to explore the regime @xmath505 .
the closest material to the stoner instability we are aware of , yfe@xmath506zn@xmath507 , @xcite has the exchange interaction @xmath508 which is near the border of the regime with strong level fluctuations at low temperatures .
we acknowledge useful discussions with y. fyodorov , y. gefen , a. ioselevich , a. shnirman and m. skvortsov .
the research was funded in part by rfbr grant no .
14 - 02 - 00333 , the council for grant of the president of russian federation ( grant no .
mk-4337.2013.2 ) , dynasty foundation , ras programs `` quantum mesoscopics and disordered systems '' , `` quantum physics of condensed matter '' and `` fundamentals of nanotechnology and nanomaterials '' , and by russian ministry of education and science .
in this appendix we present a derivation of the partition function for the hamiltonian . for simplicity , we consider the case of zero magnetic field .
we use the notation of ref . [ ] .
we start from the hamiltonian @xmath510 .
then the corresponding partition function can be written as @xmath511 , where @xmath66 is given by eq .
and @xmath512 denotes the averaging over all many - particle states with the weight @xmath513 . to get rid of terms of the fourth order in electron operators in the exponent
we apply the hubbard - stratonovich transformation , @xmath514 } = \lim_{n\rightarrow \infty}\int \bigl[\prod_{n=1}^{n}d\bm\theta_n\bigr ] \prod_\alpha \mathcal{t}e^{it{\bm\theta_n \bm s_\alpha}/n}\notag \\
\times \exp \left [ -\frac{i \delta}{4}\sum\limits_{n=1}^{n}\left ( \frac{\theta^2_{x , n}+\theta^2_{y , n}}{j_\perp}+\frac{\theta^2_{z , n}}{j_z}\right ) \right ] , \label{evol}\end{gathered}\ ] ] where @xmath515 . here and further we omit the normalization factors .
we restore the correct normalization factor ( depending on @xmath516 , and @xmath22 ) in the final result . to calculate the time - ordered exponent ( @xmath315 ) of non - commuting operators it is useful to apply the wei - norman - kolokolov transformation @xcite allowing us to rewrite @xmath315-exponent as a product of usual exponents : @xmath517 where @xmath518 and @xmath519 .
we employ the initial condition @xmath520 . the new variables @xmath521 and @xmath522 are related to the variables @xmath523 as follows :
@xmath524 the vector variables @xmath525 are real but the transformation assumes that the contour of integration in eq . has been rotated . in order to preserve the number of variables we impose the following constraints on the new variables : @xmath526 and @xmath527 .
we mention that the transformation assumes such a discretization of time that the quantity @xmath528 corresponds to @xmath529 in the continuous limit .
in general , there are a lot of discrete representations of @xmath529 , e.g. of the form @xmath530 with @xmath531 .
however , the choice of the symmetric one ( with @xmath532 ) is optimal since it allows us to work within the first order in @xmath533 in eq . .
we note that the jacobian of the transformation is equal to @xmath534 .
having in mind the further usage of the results , we rewrite @xmath535 as the product @xmath536 with @xmath537 .
now , rewriting two exponents in terms of two sets of new variables , we obtain @xmath538 ^ 2}\biggr \ } \prod\limits_\alpha \operatorname{tr}\prod_{p=\pm } \left [ e^{-ipt_p \epsilon_\alpha n_\alpha } e^{p s^{-p}_\alpha \kappa_{p , n_p}^p } e^{i s_\alpha^z \delta \sum_{n=1}^{n_p } \rho_{p , n } } e^{i s_\alpha^p \delta \sum_{n=1}^{n_p } \kappa_{p , n}^{-p } \exp(-
i p \delta \sum_{j=1}^n \rho_{p , j } ) } \right ] , \label{eq2}\end{aligned}\ ] ] where @xmath539 and @xmath540 .
let us introduce a set of auxiliary variables @xmath541 to get rid of terms of the fourth order in @xmath542 s : @xmath543 ^ 2}=\int d\eta_{p , n_p } e^ { \frac{ip \delta\varkappa}{4 j_\perp } \eta_{p , n_p}^2 } \notag \\
\times e^ { -\frac{ip \delta\varkappa}{2 j_\perp } \eta_{p , n_p}\kappa_{p , n_p}^{-p}(\kappa_{p , n_p}^p+\kappa_{p , n_p-1}^p ) ] } .\end{gathered}\ ] ] to proceed with the evaluation of @xmath544 we need to calculate the following integrals over @xmath545 s : @xmath546 following ref .
[ ] , we introduce the new variables @xmath547 where @xmath548 such choice of @xmath549 and @xmath550 allows us to get rid of terms of the third order ( second order in @xmath551 s and first order in @xmath552 ) in eq . within the first order in @xmath533 .
the last term in the right hand side of the second equation in determines the jacobian @xmath553 of the transformation , @xmath554 $ ] .
we emphasize that it can be missed in the continuous representation .
evaluating the single - particle trace @xmath555 in the expression explicitly , one can obtain ( the limit @xmath556 is assumed ) @xmath557 } e^{-\frac{i p \delta}{4j_z } [ \rho_{p , n_p}^2+\frac{\varkappa}{1-\varkappa } \eta_{p , n_p}^2 ] } e^{-\frac{1}{j_\perp}\chi_{p , n_p}^{-p}(\chi_{p , n_p}^p-\chi_{p , n_p-1}^p ) } \biggr \ } \notag \\ \times \prod\limits_\alpha \biggl \{1+e^{-2i\epsilon_\alpha(t_+-t_- ) } + 2e^{-i\epsilon_\alpha(t_+-t_-)}\cos\left ( \frac{\delta}{2}\sum_{p=\pm}\sum_{n_p=1}^{n_p}\rho_{p , n_p}\right ) + \prod_{p=\pm } e^{-ip \epsilon_\alpha t_p } \exp\left ( \frac{i p \delta}{2 } \sum_{n_p=1}^{n_p}\rho_{p , n_p}\right ) \notag \\
\times \biggl ( p \chi_{p , n_p}^p e^{- i p \delta \varkappa \sum_{n_p=1}^{n_p } ( \rho_{p , n_p}-\eta_{p , n_p } ) } + i \delta \sum_{n_{-p}=1}^{n_{-p } } \chi_{-p , n_{-p}}^p e^{- i p \delta \varkappa \sum_{n=1}^{n_{-p } } ( \rho_{-p , n}-\eta_{-p , n } ) } e^{i p \delta \sum_{n=1}^{n_{-p } } \rho_{-p , n } } \biggr ) \biggr \ } .\end{gathered}\ ] ] now the integration over variables @xmath558 can be performed ( see details in appendix b of ref . [ ] )
. then we find @xmath559 } e^{-\frac{i p \delta}{4j_z } [ \rho_{p , n_p}^2+\frac{\varkappa}{1-\varkappa } \eta_{p , n_p}^2 ] } \biggr \ } \prod_\alpha\left ( \oint\limits_{|z_\alpha|=1}\frac{i dz_\alpha}{2\pi z_\alpha^2}\right ) e^{-w } \notag \\
\times \exp \biggl ( - 2 v \cos\left [ \frac{\delta}{2}\sum_{p=\pm } \sum_{n_p=1}^{n_p } \rho_{p , n_p}\right ] \biggr ) \int\limits_0^\infty dy \ ,
e^{-y } \exp \biggl \
{ -i j_\perp v y \left
( \prod \limits_{p=\pm } e^{i \frac{p \delta}{2 } \sum_{n_p=1}^{n_p } \rho_{p , n_p}}\right ) \notag
\\ \times \left ( \sum_{p=\pm } p\ , e^{- i p \delta \varkappa \sum_{n_p=1}^{n_p } ( \rho_{p , n_p}-\eta_{p , n_p } ) } \delta \sum_{n_p=1}^{n_p } e^{- i p
\delta \sum_{n=1}^{n_p } [ ( 1-\varkappa)\rho_{p , n}+\varkappa\eta_{p , n } ] } \right ) \biggr \ } . \label{eq : zj1}\end{gathered}\ ] ] here we introduce the following notation @xmath560 let us introduce new variables to make the expression more standard : @xmath561 + \xi_{p}(0 ) . \ ] ] here we switch to continuous representation .
we obtain @xmath562 \notag \\
\times e^{\frac{1}{2}[\xi_p(t_p)-\xi_p(0 ) ] } e^{-\frac{i p}{4j_z } \int_0^{t_p } dt [ \frac{(ip \dot{\xi}_p+\varkappa \eta_p)^2}{(1-\varkappa)^2 } + \frac{\varkappa \eta_p^2}{1-\varkappa } ] } \biggr \ } \notag \\ \times e^{- 2 v
\cosh [ \frac{1}{2(1-\varkappa)}\sum_{p=\pm}p ( \xi_{p}(t_p)-\xi_p(0 ) - i p \varkappa\int_0^{t_p } dt^\prime \eta_p(t^\prime ) ) ] } \notag \\
\times \exp \biggl \ { -i j_\perp v y \left
( \prod \limits_{p=\pm } e^{\frac{1}{2(1-\varkappa ) } [ \xi_p(t_p)-\xi_p(0)-ip\varkappa \int_0^{t_p } dt \eta_p(t ) ] } \right ) \notag \\
\times \left ( \sum_{p=\pm } p\ , e^{\frac{1}{1-\varkappa } [ \xi_p(0)-\varkappa \xi_p(t_p)+ip \varkappa \int_0^{t_p } dt \eta_p(t ) ] } \int\limits_0^{t_p}\!\ !
dt \ , e^{- \xi_p(t ) } \right ) \biggr \ } .
\label{eq : zj2}\end{gathered}\ ] ] there is some freedom in choosing the initial conditions for field variables @xmath563 .
it is convenient to choose them such that the following relations hold : @xmath564 = 0 .
\label{eq : bc}\end{gathered}\ ] ] then eq . can be rewritten as @xmath565 \notag \\
\times \int\limits_{-\infty}^\infty dx\ , e^{\frac{1}{2}[\xi_p(t_p)-\xi_p(0 ) ] } e^{-\frac{i p}{4j_z } \int_0^{t_p } dt [ \frac{(ip \dot{\xi}_p+\varkappa \eta_p)^2}{(1-\varkappa)^2 } + \frac{\varkappa \eta_p^2}{1-\varkappa } ] } \notag \\
\times e^{\frac{i x p}{1-\varkappa } [ \xi_p(0)-\varkappa \xi_p(t_p ) + i p
\varkappa \int_0^{t_p}dt \eta_p(t ) ] } \biggr \ } e^{- 2 v \cosh\frac{\xi_{+}(t_+)-\xi_{-}(t_-)}{2 } } \notag \\
\times e^{-\frac{i j_\perp}{4 } \sum_{p=\pm } p \int_0^{t_p } dt\ , e^{- \xi_p(t ) } } \delta \left ( \sum_{p=\pm } \xi_p(t_p ) + 2 \ln ( 4v y)\right ) .
\label{eq : zj3}\end{gathered}\ ] ] integrating over the variables @xmath566 we find @xmath567 \,e^{i p \int_0^{t_p } dt \mathcal{l}_p } e^{-(1 - 2i p x ) \xi_p(0)/2 } \biggr \ } \notag \\ \times \int\limits_0^\infty \frac{dy}{4 y v } \ , e^{-y - w } \ , \exp \left ( - 2 v \cosh\bigl [ \frac{1}{2}\sum_{p=\pm}p \xi_{p}(t_p ) \bigr ] \right ) \notag \\
\times \delta \left ( \sum_{p=\pm } \xi_p(t_p ) + 2 \ln ( 4v y)\right ) .
\label{eq : zj4}\end{gathered}\ ] ] the functional integral is of feynman - kac type with the lagrangian @xmath568 then the calculation of the partition function can be reduced to an evaluation of two matrix elements : @xmath569 } \notag \\
\times \langle \xi_+ | e^{-i \mathcal{h}_j t_+ } | \xi_+^\prime \rangle \langle \xi_-^\prime | e^{i \mathcal{h}_j t_- } | \xi_- \rangle .
\label{eq : zj5}\end{gathered}\ ] ] here the one - dimensional quantum mechanical hamiltonian @xmath570 its eigenfunctions are given by the modified bessel functions @xmath571 where @xmath572 is a real number : @xmath573 the eigenvalues of @xmath574 are equal to @xmath575 : @xmath576 .
after integration over @xmath218 we obtain @xmath577 here we use the following result ( see formula 6.794.11 on p. 794 of ref . [ ] ) @xmath578 integration over @xmath579 can be now easily performed , and we obtain @xmath580 using the identity ( see formula 6.521.3 on p. 658 of ref . [ ] ) @xmath581 we can perform the integration over @xmath582 . with the help of the integral representation of the modified bessel function
@xmath583 we integrate over @xmath572 . finally , integration over @xmath217 yields
@xmath584 here we restored the correct numerical factor using the normalization condition @xmath585 at @xmath586 . in order to derive the partition function for the hamiltonian from eq .
one needs to make the substitution @xmath587 and to integrate over the variable @xmath87 : @xmath588 then we obtain eq . with @xmath189 .
it is easy to obtain the partition function with non - zero magnetic field .
the field shifts the @xmath80 projection of the total spin in the evolution operator : @xmath589 .
this shift affects only the boundary conditions on @xmath294 in .
at @xmath591 , the value of the integral in eq . is determined by the region @xmath592 .
thus one can expand @xmath197 in the denominator into a series in @xmath593 and obtain @xmath594 here we performed the expansion to the second order in @xmath595 , having the further calculation for the spin susceptibility in mind . at @xmath596 , the argument of the @xmath197 in the denominator is large and one can make the following replacement : @xmath597 .
then we find @xmath598 .
\label{eq : app : f1:zzf1r2}\ ] ] the same simplification for @xmath599 can be used in the limit @xmath600 and @xmath601 .
then we obtain @xmath602 .
\label{eq : app : f1:zzf1r3}\ ] ] at @xmath600 and @xmath603 the relevant region of integration in eq .
is determined by the denominator .
thus one can omit @xmath604 , expand @xmath605 in the numerator and find @xmath606 the only relevant case for our values of @xmath217 and @xmath218 in eq .
is the case with @xmath607 . in this regime
the denominator can be substituted by @xmath608 and the region of integration can be extended to infinity .
then the following result can be obtained from by replacement @xmath609 : @xmath610
in this appendix we present brief arguments why eqs . and are correct . since eq . can be obtained from eq . , we consider only the latter .
we start from the following exact expression @xmath611 where @xmath612 . as usual , at @xmath138 the integral over @xmath613 can be performed in the saddle - point approximation .
this yields @xmath614 where @xmath615 the function @xmath616 can be rewritten as @xmath617 now we bound @xmath618 from above . using that random function @xmath619 depends , in fact , on @xmath620 ( cf .
, we obtain the following set of inequalities : @xmath621 hence we demonstrate that @xmath622 is independent of @xmath423 . therefore we can write @xmath616 at @xmath603 as follows : @xmath623
in this appendix we present a brief derivation of eq . .
the correlation function of the single - particle density of states is given by @xcite @xmath625 .
\label{app_dnu1}\ ] ] here the function @xmath626 depends on the statistics of the ensemble of single - particle energies . using eq . ,
the identity @xmath627 and the definition of @xmath112 we obtain @xmath628 , \label{eqc_app}\end{aligned}\ ] ] where @xmath629 .
\label{eq : app : g}\ ] ] the function @xmath630 has the following fourier transform with respect to variable @xmath631 : @xmath632 since the function @xmath630 is even in @xmath631 , the function @xmath633 is even in @xmath196 .
the function under the integral in the r.h.s . of eq .
has poles at @xmath634 where @xmath428 and @xmath423 are integers .
computation of the residues yields @xmath635 substitution into eq .
leads to @xmath636 .\end{aligned}\ ] ] at @xmath637 the function @xmath626 has the following asymptotic behavior : @xmath638 recall that @xmath379 for the orthogonal wigner - dyson ensemble , @xmath212 for the unitary wigner - dyson ensemble and @xmath639 for the symplectic wigner - dyson ensemble .
then at @xmath206 we find @xmath640[1-\cos(h_2 t)]}{t \sinh^2(\pi t ) } \notag \\ & = \sum_{\sigma=\pm } l(h_1+\sigma h_2)- 2 l(h_1 ) - 2l(h_2 ) , \end{aligned}\ ] ] where @xmath641 is even in @xmath88 .
next , for @xmath642 @xmath643 e^{-2\pi n t } \notag \\ & = \frac{2 h}{\pi^2\bm{\beta } } \left [ \operatorname{re}\psi\left ( 1+\frac{i h}{2\pi}\right ) - \psi(1)\right ] .\end{aligned}\ ] ] this is the eq . of the paper . using the well - known asymptotic expressions for the euler digamma function @xmath644 at small and large values of its argument one arrives at eq . .
in this appendix we present the derivation of the perturbative results and for @xmath189 . in addition
, we compute the next order in @xmath237 for the correction to @xmath215 . the contribution of the second order in @xmath207 is given by @xmath647 and @xmath648 .
we find @xmath649 .
\label{eq : f2_11}\ ] ] here we remind @xmath650 .
it is instructive to compare the second order contribution with the second order contribution to the variance of @xmath299 : @xmath651 .
\label{eq : var}\end{aligned}\ ] ] in the regime @xmath652 the arguments of @xmath237 in the right hand side of eqs . and
are small . using the asymptotic expression for @xmath213 at @xmath465
, we obtain @xmath653 the result for @xmath654 is translated into eq . of the paper . from eq .
we find that @xmath655 at low temperatures @xmath656 the asymptotic expression of @xmath213 for @xmath470 must be used in eq . .
we find @xmath657 from eq .
it follows that @xmath658 in view of the result we can expect that @xmath299 has a normal distribution with mean @xmath659/2 $ ] and variance @xmath648 in the regime @xmath660 . for @xmath661 and @xmath662 the complementary cumulative distribution function for the normal distribution and the complementary cumulative distribution function obtained numerically for the process
@xmath112 are compared in fig .
[ fig : ccdf ] .
we note that for @xmath661 and @xmath663 numerical integration of eqs . and yields @xmath664 and @xmath665 .
these values are still different from the asymptotic estimates . in the regime
@xmath652 the fourth order contributions are proportional to @xmath666 and therefore negligible . for low temperatures @xmath656 the contributions of the fourth order in @xmath207 are listed below : @xmath667 ^ 2 = -36 \ln^2 2 \ , z^4 , \ ] ] @xmath668 ^ 2 \notag \\
& + 2 \int_{-\infty}^\infty \frac{dh_1d h_2}{\pi } \ ,
e^{-h_1 ^ 2-h^2_2 } \bigl [ \overline{v(h_1\sqrt{y})v(h_2\sqrt{y})}\bigr ] ^2 \notag \\ & = \left ( 4 \ln^2 2 + 8 b_{2,2 } \right ) \ , z^4 , \\ b_{2,2 } & = \frac{1}{2 } \int_0^{2\pi } \frac{d\phi}{2\pi } \bigl ( \overline{v(\cos\phi)v(\sin\phi)}\bigr ) ^2 \approx 0.35 , \end{aligned}\ ] ] @xmath669 @xmath670 \notag \\ = & - \left ( 2\ln^2 2 + 2 b_{2,1,1 } \right ) \ , z^4 , \\ b_{2,1,1 } = & \frac{15}{4 } \int_0^{2\pi}\frac{d\phi}{4\pi}\int_0^\pi d\theta \ , \sin^3\theta \,\ , \overline{v(\cos\phi)v(\sin\phi)}\notag \\ & \times \overline{v(\cos\theta)v(\sin\theta\cos\phi)}\approx 0.79 , \end{aligned}\ ] ] @xmath671 ^ 2 = 3 \ln^2 2 \ , z^4 .\ ] ] here we recall that @xmath672 .
summing up , for @xmath656 we obtain @xmath673 where @xmath674 using eq . and the definition of the spin susceptibility we obtain for @xmath189 @xmath675 .\ ] ] j. a. folk , c. m. marcus , r. berkovits , i. l. kurland , i. l. aleiner , b. l. altshuler , phys .
t*90 * , 26 ( 2001 ) ; g. usaj , h.u .
baranger , phys .
b * 67 * , 121308 ( 2003 ) ; y. alhassid , t. rupp , phys . rev . lett . *
91 * , 056801 ( 2003 ) ; y. alhassid , t. rupp , a. kaminski , l. i. glazman , phys .
b * 69 * , 115331 ( 2004 ) .
m. schechter , phys .
b * 70 * , 024521 ( 2004 ) ; zu - jian ying , m. cuoco , c. noce , huan - qiang zhou , phys . rev .
b * 74 * , 012503 ( 2006 ) ; zu - jian ying , m. cuoco , c. noce , huan - qiang zhou , phys . rev .
b * 74 * , 214506 ( 2006 ) ; s. schmidt , y. alhassid , k. van houcke , europhys . lett .
* 80 * , 47004 ( 2007 ) ; s. schmidt , y. alhassid , phys .
* 101 * , 207003 ( 2008 ) ; k. van houcke , y. alhassid , s. schmidt , s. m. a. rombouts , arxiv:1011.5421 ; y. alhassid , k. n. nesterov , s. schmidt , phys .
t * 151 * , 014047 ( 2012 ) ; k. n. nesterov , y. alhassid , phys . rev .
b * 87 * , 014515 ( 2013 ) .
s. guron , m.m .
deshmukh , e.b .
myers , and d.c .
ralph , phys .
lett . * 83 * , 4148 ( 1999 ) ; m.m .
deshmukh , s. kleff , s. guron , e. bonet , a.n .
pasupathy , j. von delft , and d.c .
ralph , phys .
* 87 * , 226801 ( 2001 ) .
canali and a.h .
macdonald , phys .
85 * , 5623 ( 2000 ) ; s. kleff , j. von delft , m.m .
deshmukh , and d.c .
ralph , phys .
b * 64 * , 220401 ( 2001 ) ; s. kleff and j. von delft , phys . rev .
b * 65 * , 214421 ( 2002 )
. l.d .
graham and d. s. schreiber , j. appl .
phys . * 39 * , 963 ( 1968 ) ; l. shen , d. s. schreiber , and a. j.arko , phys . rev . * 179 * , 512 ( 1969 ) ; p. gambardella , s. rusponi , m. veronese , s. s. dhesi , c. grazioli , a. dallmeyer , i. cabria , r. zeller , p. h. dederichs , k. kern , c. carbone , h. brune , science * 300 * , 1130 ( 2003 ) ; a.m.clogston , b. t. matthias , m. peter , h. j. williams , e. corenzwit , and r. c. sherwood , phys . rev . * 125 * , 541 ( 1962 ) ; d. shaltiel , j. h.wrenick , h. j.williams , and m. peter , phys . rev . * 135 * , a1346 , ( 1964 ) ; j. w. loram and k. a. mirza , j. phys .
phys . * 15 * , 2213 ( 1985 ) ; j. w. loram , k. a. mirza , and z. chen , j. phys .
* 16 * , 233 ( 1986 ) ; g. mpourmpakis , g.e .
froudakis , a.n .
andriotis , m. menon , phys .
b * 72 * , 104417 ( 2005 ) .
i. v. kolokolov , phys . lett .
* a * 114 , 99 ( 1986 ) ; ann .
( n.y . ) * 202 * , 165 ( 1990 ) ; m. chertkov and i. v. kolokolov , phys . rev .
b * 51 * , 3974 ( 1995 ) ; sov .
jetp * 79 * , 824 ( 1994 ) ; for a review see i. v. kolokolov , int .
b * 10 * , 2189 ( 1996 ) . | we explore the static longitudinal and dynamic transverse spin susceptibilities in quantum dots and nanoparticles within the framework of the hamiltonian that extends the universal hamiltonian to the case of uniaxial anisotropic exchange . for the limiting cases of ising and heisenberg exchange interactions
we ascertain how fluctuations of single - particle levels affect the stoner instability in quantum dots .
we reduce the problem to the statistics of extrema of a certain gaussian process .
we prove that , in spite possible strong randomness of the single - particle levels , the spin susceptibility and all its moments diverge simultaneously at the point which is determined by the standard criterion of the stoner instability involving the mean level spacing only .
= 2000 |
there has been much recent interest in the properties of active galactic nuclei ( agn ) and the interaction between these objects and their host galaxies .
observationally , strong evidence supporting such interaction exists .
up to 70% of cd galaxies contain radio sources at their centres @xcite .
these radio sources drive shocks into the intra - cluster medium ( icm ) , heating and transporting outward the cluster gas .
the spatial coincidence of radio emission with cavities in the hot x - ray emitting gas @xcite agrees well with analytical and numerical models @xcite , and is required to explain the lack of catastrophic cooling expected in the absence of a feedback mechanism in dense cluster cores @xcite .
another piece of evidence comes from the field of galaxy formation .
the growth of structure through the gravitational instability provides a simple and accurate description of the build - up of dark matter scaffolding through cosmic time , however agn feedback is required to prevent runaway cooling from forming too many massive galaxies @xcite . the tight correlation between galaxy and black hole masses @xcite provides further evidence for the tight coupling between agn and galaxy evolution .
broadly speaking , two types of agn are often discussed .
powerful outbursts observable at both optical and radio wavelengths are believed to be triggered by gas - rich mergers , facilitating a sudden influx of large quantities of fuel to the agn .
these are often observed in the local volume as seyfert agn , and as quasars at higher redshifts .
they are almost invariably accompanied by strong radio emission . by contrast
, the vast majority of radio - loud agn at @xmath1 do not have a corresponding seyfert phase .
the number counts of these objects are consistent with low - luminosity radio agn activity being triggered by the cooling out of hot gas from the galactic halo @xcite . in a seminal work , fanaroff &
riley showed that the luminosity of a radio agn is intrinsically tied to its morphology .
the most powerful radio sources are invariably edge - brightened ( fanaroff & riley referred to these as type ii objects , or fr - iis ) , propagating bow shocks through the surrounding gas as they expand .
the less powerful fr - i sources , on the other hand , have radio emission that is core - dominated .
these objects dominate radio source counts in the local volume , and are typically interpreted as agn that start out as fr - iis but are not powerful enough to traverse the dense galaxy or cluster core without being disrupted by rayleigh - taylor and kelvin - helmholtz instabilities .
agn feedback influences galaxy formation and evolution by affecting the gas available for star formation .
recently , kaviraj et al
. showed that agn feedback is required to explain the rapid transition of early - type galaxies from the blue cloud to the red sequence .
there are a number ways in which agn feedback can affect the gas .
kinetic feedback , typically associated with powerful radio sources , can shock heat and uplift the gas to large radii .
the agn radiation field can also heat the gas and drive outflows .
finally , turbulence induced by disruption of the jet - inflated radio cocoons , a process mostly associated with fr - i sources , can either trigger star formation ( by aiding gravitational collapse ; e.g. silk & nusser 2010 ) or suppress it ( by heating the gas further ) .
the relative importance of these processes is unclear . in particular , it is not clear how important feedback from the powerful but rare fr - iis is compared with the less powerful but omnipresent fr - is .
we aim to address this issue in the present paper .
the paper is structured as follows . in section [ sec : sample ] we describe our sample .
results are presented in section [ sec : results ] , and discussed in section [ sec : discussion ] .
we conclude in section [ sec : summary ] .
in this section we describe the construction of a sample of galaxies around radio sources .
the radio source list was compiled from a combination of catalogues .
the largest radio agn in terms of angular size were selected from the 3crr sample of laing et al . , and complemented by additional data from machalski et al . .
smaller radio sources were identified from two large - area 1.4 ghz surveys , faint images of the radio sky at twenty centimetres ( first ; becker et al .
1995 ) and the nrao vla sky survey ( nvss ; condon et al .
1998 ) . these are complementary in the sense that first has higher sensitivity , but nvss is better at picking up extended objects .
shabala et al .
identified optical sloan digital sky survey ( sdss ; strauss et al .
2002 ) counterparts for first / nvss radio sources to @xmath2 .
we limit the present work to @xmath0 , which provides a good balance between statistically meaningful number counts of galaxies and relatively low contamination in radio / optical positional matching due to line - of - sight effects ( which increases with redshift ) . for @xmath3 , we selected radio sources from the config sample @xcite .
this approach allows construction of a sample containing both the more powerful , edge - brigtened fr - ii sources , as well as the less powerful edge - darkened fr - is .
details are given in tables [ tab : fr2s ] and [ tab : fr1s ] .
we note that while the sample is not complete , we expect it to be representative as it spans a wide range of radio source sizes , luminosities , morphologies and ages .
lllllll iau & other & redshift & frequency & las & axial & @xmath4 + name & name & & ( ghz ) & ( arcsec ) & ratio & ( myr ) + + iau & other & redshift & frequency & las & axial & @xmath4 + name & name & & ( ghz ) & ( arcsec ) & ratio & ( myr ) + 0010 - 1108 & & 0.077 & 1.4 & 200 & @xmath5 & + 0057 - 0052 & & 0.044 & 1.4 & 17 & @xmath6 & + 0131 + 0033 & & 0.079 & 1.4 & 46 & @xmath7 & + 0739 + 3947 & & 0.098 & 1.4 & 33 & @xmath5 & + 0745 + 3357 & & 0.063 & 1.4 & 14 & @xmath8 & + 0756 + 3703 & & 0.077 & 1.4 & 48 & @xmath9 & + 0758 + 3747 & & 0.041 & 1.4 & 43 & @xmath10 & + 0805 + 2409 & 3c192 & 0.060 & 1.41 & 201 & @xmath11 & + 0819 + 5232 & & 0.189 & 1.4 & 49 & @xmath12 & + 0821 + 4702 & & 0.130 & 1.4 & 38 & @xmath13 & + 0902 + 5203 & & 0.099 & 1.4 & 28 & @xmath14 & + 0911 + 3724 & & 0.105 & 1.4 & 60 & @xmath15 & @xmath16 + 0921 + 4538 & 3c219 & 0.174 & 1.52 & 190 & @xmath17 & @xmath18 + 0930 + 0348 & & 0.089 & 1.4 & 33 & @xmath19 & + 0939 + 3553 & 3c223 & 0.138 & 1.50 & 306 & @xmath20 & @xmath21 + 0941 + 3944 & & 0.107 & 1.4 & 190 & @xmath22 & + 0947 + 0725 & & 0.087 & 1.4 & 310 & @xmath23 & + 0949 - 0050 & & 0.081 & 1.4 & 25 & @xmath24 & + 0955 + 0135 & & 0.099 & 1.4 & 44 & @xmath25 & + 1001 + 2847 & 3c234 & 0.185 & 1.41 & 113 & @xmath26 & @xmath27 + 1006 + 3454 & & 0.099 & 1.4 & 1896 & @xmath28 & + 1007 + 0030 & & 0.095 & 1.4 & 17 & @xmath29 & + 1016 + 6014 & & 0.031 & 1.4 & 100 & @xmath30 & + 1016 + 4046 & & 0.128 & 1.4 & 139 & @xmath31 & + 1020 + 4832 & & 0.053 & 1.4 & 675 & @xmath32 & + 1032 + 5644 & & 0.045 & 1.4 & 389 & @xmath33 & + 1036 + 0006 & & 0.097 & 1.4 & 50 & @xmath34 & + 1039 + 0510 & & 0.068 & 1.4 & 24 & @xmath35 & + 1059 + 0517 & & 0.035 & 1.4 & 239 & @xmath36 & + 1111 + 4050 & & 0.074 & 1.4 & 322 & @xmath37 & + 1116 + 2915 & & 0.049 & 1.4 & 130 & @xmath38 & + 1137 + 6120 & & 0.111 & 1.4 & 230 & @xmath39 & + 1144 + 3710 & & 0.115 & 1.4 & 75 & @xmath40 & + 1153
+ 0329 & & 0.079 & 1.4 & 13 & @xmath41 & + 1155 + 5454 & & 0.050 & 1.4 & 216 & @xmath42 & + 1214 + 0528 & & 0.078 & 1.4 & 25 & @xmath43 & + 1217 + 0336 & & 0.077 & 1.4 & 72 & @xmath44 & + 1217 + 0339 & & 0.078 & 1.4 & 24 & @xmath45 & + 1218 + 5026 & & 0.200 & 1.4 & 210 & @xmath46 & @xmath47 + 1252 + 0315 & & 0.099 & 1.4 & 24 & @xmath48 & + 1254 + 5305 & & 0.054 & 1.4 & 33 & @xmath49 & + 1254 + 2737 & & 0.086 & 1.4 & 69 & @xmath50 & + 1302 + 6229 & & 0.076 & 1.4 & 18 & @xmath51 & + 1321 + 4235 & 3c285 & 0.079 & 1.65 & 184 & @xmath52 & @xmath53 + 1328 - 0307 & & 0.085 & 1.4 & 289 & @xmath54 & + 1330 - 0206 & & 0.087 & 1.4 & 20 & @xmath55 & + 1331 - 0252 & & 0.087 & 1.4 & 13 & @xmath56 & + 1350 + 2816 & & 0.072 & 1.4 & 60 & @xmath57 & @xmath58 + 1352 + 3126 & 3c293 & 0.045 & 1.52 & 256 & @xmath59 & @xmath60 + 1354 + 0528 & & 0.077 & 1.4 & 14 & @xmath61 & + 1400 + 3021 & & 0.206 & 1.4 & 120 & @xmath62 & @xmath63 + 1443 + 5201 & 3c303 & 0.141 & 1.45 & 47 & @xmath64 & + 1454 + 1620 & & 0.046 & 1.4 & 660 & @xmath65 & + 1504 + 2600 & 3c310 & 0.054 & 1.45 & 305 & @xmath66 & @xmath67 + 1508 + 5415 & & 0.096 & 1.4 & 15 & @xmath68 & + 1516 + 0015 & & 0.052 & 1.4 & 328 & @xmath69 & + 1524 + 5428 & 3c319 & 0.192 & 1.65 & 109 & @xmath70 & @xmath71 + 1531 + 2404 & 3c321 & 0.096 & 1.51 & 307 & @xmath72 & + 1552 + 2005 & 3c326 & 0.089 & 1.40 & 1206 & @xmath73 & + 1557 + 5440 & & 0.047 & 1.4 & 200 & @xmath74 & + 1602 + 0158 & & 0.104 & 1.4 & 606 & @xmath13 & + 1617 + 3222 & 3c332 & 0.152 & 1.4 & 800 & @xmath75 & @xmath76 + 1700 + 3008 & & 0.035 & 1.4 & 135 & @xmath15 & @xmath77 + 1713 + 6402 & & 0.079 & 1.4 & 13 & @xmath78 & + 1728 + 3146 & 3c357 & 0.166 & 1.4 & 60 & @xmath79 & @xmath80 + 1729 + 5415 & & 0.084 & 1.4 & 28 & @xmath81 & + 2153 - 0711 & & 0.059 & 1.4 & 183 & @xmath82 & + 2157 - 0750 & & 0.062 & 1.4 & 28 & @xmath83 & + 2214 + 1350 & 3c442a & 0.026 & 1.38 & 597 & @xmath84 & + 2231 - 0824 & & 0.083 & 1.4 & 111 & @xmath56 & + 2315 - 0026 & & 0.091 & 1.4 & 25 & @xmath85 & + 2333 - 0027 & & 0.059 & 1.4 & 81 & @xmath86 & + llllll iau & other & redshift & frequency & las & @xmath4 + name & name & & ( ghz ) & ( arcsec ) & ( myr ) + + iau & other & redshift & frequency & las & @xmath4 + name & name & & ( ghz ) & ( arcsec ) & ( myr ) + 0318 + 4151 & ngc 1265 & 0.025 & 1.38 & @xmath87 & + 0319 + 4130 & 3c84 & 0.017 & 1.38 & @xmath88 & + 0847 + 5352 & & 0.045 & 1.4 & @xmath89 & + 0901 + 2901 & & 0.194 & 1.4 & @xmath90 & + 1002 + 1951 & & 0.168 & 1.4 & @xmath91 & + 1031 + 5225 & & 0.166 & 1.4 & @xmath92 & + 1140 + 1203 & & 0.081 & 1.4 & @xmath93 & + 1145 + 1936 & 3c264 & 0.022 & 1.40 & @xmath94 & @xmath95 + 1225 + 1253 & 3c272.1 & 0.005 & 1.41 & @xmath96 & + 1229 + 1144 & a1552 & 0.086 & 1.49 & @xmath97 & + 1230 + 1223 & 3c274 & 0.005 & 1.46 & @xmath98 & + 1236 + 1632 & & 0.068 & 1.4 & @xmath99 & + 1316 + 0702 & & 0.051 & 1.4 & @xmath100 & + 1342 + 0504 & & 0.136 & 1.4 & @xmath92 & + 1416 + 1048 & 3c296 & 0.025 & 1.45 & @xmath101 & + 1449 + 6316 & 3c305 & 0.042 & 1.42 & @xmath102 & + 1504 + 2835 & & 0.043 & 1.4 & @xmath103 & + 1513 + 2607 & 3c315 & 0.108 & 1.42 & @xmath104 & @xmath105 + 1516 + 0701 & & 0.034 & 1.4 & @xmath92 & + 1617 + 3500 & ngc 6109 & 0.030 & 0.609 & @xmath106 & + 1628 + 3933 & 3c338 & 0.030 & 4.89 & @xmath107 & @xmath71 + galaxies located near radio sources were selected from sdss . in this paper
we are interested in the effects of radio sources on photometric properties of galaxies , and a photometric ( rather than spectroscopic ) sample was therefore employed . while this greatly increases sample statistics ,
care must be taken in assessing the accuracy of photometric redshift estimates . in figure
[ fig : photoz ] we plot spectroscopic and photometric redshifts for a subsample of 161 galaxies , derived from the sdss pipeline .
the photometric redshifts are estimated using template fitting .
these spectral energy distribution templates are obtained using galaxies with spectroscopic identifications , and extended with spectral synthesis models @xcite .
photometric redshifts are in good agreement with spectroscopic values for @xmath108 , with most values agreeing within the uncertainties of photometric measurements .
this agreement disappears at low redshifts . in our sample ,
12 of 21 fr - i radio sources have @xmath109 , compared to only 11 of 72 fr - iis .
this is a known issue with the sdss pipeline @xcite , and is related to the inherent difficulty in getting spectroscopic redshifts to be more accurate than @xmath110 .
it is therefore important to consider whether any systematic effects are introduced into the analysis by adopting photometric redshifts .
specifically , there are three crucial considerations : completeness of the sample , contamination , and selection biases . , but diverge at lower redshift.,scaledwidth=45.0% ] _ completeness _ an estimate of how many bona fide group members are not picked up by our photometric identification procedure comes from comparing galaxy counts obtained using spectroscopic and photometric redshifts .
we selected all spectroscopically identified galaxies projected to be in or near the 26 largest radio sources in our sample .
the true number of galaxies associated with the group was estimated by requiring that the spectroscopic redshifts of group members were not offset from the redshift of the radio source host by more than @xmath111 .
this @xmath112 is the optimal tolerance value : relaxing the redshift restriction ( by increasing @xmath112 ) in an individual group does not change the number counts appreciably ; however , decreasing @xmath112 gives an appreciable decrease in the number counts , indicating that some group members are being excluded .
the number of photometric matches was then obtained by requiring that the spectroscopic redshift of each galaxy lay within the photometric @xmath113 uncertainties .
figure [ fig : spectrophotoratio ] shows the redshift dependence of the photometric - to - spectroscopic matching rates . at @xmath114
the completeness is close to unity . at @xmath115
the completeness can be as low as 30 percent . .
the photometric samples can be significantly incomplete ( as low at 30% ) at lower redshift .
only groups with more than one spectroscopic identification are shown.,scaledwidth=45.0% ] _ contamination _
incorrect redshifts can also result in inclusion of unrelated galaxies seen in projection .
the largest radio source in our sample , 3c236 , is 32 arcmin in size . at a redshift of 0.1 , and with a redshift uncertainty of 0.02
, this corresponds to a co - moving volume of 960 mpc@xmath116 . for comparison
, the local number density of groups has been estimated by yang et al .
to be around @xmath117 mpc@xmath118 .
a similar number is obtained from simulations for the number density of local dark matter haloes with masses in excess of @xmath119 @xmath120 .
most of these groups contain only a single member , with only 3 percent harbouring four or more galaxies @xcite , and thus the expected contamination is of order one to a few galaxies .
since most groups subtend a much smaller solid angle on the sky than 3c236 , contamination is not expected to be a major issue in the present work .
_ selection effects _
it is important to examine whether adopting photometric redshifts preferentially selects against galaxies with certain properties .
galaxy properties are often a strong function of mass , and galaxy colours are of particular interest in the present work .
we therefore investigated whether the offset between photometric and spectroscopic redshift values correlates with either colour or @xmath121-band magnitude ( a proxy for stellar mass ; e.g. shabala et al .
we found no such correlations , suggesting that despite the lack of completeness at low redshift , photometric redshift identifications are suitable for our analysis .
it is further worth noting that because in our analysis below we compare similar galaxies inside and outside the radio contours , any major redshift - dependent systematic effects are unlikely . in the following
we therefore include all galaxies for which the redshift of the target radio source lies within the uncertainty of the galaxy photometric redshift .
sdss photometry was obtained for objects within a radius equal to 1.5 times the maximum angular extent of the radio source ( column 5 in tables [ tab : fr2s ] and [ tab : fr1s ] ) from the centre of the radio source .
galaxies with photometric redshifts matching the spectroscopic redshift of the radio source host ( i.e. with @xmath122 ) were retained for analysis .
radio maps were used to detemine whether individual galaxies are ( in projection ) located inside or outside the radio contours . where available , high resolution radio source maps at 1.4 - 1.6 ghz ( leahy , bridle & strom 2000 ) were used .
first / nvss maps at 1.4 ghz were adopted for the rest of the sources , with the exception of fr - i sources 3c338 ( 4.89 ghz map used ) and ngc6109 ( 608 mhz ) .
for these two objects , the lower resolution first maps at 1.4 ghz show similar source size and structure to the above maps , and the higher resolution versions were therefore used .
examples are shown in figure [ fig : contoursexample ] . _
fr - ii sources _ galaxies lying within the radio source contours are subject to projection effects due to the uncertainty in radio source orientation . in other words , these galaxies could lie either immediately behind or in front of the radio emission , but appear to be coincident with this emission . this problem can be mitigated for the fr - ii ( classical double ) radio sources .
these objects typically have cylindrical morphologies , at least in regions away from the hotspots . the aspect ratio @xmath123 between semi - major and semi - minor axes
is given for each source in column 5 of table [ tab : fr2s ] .
for @xmath124 being the maximum linear extent of the fr - ii source , galaxies lying within @xmath125 of the radio cource centre will be immune from such a projection effect , since they fall within the radio cocoon regardless of the projection angle .
projection effects due to uncertainty in galaxy position in redshift space are much harder to quantify . however , as most radio agn reside in cluster centres and are quite extended in the direction perpendicular to the jet ( i.e. @xmath126 is comparable with the cluster core radius ) , we expect the majority of the galaxies that appear to lie within @xmath126 of the cluster centre to do so . motivated by these considerations , we separated galaxies associated with fr - ii radio sources into three categories .
galaxies lying outside the radio emission ( in projection ) were allocated to the `` outside '' group .
galaxies projected to lie within the radio source contours but outside @xmath126 of the cluster centre were placed in the `` mixed '' group . some of these would presumably be physically coincident with the radio emission , while others wo nt .
finally , galaxies located within @xmath126 of the cluster centre were placed in the `` inside '' group .
these correspond to objects that are very likely to have been overrun by the radio source- only galaxies located well away from the cluster core but seen close to the line of sight to the core can contaminate this sample . _ fr - i sources _ it is much more difficult to perform a similar analysis for the edge - darkened fr - i sources .
these primarily consist of two types of objects which are often referred to in the literature as twin - jet radio agn and relaxed doubles .
the former initially appear as two jets emanating from the central engine .
the jets are eventually disrupted ( usually at a well - defined flare point ) , and at larger radii the radio emission turns into a plume due to interaction with the intra - cluster medium .
3c296 is a classical example of this type of source .
we include narrow and wide - angled tail galaxies ( e.g. 3c264 ) in this subclass of fr - is , since these are most likely twin - jet radio sources that have been swept up by interaction with the icm through which they are travelling .
relaxed doubles are sources in which radio emission declines gradually with radius , in a quasi - radially symmetric fashion .
these objects do not exhibit any compact structure , with 3c84 being a famous example .
we identify @xmath126 with the flare point for twin - jet sources , and with half the maximal radial extent ( i.e. @xmath127 ) for relaxed doubles .
the galaxies associated with fr - i radio sources are then split into `` outside '' , `` inside '' and `` mixed '' regions in exactly the same way as those near fr - iis . as we show below ,
fr - is do not appear to exhibit any difference in properties between the three classes , suggesting our classification is adequate .
the major purpose of this work is to study the effects of radio sources on photometric properties of galaxies , and in particular galaxy colours .
we therefore stacked all galaxies into the three groups outlined above . around
fr - ii radio sources this yielded 58 galaxies in the `` inside '' category , 6678 in the `` outside '' group , and 318 in the `` mixed '' .
for fr - is , 27 galaxies were found in the `` inside '' group , 2149 in the `` outside '' , and 36 in `` mixed '' .
galaxy properties such as colour are a strong function of mass , with massive galaxies preferentially being redder .
we therefore restricted our analysis to objects within a well - defined mass range @xmath128 @xmath120 .
stellar mass was approximated via the @xmath121-band absolute magnitude , assuming that the solar bolometric correction is applicable ( see shabala et al .
2008 for details ) . in figure
[ fig : coloursall ] we plot the cumulative distribution of @xmath129 @xmath130-corrected colours for the three groups , stacking galaxies around all radio sources in our sample .
there is a clear offset between the `` inside '' and `` outside '' groups , with the `` inside '' galaxies exhibiting redder colours .
colours , stacked for all radio sources .
the three regions are defined in the text .
galaxies within the path of the radio source show a clear offset to redder colours.,scaledwidth=45.0% ] splitting up the sample into fr - i and fr - ii radio sources ( figure [ fig : coloursfrtypes ] ) , it can be seen that this colour difference is driven exclusively by the fr - ii population .
in other words , fr - iis appear to affect their environment , while fr - is do not .
these conclusions are confirmed by komogorov - smirnov ( ks ) tests , with the observed fr - ii offset being statistically significant at the 2.5% level , and no statistically significant differences found for fr - is .
care must be taken in interpreting these results .
all radio sources in our sample are located at the centres of groups or clusters .
mass segregation means that galaxies near the cluster centres ( and therefore more likely to have been classified in the `` inside '' category ) will be more massive , and therefore redder .
figure [ fig : massdist ] shows that this is indeed the case for both the fr - i and fr - ii samples .
we correct for this effect in two ways .
firstly , our analysis is restricted to galaxies with @xmath131 @xmath120 .
in other words , the most massive , red galaxies that would drive redder colours for the `` inside '' category are excluded .
furthermore , we break up the @xmath128 @xmath120 mass range into bins and weigh the colours by the ratio of number counts between the category of interest ( `` inside '' or `` mixed '' ) and the control sample ( `` outside '' ) . to put it another way ,
we mimic the mass profile of the `` outside '' category .
figure [ fig : coloursfrtypescorr ] , confirmed by ks tests , shows that our findings are robust to these corrections .
the results of section [ sec : results ] appear to suggest that fr - ii and fr - i radio sources interact with their environments in entirely different ways .
a large body of observational evidence suggests that fr - ii sources grow by driving large - scale bow - shocks ( e.g. fabian et al .
2003 , croston et al .
2009 ) , sweeping up the ambient gas as they proceed .
shock heating and gas uplifting away from the centre of the potential well are seen in analytical models and numerical simulations @xcite , and are in fact necessary to solve the cooling flow problem @xcite and reconcile observations of the local stellar mass function with galaxy formation models @xcite . while the impact of powerful ( i.e. fr - ii ) radio source feedback on cluster gas is well understood
, it is not immediately obvious that such feedback can affect individual galaxies in a similar way .
a propagating bow shock will compress and shock heat both the atomic and molecular gas reservoirs that it overruns .
if the local gas density is above some critical value , the gas clumps will become radiative ( e.g. sutherland & bicknell 2007 ) and the shock can in fact _ enhance _ star formation , rather than suppress it @xcite .
it is a different story if one is concerned with diffuse hi gas reservoir , however , which is more susceptible to both shock heating ( due to longer cooling times ) and uplifting .
fr - i radio sources start their lives as fr - iis , but have their jets disrupted by interaction with the dense interstellar / intergalactic medium . as a result
, the bow shocks typically do not propagate far outside the host galaxy , and it is difficult for fr - is to affect their environment other than by turbulent mixing of the radio plasma with the cluster gas .
the findings of the previous section strongly suggest that it is the propagation of the bow shock that is responsible for the redder colours of galaxies . the exact feedback mechanism ( i.e. heating or gas expulsion )
is not important to the present discussion . either way
, fr - iis provide a way of suddenly depleting a reservoir of gas previously available for star formation .
we set out to test whether gas heating / removal associated with an fr - ii radio source can explain the observed distribution of galaxy colours .
the first step is to construct a star formation history for a galaxy unaffected by agn feedback .
our modelling follows the prescriptions of kaviraj et al .
, which have proved very successful at describing the colour evolution of the early - type galaxy population .
the star formation histories of galaxies are represented by an instantaneous burst of star formation at high redshift , followed by more gradual recent star formation .
the old burst , in which most of the stars are formed , is characterized by its age @xmath132 , typically @xmath133 gyrs ( e.g. kaviraj 2009 ) .
recent star formation is parametrized by the age of the burst @xmath134 , the timescale associated with star formation @xmath135 and the initial gas fraction @xmath136 available for recent star formation at time @xmath134 .
the star formation rate is given by the schmidt - kennicutt law , @xmath137 with star formation efficiency @xmath138 and dynamical timescale @xmath139 gyrs .
apart from being physically sensible , kaviraj et al . find that these values reproduce the observed number counts of blue cloud and red sequence galaxies .
we note that , being an empirical relation , the schmidt - kennicutt law implicitly includes the effects of supernova feedback on star formation .
equating the star formation rate with the rate of change of gas mass yields @xmath140 where @xmath136 is the initial gas fraction in the galaxy at time @xmath134 .
stellar population synthesis models can then be used to infer the colour evolution of the galaxy .
we adopt the bruzual & charlot models with a salpeter initial mass function ( imf ) .
galaxy colours are computed from luminosities , adding up contributions from the old underlying stellar population and the more recently formed stars .
figure [ fig : examplecolourtracks ] shows example tracks for @xmath129 colour evolution with @xmath141 gyrs , @xmath142 gyrs , @xmath143 and metallicities of @xmath144 and @xmath145 .
colour evolution for a star - forming galaxy with an underlying 9 gyr old population , and recent star formation described by the schmidt - kennicutt law ( equation [ eqn : sk_sfr ] ) with @xmath146 gyrs and @xmath143 .
solid lines are for metallicity of 0.2 z@xmath147 , and dashed for 1 z@xmath147.,scaledwidth=45.0% ] comparison with figure [ fig : coloursfrtypes ] shows that the model and observations span a similar range in colour . in reality , the parameters @xmath132 , @xmath134 , @xmath148 and @xmath149 will differ from galaxy to galaxy .
we follow the study of local lirgs by kaviraj in adopting distributions for @xmath132 and @xmath134 . the probability distribution in the old burst timescale
@xmath132 spans the range @xmath150 gyrs , rising linearly from these values towards a peak at @xmath151 gyrs .
the timescale @xmath134 associated with the onset of recent star formation is modelled with a flat prior to an age of @xmath152 gyrs .
while it is true that these distributions found for lirgs may not apply to our more general galaxy population , we believe that we can be justified in adopting these .
the @xmath129 colours are primarily driven by the @xmath153-band luminosity evolution of the young stellar population , and are therefore not very sensitive to the old burst age . as figure
[ fig : examplecolourtracks ] shows , the colour evolution is relatively slow for @xmath154 gyr , and we therefore do not expect the exact form of the prior on @xmath134 to affect our results either .
importantly , while we adopt a log - normal form for the distribution of the initial gas graction @xmath136 , the normalization is left as a free parameter used to match the observed colours in unaffected ( i.e. `` outside '' ) galaxies in our sample .
model colour distributions without agn feedback .
thick black solid line corresponds to observations of galaxies lying outside the fr - ii radio sources .
thick red solid line shows observations of galaxies that have been overrun by fr - ii radio sources . dashed and dotted lines denote various models , with mean initial gas fraction fixed at @xmath155 but varying scatter about this value and metallicity .
while a good fit can be obtained for the bulk of the `` outside '' galaxy population , no reasonable gas fraction distribution can reproduce the colours of galaxies affected by fr - ii feedback.,scaledwidth=45.0% ] figure [ fig : outsidefit ] shows the best - fit models to @xmath129 colours for the `` outside '' galaxies near fr - ii radio agn .
a range of metallicities is considered .
reasonable gas fractions of @xmath156 and @xmath157 reproduce the bulk of the observations .
these values are consistent with the observed gas fraction @xmath158 in local early - type galaxies @xcite .
the very blue and very red galaxies are not very well reproduced by the models , most likely due to a number of contributing factors .
it is possible that we have not encompassed the full range of metallicities or old starburst ages .
as figure [ fig : outsidefit ] shows , up to a 0.5 dex offset in @xmath129 colour can arise due to a change in metallicity .
alternatively , tidal stripping effects not considered here may be important : accreting galaxies would exhibit enhanced star formation , while the stripped companions would appear redder . an intriguing possibility for the origin of the reddest galaxies , discussed in section [ sec : feedbackmodes ] , is that these carry an imprint of previous agn outbursts .
we note that , overall , the models fit the bulk of the sample well . modifying the scatter changes the overall slope as well as shape in the middle of the distribution , but does not systematically shift all galaxies to either redder or bluer colours .
the redder colours of `` inside '' galaxies in our sample can only be achieved with very low gas fractions , a requirement that is both at odds with the observed gas fraction values in the local volume @xcite and the gas fractions required to explain the colours of unaffected galaxies ( lying in the `` outside '' group ) in the same clusters . in our simple model
we assume that the effect of a poweful agn ( such as an fr - ii ) is to starve the galaxy of gas previously available for star formation .
the galaxy star formation history is therefore identical to that described above , except for a sudden truncation a time @xmath159 ago , where @xmath159 is the time since an agn - driven shock has overrun the galaxy . since typical lobe expansion speeds for classical double radio sources are of the order of a few percent of the speed of light @xcite , and most galaxies that are definitely affected by the agn lie relatively close to cluster centre by construction ( see section [ sec : separatesamples ] ) , @xmath159 approximately corresponds to the current age of the radio source . but with agn feedback .
colours of galaxies overrun by expanding fr - ii shocks ( red line ) are well explained by the model.,scaledwidth=45.0% ] in figure [ fig : insidefit ] we show the impact of including this mode of feedback in our best - fit models .
models with agn ages of @xmath160 myrs do a good job of reproducing the data .
these timescales are consistent with a median fr - ii age of 44 myrs in our sample , and provide strong evidence that it is the depletion of the gas reservoir by an expanding fr - ii radio source that drives the colour evolution of affected galaxies .
in preceding sections we have shown that powerful fr - ii radio sources can ( and do ) affect the evolution of galaxies outside their hosts .
it is important to understand how long - lasting such effects are , and whether they are important to the interplay between gas heating and cooling . to address this question , we need to know how often powerful radio sources are triggered .
shabala et al .
studied a complete radio / optical sample of local galaxies , and concluded that around 1 percent of the most massive ( @xmath161 @xmath120 ) galaxies host agn with 1.4 ghz luminosity in excess of @xmath162 @xmath163 , which approximately corresponds to the separation luminosity between fr - is and fr - iis @xcite .
given maximum fr - ii lifetimes of @xmath160 myrs , this suggests that a typical massive galaxy will launch between one and a few fr - ii outbursts in a hubble time , with the typical time between outbursts of order a few gyrs .
the passing bow shock will both sweep up and heat the gas . the time delay before this gas is once again available for star formation is therefore equal to the sum of the cooling time @xmath164 and the timescale associated with reincorporation of this gas in the galaxy @xmath165 .
the cooling time is a function of gas temperature and density , @xmath166 myrs .
the reincorporation timescale is typically longer than this , being comparable to the dynamical timescale for the dark matter halo hosting the cluster , @xmath167 gyrs at the present epoch .
these numbers suggest that fr - ii outbursts in clusters can indeed prevent galaxies from forming stars despite the relatively infrequent , short - term nature of these outbursts .
the timescales of star formation suppression can also be examined observationally . for this purpose
we employed a sample of 16 clusters with no detectable radio emission at 1.4 ghz , as described by dunn & fabian and dunn et al . .
we split these clusters into regions in an attempt to mimic the process for fr - ii host clusters .
this is done by weighing the @xmath168 and @xmath124 values ( see section [ sec : sample ] ) for each cluster hosting an fr - ii by the number of galaxies contained within that radius , yielding mass - weighted averages for these quantities .
colours for clusters with no detected 1.4 ghz radio agn or bubbles .
the three regions are chosen to match the fr - ii sample , and mass corrections made .
there is still a clear offset to redder colours for innermost galaxies.,scaledwidth=45.0% ] figure [ fig : coloursquiescent ] shows the resultant colour distribution for the three regions . there is a clear difference between galaxies located close to cluster centre and those on the periphery , even after the usual mass corrections are performed .
it is interesting to compare the radio source and quiescent samples . in figure
[ fig : coloursquiescentvsfrii ] we plot colour distributions for the same regions in fr - i , fr - ii and quiescent clusters , adjusting the mass distributions in clusters with radio sources to mimic those for the appropriate quiescent samples .
galaxies in clusters which contain fr - i radio sources are consistently bluer than those in the quiescent cluster sample .
by contrast , fr - ii galaxies are marginally redder that the quiescent sample ( and both much redder than the fr - i sample ) in the `` inside '' region ; bluer in the `` mixed '' region ; and bluer still in the `` outside '' region , where the fr - i and fr - ii distributions are statistically indistinguishable . at face value , these findings are somewhat surprising . if galaxies in the quiescent cluster sample are unaffected by agn feedback , we would expect these to have bluer colours that the fr - ii sample at all radii , and perhaps even bluer colours than the fr - i galaxies if these suffer any ( perhaps low - level ) feedback .
the results make perfect sense , however , if quiescent clusters represent a population which has undergone past agn outbursts and has not yet regained all of the ejected / heated gas . in this picture ,
an fr - ii agn is triggered , heating and ejecting gas from galaxies within the cluster .
this gas cools and returns on timescale @xmath169 .
since fr - ii lifetimes are at least an order of magnitude less than the gas return timescale ( section [ sec : recurrence ] ) , it is possible that the agn activity terminates long before the cool gas is returned to the cluster .
the cluster can then appear quiescent , but still be very much affected by the recent agn outburst , consistent with the idea that radio sources can affect their host cluster on timescales much longer than the duration of the active phase ( e.g. fabian et al . 2003 , shabala & alexander 2009a ) .
alternatively , the observable agn signature could disappear while still undergoing feedback .
by contrast , clusters hosting radio sources must have had sufficient fuel to feed the central agn for the past @xmath159 . since @xmath170 , the radio source triggering- and , by extension , the return of the cool gas- must have happened recently .
this explains the bluer colours of galaxies in clusters which host a radio source in the `` outside '' and `` mixed '' regions .
the `` inside '' region of fr - ii lobes is by definition affected by feedback ( since it is the volume which has definitely been overrun by the radio lobes ) , and is therefore devoid of gas .
it is likely that these galaxies simply have not had the time to form new stars before the gas was blasted out again by the latest agn outburst , explaining why their colours are similar to those for the quiescent sample .
the marginally redder colours of galaxies in fr - ii clusters are likely due to a slightly longer time since the last burst of star formation than in the quiescent sample .
the quiescent cluster sample was selected on the basis of a lack of 1.4 ghz radio emission .
it is not immediately clear what physical condition this corresponds to .
radio sources undergo significant dynamical evolution , increasing in size and typically decreasing in luminosity after a few myrs @xcite .
radio telescopes are surface brightness limited , a quantity that depends on both source size and luminosity , @xmath171 where @xmath172 is the beam fwhm . at some point in its evolution
a radio source can become dim and diffuse enough to no longer be detected in surveys .
this can happen both as the agn jet causes expansion of the radio cocoon , and after the jet terminates .
many clusters host radio bubbles ( e.g. fabian et al . 2003 , forman et al .
these are remnants of past agn activity , formed as the radio lobes are pinched by kelvin - helmholtz and rayleigh - taylor instabilties .
the radio plasma in the bubbles is underdense , and these therefore rise buoyantly through the cluster at a speed comparable with the sound speed , sweeping up cluster gas as they do so @xcite . if undisturbed , these bubbles can take a few gyrs to reach cluster periphery .
it is therefore important to ask is over what fraction of their lifetime the active radio agn and bubbles are detectable in radio emission .
the electron population emitting at 1.4 ghz is subject to energy losses via synchrotron emission , adiabatic expansion , and inverse - compton upscattering of cmb photons .
radio bubbles are typically @xmath173 myrs old , a time when inverse - compton losses begin to dominate the radio luminosity .
kaiser & best give @xmath174 in this regime , where the density profile of the icm goes as @xmath175 .
the largest cocoon / bubble that can be detected in nvss ( @xmath176 arcsec , @xmath177 mjy ) is then @xmath178 . using the luminosity - size relation , @xmath179^{6/(10+\beta ) } \label{eqn : rbubble}\ ] ] where @xmath180 and @xmath181 are the radius and luminosity at reference time @xmath182 .
source age is related to size via an appropriate dynamical model ; this age can be written as @xmath183 . in the case of an active radio source , kaiser & alexander
give @xmath184 , while for an inactive bubble @xmath185 @xcite .
this yields the maximum observable lifetime of @xmath186^{\frac{6}{x(10+\beta ) } } \label{eqn : tmax}\ ] ] shabala et al .
derive ages and jet power for a complete sample of local radio galaxies . here , we identify the largest , oldest sources with the jet termination phase , and therefore set @xmath187 myrs , @xmath188 kpc , @xmath189 @xmath163 .
this yields @xmath190 for the active , and @xmath191 for the bubble phases .
in other words , the radio source would become undetectable after a few hundred myrs .
this is significantly shorter than the gas return timescale , and therefore it is entirely possible that the `` quiescent '' clusters in our sample contain radio cocoons and/or bubbles undetectable in current large - scale surveys , explaining the red colours of their galaxies .
the bubbles can be disrupted by a number of instabilities .
however , even if they are not long - lived enough to fade and avoid detection , the long gas return timescales imply that galaxies will appear red in such clusters for an appreciable fraction of the hubble time . in considering fr - ii and fr - i radio sources separately ,
we have shown that two distinct modes of agn feedback exist .
powerful events , which we identify with the fr - ii phase , are relatively rare , but can affect their environment on cosmological timescales through gas heating and expulsion . on the other hand , the less poweful fr - i events are much more frequent ( by as much as two orders of magnitude ; shabala et al .
2008 ) , but can not affect the evolution of galaxies outside their own host .
this is consistent with a picture in which fr - i sources start out as fr - iis and are simply disrupted within the dense galaxy core .
agn counts in the local volume are dominated by these low - luminosity events , and are consistent with the agn being triggered by cooling of gas out of the hot halo @xcite .
recently shabala & alexander have shown that this low - level mode of feedback is sufficient to stop star formation in massive galaxies at late times and match the local stellar mass function .
observations of the agn fraction are insensitive to the duration of the active and quiescent phases of radio activity , instead only providing information on the ratio @xmath192 .
as discussed above , both these timescales will be significantly longer for fr - ii events than for fr - is .
galaxy colours , on the other hand , are sensitive to the _ absolute _ timescales , rather than just this ratio .
it appears likely that the few very red galaxies in figure [ fig : insidefit ] that can not be explained by our simple model may reside in clusters which have hosted more than one powerful fr - ii outburst . in this scenario ,
if re - triggering of the agn in the fr - ii host is facilitated by a sudden influx of cold gas ( as would happen , for example , in a gas - rich minor merger ) , it can proceed without catastrophic cooling in other galaxies within the cluster .
repeated gas ejection outbursts are therefore possible without associated star formation . on the other hand ,
fr - i mode of feedback is quite different in that only the host galaxy is affected by the radio source .
this means that gas must be supplied to the galaxy for another agn outburst to take place , and the outburst will therefore usually be accompanied by circumnuclear star formation .
these galaxies will typically appear bluer .
kaviraj et al .
find that gas ejection is required to explain the colour evolution of early - type galaxies from the blue cloud to the red sequence .
this could be provided by either of the two modes of feedback discussed here .
however , it is difficult to envisage a scenario in which small - scale agn outbursts can explain the colours of the reddest observed galaxies .
galaxies recently overrun by fr - ii radio sources show no correlation between the radio source age and @xmath129 colour , suggesting that the colours are only sensitive to agn feedback on timescales exceeding typical radio source lifetimes of @xmath193 myrs .
we present an analysis of galaxy colours around fr - i and fr - ii type radio sources . we find that the less powerful fr - i radio sources can not affect the colours of galaxies near agn hosts ; while galaxies overrun by expanding fr - ii radio sources exhibit redder colours that are consistent with truncation of star formation following the passage of a radio source - driven bow - shock .
we compare the fr - i and fr - ii samples with galaxies located in clusters that show no evidence of current agn activity , and find that these quiescent clusters are redder than expected for a secular evolution scenario .
galaxy colours can be affected on timescales significantly exceeding the detectable agn lifetime , and rare powerful agn events thus play an important role in the colour evolution of local galaxies .
we thank the referee , elaine sadler , for insightful , thorough and constructive comments that have undoubtedly improved the manuscript .
sss thanks new college , oxford for a research fellowship and the bipac institute at oxford for support .
sk acknowledges a research fellowship from the royal commission for the exhibition of 1851 , an imperial college junior research fellowship , a senior research fellowship form worcester college , oxford and support from the bipac institute .
sss thanks anna scaife for assistance with the radio data , and dominic ford for creating pyxplot . funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the us department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society and the higher education funding council for england .
the sdss web site is http://www.sdss.org .
the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory and the university of washington .
the nvss and first surveys were carried out using the national radio astronomy observatory vla .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . | we investigate the effects of agn feedback on the colour evolution of galaxies found in local ( @xmath0 ) groups and clusters .
galaxies located within the lobes of powerful fanaroff - riley type ii ( edge - brightened ) sources show much redder colours than neighbouring galaxies that are not spatially coincident with the radio source .
by contrast , no similar effect is seen near fanaroff - riley type i ( core - dominated ) radio sources .
we show that these colours are consistent with fr - iis truncating star formation as the expanding bow shock overruns a galaxy .
we examine a sample of clusters with no detectable radio emission and show that galaxy colours in these clusters carry an imprint of past agn feedback .
agn activity in the low - redshift universe is predominantly driven by low - luminosity radio sources with short duty cycles .
our results show that , despite their rarity , feedback from powerful radio sources is an important driver of galaxy evolution even in the local volume .
galaxies : evolution galaxies : photometry galaxies : active |
a recent trend in computer networks and cloud computing is to virtualize network functions , in order to provide higher scalability , reducing maintenance costs , and increasing reliability of network services .
virtual network functions as a service ( vnfaas ) is currently under attentive study by telecommunications and cloud stakeholders , as a promising business and technical direction consisting of providing network functions ( i.e. , firewall , intrusion detection , caching , gateways ... ) as a service instead of delivering standalone network appliances . while legacy network services are usually implemented by means of highly reliable hardware specifically built for a single purpose middlebox , vnfaas moves such services to a virtualized environment @xcite , named _
nfv infrastructure ( nfvi ) _ and based on commercial - off - the - shelf hardware @xcite .
services implementing network functions are called _ virtual network functions ( vnfs)_. one of the open issues for nfvi design is indeed to guarantee high levels of vnf availability @xcite , i.e. , the probability that the network function is working at a given time .
in other words , a higher availability corresponds to a smaller downtime of the system , and it is required to satisfy stringent _
service level agreements ( sla)_. failures may result in a temporary unavailability of the services , but while in other contexts it may be tolerable , in nfvi network outages are not acceptable , since the failure of a single vnf can induce the failure of all the overlying services @xcite . to achieve high availability , backup vnfs can be placed into the nfvi , acting as replicas of the running vnfs , so that when the latter fail , the load is rerouted to the former . however , not all vnfs are equal ones : the software implementing a network function of the server where a vnf is running may be more prone to errors than others , influencing the availability of the overall infrastructure .
also , client requests may be routed via different network paths , with different availability performance . therefore to guarantee high levels of availability
it is important not only to increase the number of replicas placed on an nfvi , but it is also crucial to select where they are placed and which requests they serve . in this context , we study and model the _ high availability virtual network function placement ( ha - vnfp ) _ problem , that is the problem of placing vnfs on an nfvi in order to serve a given set of clients requests guaranteeing high availability .
our contribution consist of : a quantitative probabilistic model to measure the expected availability of vnf placement ; a proof that the problem is @xmath0-hard and that it belongs to the class of nonlinear optimization problems ; a linear mathematical programming formulation that can be solved to optimality for instances of limited size ; a _ variable neighborhood search ( vns ) _
heuristic for both online and offline planning ; an extensive simulation campaign , and algorithm integration in a decision support system ( dss ) tool @xcite .
the paper is organized as follows : in [ sec : description ] we present the ha - vnfp , and in [ sec : literature ] we briefly describe previous works on vm / vnf placement in cloud / nfvi systems . in [ sec : modelling ] we formally describe the optimization problem and propose a linearization of the problem and a mathematical programming formulation that can solve it to optimality . in [ sec : algorithms ] we describe our heuristic methodologies , which are then tested in an extensive simulation campaign in [ sec : simulation ] .
we briefly conclude in [ sec : conclusion ] .
we consider an nfvi with several geo - distributed datacenters or _ clusters _ ( see [ figure : nfv - abstract - infrastructure ] ) .
each cluster consists of a set of heterogeneous servers with limited available computing resources .
several instances of the same vnf type can be placed on the nfvi but on different servers .
each vnf instance can be assigned to a single server allocating some of its computing resources .
indeed each server has a limited amount of computing resources that can not be exceeded .
a network connects together all servers of the nfvi : we suppose that the communication links inside a cluster are significantly more reliable than those between servers in different clusters .
an access network with multiple access points connects servers to clients .
links connecting access points to clusters can differ in the availability level , depending on the type of the connection or the distance from the cluster . in this article
, we assume that the total amount of resources and network capacities are sufficient to manage the expected client requests at any time .
however assignment decisions may artificially produce congestion over the servers .
we analyze how to find assignments providing a trade - off between nfvi availability and system congestion .
we are given an estimation of the expected client vnf requests , each characterized by a computing resource demand .
an assigned request consumes part of the resources reserved by a vnf instance . indeed the consumed resources must not exceed the reserved ones .
requests can be assigned using two different policies : _ demand load balancing _ and
_ split demand load balancing_. in the former , a client request is always fully assigned to a single server , while in the latter it may be split among different ones . splitting a request also splits proportionally its demand of computing resources .
indeed , when a demand is split it relies on the availabilities of many vnf instances , decreasing the expected availability of the service , but increasing the chance of finding a feasible assignment in case of congestion .
we suppose a multi - failure environment in which vnfs , servers , clusters , and networks may fail together .
our aim is to improve the vnf availability by replicating instances on the nfvi .
we distinguish between _ master _ and _ slave _ vnfs : the former are active vnf instances , while the latter are idle until masters fail .
an example of vnf placement is depicted in [ figure : nfv - abstract - placement ] .
each master may be _ protected _ by many slaves - we assume in this article that a slave can protect only a master , must be placed on a different server , and must allocate at least the same amount of computing resources of its master .
each master periodically saves and sends its state to its slaves , e.g. using technologies such as the one presented in @xcite , in such a way that the latter has always an updated state and can consistently restore the computation in case of failure of the former .
we suppose that if a master is unreachable , a searching process is started in order to find a slave that can complete the computation .
if the searching process fails and all slaves are unreachable , then the service is considered unavailable .
a representation of vnf protection is in [ figure : nfv - dp ] . and
@xmath1 , while slave are running on @xmath2 and @xmath3 .
each link between vnf instances represents the connection between a master and its slave . ]
even if vm and vnf resource placement in cloud systems is a recent area of research ( see @xcite for a high - level comprehensive study ) , however there already exists orchestrators that are driven by optimization algorithms for the placement , such as @xcite .
we now present few works in literature studying the optimization problems that arises in this context . [
[ placement - of - virtual - machines ] ] placement of virtual machines + + + + + + + + + + + + + + + + + + + + + + + + + + + + + @xcite studies the problem of placing vms in datacenters minimizing the average latency of vm - to - vm communications .
such a problem is @xmath0-hard and falls into the category of _ quadratic assignment problems_. the authors provide a polynomial time heuristic algorithm solving the problem in a _
`` divide et impera '' _ fashion . in @xcite
the authors deal with the problem of placing vms in geo - distributed clouds minimizing the inter - vm communication delays .
they decompose the problem in subproblems that they solve heuristically .
they also prove that , under certain conditions , one of the subproblems can be solved to optimality in polynomial time .
@xcite studies the vm placement problem minimizing the maximum ratio of the demand and the capacity across all cuts in the network , in order to absorb unpredictable traffic burst .
the authors provide two different heuristics to solve the problem in reasonable computing time .
[ [ placement - of - virtual - network - functions ] ] placement of virtual network functions + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + @xcite applies nfv to lte mobile core gateways proposing the problem of placing vnfs in datacenters satisfying all client requests and latency constraints while minimizing the overall network load .
instead , in @xcite the objective function requires to minimize the total system cost , comprising the setup and link costs .
@xcite introduces the vnf orchestration problem of placing vnfs and routing client requests through a chain of vnfs . the authors minimize the setup costs while satisfying all client demands .
they propose both an ilp and a heuristic to solve such problem .
also @xcite considers the vnf orchestration problem with vnf switching piece - wise linear latency function and bit - rate compression and decompression operations .
two different objective functions are studied : one minimizing costs and one balancing the network usage .
[ [ placement - with - protection ] ] placement with protection + + + + + + + + + + + + + + + + + + + + + + + + + in @xcite vms are placed with a protection guaranteeing @xmath4-resiliency , that is at least @xmath4 slaves for each vm .
the authors propose an integer formulation that they solve by means of constraint programming . in @xcite the recovery problem of a cloud system
is considered where slaves are usually turned off to reduce energy consumption but can be turned on in advance to reduce the recovery time .
the authors propose a bicriteria approximation algorithm and a greedy heuristic . in @xcite
the authors solve a problem where links connecting datacenters may fail , and a star connection between vms must be found minimizing the probability of failure .
the authors propose an exact and a greedy algorithms to solve both small and large instances , respectively . within disaster - resilient vm placement
, @xcite proposes a protection scheme in which for each master a slave is selected on a different datacenter , enforcing also path protection . in @xcite
the authors solve the problem of placing slaves for a given set of master vms without exceeding neither servers nor link capacities .
their heuristic approaches decompose the problems in two parts : the first allocating slaves , and the second defining protection relationships . in a recent work
@xcite , the authors model the vm availability by means of a probabilistic approach and solve the placement problem over a set of servers by means of a nonlinear mathematical formulation and greedy heuristics . this is the only work offering an estimation of the availability of the system .
however , it considers only the availability of the servers , while in our problem we address a more generic scenario : when datacenters are geo - distributed , a client request shall be assigned to the closest datacenter , since longer connections may have a higher failure rate .
therefore , the source of the client requests may affect the placement of the vnfs on the nfvi , and must be taken into account in the optimization process and in the estimation of the availability .
in the following we propose a formal definition to the ha - vnfp and a mathematical programming formulation .
[ [ clusters - and - servers ] ] clusters and servers + + + + + + + + + + + + + + + + + + + + we are given the set of clusters @xmath5 and the set of servers @xmath6 .
each server @xmath7 belongs to a cluster @xmath8 , and we define as @xmath9 the set of servers of cluster @xmath10 .
we represent the usual distinct types of computing resources ( cpu , ram , ... ) of server @xmath11 by the same global amount @xmath12 of available resources .
[ [ virtual - network - functions ] ] virtual network functions + + + + + + + + + + + + + + + + + + + + + + + + + a set @xmath13 of vnf types is given .
each vnf instance runs on a single server .
each server can host multiple vnf instances , but at most one master for each type .
[ [ networks ] ] networks + + + + + + + + an inter - cluster network allows synchronization between clusters , while an access network connects clusters to a set of access points @xmath14 .
we are given sets @xmath15 and @xmath16 of logical links @xmath17 connecting clusters @xmath18 , and logical links @xmath19 connecting cluster @xmath20 to access point @xmath21 , respectively . [ [ clients - requests ] ] clients requests + + + + + + + + + + + + + + + + a set of clients requests @xmath22 is given .
each request @xmath23 is a tuple @xmath24 of the requested vnf type @xmath25 , a subset of available access points @xmath26 , and the resources demand @xmath27 . [ [ availability ] ] availability + + + + + + + + + + + + taking into account explicit availability in nfvi design becomes necessary to ensure slas @xcite .
we suppose that the availabilities of each component ( server , cluster , vnf , link ) are given ( see [ tab : data ] ) , each corresponding to the probability that a component is working . [ [ objective - function ] ] objective function + + + + + + + + + + + + + + + + + + all clients requests must be assigned to servers maximizing the availability of the system , we measure as the minimum availability among all requests .
c|l|c|l @xmath5 & set of clusters & @xmath6 & set of servers + @xmath28 & set of servers in cluster @xmath10 & @xmath13 & set of vnf types + @xmath22 & set of requests & @xmath14 & set of access points + @xmath29 & set of synchro .
links & @xmath16 & set of access links + @xmath30 & set of request access points & + @xmath31 & capacity of server @xmath7 & @xmath32 & demand of request @xmath33 + @xmath8 & cluster of server @xmath7 & @xmath34 & vnf of request @xmath33 + @xmath35 & availability of vnf @xmath36 & @xmath37 & availability of server @xmath7 + @xmath38 & availability of synchro . link @xmath39 & @xmath40 & availability of access link @xmath41 + @xmath42 & availability of cluster @xmath10 & + concerning the assignment of requests , we can prove that : [ theo : feasibility - split ] when demand split is allowed and @xmath43 , ha - vnfp has always a feasible solution that can be found in polynomial time .
in fact since the requests can be split among servers , the feasibility of an instance can be found applying a next - fit greedy algorithm for the bin packing problem with item fragmentation ( bppif ) @xcite : servers can be seen as bins , while requests as items that must be packed into bins .
the algorithm iteratively pack items to an open bin . when there is not enough residual capacity , the item is split , the bin is filled and closed , and a new bin is open packing the rest of the item .
when requests can be split , such algorithm produces a feasible solution for the ha - vnfp : if a request is assigned to a server , then a master vnf serving such a request is allocated on that server too .
the next - fit algorithm runs in @xmath44 and therefore a feasible solution can be found in polynomial time .
[ theo : feasibility - nosplit ] the feasibility of a ha - vnfp instance without demand split is a @xmath0-hard problem .
indeed we can see again the feasibility problem as a bin packing problem ( bpp ) .
however , without split each item must be fully packed into a single bin .
therefore , finding a feasible solution is equivalent to the feasibility of a bpp , which is @xmath0-hard , and it directly follows that : [ theo : complexity - nosplit ] the ha - vnfp without demand split is @xmath0-hard .
that is , for unsplittable demands , it is @xmath0-hard finding both a feasible solution and the optimum solution .
it is less straightforward to also prove that : [ theo : complexity - split ] the ha - vnfp with demand split is @xmath0-hard .
in fact , let us suppose a simple instance where all components ( servers , clusters , links , ... ) are equal ones and where @xmath45 , which means that there will be no slaves in our placement .
the problem can be seen again as a bppif in which the objective is to minimize the number of splits of the item that is split the most : in fact , every time a request is split , the availability of the system decreases . in such scenarios
the best solution is the one in which no request is split at all - however , if we could solve such a problem in polynomial time , then we could solve also the feasibility problem of a bpp in polynomial time , which instead is @xmath0-hard . therefore , since we can reduce a feasibility problem of bpp to an instance of bppif , and the latter to an instance of ha - vnfp , the ha - vnfp with split is @xmath0-hard . in the following we propose a mathematical programming formulation of ha - vnfp starting from the definition of the set of the solutions : a _ request assignment _
@xmath46 is a pair @xmath47 indicating the subset of servers @xmath48 running either the master or the slaves of a vnf instance , and the server @xmath49 where the master is placed .
we also define @xmath50 as the set of all request assignments . an _ assignment configuration _
@xmath51 ( see [ fig : assignment - configurations ] ) is a set of all request assignments @xmath46 for all the fragments of a request .
we define as @xmath52 the set of all assignment configurations @xmath51 , that is @xmath53 , where request @xmath54 is split and assigned to two different master vnfs on servers @xmath55 and @xmath56 . both masters have slaves : the master vnf on server @xmath55 has slaves on servers @xmath2 and @xmath57 , while the one on server @xmath56 has slaves on servers @xmath3 and @xmath58 . ]
[ [ availability - computation ] ] availability computation + + + + + + + + + + + + + + + + + + + + + + + + we compute the nfvi availability for a request @xmath33 by means of a probabilistic approach @xcite .
given a cluster and a set of access points , @xmath59 is the function computing the probability that at least one of the access links is working : @xmath60 given a vnf and a set of servers , @xmath61 is the function computing the probability that at least one instance of vnf is working : @xmath62 given a request @xmath33 and a request assignment @xmath63 , @xmath64 is the function computing the probability that at least one of the instances of @xmath46 is working : @xmath65\end{gathered}\ ] ] when a request @xmath33 is split , we compute its availability @xmath66 as the probability that all of its parts succeed : @xmath67 we remark that such formula is nonlinear and produces a integer nonlinear programming formulation which can not be solved by common integer solvers like cplex .
therefore we propose a mip linearization of such nonlinear formulation in which for each assignment configuration @xmath68 we have a binary variable stating if such configuration is selected in the solution . [ [ variables ] ] variables + + + + + + + + + the following variables are needed : @xmath69 [ [ model ] ] model + + + + + ha - vnfp can be modeled as follows : @xmath70 constraints and ensure that each request is fully assigned and selects an assignment configuration , respectively . constraints and set the allocated resources of masters and slaves , respectively .
constraints ensure that servers capacities are not exceeded .
constraints impose that at most one assignment configuration is selected for each request .
constraints compute the minimum availability .
our formulation can model both the ha - vnfp with and without split : in fact by simply setting @xmath71 for each configuration @xmath51 we forbid configurations splitting a request .
solving ha - vnfp as a mip using an integer solver works only for small nfvi , since the number of variables is exponential w.r.t the size of the instances .
therefore we propose two different heuristic approaches for ha - vnfp : the first is an adaptation of well - known greedy policies for the bpp that will serve as comparison , while the second is a _ variable neighborhood search _ heuristic using different algorithmic operators to explore the neighborhood of a starting point .
most of the heuristics for the placement of vms or vnfs are based on a greedy approach , and bpp heuristics are often exploited to obtain suitable algorithms for the placement , such as in @xcite .
we also exploit bpp heuristics to obtain three different greedy approaches for the ha - vnfp : _ best availability _ , _ best fit _ , and _ first fit _ greedy heuristics .
the algorithm , reported in [ algo : greedy ] , starts from an empty initial placement and for each request @xmath33 it looks for a server having enough residual capacity to satisfy the demand @xmath32 .
if such a server is found , then the request is assigned to it , otherwise the algorithm fails without finding a feasible solution .
however , we can observe that : when @xmath72 and split is allowed , our greedy heuristic always finds a feasible solution .
in fact we can always split a request between two servers , as stated also in [ theo : feasibility - split ] .
the selection of the server is performed by the procedure @xmath73 which discards the servers without sufficient resources to satisfy demand @xmath74 , and selects a server depending on the chosen policy : * _ best fit _ : the server whose capacity best fits the demand ; * _ first fit _ :
the first server found ; * _ best availability _ : the server with the highest availability
. while the first three policies are well - know for the bpp , the fourth one is designed for the ha - vnfp .
master vnfs are placed during the assignment of the requests .
then , in a similar way , the algorithm places additional slaves : for each master the algorithm looks for a server having enough capacity for a slave still using @xmath75 procedure . after a server is found
, the slave is placed .
such a procedure is repeated until no additional slave is placed .
@xmath76 assignment of requests create vnf @xmath34 if it does not exists in @xmath77 assign request @xmath33 to server @xmath78 in @xmath77 demand @xmath32 is decreased infeasible add slaves @xmath79 servers of @xmath6 without @xmath80 and its slaves create slave of vnf @xmath80 on server @xmath78 in @xmath77 @xmath77 the _ variable neighborhood search ( vns ) _ is a meta - heuristic that systematically changes the neighborhood within the local search algorithm , in order to escape from local optima . in other words
, it starts from an initial solution , applies a local search algorithm until it improves , and then changes the type of local search algorithm applied to change the neighborhood .
our vns algorithm explores @xmath81 different neighborhoods and it is initialized with several starting points , each obtained using a different greedy algorithm .
@xmath82 , , @xmath83 @xmath84 @xmath85 @xmath86 @xmath87 apply @xmath88 to @xmath77 @xmath89 * break * @xmath90 the main logic of our vns algorithm is sketched in [ algo : vns ] : we generate @xmath91 starting points by using the greedy heuristics of [ sec : greedy ] and we explore their neighborhood for a placement improving the availability .
if no improvement can be found , the algorithm switches the neighborhood .
indeed applying local search is time expensive , but we can observe that a max - min objective function divides the requests in two sets : a set of requests having an availability equal to the objective function and another set having a better availability .
we refer to the former as the set of the _ worst requests _ , since they are the ones whose improvement will also improve the availability of the entire solution . to reduce the computing time and focus our algorithm we found to be profitable to restrict the explored neighborhood to the worst requests only .
also , after applying each operator we look for new slaves , as in the greedy procedure [ algo : greedy ] . given two feasible placements , we say that one is improving if it has a higher availability or if it has the same availability but fewer worst requests . in the following
we describe the neighborhoods of our vns .
[ [ vnfs - swap ] ] vnfs swap + + + + + + + + + the first neighborhood consists of swapping vnfs ( see [ fig : vnf - swap - n ] ) : given a vnf , we swap it with a subset of vnfs deployed on a different server . if the placement is improved , then we store the result as the best local change .
in general our operator is @xmath92 but we found profitable to set an upper bound of @xmath93 to the cardinality of the set of swapped vnfs , obtaining a @xmath94 operator . and
@xmath3 are swapped .
if a vnf is a master , then all its assigned requests are redirected to the new server.,title="fig : " ] and @xmath3 are swapped .
if a vnf is a master , then all its assigned requests are redirected to the new server.,title="fig : " ] and @xmath3 are swapped .
if a vnf is a master , then all its assigned requests are redirected to the new server.,title="fig : " ] [ [ slave - vnfs - swap ] ] slave vnfs swap + + + + + + + + + + + + + + + we explore the neighborhood where a slave vnf is removed to free resources for an additional slave of a different master vnf ( see [ fig : b - swap - n ] ) .
the complexity of this operator is @xmath94 .
[ [ requests - swap ] ] requests swap + + + + + + + + + + + + + we also explore the neighborhood where requests are swapped ( see [ fig : r - swap - n ] ) : given a request we consider a subset of requests assigned to a different server and then swap the former with the latter . similarly to the swap of vnfs , the complexity of this operator is @xmath95 .
however , by setting an upper bound of @xmath93 to the cardinality of the swapped requests set we obtain a @xmath44 operator . and @xmath96 are swapped changing the respective servers . when swapping a request , a new vnf instance is created if none existed on the new server .
, title="fig : " ] and @xmath96 are swapped changing the respective servers . when swapping a request , a new vnf instance
is created if none existed on the new server .
, title="fig : " ] and @xmath96 are swapped changing the respective servers . when swapping a request , a new vnf instance
is created if none existed on the new server .
, title="fig : " ] [ [ request - move ] ] request move + + + + + + + + + + + + in the last exploration we consider the neighbors where a request is simply moved to a different server ( see [ fig : r - move - n ] ) .
the complexity of this operator is @xmath97 .
is moved and assigned to a different server .
, title="fig : " ] is moved and assigned to a different server .
, title="fig : " ] is moved and assigned to a different server .
, title="fig : " ] in principle , even if all the operators polynomial time our vns algorithm is not .
however , an upper bound @xmath4 to the number of iterations can be set , obtaining a @xmath98 heuristic .
also , in the following we show that our vns requires small computing time for nfvi of limited size and it can be parameterized to end within a time limit , making it suitable for both online and offline planning .
we evaluate empirically the quality of our methodologies : the greedy heuristic using four different policies ( best fit , first fit , and best availability ) , the vns algorithm , and the mathematical programming formulation as a mip .
however we could run our mip only on small instances with @xmath91 or @xmath81 servers and @xmath99 requests . in our framework
we first run the algorithms using the demand load balancing policy , and allowing split only if the former fails to assign all the requests .
all methodologies are implemented in c++ , while cplex 12.6.3 @xcite is used to solve the mip .
the simulations have been conducted on intel xeon x5690 at 3.47 ghz .
we also produced a graphical dss tool integrating the vns and the greed algorithms ( in python ) working on arbitrary 2-hop topologies and made it available in @xcite .
we generated a random dataset consisting of instances that differ for the number of requests , total amount of computing resources , and availabilities of the network components .
we set the number of vnf types provided by our nfvi to @xmath100 .
we assumed an nfvi with @xmath91 clusters ( @xmath101 ) and @xmath91 access points ( @xmath102 ) .
each request has a random demand @xmath103 $ ] , while each server has a random capacity @xmath104 $ ] .
the availabilities of all the components of our nfvi are selected between @xmath105 as in @xcite .
we generated @xmath106 instances for each combination of : * number of requests @xmath107 ; * number of access points for each request @xmath108 .
the number of servers depends on the number of requests , the total amount of the demands , and the random capacities : we generated a set of servers such that their capacities are enough to serve all the demands @xmath109 .
note that such condition guarantees the feasibility only when splitting requests is allowed .
servers are randomly distributed among all the clusters , in such a way that for each pair of clusters @xmath10 and @xmath110 we have @xmath111 under these conditions we obtained instances with around @xmath91 servers when @xmath112 and @xmath113 servers when @xmath114 .
we first evaluate the quality of our vns heuristic against the solutions obtained by the mip solver and the greedy heuristics . in order to study
how the nfvi behaves on different levels of congestion , we let its computing resources to grow from an initial value of @xmath115 , to @xmath116 , with a step of @xmath117 .
due to the exponential number of the variables of our formulation , the mip solver could handle only small instances with @xmath99 requests .
all tests have been performed setting a time limit of two hours each and all the algorithms managed to assign all requests without splits . in [ fig : test - time - random ] we show the average computing time of the algorithms : while the mip hits several times the time limit , computing times are negligible for all the heuristics , and vns can be considered as a good alternative for online orchestration when the set of the requests is small .
the optimization problem seems harder when the amount of computing resources is scarce : in fact , the average computing time of the mip is closer to the time limit when the overall capacity is less than @xmath118 . instead , with higher quantities of resources the mip always find the optimal solution within the time limit . in [ fig : test - availability-50 ] we show that the results of our vns heuristic are close to the mip ones , while there is a significant gap between the latters and the greedy heuristic ones . in fact , both the mip and the vns succeed in finding solutions with an availability of three nines even with scarce resources .
eventually all the algorithms reach an high level of availability when computing resources are doubled . in [ fig : availability-50-aps ] we show the variation of the availability when the number of access points for each request increases : in [ fig : availability-50 - 11 ] , [ fig : availability-50 - 22 ] , and [ fig : availability-50 - 33 ] we report the average availability when requests can be routed to the nfvi using @xmath93 , @xmath119 , and @xmath91 access points , respectively .
the path protection obtained by using more than one access point substantially increases the level of the availability .
however , having more than @xmath119 access points does not provide additional benefits .
[ fig : availability-100-vns ] in a second round of experiments , we evaluated how our vns algorithm behave when scaling up the number of requests .
however , when the number of servers increases its is not possible to use the mip .
therefore , in the following analysis we compare the results of our vns algorithm to the greedy ones only .
in addition , since our vns algorithm has not polynomial time complexity , we include in the comparison the results obtained by setting a time limit of @xmath120 at the exploration of each starting solution
. we can first observe in [ fig : test - time - big ] that the computing time of the vns algorithm grows exponentially when the number of requests increases . indeed setting a time limit
reduces the overall computing time , which is always less than a minute . in [ fig : test - availability - big ]
we show that on average there is always a substantial gap between the results obtained by our vns algorithms and the greedy heuristics .
we can also observe that on average the vns time limit does not penalize substantially the results .
therefore our vns algorithm with time limit can reduce computing times with minimal loss in availability . from [ fig : availability-100 ] to [ fig : availability-500 ] we show the results on instances having from @xmath121 to @xmath122 requests individually .
we can observe that the greedy heuristics progressively loose in quality of the solutions , and the gap with the vns algorithms increases with the number of the requests . finally in [ fig : availability - vns ]
we show only the results concerning the vns algorithm without time limit and how it behaves when both the number of requests and the overall capacity increase .
we can observe that the curves are similar and the quality of the solutions provided by our algorithm is not affected by the increasing of the size of the instances .
our vns algorithm always provides placements with an availability of three nines even when resources are scarce , and it always reaches four nines when the capacity is doubled .
we defined and modeled the ha - vnfp , that is the problem of placing vnfs in nfvi guaranteeing high availability .
we provided a quantitative model based on probabilistic approaches to offer an estimation of the availability of a nfvi .
we proved that the arising nonlinear optimization problem is @xmath0-hard and we modelled it by means of a linear formulation with an exponential number of variables .
however , to solve instances with a realistic size we designed both efficient and effective heuristics . by extensive simulations ,
we show that our vns algorithm finds solution close to the mip ones , but in computing times smaller of orders of magnitude .
we highlighted the substantial gap between the availability obtained using classic greedy policies , and the one obtained with a more advanced vns algorithm , when the nfvi is congested .
our vns algorithm showed to be a good compromise to solve ha - vnfp in reasonable computing time , proving to be a good alternative for both online and offline planning .
we integrated our ha - vnfp algorithms in a graphical simulator made available with tutorial videos in @xcite .
this article is based upon work from cost action ca15127 ( `` resilient communication services protecting end - user applications from disaster - based failures - recodis '' ) supported by cost ( european cooperation in science and technology ) .
this work was funded by the anr reflexion ( contract nb : anr-14-ce28 - 0019 ) and the fed4pmr projects . | virtual network functions as a service ( vnfaas ) is currently under attentive study by telecommunications and cloud stakeholders as a promising business and technical direction consisting of providing network functions as a service on a cloud ( nfv infrastructure ) , instead of delivering standalone network appliances , in order to provide higher scalability and reduce maintenance costs .
however , the functioning of such nfvi hosting the vnfs is fundamental for all the services and applications running on top of it , forcing to guarantee a high availability level . indeed
the availability of an vnfaas relies on the failure rate of its single components , namely the servers , the virtualization software , and the communication network .
the proper assignment of the virtual machines implementing network functions to nfvi servers and their protection is essential to guarantee high availability .
we model the high availability virtual network function placement ( ha - vnfp ) as the problem of finding the best assignment of virtual machines to servers guaranteeing protection by replication .
we propose a probabilistic approach to measure the real availability of a system and design both efficient and effective algorithms that can be used by stakeholders for both online and offline planning . |
recently , bicknell , dopita , & odea ( 1997 ) developed a model which unifies the gigahertz peak spectrum ( gps ) sources and the related active galactic nuclei ( agn ) classes known as compact steep spectrum ( css ) sources and compact symmetric objects ( csos ) .
their model attributes the observed radio spectra and optical emission to the interaction of a jet driven , non thermal lobe with the ambient galactic interstellar medium ( ism ) . as the radio lobe forces its way through the ism , strong radiative shocks create shells of shock ionized and photo ionized gas which produce a low frequency break in the power law spectrum through free free absorption ( ffa ) .
although synchrotron self absorption ( ssa ) provides a natural mechanism to produce a low frequency break in a non thermal spectrum and is successful in explaining the spectra of agn with distinct core
jet morphologies ( @xcite ) , we expect that it is unimportant in the more extended jet lobes of gps sources , where the non thermal plasma is more likely to be optically thin to its own radiation .
however , the spectrum emitted by the non thermal lobe plasma can still exhibit a low frequency break not only as a result of ffa in an external shell of thermal gas surrounding the lobe , as suggested by bicknell et al .
( 1997 ) , but also as a result of the competing low frequency process of induced compton scattering ( ics ) .
ics becomes important in a high brightness temperature radiation field , such as that which is readily generated by synchrotron radio sources and which has a thomson scattering optical depth @xmath1 and a brightness temperature @xmath2 , such that @xmath3 where the parameter @xmath4 quantifies the strong effect of the angular distribution of the radiation and is given by : @xmath5 where @xmath6 is the scattering angle and the integral is over the solid angle of the beam incident on the scattering region ( @xcite ) . for a conical beam of radiation with half angle @xmath7 and solid angle @xmath8 , @xmath9 ; for complete isotropy of the radiation field @xmath10 ; for @xmath11 near unity , @xmath12 . in both cases @xmath13 .
ics redistributes photon energies from frequencies near where @xmath14 peaks to lower frequencies where they are absorbed ( e.g. by ssa or ffa ) effectively reducing the peak @xmath14 observed .
as first demonstrated by @xcite , the key observational consequence is a break in synchrotron radio spectra
. equation ( [ eq : opdepth ] ) implies that induced scattering can produce spectral distortions when @xmath15 and hence , unlike ` ordinary ' ( spontaneous ) compton scattering , it can be important even when @xmath16 .
indeed , for a typical synchrotron radio source , ics can be more efficient than ssa in producing a spectral turnover and can limit the brightness temperature at ghz frequencies to @xmath17 k , where @xmath18 is the low energy cutoff to the non thermal electron distribution , which is a fiducial value for the gps sources we are concerned with here ( bicknell et al . 1997 ) . ]
( @xcite ) .
note that this limit is only weakly dependent on the source parameters and ( for @xmath19 ) is lower than the theoretical inverse compton limit , @xmath20 k , for typical agn sources ( @xcite ) . as argued by sincell & krolik ( 1994 )
, ics may explain why the inverse compton limit is , in practice , rarely observed .
such is the case for gps sources , where the observed brightness temperatures of lobe subcomponents are typically @xmath21 k ( @xcite ) .
since ics by the non thermal , relativistic electrons in the jet lobes can not reduce the intrinsic peak @xmath22 to these observed levels , then ics may instead be operating externally . indeed ,
coppi , blandford & rees ( 1993 ) have shown that ics by relatively ` cold ' ( nonrelativistic ) electrons outside of a primary non thermal source can effectively reduce the peak brightness temperature to values such that @xmath23 and this alone makes it very tempting to adopt such a scenario to explain the observed @xmath22 values in gps sources . in this _ letter _ , we examine the process of induced scattering to show that in order to account for the observed brightness temperatures in gps sources , this process must be operating at some level in the shock ionized shell that has been postulated to surround the non thermal lobes .
in the scenario proposed by @xcite , the non thermal jet lobes in gps sources are surrounded by a shell of thermal , nonrelativistic plasma formed from shocked and photoionized material in the ambient ism .
although ffa is likely to be operating at some level in this external shell , there will be strong competition from ics , especially if the shell is tenuous . indeed , with the expected brightness temperatures , @xmath24 k , of the primary non thermal radiation which impinges upon the shell , ics is more efficient than ffa in producing a break at @xmath0 ghz frequencies when the electron number density , @xmath25 , satisfies @xmath26 where the electron temperature , @xmath27 k. this is precisely the range of densities relevant to shock
ionized ism material surrounding jet lobes in gps sources , as found by @xcite .
although , in general , collective plasma effects will not have an overall affect on ics ( see @xcite ) , this is not strictly the case in the presence of a high brightness temperature radiation field , which can trigger stimulated raman scattering , whereby the radio waves scatter off plasmons ( langmuir waves ) rather than off the individual electrons themselves . since , at the densities relevant here , this process requires brightness temperatures @xmath28 k ( @xcite ) , it is unlikely to be important in this context . a definitive test for any model which attempts to explain the spectral peak in gps / css sources is the observed inverse correlation between the frequency at which the spectrum peaks , @xmath29 , and the source size .
in the ffa model of @xcite , a theoretical relationship between these two parameters readily emerges from the shock dynamics .
we follow the same method here in constructing an analogous ics model , but as we explain below , the predicted relationship has different dependences on the physical parameters
. a spectral break due to ics is expected at the frequency where the optical depth , given by the left hand side of equation ( [ eq : opdepth ] ) , reaches unity . however , as equations ( [ eq : opdepth ] ) and ( [ eq : f ] ) show , induced scattering depends strongly on the angular distribution of the high brightness temperature radiation field and we have allowed for this through the parameter @xmath4 .
the flux density corresponding to the condition on the brightness temperature is obtained by simply integrating the surface brightness over the solid angle @xmath30 subtended by the source at the observer , giving @xmath31 in the model developed by @xcite , @xmath1 is approximately uniform over the sides of the lobe and the solid angle can be expressed as @xmath32 , where @xmath33 is the radius of the semi
ellipsoidal cocoon which forms as the jet driven lobe expands into the ism , @xmath34 is the distance from the core at the base of the jet to the lobe hotspot ( i.e. the semi major axis of the cocoon ) , @xmath35 is the distance to the source and @xmath36 typically ) is the ratio of the averaged hotspot pressure to cocoon pressure . by taking a power law spectrum , @xmath37 , where @xmath38 is a fiducial frequency ( 5 ghz in this case ) , and using the monochromatic power @xmath39 , equation ( [ eq : flux ] ) predicts a peak frequency given by @xmath40^{\frac { 1}{\alpha+2}}\ ; , \label{eq : nupeak}\ ] ] where @xmath39 is the monochromatic power .
the thomson optical depth is assumed to be approximately uniform over the sides of the lobe and is calculated with the mappingsii code ( @xcite ) for radiative shocks , giving @xmath41 where @xmath42 is the shock velocity corresponding to the sideways expansion of the radio lobe .
the two terms correspond to the shocked gas and photoionized precursor .
specifically , the shock velocity is given by ( see equations 2 - 10 and 2 - 11 in bicknell et al .
1997 ) @xmath43 where @xmath44^{1/3 } \label{eq : velocity}\ ] ] and where @xmath45 is the jet energy flux and @xmath46 is the power
law index of the hydrogen density distribution , given by @xmath47 .
equation ( [ eq : nupeak ] ) predicts a relationship between peak frequency and source size that derives quite differently from that predicted by the ffa model ( see equation 5 - 10 in @xcite ) .
since the opacity of ics is proportional to the radiation brightness temperature , the chief dependence of @xmath29 derives from the integrated surface brightness ( @xmath48 ) .
there is also a weaker dependence on @xmath34 through the shock velocity , as given in equation ( [ eq : vel_scale ] ) , which in turn determines @xmath1 . in the case of the ffa model , on the other hand , while there is also a similar weak dependence through @xmath49 , the predicted @xmath29
@xmath34 relationship derives chiefly from the dependence on density ( which has a power
law scaling with source size ) . this density dependence arises simply because the opacity due to ffa is @xmath50 , whereas for ics , it is @xmath51 where both integrals are through the ionized screen .
allowing for the inverse dependence of the shocked cooling zone and the photoionized precursor on density , the dependences on density are respectively on the first and zeroth powers of @xmath52 .
this is reflected , for instance , in the dependence of the thomson optical depth on velocity alone and this is where a dependence on density does enter , since the expansion velocity is a function of the ambient density .
although these two different models predict a @xmath53 @xmath34 relationship that derives from intrinsically different physical processes ( i.e. two body interactions versus particle photon interactions ) , the overall differences are surprisingly small : both ffa and ics models predict a negative slope in the @xmath29
@xmath34 plane which is close to unity and which varies very weakly with the relevant range of parameters .
let us now examine the predicted peak frequency - source size relationship more quantitatively .
the relationship between peak frequency , @xmath29 , and source size , @xmath34 , as predicted by equation ( [ eq : nupeak ] ) for the anisotropy parameter @xmath54 is plotted and compared with observational data ( @xcite ) in figure 1 .
the parameters involved in the theoretical lines are as follows : values of the hydrogen densities at 1 kpc , @xmath55 , are 1 and @xmath56 ; values of the density power law index , @xmath46 , are 1.5 and 2 ( these are the lower and upper limits suggested by @xcite in his fit to the luminosity size relation for csos ) ; the total 5 ghz radio power is @xmath57 , corresponding to the average radio power of gps sources ( @xcite ) ; the jet energy flux is @xmath58 and the spectral index is @xmath59 ( bicknell et al . 1997 ) .
table 1 lists the values of the parameter @xmath60 , which provides the scale for the size of the velocity as a function of source size ( see eq .
[ [ eq : velocity ] ] ) .
the slopes of both the @xmath61 and @xmath62 lines are adequate fits to the slope of the data points and the overall fits to the data are obviously better for the lower values of @xmath55 .
the fits could also be improved by increasing the radio power and thomson optical depth .
increasing the radio power increases the brightness temperature of the lobe and hence the frequency at which induced compton scattering becomes important .
we have not decreased the density still further to produce a more acceptable fit since it is evident from table 1 that lower densities would involve an unacceptable extrapolation of the fit to the thomson opacity , well beyond the velocities of @xmath63 that the mappingsii code can adequately handle .
indeed , the velocity in even the @xmath61 , @xmath64 model is greater than @xmath65 for an overall size @xmath66 .
it can be reasonably assumed , however , that the lower densities favoured by our ics model imply cocoon velocities in excess of @xmath65 , even if the extrapolation of the shock modeling beyond such a velocity is uncertain .
our model assumes a uniform distribution of radio emitting plasma in the jet lobes , whereas in a number of imaged gps sources , the emitting region is clumpy ; the highest brightness temperatures are typically identified with a small number of localized components ( @xcite ) . if these clumps were embedded in a lower surface brightness cocoon outside of which the ionizing screen were located then the radiation intercepting this screen would be anisotropic .
this effect of anisotropy is countered however , by the power ( typically @xmath67 ) entering in the expression for @xmath68 .
thus , @xmath68 would be reduced to approximately 40% of the isotropic value for @xmath69 . moreover , as inspection of the images of gps and cso sources ( the latter all being gps ) reveals , the `` clumps '' are usually situated at increasing distances from the nucleus with a local filling factor of order unity .
the beam solid angle corresponding to each clump should therefore be of order @xmath70 .
anisotropy need not be an important factor especially in view of the insensitivity of the peak frequency to it . on the other hand , non - uniformity in
the emission increases the brightness temperature . in our uniform model ,
the brightness temeprature at 5 ghz , is given by : @xmath71 and is about an order of magnitude lower than the values observed in subcomponents by @xcite , for example .
a larger brightness temperature of course _ enhances _ the prospect of induced compton scattering .
interestingly , the stanghellini et al .
study of a complete sample of gps sources has revealed a pronounced peak in the distribution of low frequency spectral indices near @xmath72 , which is in remarkabe agreement with the theoretically predicted spectra resulting from induced compton scattering by external thermal electrons ( e.g. @xcite ) .
this feature , combined with the successful ics explanation for the peak frequency source size relationship , shows that gps sources may indeed constitute the first real instance where induced compton scattering plays a relatively clear role in extragalactic radio sources .
we therefore conclude that induced compton scattering may be an important process in the shocked and photoionized plasma surrounding an expanding lobe in gps sources ; it can contribute to the formation of the characteristic peak in the radio spectrum when the density falls below @xmath73 and free free absorption becomes inefficient .
expansion velocities of @xmath74 are then implied when induced scattering is the dominant process .
induced scattering may thus be particularly relevant in the high redshift gps sources ( @xcite ) , which have column depths of the order of a few @xmath75 @xmath76 , at the lower end of the range estimated by @xcite on the basis of the ffa model .
in general , the introduction of induced compton scattering not only increases the range of plasma densities over which a low frequency break can be produced , but also increases the efficiency of ffa at the lowest frequencies , where photons are being continuously replenished as a result of the downscattering of ghz photons .
indeed , the combined effect of induced scattering and free - free absorption could also be important in some other sources , such as ngc 1275 , in which one of the jets appears to be significantly weaker than the other at low frequencies .
this has been attributed to free free absorption ( @xcite ) and although the observations are on parsec scales in this particular source , the brightness temperatures are similar to the values considered here , suggesting that induced compton scattering may well be operating on various jet scales , at the very least in conjuction with the more conventional process of free free absorption .
, m. c. 1996 , cygnus a : study of a radio galaxy ( cambridge : cup ) , 209 , g. v. , dopita , m. a. , & odea , c. p. 1997 , preprint , r. d. & knigl , a. 1979 , apj , 232 , 34 conway j.e . , myers s.t . , pearson t.j .
, readhead a.c.s .
, unwin s.c . and xu w. , 1994 , apj , 425 , 568 , p. , blandford , r. d. , & rees , m. j. 1993 , mnras , 262 , 603 , m. , fiore , f. , giommi , p. , & padovani ,
p. 1997 , apjl , in press , r. , fanti , c. , schilizzi , r. t. , spencer , r. e. , rendong , n. , parma , p. , van breugel , w. j. m. , & venturi , t. 1990 , astron .
, 231 , 333 , f. , burbidge , g. r. & sargent , w. l. w. 1966 , nature , 209 , 751 , a. & blandford , r. d. 1995 , mnras , 274 , 717 , a. , laor , a. & vermeulen , r. c. 1995 , apj , 448 , 589 , c. p. & baum , s. a. 1997 , astron .
, submitted , m. w. & krolik , j. h. 1994 , apj , 430 , 550 , m. w. & coppi , p. s. 1996 , apj , 460 , 163 , c. , odea , c. p. , baum , s. a. , & fanti , r. 1997 , in preparation , r. a. 1970 , astrophys .
, 7 , 19 , r. a. & dopita , m. a. 1993 , apjs , 88 , 253 wilkinson p.n.,polatidis a.g.,readhead a.c.s . , xu w. and pearson t.j . , 1994 ,
apj , 432 , l87
* table 1 : parameters of expanding cocoon * 11 pt | we revisit the shocked shell model for the class of active galactic nuclei known as gigahertz peak spectrum sources , incorporating new observational data on the radiation brightness temperatures .
we argue that in addition to free free absorption , induced compton scattering will also have an important effect in forming the @xmath0 ghz peak and in shaping the radio spectra that characterize these sources . indeed , our arguments suggest that gps sources may provide the first real evidence for the role of induced compton scattering in extragalactic radio sources . |
pre - planetary nebulae ( ppns ) , objects in transition between the agb and planetary nebula ( pn ) evolutionary phases , hold the key to our understanding of the relationship between these late evolutionary stages of low and intermediate mass ( @xmath11 ) stars . the hydrodynamic interaction of one or more fast , collimated post - agb outflows , believed to occur during the ppn phase , with the dense , massive , slowly - expanding spherical agb wind is believed to be responsible for shaping planetary nebulae ( sahai & trauger 1998 , sahai 2002 ) .
imaging surveys with the hubble space telescope have been crucial in revealing a wide variety of aspherical morphologies in ppns ( e.g. sahai 2004 ; ueta et al . 2000 ) , however the bulk of the circumstellar mass , which often resides in a spherical , dusty component surrounding the central aspherical nebula , is not well - probed in these data . thus a crucial ingredient for theoretical studies of such interactions , namely the mass and density of the ambient circumstellar medium ( e.g. , lee & sahai 2003 , 2004 ) , is lacking for a majority of the ppns discovered in hst surveys .
although attempts have been made to estimate the mass from submillimeter data ( e.g. gledhill et al .
2002 ) , these are compromised by the simplifying assumptions about the dust temperature .
the mineralogy of circumstellar dust shells around ppns became possible with the wavelength coverage and spectral resolution offered by the infrared space observatory ( iso ) ( see e.g. hrivnak et al .
2000 ; molster et al . 2002a , b , c ; gauba & parthasarathy 2004 and references therein ) .
in order to understand the grain temperatures , mass - loss history and morphology responsible for the formation of these minerals in the circumstellar environment of ppns , detailed modelling of the spectral energy distributions ( seds ) of a few ppns ( e.g. , hd161796 , hoogzaad et al .
2002 ; iras 16342 - 3814 , dijkstra et al . 2003a
; iras 16279 - 4757 , matsuura et al .
2004 ; gauba & parthasarathy 2004 ) have been undertaken . in this paper
, we examine the composition and distribution of dust in the ppn , iras19475 + 3119 ( hereafter i19475 ) and model its sed from the optical to the far - infrared to set constraints on the physical properties of its dust shell ( mass , temperature , size ) .
i19475 is listed as a f3ia star in the case - hamburg luminous stars survey ( stock et al .
high resolution optical spectra of the star have confirmed its post - agb nature ( klochkova et al .
2002 ; sivarani et al . 2001 ; arellano ferro et al .
2001 ) . a point - symmetric nebula with an extent of 4.9@xmath12 3.4 has been seen in j - band polarized flux images ( gledhill et al . 2001 ) .
co j=10 and 21 emission typical of agb and post - agb objects was detected from the circumstellar envelope of the star ( loup et al .
hrivnak & bieging ( 2005 ) noticed asymmetry and possible structure in the co j=21 and 43 lines .
recent imaging of i19475 with the hubble space telescope at optical wavelengths reveals a dusty quadrupolar nebula and spherical halo , surrounding the central star ( sahai 2004 ; sahai et al .
2006 ) .
the rest of the paper is organized as follows . in 2 ,
we provide observational details of the iso data , in 3 we describe the iso spectra , and the re - construction of i19475 s spectral energy distribution ( sed ) from these as well as broad - band photometric data , in 4 we describe our modelling of the sed using a spherical dust radiative transfer code , in 5 we discuss our models , and in 6 , we present our conclusions .
infrared space observatory ( iso ) observations of i19475 were extracted from the iso data archive .
these include spectroscopic observations made with the short wavelength spectrometer ( sws ) and the long wavelength spectrometer ( lws ) and spectrophotometric observations made with the imaging photopolarimeter ( isophot ) onboard iso .
an off - source measurement made with the lws in order to estimate the background was also extracted from the archive .
a log of the observations is given in table 1 .
iso sws spectra have a wavelength coverage of [email protected] @xmath8 m . our spectra were obtained in the low resolution mode ( aot 01 ) of the sws instrument ( de graauw et al .
1996 ) with a 33@xmath12 20aperture .
each sws spectrum contains 12 subspectra , that each consist of two scans , one in the direction of decreasing wavelength ( ` up ' scan ) and the other in the direction of increasing wavelength ( ` down ' scan ) .
there are small regions of overlap in wavelength between the sub - spectra .
each sub - spectrum is recorded by 12 independent detectors .
lws spectra extend from 43@xmath13197@xmath8 m .
the lws observations were obtained in lws01 mode , covering the full spectral range at a resolution ( @xmath14 ) of @xmath10 200 .
the lws circular field of view had a diameter of 84 .
the iso lws instrument and its calibration are described in clegg et al .
( 1996 ) and in swinyard et al .
( 1996 ) respectively .
isophot observations were carried out using the spectrophotometer subsystem pht - s .
pht - s consists of two low - resolution grating spectrometers covering the wavelength ranges [email protected] @xmath8 m ( pht - ss ) and [email protected] @xmath8 m ( pht - sl ) and having a common entrance aperture of 24@xmath12 24 .
each channel has a linear 64-element array of si : ga detectors .
the spectral band width ( fwhm ) of a single detector is 0.0383 @xmath8 m for pht - ss and 0.0918 @xmath8 m for pht - sl , resulting in a mean @xmath14 of about 95 for both channels .
a more detailed description of pht - s has been given by klaas et al .
( 1997 ) . in the pht40
observing mode spectrophotometry is performed simultaneously at wavelengths [email protected] @xmath8 m and [email protected] @xmath8 m .
arellano ferro , a. , giridhar , s. , mathias , p. , 2001
, a&a , 368 , 250 blcker , t. , 1995 , a&a , 299 , 755 castro - carrizo , a. , fong , d. , bujarrabal , v. , et al . , 2001 , in _ the promise of the herschel space observatory _ , eds .
pilbratt , j. cernicharo , a.m. heras , t. prusti & r. harris , esa - sp 460 , p. 257
clegg , p.e . ,
ade , p.a.r . ,
armand , c. , et al .
, 1996 , a&a , 315 , de graauw , th . , haser , l. n. beintema , d. a. , et al . , 1996 ,
a&a , 315 , l49 cohen , m. , walker , r.g . ,
witteborn , f.c . , 1992 ,
aj , 104 , 2030 dijkstra , c. , waters , l.b.f.m . , kemper , f. et al .
, 2003a , a&a , 399,1037 dijkstra , c. , dominik , c. , hoogzaad , s.n . , de koter , a. , min , m. , 2003 , a&a , 401 , 599 draine , b.t . , &
lee , h.m . , 1984 , apj , 285 , 89 ferrarotti , a. , gail , h .- p . ,
degiorgi , l. & ott , h.r . , 2000 ,
a&a , 357 , l13 garca - laro , p. , manchado , a. , pych , w. , pottasch , s.r . , 1997 , a&as , 126 , 479 gauba , g. & parthasarathy , m. , 2004 , a&a , 417 , 201 gledhill , t. m. , chrysostomou , a. , hough , j. h. , yates , j. a. , 2001 , mnras , 322 , 321 gledhill , t. m. , bains , i. , and yates , j. a. , 2002 , mnras , 332 , l55 hakkila , j. , myers , j.m .
, stidham , b.j . & hartmann , d.h . 1997 , aj , 114 , 2043 hanner , m.s . , 1988 ,
nasa cong . pub .
, 3004 , 22 hoogzaad , s.n . , molster , f.j . , dominik , c. et al . , 2002 , a&a , 389 , 547 hony , s. , waters , l.b.f.m . &
tielens , a.g.g.m .
, 2002 , a&a , 390 , 533 hrivnak , b.j . ,
volk , k. & kwok , s. , 2000 , apj , 535 , 275 hrivnak , b.j . & bieging , j.h . , 2005 ,
apj , 624 , 331 ivezi , z. , nenkova , m. , elitzur , m. , 1999 , _ user manual for dusty _ , university of kentucky internal report klochkova , v. g. , panchuk , v. e. , tavolzhanskaya , n. s. , 2002 , astl , 28 , 49 jura , m. 1986 , apj , 303 , 327 klaas , u. , acosta - pulido , j.a . , brahm , p. , et al . , 1997 , isophot - s : capabilities and calibration , in proc .
first iso workshop on analytical spectroscopy , ed .
a.m. heras , k. leech , n. r. trams & m. perry ( esa sp-419 ) , 113 klaas , u. , brahm , p. , laureijs , r.j .
et al . , 2002 , _ isophot calibration accuracies , version 5.0 _ , + http://www.iso.vilspa.esa.es/instr/pht/ laureijs , r.j . ,
ulrich , k. , richards , p.j . , schulz , b. & brahm , p. , 2003
, _ the iso handbook , volume iv : pht - the imaging photo - polarimeter _ , sai-99 - 069/dc , version 2.0.1 lee , chin - fei , sahai , r. , 2003 , apj , 586 , 319 lee , chin - fei , sahai , r. , 2004 , apj , 606 , 483 likkel , l. , omont , a. , morris , m. , forveille , t. , 1987 , a&a , 173 , l11 lim , t. , _ private communication _ loup , c. , forveille , t. , omont , a. , paul , j. f. , 1993 , a&as , 99 , 291 mathis , j.s . ,
rumpl , w. & nordsieck , k.h . , 1977 ,
apj , 17 , 25 matsuura , m. , zijlstra , a.a . , molster , f.j . , et al
. , 2004 , apj , 604 , 791 molinari , s. , _ data reduction recipe for an lws01 aot .
the case of a faint source _
molster , f.j . ,
lim , t.l . ,
sylvester , r.j .
, et al . , 2001 , a&a , 372 , 165 molster , f.j . ,
waters , l.b.f.m . , tielens , a.g.g.m . ,
barlow , m.j .
, 2002a , a&a , 382 , 184 molster , f.j . , waters , l.b.f.m . , tielens , a.g.g.m . , 2002b , a&a , 382 , 222 molster , f.j . , waters , l.b.f.m . , tielens , a.g.g.m .
et al . , 2002c , a&a , 382 , 241 monet , d.g . , levine , s.e . ,
casian , b. , et al .
, 2003 , aj , 125 , 984 omont , a. , loup , c. , forveille , t. , et al . , 1993 , a&a , 267 , 515 ossenkopf , v. , henning , th . & mathis , j.s . , 1992 ,
a&a , 261 , 567 parthasarathy , m. , pottasch , s.r . , 1986 ,
a&a , 154l , 16 pgouri , b. , 1988 , a&a , 194 , 335 reed , c. , 2001 , jrasc , 95 , 268 rieke , g.h . & lebofsky , m.j .
, 1985 , apj , 288 , 618 sahai , r. , & trauger , j. t.1998 , aj , 116 , 1357 sahai , r. , 2002 , astr.sp.sc.lib .
265 , 53 sahai , r. , 2004 , in asymetrical planetary nebulae iii , asp conf .
313 , m. meixner , j. kastner , b. balick & n. soker eds . , p.141 sahai , r. , et al . , 2006 , in preparation schlegel , d.j . ,
finkbeiner , d.p . ,
& davis , m. , 1998 , apj , 500 , 525 schmidt - kaler , th . , 1982 , in landolt - brnstein : numerical data and functional relationships in science and technology , vol .
k. schaifers & h. h. voigt ( berlin : springer - verlag ) seaquist , e. , yao , l. , dunne , l. and cameron , h. 2004 , mnras , 349 , 1428 sivarani , t. , parthasarathy , m. , garca - lario , p. , manchado , a. , 2001 , in _ post - agb objects as a phase of stellar evolution _
eds . , r. szczerba & s. k. grny , astrophysics & space science library , vol .
265 , p. 301
smith , r.g .
, robinson , g. , hyland , a.r . , carpenter , g.l . , 1994 , mnras , 271 , 481 stock , j. , nassau , j.j . ,
stephenson , c.b .
, 1960 , _ luminous stars in the northern milky way ii _ ( hamburger sternwarte and warner and swasey obs .
: hamburg - bergedorf ) sturm , e. , 2000 , _ how to reduce sws aot01 data in the isap gui _ , iso explanatory library doc .
, + http://www.ipac.caltech.edu/iso/sws/new_recipes/sws01_recipe.ps sylvester , r.j . , kemper , f. , barlow , m.j . , et al .
, 1999 , a&a , 352 , 587 swinyard , b.m . ,
clegg , p.e .
, ade , p.a.r .
, et al . , 1996 , a&a , 315 ,
l43 ueta , t. , meixner , m. , & bobrowsky , m. 2000 , , 528 , 861 waters , l.b.f.m . ,
molster , f.j .
, de jong , t. , et al . , 1996 ,
a&a , 315 , l361 waters , l.b.f.m . & molster , f.j . , 1999 ,
iau symp . , 191 , 209 , ed . t. le bertre , a. lbre & c. waelkens | we present the spectral energy distribution ( sed ) of the pre - planetary nebula , iras 19475 + 3119 ( i19475 ) , from the optical to the far - infrared .
we identify emission features due to crystalline silicates in the iso sws spectra of the star .
we have fitted the sed of i19475 using a 1-d radiative transfer code , and find that a shell with inner and outer radii of 8.8@xmath0 and 4.4 @xmath1 cm , and dust temperatures ranging from about 94k to 46k provide the best fit .
the mass of this shell is @xmath21[34 @xmath3 g@xmath4/@xmath5m)][@xmath6/200 ] , where @xmath7(100@xmath8 m ) is the 100@xmath8 m dust mass absorption coefficient ( per unit dust mass ) , and @xmath6 is the gas - to - dust ratio . in agreement with results from optical imaging and millimeter - wave observations of co emission of i19475
, our model fits support an r@xmath9 density law for its dust shell , with important implications for the interaction process between the fast collimated post - agb winds and the dense agb envelopes which results in the observed shapes of ppns and pns .
we find that the observed jcmt flux at sub - millimeter wavelengths ( 850 ) is a factor @xmath102 larger than our model flux , suggesting the presence of large dust grains in the dust shell of i19475 which are not accounted for by our adopted standard mrn grain size distribution . |
accurate stellar radii are important for the study of cepheid mass , pulsation and distance .
direct radius measurements of bright , nearby cepheids allows for comparison to radii found by indirect and/or theoretical methods such as numerical models @xcite , the infrared flux method @xcite and surface brightness relations @xcite .
these methods are easily applied to distant cepheids including those in nearby galaxies @xcite .
each of these indirect methods results in period - radius and period - radius - mass relations which yield different radii , and different masses , at very small and very large periods . from
directly measured radii we may make comparisons with these relations . since there will always be cepheids too small or too faint for direct measurement ,
the comparison between these indirect measurements and relations is crucial for the radius estimation of ever more distant cepheids . at its current magnitude limit ( m@xmath6 @xmath75 ) and longest baseline ( 38 meters )
the angular diameter of four cepheids are measurable with the npoi : @xmath2 cephei , @xmath1 geminorum , @xmath3 aquilae and @xmath0 ursae minoris ( hereafter polaris ) . in this paper
we present mean angular diameters and compare linear radii for all four cepheids with those in the literature and with published period - radius , period - mass - radius , and period - mass relations . even though the sample is small , these four cepheids span an interesting range in pulsation period and characteristics .
polaris and @xmath1 gem were observed over the course of two years .
polaris was observed on 10 nights from september to november 1997 , while @xmath1 gem was observed for four nights : 12 october 1998 , and 20 , 23 and 24 february 1999 .
the detailed observing strategy and data reduction techniques for obtaining mean angular diameters at the npoi are described in @xcite .
@xcite present the specific observations and data reduction of @xmath2 cep and @xmath3 aql .
comparisons between the reduction method employed by @xcite and that used in this work are made at the end of this section .
briefly , as described in @xcite squared - visibilities are measured in each of 10 spectral channels , spaced evenly in wavenumber , ranging from 649 nm to 849 nm .
a uniform - disk model is fit to the visibility data from which a uniform - disk diameter is derived .
the uniform - disk diameters of polaris and @xmath1 gem are found to be 3.14 @xmath4 0.02 mas and 1.48 @xmath4 0.08 mas respectively . as reported in @xcite the uniform - disk diameters for @xmath2
cep and @xmath3 aql are 1.46 @xmath4 0.02 mas and 1.65 @xmath4 0.04 mas respectively .
figure 1 shows visibility data for the npoi s longest baseline ( east - west ) for each of the four cepheids .
the data shown in figure 1 ( a ) - ( d ) are for the night listed in each .
the mean uniform - disk diameter for each is the overall mean diameter determined for that cepheid .
although limb - darkening of evolved stars has been directly observed with the npoi @xcite , those stars are three times larger than the cepheids in this study . at the spatial frequencies currently available to the npoi the visibility differences between limb - darkened and uniform - disks for such small stars
is less than the scatter in the data . until the availability of longer baselines , limb - darkened diameters , @xmath8 , can be derived from uniform - disk diameters using a multiplicative conversion factor .
this conversion factor is a single quadratic coefficient from @xcite interpolated for the cepheid s average specific gravity ( @xmath9 ) , average effective temperature , and for the mean central wavelength of the npoi bandwidth ( 740 nm ) .
the bright star catalogue @xcite categorizes both @xmath1 gem and polaris as spectral type f7ib . for this spectral type @xcite give a @xmath9 of 1.71 and an effective temperature of 6000 k. using these values and the technique described in @xcite we derive a limb - darkened conversion factor ( ratio of limb - darkened diameter to uniform - disk diameter ) of 1.046 for both cepheids at 740 nm
the uncertainty in this conversion factor is estimated to be on the order of 0.5% @xcite even for zeta gem , whose spectral type is very uncertain . with this derived limb - darkening coefficient
we find a limb - darkened diameter of 3.28 @xmath4 0.02 mas for polaris and 1.55 @xmath4 0.09 mas for @xmath1 gem . using this method in @xcite resulted in a limb - darkened diameter of 1.52 @xmath4 0.02 mas for @xmath2 cep and 1.65 @xmath4 0.04 mas for @xmath3 aql .
@xcite use a different reduction method for the calibration of the raw visibility data for these two cepheids ( as well as two non - variable `` check - stars '' ) .
in addition limb - darkened diameters are fit directly to the squared - visibility data without first calculating uniform - disk diameters .
@xcite measure a limb - darkened angular diameter of 1.520 @xmath4 0.014 mas for @xmath2 cep and 1.69 @xmath4 0.04 mas for @xmath3 aql . for
the non - variable star @xmath10 lac , @xcite derive a limb - darkened diameter of 1.909 @xmath4 0.011 mas while @xcite finds 1.92 @xmath4 0.02 mas . since these two different reduction methods produced diameters equal within the errors there is strong confidence in the robustness of the final results .
the diameters for @xmath2 cep and @xmath3 aql used throughout the rest of this work are those of @xcite .
finally , each of the four cepheids is part of a multiple system .
if the npoi should detect light from more than one star , the visibilities measured will be depressed depending upon the position angle and separation of the system .
if not taken into account this variation will have the effect of changing the model diameter that best fits the observed data .
fortunately , each of the companions is either several magnitudes fainter than the cepheid being observed ( placing it well below the npoi s detection threshold ) , at a large enough separation ( @xmath1118 arcseconds ) that it is outside the npoi s photometric field of view , or both .
for example , the companion to @xmath3 aql is 4.6 magnitudes fainter @xcite while at the same time being substantially bluer ( spectral type a1v compared to f6ib - g4ib for @xmath3 aql ) . since the visibilities from only the ten reddest channels are used to fit diameters , further chances of contamination by the companion are reduced .
there is therefore no indication that there has been contamination of the measured diameters for any of the cepheids due to stellar companions . where there is a measured trigonometric parallax , @xmath12 , the distance , @xmath13 , to the cepheid is the reciprocal of @xmath12 , and the linear radius is simply : @xmath14 all four cepheids have parallaxes measured by hipparcos @xcite , while @xmath2 cep and @xmath3 aql have additional parallaxes measured at the u.s .
naval observatory flagstaff station ( h. harris , 1999 , personal communication ) .
for these two cepheids , the distance and linear radius , are derived from the weighted mean of the two measured parallaxes : 3.60 @xmath4 0.53 mas for @xmath2 cep and 2.62 @xmath4 0.74 mas for @xmath3 aql .
these linear radii are nearly model - independent ; what dependence there is enters from the conversion between uniform - disk and limb - darkened angular diameters , and as previously noted , is estimated to be at the level of @xmath70.5% of the mean radius . when the @xmath770 meter baseline at the npoi becomes operational spatial frequencies of @xmath7400 cycles per arcsecond will be accessible and at that time limb - darkened angular diameters for these stars can be measured directly .
lrcccll polaris & 3.9729 & 3.28 @xmath4 0.02 & 7.56 @xmath4 0.48 & ... & 132@xmath15 & 46@xmath16 + @xmath2 cep & 5.3663 & 1.52 @xmath4 0.01 & 3.32
@xmath4 0.58 & 5.0 @xmath4 1.3 & 278@xmath17 & 45@xmath18 + @xmath3 aql & 7.1766 & 1.69 @xmath4 0.04 & 2.78 @xmath4 0.91 & 2.3 @xmath4 1.3 & 382@xmath19 & 69@xmath20 + @xmath1 gem & 10.1507 & 1.55 @xmath4 0.09 & 2.79 @xmath4 0.81 & ... & 358@xmath21 & 60@xmath22 + table 1 lists the four cepheids and includes the npoi limb - darkened angular diameter , the hipparcos parallax , @xmath23 , usno parallax , @xmath24 , the distance ( found from the weighted mean parallax for @xmath2 cep and @xmath3 aql ) , and the npoi s direct linear radius , @xmath25 .
the most common method for estimating cepheid radii and distances is the baade - wesselink , @xcite or surface brightness method @xcite . this method ,
of which there are several variations , relies upon observations of color and radial velocity changes @xcite . for
@xmath1 gem , @xcite derive a radius of 65 @xmath4 12 r@xmath5 while @xcite derives a radius of 68 @xmath4 3 r@xmath5 .
the average difference between these and the npoi result ( table 1 ) is 10% whereas the error of @xmath25 towards higher values is 42% .
the difference between the percent error towards lower and higher values arises from unequal error bars for the distance and linear radius in columns 6 and 7 of table 1 .
similarly for @xmath2 cep , @xcite and @xcite derive a radius of 41 @xmath4 2 and 37 @xmath4 4 r@xmath5 respectively .
the average percent difference between @xmath25 and these is 12% which is slightly smaller than the percent error toward lower values of the npoi radius ( 14% ) .
given the uncertainties in @xmath25 , the measured radii for both these cepheids are consistent with values in the literature .
@xcite derive a radius of 55 @xmath4 4 r@xmath5 for @xmath3 aql , while @xcite calculate 53 @xmath4 5 r@xmath5 .
the average percent difference with @xmath25 is 24% , slightly larger than the percent error towards lower values of the npoi radius ( 22% ) .
@xcite , however , use high - resolution infrared spectroscopy to find a radius of 63 @xmath4 6 r@xmath5 for @xmath3 aql , almost 10% larger than those found using optical spectroscopy .
the photospheric lines in the high - resolution infrared spectra show asymmetries and line splitting which are interpreted to be pulsationally driven shock waves in the atmosphere . the larger radius results from the new interpretation of these spectra and projection factors derived from them specifically for the degree of limb - darkening expected in the infrared .
the difference between the diameter derived from the ir spectra and the npoi diameter ( table 1 ) is only 9% .
table 2 lists these previously published radii for three of the cepheids in this paper ( not including polaris ) .
the method which shows the least agreement with the observations reported here is the cors method , a variation of the surface brightness technique which is different in its mathematical computation @xcite .
table 2 shows that for these three cepheids the cors method produces radii consistently larger than those produced by other optical surface brightness methods .
lllccccc @xmath2 cep & 45@xmath18 & 41 @xmath4 2 & 37 @xmath4 4 & 53 @xmath4 3 & ... & ... + @xmath3 aql & 69@xmath20 & 55 @xmath4 4 & 53 @xmath4 5 & 57 @xmath4 3 & 62 @xmath4 6 & ... + @xmath1 gem & 60@xmath22 & 65 @xmath4 12 & ... & 86 @xmath4 4 & ... & 68 @xmath4 3 + for polaris , it has been observed that the amplitude of the photometric and radial velocity variations has decreased steadily @xcite and although there is indication that this decrease has stopped , the amplitudes of these variations are currently at the level of only 0.032 mag and @xmath71.7 km s@xmath26 @xcite .
a surface brightness analysis based on such small variations is impractical , the radius change would be on the order of 0.2 r@xmath5 , representing an angular diameter change less than 0.5% . as a result ,
there are no published radius estimates with which we may make a comparison .
for an evaluation of the accuracy of the npoi linear radius for polaris we make a comparison in the following section to various published period - radius relations derived from both theory and the application of surface brightness methods to large samples of cepheids .
once a star is identified as a cepheid , the pulsation period is the one quantity that is always known . period - radius relations ( hereafter p - r ) are therefore powerful tools for determining the radius of even the most distant cepheid . typically , p - r relations are of the form : @xmath27 where @xmath28 is the radius in units of solar radii , @xmath29 is the period in days , and @xmath30 and @xmath31 are determined through observation of cepheids for which the radius can be estimated
. different methods of determining cepheid radii have in the past tended to yield somewhat different p - r relations @xcite .
we present here a few representative methods from the literature .
table 3 lists the derived @xmath30 and @xmath31 coefficients for each method .
lrrcc theory & 1.188 @xmath4 0.008 & 0.655 @xmath4 0.006 & 47.6 @xmath4 1.0 & 1 + opt .
sb & 1.146 @xmath4 0.025 & 0.680 @xmath4 0.017 & 45 @xmath4 3 & 2 + ir sb & 1.070 @xmath4 0.027 & 0.751 @xmath4 0.026 & 43 @xmath4 3 & 3 + @xcite calculate theoretical p - r relations using full - amplitude , nonlinear convective models for a variety of metallicities and stellar masses .
the coefficients for a metallicity representing galactic cepheids ( z = 0.02 ) are given in table 3 .
@xcite use the baade - wesselink ( or surface brightness ) technique employing v and v - r photometry , with calibration by @xcite , to derive the radii of 116 cepheids in both the galaxy and magellanic clouds .
they find no evidence for a difference between galactic and magellanic relations and so calculate a single relation for both .
in addition , @xcite find an intrinsic width to their p - r relation of @xmath32 which allows for radii consistent with @xcite over the range of periods in this paper .
while still using the surface brightness technique for estimating cepheid diameters , @xcite find k and j - k ( as well as v - k ) photometry yield more accurate results than optical photometry , due to minimal effects of gravity and microturbulence on infrared fluxes . for periods
less than 11.8 days , @xcite derive smaller radii than the other two methods . for periods less than 48 days , which is the range within which all of the cepheids in this paper are found , the theoretical relation predicts a larger radius ( @xmath77% ) than that found from surface brightness relations .
these representative p - r relations are shown in figure 2 along with the four cepheid radii measured at the npoi as given in table 1 .
figure 2 shows that although polaris has the highest radius precision ( owing to the most precise parallax ) its radius is larger than predicted by any of the published p - r relations .
the difference between the theoretical curve of @xcite , the relation which predicts the largest radius , and the measured radius of polaris is 2.6@xmath33 given the uncertainty of only 3@xmath34 in @xmath25 .
even with the intrinsic width of the @xcite p - r relation , the observed radius for polaris is too large .
this problem is resolved if polaris is a first overtone pulsator rather than a fundamental mode pulsator .
since the ratio of the first overtone period to the fundamental period is 0.71 , an overtone cepheid plotted on a p - r diagram using the log of the first overtone period instead of the fundamental period , will result in a radius larger than what the p - r relation would predict @xcite .
the overtone nature of polaris has been noted recently in the literature @xcite .
@xcite first used hipparcos parallaxes and visual magnitudes for 220 cepheids to calculate the cepheid period - luminosity zero - point .
they find that the zero - point derived from polaris alone is brighter than that produced by the rest of the sample if polaris is considered as a fundamental pulsator .
the arrow and diamond in figure 2 places the measured radius of polaris relative to its fundamental period : 3.9729/0.71 = 5.5957 days .
column 4 of table 3 lists the radius predicted by each p - r relation for a p = 5.5957 day cepheid .
the percent difference between the observed radius of polaris ( 46 @xmath4 3 r@xmath5 ) and the fundamental period radius predicted by @xcite and @xcite is 3% and 2% respectively .
the excellent agreement of polaris with these radii derived from published p - r relations is evidence of the overtone nature of polaris suggested by @xcite .
in the same way that p - r relations yield radii from a known period , there are period - mass and period - radius - mass relations from which masses can be derived .
the period - radius - mass relation of @xcite , @xmath35 when solved for mass yields @xmath36 from this equation @xcite calculates masses for a sample of 101 cepheids using the radii of @xcite .
these masses , which we refer to as @xmath37 in this paper , are used by @xcite to derive the radius independent wesselink period - mass ( hereafter p - m ) relation : @xmath38 using equation 3 and the radii in table 1 we calculate masses , @xmath39 , for the four cepheids in our sample .
these masses are presented in table 4 with masses for the four cepheids from @xcite .
lrcccc polaris & 5.3@xmath40 & ... & 5.3 @xmath4 0.9 & 5.8 @xmath4 0.5 & 4.5 @xmath4 2.0 + @xmath2 cep & 5.0@xmath41 & 3.9 & 5.2 @xmath4 0.9 & 5.7 @xmath4 0.5 & 4.5 @xmath4 2.0 + @xmath3 aql & 10@xmath42 & 5.1 & 5.7 @xmath4 0.9 & 6.2 @xmath4 0.5 & 5.0 @xmath4 2.0 + @xmath1 gem & 4.2@xmath43 & 4.7 & 6.6 @xmath4 0.9 & 7.0 @xmath4 0.5 & 6.0 @xmath4 2.0 + as can be seen from equation 3 and table 4 , large uncertainties in the linear radius propagate into even larger uncertainties in the mass , resulting in errors of almost 50% for @xmath2 cep and 100% towards larger masses for @xmath3 aql .
compare this to polaris which has the highest radius precision and thus a mass uncertainty of only 17% .
this percent error is slightly less than the 20% `` accidental '' error @xcite estimates for @xmath37 based on errors in their radii of @xmath77 - 8% .
masses from two theoretical p - m relations are also listed in table 4 : evolution mass and pulsation mass @xcite .
the evolution mass , @xmath44 , is calculated from stellar evolution theory : @xmath45 whereas the pulsation mass , @xmath46 , is calculated from period - effective temperature relations : @xmath47 figure 3 shows the three p - m relations relative to @xmath39 . as with the estimated error of @xmath37 the uncertainty in @xmath44
is estimated to be on the order of 15 - 20% , while uncertainties in the effective temperature scale for cepheids are capable of bringing the pulsation mass relation into agreement with the evolution mass @xcite .
direct diameter measurements will be able to address the uncertainties in this scale once one can directly measure cepheid diameter variations and thus calculate effective temperature as a function of pulsation phase to a precision limited only by the photometry . within the present uncertainties
then , all three p - m relations are consistent with each other and the npoi masses .
the cepheid @xmath3 aql , with a period @xmath77 days , possesses a bump in the descending phase of its radial velocity and light curves .
radii for bump cepheids ( 6 days @xmath48 p @xmath48 20 days ) can be calculated using the bump phase and pulsation theory @xcite .
there is a discrepancy , however , between the radius derived in this manner and the radius ( and hence mass ) derived from surface brightness techniques @xcite . for a cepheid such as @xmath3 aql
, @xmath37 will be larger than the predicted bump mass .
since the linear radius observed by the npoi for @xmath3 aql is larger than derived by @xcite , @xmath39 will be larger than @xmath37 and therefore in even worse agreement with the mass from the predicted bump phase method ( although the error in @xmath39 is quite large owing to the large parallax uncertainty ) .
recent theoretical models of bump cepheids by @xcite using full amplitude , nonlinear , convective models ( with no convective core overshooting ) result in a cepheid mass of 6.9 @xmath4 0.9 m@xmath5 for a period of 11.2 days .
this result agrees well with the wesselink mass of equation ( 4 ) for a 11.2 day cepheid which yields m@xmath49 = 6.85 m@xmath5 .
although @xcite calculate the mass , and thus make their comparison to @xcite for a p = 11.2 day cepheid , the implication is that the disagreement between observation and theory has been resolved in the matter of bump cepheid masses with a resolution in favor of the larger wesselink mass , and thus closer to @xmath39 in table 4 .
the agreement between polaris and @xmath2 cep and the curve for @xmath50 is less significant than it would at first seem since the method for calculating @xmath39 is based upon the same theory as that used for calculating @xmath50 . in the same way that it is desirable to compare the model - independent cepheid radii of table 1 to radii derived through indirect methods ,
the same should be done for mass . a number of cepheids are located in binary systems ( including all four cepheids in this study as noted earlier ) .
already @xcite have used spectroscopy to calculate the masses of five cepheids in binaries : u aql , s mus , v350 sgr , y car , su cyg .
using ground based optical and satellite ultraviolet spectra the mass ratio of the two members of the binary were found .
inferring the mass of the companion based on the spectral type yielded the mass of the cepheid .
these five cepheids are plotted in figure 3 as open circles and show good agreement with the p - m relations and the npoi observations . with long enough baselines ,
optical interferometry will be able to image the orbits of cepheids as has already been done for binaries of non - cepheids @xcite . in conjunction with radial velocities from spectroscopy ,
all the orbital elements of the system , including the mass and distance , will be directly determined and independent of all models .
at the present time only four cepheids have had their diameters measured with the npoi . over the next two years
as the longest baseline available increases from 38 meters to 440 meters , the number of cepheids resolvable and the precision of their measurements will increase by at least a factor of five .
figure 2 shows that at the present the measured linear radii are consistent with each of the published p - r relations . with the increased precision of the angular diameter measurements , and the increased precision of parallax observations ( also undertaken at the usno ) it will be possible to differentiate between the various p - r relations which are seen to diverge in figure 2 for periods shorter than 30 days
this is precisely the range of periods for which cepheids observable by the npoi are located .
optical long baseline interferometry has successfully measured the mean angular diameters of the four brightest cepheid variables in the northern sky .
these angular diameters coupled with trigonometric parallaxes have produced virtually model - independent linear radii .
these radii are compared to radii in the literature which have been derived from a variety of baade - wesselink , or surface brightness , methods .
the agreement between the direct radius determinations presented here and published indirect radius estimates is quite good .
the differences are @xmath710% , better than the error in @xmath25 which is on the order of 20 - 40% . for @xmath3 aql ,
the derived linear radius is in marginal agreement with the optical surface brightness results but it is in very good agreement with the radius estimated from infrared spectroscopy by @xcite . for polaris the radius precision ( 6% ) is high enough that we are able to confirm its overtone nature . at a period of 3.97 days a radius of 46 @xmath4 3 r@xmath5
is inconsistent with the published p - r relations of @xcite and @xcite . only as an overtone pulsator with a fundamental period of 5.59 days
is polaris in agreement with these p - r relations , confirming the findings of @xcite using completely independent means . at a period of 5.59 days
, the wesselink mass of polaris is found to be in excellent agreement with the period - mass relation of @xcite .
the authors would like to thank hugh harris for providing results from the usno variable star parallax study .
thanks are also due to siobahn morgan for bringing the polaris overtone pulsator discussion to our attention .
don fernie was also enormously helpful in providing the period and rate of change for polaris .
benson , j. a. , hutter , d. j. , elias ii , n. m. , bowers , p. f. , johnston , k. j. , hajian , a. r. , armstrong , j. t. , mozurkewich , d. , pauls , t. a. , rickard , l. j , hummel , c. a. , white , n. m. , black , d. , & denison , c. s. 1997 , , 114 , 1221 nordgren , t. e. , germain , m. e. , benson , j. a. , mozurkewich , d. , sudol , j. j. , elias ii , n. m. , hajian , a. r. , white , n. m. , hutter , d. j. , johnston , k. j. , gauss , f. s. , armstrong , j. t. , pauls , t. a. , & rickard , l. j 1999 , , 118 , 3032 perryman , m. a. c. , lindegren , l. , kovalevsky , j. , hg , e. , bastian , u. , bernacca , p. l. , creze , m. , donati , f. , grenon , m. , grewing , m. , van leeuwen , f. , van der marelh , h. , mignard , f. , murray , c. a. , le poole , r. s. , schrijver , h. , turon , c. , arenou , f. , froeschle , m. , & petersen , c. s. 1997 , , 323 , l49 | we present mean angular diameters for two cepheid variables , @xmath0 ursae minoris and @xmath1 geminorum , determined with the navy prototype optical interferometer ( npoi ) .
we present linear radii for these cepheids and two additional cepheids , @xmath2 cephei and @xmath3 aquilae , previously observed at the npoi .
we find the limb - darkened angular diameter of @xmath0 ursae minoris and of @xmath1 geminorum to be 3.28 @xmath4 0.02 and 1.55 @xmath4 0.09 milliarcseconds respectively .
using trigonometric parallaxes , we find the linear radii of @xmath0 ursae minoris , @xmath1 geminorum , @xmath2 cephei and @xmath3 aquilae to be 46 ( @xmath4 3 ) r@xmath5 , 60 ( + 25 , -14 ) r@xmath5 , 45 ( + 8 , -6 ) r@xmath5 , and 69 ( + 28 , -15 ) r@xmath5 respectively .
we compare the pulsation periods and linear radii of this sample of cepheids , which range in period from three to 11 days , to theoretical and empirical period - radius and period - radius - mass relations found in the literature .
we find that the observed diameter of @xmath0 ursae minoris is in excellent agreement with the predicted diameter as determined from both surface brightness techniques and theory only if @xmath0 ursae minoris is a first overtone pulsator . |
the _ chandra _ x - ray observatory was launched on july 23 1999 , carrying on board a revolutionary high resolution mirror assembly , with a point spread function of 0.5 arcsec ( half power radius ) over the broad 0.1 to 10 kev band ( van speybroeck , et al .
this , together with the aspect camera which at the moment provides attitude solutions with errors of the order of 1 - 2 arcsec arcsec and the source positions should be better than 1 arcsec ] , allows the study of spatial extent of x - ray sources on similar scales , i.e. smaller or similar to the size of a l@xmath2 galaxy at any redshift ; and gives x - ray source positions at least as good as 23 arcsec , immediately allowing the unambiguous identification of the optical counterparts of faint x - ray sources .
consequently , the determination of the source redshifts via optical spectroscopy becomes highly efficient .
the improvement provided by _
chandra _ is especially significant in the hard ( 210 kev ) x - ray band .
surveys of the hard x - ray sky have been performed in the past by asca and bepposax ( ueda et al .
1998 , della ceca et al . 1999 ,
fiore et al
. 2000a , giommi et al .
2000 , comastri et al .
2000 ) . however , the large error boxes ( 1 - 2 arcmin ) limited the optical identification process to classes of objects with low surface density , at a given optical magnitude limit . as a result ,
most of the identified sources are emission line agn ( fiore et al .
1999 , akiyama et al .
2000 , and la franca et al . in preparation ) .
about 30 % of the bepposax hellas survey sources studied spectroscopically down to r=20.5 have escaped a secure identification ( fiore et al .
2000b , la franca et al . in preparation ) , although many normal galaxies and stars have been observed in these error - boxes .
_ chandra _ s unprecedented capabilities make identifications unambiguous , and open up the possibility of searching for and studying classes of sources not previously recognized as strong hard x - ray emitters , and of assessing their contribution to the hard x - ray cosmic background ( xrb ; e.g. griffiths & padovani 1990 )
. in particular , it will be possible for the first time to begin studying normal galaxies at @xmath3 , as well as possible `` minority '' hard x - ray source populations ( kim & elvis 1999 ) .
we have started a pilot project of spectroscopic identification of _ chandra _ sources in two medium deep fields that were visible from la silla in january 2000 , with the aim of verifying the feasibility of such studies with 4 m class telescopes .
the results on 10 x - ray sources are very encouraging and are described in the following .
the _ chandra _ x - ray observatory consists of four pairs of concentric wolter i mirrors reflecting 0.1 - 10 kev x - rays into one of the four focal plane detectors : acis - i , acis - s , hrc - i or hrc - s ( weisskopf , odell & vanspeybroeck 1996 ) .
all the results presented in the following were obtained with the @xmath4 acis - i ccd instrument .
table 1 gives the log of the chandra observations .
one field is centered on the @xmath1=0.6 quasar pks0312@xmath577 , the other is located @xmath6 away from the radiant point of the leonid meteor shower .
level 2 processed data ( fabbiano et al . in preparation ) were obtained from the chandra public archive .
data were cleaned and analyzed using the ciao software ( release v1.1 , elvis et al . in preparation ) .
time intervals with large background rate were removed , as well as hot pixels and bad columns . only the standard event grades ( 0 , 2 , 3 , 4 and 6 ) were used .
sources were detected in images accumulated in the 210 kev band ( channels 136 - 680 ) .
robust sliding - cell algorithms were used to locate the sources .
we used both the @xmath7 program available in the cxc data analysis package ciao ( dobrzycki et al .
1999 ) and a variation of the detect routine included in the ximage package ( giommi et al .
the method consists in first convolving the x - ray image with a wavelet function , in order to smooth the image and increase contrast , and then running a standard sliding cell detection method on the smoothed image .
the quality of each detection was checked interactively .
final net counts are estimated from the original ( un smoothed ) image , to preserve poisson statistics .
a binning factor of 2 ( i.e. pixels of @xmath8 1 arcsec ) and source box sizes of 6 arcsec ( offaxis angle @xmath9 ) , 10 arcsec ( @xmath10 offaxis angle @xmath11 ) and 30 arcsec ( offaxis angle @xmath12 ) maximize the signal to noise ratio , given the local background , psf and source intensity .
the background is calculated using source - free boxes near the sources .
this is usually very low in the 210 kev band , i.e. 0.150.3 counts per 6 arcsec detection box side , per 10 ks . for this study
only sources with more than 10 , 15 and 40 counts in the 6 arcsec , 10 arcsec and 30 arcsec detection cells respectively , are considered .
given a conservative background of 0.6 , 1.5 and 15 counts in the detection cells this corresponds to a poisson probability for a background fluctuation of @xmath13 in all cases .
the total number of detection cells can be estimated in about 7800 6@xmath14 cells plus 5000 10@xmath14 cells plus a few tens of 30@xmath14 cells in each observation ( 4 acis - i chips ) .
therefore , the total number of spurious sources in the two fields ( 0.14 deg@xmath15 ) at the chosen detection thresholds is absolutely negligible ( 10@xmath16 ) .
thirteen sources detected in the 210 kev images passed these thresholds .
all but one are also detected in the softer 0.52 kev band ( the non detected source has about 6 counts in the soft band ) .
count rates were corrected for the psf and the telescope vignetting calculated at 4.5 kev ( 210 kev band ) and 1.5 kev ( 0.52 kev band ) using figures 3.8 and 3.10 of the axaf observatory guide ( 1997 ) .
the correction is about 25 % at an off - axis angle of 12 arcmin and 10 % at 6 arcmin .
fluxes in the 210 kev band were computed assuming a count rate to flux conversion factor of @xmath17 per count s@xmath18 .
fluxes in the 0.52 kev were computed assuming factors of @xmath19 and @xmath20per count s@xmath18 for leonid anti - rad and the pks0312@xmath5770 fields respectively .
these conversion factors are appropriate for power law models with @xmath21 ( 210 kev ) and @xmath22 ( 0.52 kev ) respectively , assuming the galactic absorbing column densities quoted in table 1 .
table 2 gives for the thirteen detected sources the x - ray position and off - axis angle , the 210 kev and 0.52 kev counts ( corrected for the psf and the vignetting ) and the corresponding fluxes . [ cols="<,^,^,^,^ " , ]
optical counterparts were found in the usno catalogue down to r@xmath23 in ten cases , in the dss2 e(r ) images in two cases ( lar1 and lar3 ) , and on efosc2 r band image in the last case ( p4 ) .
optical counterparts were found for 11 sources with a typical displacement of @xmath24 arcsec . in one case ( lar4 )
the displacement is of 4 .
we consider the identification secure because a ) lar4 is observed by acis at an offaxis angle of 12 arcmin , where the psf is highly asymmetric and broader than 10 arcsec , making the x - ray position more uncertain , and b ) the optical source is a bright , broad line quasar , the surface density of which is so low that the probability to find one by chance at @xmath25 from an x - ray source is negligible .
in fact , the probability of finding _ any _ r@xmath2622 galaxy in a @xmath25 radius error - box by chance is @xmath27 0.04 , while the same probability for an agn is at least ten times smaller . therefore the number of agn misidentifications in our sample is @xmath28 .
the number of expected galaxy misidentifications at r@xmath2620 is @xmath27 0.2 , while that of r@xmath2622 is @xmath27 0.5 . in one case ( lar2 )
no optical counterpart brighter than r @xmath8 20 is present in the x ray error box .
however a r=18 galaxy , with an optical extension of about 3 arcsec , is present at 6 arcsec from the x - ray position .
therefore the galaxy outskirts touch the x - ray errorbox .
we conservatively consider this identification uncertain .
we obtained long slit spectra of ten counterparts , using efosc2 at the eso 3.6 m telescope on january 3 - 6 2000 .
spectra were obtained using grism n.6 in the 3800 - 8100 @xmath29 range and a slit 2 arcsec wide , corresponding to a resolution of 26 .
the complete set of flux - calibrated spectra is shown in figure 1 .
r magnitude , redshift , optical and x - ray luminosity , and a classification of the optical spectrum are given in table 2 .
classification of narrow line objects is done using standard line ratio diagnostics ( e.g. osterbrook 1981 , tresse et al .
1996 ) , in particular the [ oiii]@xmath30/h@xmath31 , [ sii]@xmath32/h@xmath33 and [ oiii]@xmath30/[oii]@xmath34 line ratios .
we find six broad line quasars at redshift between 0.42 and 1.19 and 210 kev x
ray luminosities in the range @xmath35 erg s@xmath18 .
the four remaining source identifications are as follows : lar5 is identified with a bright @xmath1=0.016 starburst galaxy with an x - ray luminosity of @xmath36 erg s@xmath18 , similar to that of other starburst galaxies of similar optical luminosity ( ptak et al .
1999 ) . in lar6
, although [ oiii ] and @xmath37 are detected , the spectrum is rather noisy and it is not clear if @xmath37 has broad wings .
optical classification is therefore uncertain .
the 210 kev luminosity of @xmath38 erg s@xmath18 and the x - ray softness ratio ( see below ) suggest that lar6 is an obscured agn .
p3 is identified with a normal galaxy without strong emission lines ( only a weak [ oii ] line with an equivalent width of @xmath39 is apparent in the spectrum ) .
the x - ray and optical luminosities are @xmath40 erg s@xmath18 and @xmath41 .
the galaxy continuum is very red , suggesting an early type galaxy .
the calcium break is 0.44@xmath42 , consistent with little or no dilution from non stellar light .
a faint h@xmath43 is detected in absorption ( equivalent width of @xmath44 ) .
the spectrum of p4 is rather noisy because the source lies only 9 arcsec from a 10 mag star , which strongly enhances the background .
we identify p4 with an agn at z=0.683 thanks to a rather strong [ oii ] and weaker [ neiii ] and mgii emission lines .
the mgii to [ oii ] intensity ratio of @xmath45 is much smaller than in broad line quasars and similar to those of type 2 agn ( woltjer 1990 ) .
the 210 kev luminosity of @xmath46 erg s@xmath18 and the x - ray softness ratio ( see below ) suggest that p4 is a moderately obscured quasar .
we have searched the nvss , iras faint source catalog , clusters of galaxies ( aco , abell , corwin & orowin 1989 ) , normal galaxies , stars and agn catalogs , finding only one coincidence : lar5 is identified with an apm galaxy .
the high spatial resolution of the _ chandra _ telescope allows for the first time a comparison of x - ray and optical images at similar , arcsec , resolution .
we have studied the spatial extent of the x - ray sources by comparing their count profiles with the _ chandra _ psf , taking into account its dependence on the off - axis angle ( as calibrated on the ground ) .
the spatial extent is smaller than a few arcsec in all cases and therefore consistent with their being point sources .
conversely , at least three of the optical counterparts are extended galaxies ( `` gal . '' in table 2 ) .
figures 2a , b show the x - ray ( 210 kev ) contours overlaid on the optical r band images of the @xmath1=0.158 galaxy ( p3 ) and the @xmath1=0.016 starburst galaxy ( lar5 ) obtained with efosc2 . galaxy p3 is bulge dominated , the size of the bulge being 23 arcsec .
also galaxy lar5 has a bright bulge , of similar size .
these sizes are similar to or smaller than the chandra psf at the offaxis angle where the sources were detected in acis - i ( 2.6 arcmin and 6 arcmin respectively ) .
both galaxies are extended ( up to 1020 arcsec ) .
if the x - ray emission were connected to the outer parts of the galaxy , it would have been resolved by _
chandra_. both x - ray sources are centered on the galaxy nuclei .
the x - ray source in lar5 appears slightly elongated in a direction perpendicular to the galaxy major axis , as seen in several nearby starburst galaxies ( e.g. fabbiano et al .
1992 , dahlem et al .
however , this elongation may also be due to the degradation of the _ chandra _ psf at the offaxis - angle of this source ( 6 arcmin ) .
figure 3 shows the x - ray to optical flux ratios of the thirteen _ chandra _ sources as a function of the 210 kev flux , and compares them with those of local agn and of the agn found in the bepposax hellas survey .
this figure shows that the x - ray to optical ratio of most _ chandra _ sources is similar to that of local and hellas agn .
their optical magnitude is bright enough to allow redshift determination with 4 m or 8 m class telescopes . at fainter fluxes figure 3
shows the sources recently optically identified by mushotzky et al .
r band fluxes have been computed assuming r - i=0.3 . the flux ratio of about 60% of the mushotzky et al .
_ chandra _ sources is within the range of values covered by brighter agn ( i.e. log@xmath47 from @xmath52 to 1 ) .
the remaining 40 % show an x - ray to optical ratio higher than that of the brighter _ chandra _ sources presented here and of the local agn ( see figure 3 ) . as discussed by mushotzky et al .
( 2000 ) the nature of these optically faint sources is mysterious .
many of them have optical magnitude fainter than r@xmath824 which make difficult to obtain a precise redshift through optical spectroscopy .
the @xmath1=0.158 normal galaxy has an x - ray to optical ratio of 0.36 , much higher than the typical value of nearby normal galaxies ( see section 5.2 ) . for most of the _ chandra _ sources
the total number of detected counts is @xmath27100 , preventing the use of proper spectral fitting procedures to study their spectrum .
the broad band x - ray spectral properties can only be investigated using count ratios .
if the spectrum is parameterized as an absorbed power law and the redshift of the source absorber is known , the column density can be evaluated .
following fiore et al .
( 1998 ) , we assume here that the x - ray absorber redshift coincides with the optical redshift .
figure 4 shows the softness ratio ( s@xmath5h)/(s+h ) ( s=0.52 kev band , h=210 kev band ) of the ten spectroscopically identified _
chandra _ sources as a function of the redshift .
errors include counting statistics only .
dashed lines show the expectation of power law models with @xmath48 absorbed by varying column densities _ in the source frame_. two of the _ chandra _ sources ( p3 and lar6 ) have a softness ratio inconsistent with that expected by an intrinsically unobscured power law at better than the 90% confidence level . in particular
, lar6 is likely to be obscured by a column of @xmath49 @xmath50 .
it is worth noting that the column densities implied by figure 4 are probably lower limits to the columns toward the nuclear hard x - ray source , because highly obscured agn usually have strong soft components ( e.g. schachter et al .
1998 , maiolino et al .
1998 , della ceca et al . 1999 ,
fiore et al .
2000a , b ) . to quantify this possible
underestimate we simulated acis - i observations of three highly obscured type 2 agn ( ngc1068 , @xmath51 @xmath50 ; the circinus galaxy @xmath52 @xmath50 ; and ngc6240 , @xmath53 @xmath50 ) , based on asca and bepposax results ( matt et al .
1997 , 1999 , iwasawa & comastri 1998 , vignati et al . 1999 ) .
the ( s@xmath5h)/(s+h ) observed by acis - i would have been 0.88 , @xmath50.08 and 0.57 respectively .
the results are strongly dependent from the sources spectral shape and redshift . at @xmath1=1 the simulated hardness ratios are 0.53 , @xmath50.30 and @xmath50.05 respectively .
we also note that relatively low values of ( s@xmath5h)/(s+h ) can be produced by intrinsically flat and unobscured power laws ( for example the p3 hardness ratio of 0.16 would correspond to @xmath54 ) .
efficient x - ray follow - up of relatively bright x ray sources with xmm will provide a relatively accurate measure of the intrinsic spectrum through proper spectral fitting allowing to remove the ambiguity between an intrinsically flat spectrum and a spectrum flattened by absorption .
one of the main goals of hard x - ray surveys is to investigate the origin of the hard x - ray cosmic background . the most popular model explains the xrb in terms of a mixture of obscured and unobscured agn ( setti & woltjer 1989 ) following a strong cosmological evolution .
the relatively small number of sources in the present survey ( only thirteen ) , implies a @xmath55 statistical error on their number density .
this limits the strength of the constraints we are able to put on agn synthesis models for the xrb . however we believe it is useful to begin addressing this issue here . the main problem when comparing model predictions and observations is that the flux limit of _ count rate _ limited surveys depends on the actual intrinsic spectrum of the sources ( zamorani et al .
harder sources would generally produce less counts ( because of the decrease of the effective area toward the higher energies ) , and therefore they would pass a detection threshold only if their flux is higher than that of softer sources with similar count rate . in order to quantify this effect we have folded a heavily absorbed ( 10@xmath56 @xmath50 ) power law spectrum ( @xmath57 = 0.8 ) with the _ chandra _ sensitivity , and computed the flux limit corresponding to the count rate threshold .
this turns out to be a factor 56 higher than that of an unabsorbed power law .
taking into account the column density distribution predicted by the comastri et al .
( 1995 ) synthesis model and the _ chandra _ sensitivity , the predicted fraction of obscured ( log@xmath58 22 ) sources is 2030 % .
the softness ratio results ( figure 4 ) suggest that at least 3 out of 10 optically identified sources may be absorbed by column densities equal or higher than @xmath59 @xmath50 .
furthermore , one of the three unidentified objects has a hard spectrum , as can be judged from the hard to soft x ray flux ratio of table 2 .
we therefore conclude that the present observations are consistent with the predictions of agn synthesis models for the xrb .
a surprising result from our pilot study is the detection of a luminous x - ray source in an otherwise normal galaxy ( p3 ) .
the x - ray luminosities of @xmath40 erg s@xmath18 , @xmath60 erg s@xmath18 are about a factor 30 higher than those expected on the basis of the optical luminosity ( @xmath61 ) for both spiral ( fabbiano et al .
1992 ) and elliptical / s0 ( eskridge , fabbiano & kim 1995 ; pellegrini 1999 ) galaxies . a few optically `` dull '' galaxies with strong x - ray emission have been reported in the past ( e.g. 3c264 and ngc4156 , elvis et al .
1981 ; j2310@xmath5437 , tananbaum et al . 1997 ; griffiths et al .
1996 ) and more recently in a deep _ chandra _ observation ( mushotzky et al .
hard power law tails have been also discovered in a few nearby elliptical galaxies ( allen , di matteo & fabian 2000 ) . the presence of relatively strong x ray emission in objects with no evidence of activity in the optical spectrum is still not well understood .
one possibility would be a large contribution from a beamed non - thermal component in the x - ray band ( elvis et al .
1981 , tananbaum et al .
1997 , worrall et al . 1999 ) as for bl lacertae objects .
if p3 hosts a bl lac then the non thermal featureless optical continuum would reduce the height of the calcium break from a typical value of 0.5 found in elliptical and s0 galaxies to @xmath62 ( stocke et al .
this threshold has been raised to about 0.4 by marcha and browne ( 1995 ) to take into account the spread of galaxy luminosity and sizes .
the p3 calcium break of [email protected] does not allow to rule out the presence of a bl lacertae object , even if it would be a rather extreme member of its class . if this is the case , the radio flux predicted from the observed x - ray flux would be in the 0.130 mjy range ( padovani & giommi 1996 ) .
the pmn ( griffith et al . 1991 ) limit of about 50 mjy is not useful to settle this issue .
an obscured agn could also provide a viable explanation .
indeed the p3 hardness ratio implies a substantial column for a typical agn x ray continuum .
the optical emission lines could also be completely hidden except for a weak [ oii ] emission feature .
it is worth remarking that examples of x ray obscured agn with neither blr nor nlr already exist ( e.g. ngc4945 , marconi et al .
2000 ; ngc6240 , vignati et al . 1999 and references therein ) , even if in these cases emission lines related to starburst emission are present .
finally the presence of a low radiative efficiency accretion flow ( adaf ) might also be tenable .
the putative central black hole mass estimated from the observed b luminosity of the galaxy s bulge following the magorrian et al .
( 1998 ) relation is about 4@xmath6410@xmath65 m@xmath66 ( this is actually an upper limit , as we have assumed for the bulge luminosity that of the entire galaxy , see fig .
if the x ray emission comes from the nucleus , and it is a sizeable fraction of the bolometric luminosity , its eddington ratio is of the order of 10@xmath67 , consistent with the adaf regime ( narayan et al .
an estimate of the 8.4 ghz flux has been obtained assuming the adaf spectral models calculated by di matteo et al .
( 2000 ) as well as the x ray to radio flux ratios of a few adaf candidates observed at the vla ( di matteo et al .
1999 ) rescaled to the p3 x ray flux . in both cases
the maximum expected radio emission is of the order of a few mjy and thus well below the present limit .
the adaf contribution to the optical uv light strongly depends on the adopted model ( see figure 2 in di matteo et al .
the p3 hardness ratio is consistent with a very flat power law ( @xmath68 [email protected] ) , and a wind model is therefore to be preferred . if this is the case , the adaf flux in the optical uv would be much fainter ( at least two order of magnitude ) than that of the host galaxy , again consistent with the present findings .
we have carried out a pilot program to study the faint hard x - ray source population using the revolutionary capabilities of the _ chandra _ satellite .
we identified ten 210 kev selected sources from two 4-chip , medium deep _ chandra _ fields covering about 0.14 deg@xmath15 of sky at fluxes in the range @xmath69 , a factor of 3 fainter than previous asca and bepposax surveys .
recently , mushotzky et al . ( 2000 ) reported detection of faint x - ray sources from a single 1chip field ( 0.0175 deg@xmath15 ) at fluxes @xmath70 .
our results fill the gap between the shallow bepposax and asca surveys and the deep _ chandra _ field .
almost all sources have an optical counterpart within 2 arcsec .
optical spectra allow us to measure their redshifts , and assess their optical classification .
we find six broad line quasars , two emission line agn ( lar6 and p4 ) , one starburst galaxy ( lar5 ) , and one apparently normal galaxy at @xmath1=0.158 ( p3 ) .
lar6 and p4 are likely to be obscured in x - ray by a column densities of @xmath71 @xmath50 and @xmath72 @xmath50 respectively .
the x - ray source in the @xmath1=0.158 normal galaxy p3 may be covered by a column density of about @xmath59 @xmath50 too .
the spatial extension of all x - ray sources is smaller than a few arcsec in all cases , and it is roughly consistent with the _ chandra _ psf .
the x - ray sources associated with the @xmath1=0.016 starburst galaxy lar5 and the @xmath1=0.158 normal galaxy p3 appear coincident with the galaxy nuclei .
the x - ray to optical luminosity ratio of p3 is higher by a factor of at least 30 than those of normal galaxies , while it is similar to those of agn .
the high x - ray luminosity and the lack of optical emission lines suggest an agn in which either continuum beaming is important or emission lines are absorbed or not efficiently produced . in any case , objects like p3 would be missed or ignored in optical surveys or in x - ray surveys with large error boxes .
it is only thanks to the new revolutionary capabilities of _ chandra _ that this kind of sources _ can _ be detected and identified .
based on the asca and bepposax surveys at fluxes @xmath73 and on our first _
_ identifications , which push the flux limit down to @xmath74 , the hard x - ray sky appears populated by a large fraction of broad line agn ( about 50% ) , by a mixture of intermediate agn ( type 1.8 - 2.0 and composite starburst / agn ) and , most intriguingly , also by x - ray luminous apparently normal galaxies .
these populations span ranges of x - ray and optical properties wider than previously thought .
the results presented in this paper are made possible by the successful effort of the entire _ chandra _ team .
in particular we thank the xrt and acis teams for building and calibrating the high resolution mirror and the ccd camera , the cxc team , and in particular a. fruscione , for the quick data reduction and archiving .
we thank p. giommi , s. molendi , g. fabbiano , e. giallongo , m. mignoli , l. stella , m. vietri and h. tananbaum for useful discussions and l.a .
antonelli for help in the eso observation preparation .
we also thank the referee , gianni zamorani , for his detailed and constructive comments which improved the quality of the paper .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this work is partly supported by the italian space agency , contract ars9975 and by the ministry for university and research ( murst ) under grant cofin980232 .
m.e . , f.n , and m.c .
acknowledge support from nasa contract nas839073 . | we report optical spectroscopic identifications of 10 hard ( 210 kev ) x - ray selected sources discovered by _
chandra_. the x - ray flux of the sources ranges between 1.5 and 25 @xmath0 , the lower value being 3 times fainter than in previous bepposax and asca surveys .
their r band magnitudes are in the range 12.822 .
six of the _ chandra _ sources are broad line quasars with redshifts between 0.42 and 1.19 , while the optical identification of the remaining four is quite varied : two are x - ray obscured , emission line agn at @xmath1=0.272 and @xmath1=0.683 , one is a starburst galaxy at @xmath1=0.016 and one , most unusually , is an apparently normal galaxy at @xmath1=0.158 .
these findings confirm and extend down to fainter x ray fluxes the bepposax results , in providing samples with a wide range of x - ray and optical properties .
the ratio between the soft x - ray and the optical luminosity of the @xmath1=0.158 galaxy is a factor at least 30 higher than that of normal galaxies , and similar to those of agn .
the high x
ray luminosity and the lack of optical emission lines suggest an agn in which either continuum beaming dominates , or emission lines are obscured or not efficiently produced .
x - rays : galaxies , cosmology : diffuse radiation , galaxies : quasars , galaxies : starburst |
folded states of globular proteins , which are evolved ( slightly ) branched heteropolymers made from twenty amino acids , are roughly spherical and are nearly maximally compact with high packing densities @xcite . despite achieving high packing densities in the folded states ,
globular proteins tolerate large volume substitutions while retaining the native fold @xcite .
this is explained in a couple of interesting theoretical studies @xcite , which demonstrated that there is sufficient free volume in the folded state to accommodate mutations .
collectively these and related studies show that folded proteins are compact .
when they unfold , which can be achieved upon addition of high concentrations of denaturants ( or applying a mechanical force ) , they swell adopting expanded conformations . the radius of gyration ( @xmath0 ) of a folded globular protein is well described by the flory law with @xmath8 @xmath9 @xcite , whereas in the swollen state @xmath10 , where @xmath11 is an effective monomer size and the flory exponent @xmath12 @xcite .
thus , viewed from this perspective we could surmise that proteins must undergo a coil - to - globule transition @xcite , a process that is reminiscent of the well characterized equilibrium collapse transition in homopolymers @xcite .
the latter is driven by the interplay of conformational entropy and intra - polymer interaction energy resulting in the collapsed globular state .
the swollen state is realized in good solvents ( interaction between monomer and solvents is favorable ) whereas in the collapsed state monomer - monomer interactions are preferred .
the coil - to - globule transition in large homopolymers is akin to a phase transition .
the temperature at which the interactions between the monomers roughly balance monomer - solvent energetics is the @xmath13 temperature . by analogy
, we may identify high ( low ) denaturant concentrations with good ( poor ) solvent for proteins . despite the expected similarities between the equilibrium collapse transition in homopolymers and the compaction of proteins
, it is still debated whether the unfolded states of proteins are more compact under folding conditions compared to their states at high denaturant concentrations .
if polypeptide chain compaction is universal , is collapse in proteins essentially the same phenomenon as in hompolymer collapse or is it driven by a different mechanism @xcite ?
surprisingly , this fundamental question in the protein folding field has not been answered satisfactorily @xcite . in order to explain the plausible difficulties in quantifying the extent of compaction ,
let us consider a protein , which undergoes an apparent two - state transition from an unfolded ( swollen ) to a folded ( compact ) state as the denaturant concentration ( @xmath14 ) is decreased . at the concentration , @xmath15 , the populations of the folded and unfolded states are equal .
a vexing question , which has been difficult to unambiguously answer in experiments , is : what is the size , @xmath0 , of the unfolded state under folding conditions ( @xmath16 ) ?
small angle x - ray scattering ( saxs ) experiments on some proteins show practically no change in the unfolded @xmath0 as @xmath14 is changed @xcite . on the other hand , from experiments based on single molecule fluorescence resonance energy transfer ( smfret ) it has been concluded that the size of the unfolded state is more compact below @xmath15 compared to its value at high @xmath14 @xcite .
the so - called smfret - saxs controversy has remained unresolved .
resolving this apparent controversy is not only important in our understanding of the physics of protein folding but also has implications for the physical basis of the evolution of natural sequences .
the difficulties in describing the collapse of unfolded states as @xmath14 is lowered could be attributed to the following reasons .
( 1 ) following de gennes @xcite , homopolymer collapse can be pictured as formation of a large number of blobs driven by local interactions between monomers on the scale of the blob size .
coarsening of blobs results in the equilibrium globule formation with the number of maximally compact conformations whose number scales exponentially with the number of monomers .
other scenarios resulting in the formation of fractal globules enroute to the formation of equilibrium maximally collapsed structures have also been proposed @xcite .
the globule formation is driven by non - specific interactions between the monomers or the blobs .
regardless of how the equilibrium globule is reached it is clear that it is largely stabilized by local interactions , because contacts between monomers that are distant along the sequence are entropically unfavorable .
in contrast , even in high denaturant concentrations proteins could have residual structure , which likely becomes prominent at @xmath16 . at low @xmath14
there are specific favorable interactions between residues separated by a few or several residues along the sequence .
as their strength grows , with respect to the entropic forces , the specific interactions may favor compaction in a manner different from the way non - specific local interactions induce homopolymer collapse .
in other words , the dominant native - like contacts also drive compaction of unfolded states of proteins .
( 2 ) a consequence of the impact of the native - like contacts ( local and non - local ) on collapse of unfolded states is that specific energetic considerations dictate protein compaction resulting in the formation of minimum energy compact structures ( mecs ) @xcite .
the number of mecs , which are not fully native , is small , scaling as @xmath17 with @xmath2 being the number of amino acid residues .
therefore , below @xmath15 their contributions to @xmath0 have to be carefully dissected , which is more easily done in single molecule experiments than in ensemble measurements such as saxs .
( 3 ) single domain proteins are finite - sized with @xmath2 rarely exceeding @xmath18 200 .
most of those studied experimentally have @xmath19 .
thus , the extent of change in @xmath0 of the unfolded states is predicted to be small , requiring high precision experiments to quantify the changes in @xmath0 as @xmath14 is changed .
for example , in a recent study@xcite , we showed that in pdz2 domain the change in @xmath0 of the unfolded states as the denaturant concentration changes from 6 m guanidine chloride to 0 m is only about 8% .
recent experiments have also established that changes in @xmath0 in helical proteins are not too large @xcite . in homopolymers
there are only two possible states , coil and globule , with a transition between the two occurring at @xmath20 . on the other hand , even in proteins
that fold in a two - state manner one can conceive of at least three states ( we ignore intermediates here ) : ( i ) the unfolded state @xmath21 at high @xmath14 ; ( ii ) the compact but unfolded state @xmath22 , which could possibly exist below @xmath15 ; ( iii ) the native state .
do the sizes of @xmath21 and @xmath22 differ ?
this question requires a clear answer as it impacts our understanding of how proteins fold , because the very nature of the unfolded states of proteins plays a key role in determining protein foldability @xcite .
given the flexibility of proteins ( persistence length on the order of @xmath23 nm ) , we expect that the size of the extended polypeptide chain must gradually decrease as the solvent quality is altered .
experiments on a number of proteins show that this is the case @xcite .
however , in some saxs experiments the theoretical expectation that @xmath24 for one protein was not borne out @xcite , precipitating a more general question : are chemically denatured proteins compact at low @xmath14 ?
the absence of collapse is not compatible with inferences based on smfret @xcite and theory @xcite . here
, we create a theory to not only resolve the smfret - saxs controversy but also provide a quantitative description of how the propensity to be compact is encoded in the native topology .
the theory , based on polymer physics concepts , includes specific attractive interactions ( mimicking interactions accounting for native contacts in the protein data bank ( pdb ) ) and a two - body excluded volume repulsion . by construction
the model does not have a native state .
there are only two states ( analogues of @xmath21 and @xmath22 ) in the model .
the formation of @xmath22 is driven by the contact map of the folded state . thus
, chain compaction is driven in much the same way as in homopolymers , altered only by specific interactions that differentiate proteins from homopolymers .
our theory predicts quantitatively how the extent of compaction ( collapsibility ) is determined by the strength and the number of the native contacts and their locations along the chain .
we use a large representative selection of proteins from the pdb to establish that collapsibility is an inherent characteristic of evolved protein sequences .
a major outcome of this work is that @xmath6-sheet proteins are far more collapsible than structures dominated by @xmath7-helices .
our theory suggests that there is an evolutionary pressure on proteins for being compact as a pre - requisite for kinetic foldability , as we predicted over twenty years ago @xcite .
we come to the inevitable conclusion that the unfolded state of proteins must be compact under native conditions , and the mechanism of polypeptide chain compaction has similarities as well as differences to collapse in homopolymers . as a by - product of this work
we also establish that certain non - coding rna molecules must undergo compaction prior to folding as their folded structures are stabilized predominantly by long - range tertiary contacts .
we start with an edwards hamiltonian for a polymer chain @xcite : @xmath25 where @xmath26 is the position of the monomer @xmath27 , @xmath28 the monomer size , and @xmath2 is the number of monomers . the first term in eq.([hamiltonian ] ) accounts for chain connectivity , and the second term represents volume interactions and favorable interactions between select monomers given by @xmath29 , @xmath30 the first term in eq.([hamiltv ] ) accounts for the homopolymer ( non - specific ) two - body interactions .
it is well established in the theory of homopolymers that in good solvents with @xmath31 the polymer swells with @xmath32 @xmath33 . in poor solvents ( @xmath34 )
the polymer undergoes a coil - globule transition with @xmath35 ( @xmath36 ) .
these are the celebrated flory laws . here
, we consider only the excluded volume repulsion case ( @xmath37 ) . the second term in eq.([hamiltv ] )
requires an explanation .
the generic scenario for homopolymer collapse is based on an observation by de gennes , who pictured the collapse process as being driven by the initial formation of blobs that arrange to form a sausage - like structure . at
later stages the globule forms to maximize favorable intra - molecular contacts while simultaneously minimizing surface tension .
compaction in proteins , although sharing many features in common with homopolymer collapse , could be different .
a key difference is that the folded states of almost all proteins are stabilized by a mixture of local contacts ( interaction between residues separated by less than say @xmath18 8 but greater than 3 residues ) as well as non - local ( @xmath38 8 residues ) contacts .
note that the demarcation using 8 between local and non - local contacts is arbitrary , and is not germane to the present argument .
these specific interactions also dominate the enthalpy of formation of the compact , non - native state @xmath22 , playing an important role in its stability .
therefore , the topology of the folded state could dictate collapsibility ( the extent to which the @xmath21 state becomes compact as the denaturant concentration is lowered ) of a given protein . in combination with the finite size of single domain proteins ( @xmath2 @xmath18 200 ) ,
the extent of protein collapse could be small . in order to assess chain compaction under native conditions
we should consider the second term in eq.([hamiltv ] ) .
we note in passing ( with discussion to follow ) that a number of studies have considered the effect of crosslinks on the shape of polymer chains @xcite .
polymers with crosslinks have served as models for polymer gels and rubber elasticity@xcite . in these studies
the contacts were either random , leading to the random loop model @xcite , or explicit averages over the probability of realizing such contacts were made @xcite , as may be appropriate in modeling gels .
these studies inevitably predict a coil - to - globule phase transition as the number of crosslinks increases .
in contrast to models with random crosslinks , in our theory attraction is only between specific residues , described by the second term in eq.([hamiltv ] ) , where the sum is over the set of interactions ( native contacts ) involving pairs @xmath39 .
we use the contact map of the protein ( extracted from the pdb structure ) in order to assign the specific interactions ( their total number being @xmath40 ) .
the contact is assigned to any two residues @xmath41 and @xmath42 if the distance between their @xmath43 atoms in the pdb entry is less than @xmath44 and @xmath45 .
we use gaussian potentials in order to have short ( but finite ) range attractive interactions . for the excluded volume repulsion ,
this range is on the order of the size of the monomer , @xmath46 . for the specific attraction ,
the range is the average distance in the pdb entry between @xmath43 atoms forming a contact ( averaged across a selection of proteins from the pbd ) .
we obtain @xmath47 . by changing the value of @xmath48 , and hence the strength of attraction
, there is a transition between the extended and compact states .
decreasing @xmath48 is analogous to chemically denaturing proteins , although the connection is not precise . at high denaturant concentrations ( @xmath49 , good solvent ) the excluded volume repulsion ( first term in eq.([hamiltv ] ) ) dominates the attraction , while at low @xmath14 ( high @xmath48 , poor solvent ) the attractive interactions are important .
the point where attraction balances repulsion is the @xmath13-point , and the value of @xmath50 .
although reserved for the coil - to - globule transition in the limit of @xmath51 in homopolymers , we will use the same notation ( @xmath13-point ) here . in our model , at the @xmath52-point , the chain behaves like an ideal chain . to describe the globular state
, a three - body repulsion needs to be added to the hamiltonian ( eq.([hamiltv ] ) ) , but we focus on the region between the extended coil and the @xmath13-point because our interest is to access only the collapsibility of proteins
. if @xmath53 is very large then significant chain compaction would only occur at very low ( @xmath54 ) denaturant concentrations , implying low propensity to collapse .
conversely , small @xmath53 implies ease of collapsibility .
note that the ground state ( @xmath55 ) of the hamiltonian in eq.([hamiltv ] ) is a collapsed chain whose @xmath0 is on the order of the monomer size .
in other words , a stable native state does not exist for the model described in eq.([hamiltv ] ) .
thus , we define protein collapse as the propensity of the polypeptide chain to reach the @xmath13-point as measured by the @xmath53 value , and use the changes in the radius of gyration @xmath56 as a measure of the extent of compaction . * assessing collapsibility : * for our model , which encodes protein topology without favoring the folded state , we calculate @xmath57 using the edwards - singh ( es ) method @xcite .
the es method is a variational type calculation that represents the exact hamiltonian by a gaussian chain , whose effective monomer size is determined as follows .
consider a virtual chain without excluded volume interactions , with the radius of gyration @xmath58 @xcite , described by the hamiltonian , @xmath59 where the monomer size in the virtual hamiltonian is @xmath60 .
we split the deviation @xmath61 between the virtual chain hamiltonian and the real hamiltonian as , @xmath62 where @xmath63 the radius of gyration is @xmath64 , with the average being , @xmath65 where @xmath66 denotes the average over @xmath67 . assuming that the deviation @xmath61 is small
, we calculate the average to first order in @xmath61 .
the result is , @xmath68 and the radius of gyration is @xmath69 ds,\ ] ] if we choose the effective monomer size @xmath60 in @xmath67 such that the first order correction ( second and third terms on the right hand side of eq.([rg ] ) ) vanishes , then the size of the chain is , @xmath58 .
this is an estimate to the exact @xmath57 , and is an approximation as we have neglected @xmath70 and higher powers of @xmath61 .
thus , in the es theory , the optimal value of @xmath60 from eq .
( [ rg ] ) satisfies , @xmath71 ds=0.\ ] ] since @xmath72 , the above equation can be written as @xmath73 ds=-\frac{1}{n } \int\limits_0^n [ \langle{\bf r}^2(s)\rangle_v \langle\mathcal{w}_2\rangle_v -\langle{\bf r}^2(s)\mathcal{w}_2\rangle_v ] ds.\ ] ] evaluation of the @xmath74 term yields , @xmath75 \\ \nonumber & = & \left(\frac{1}{a_0 ^ 2}-\frac{1}{a^2}\right ) a^2 \left(\frac{a^2 n}{6 } \right)+ \langle{\bf r}^2(s)\rangle_v \langle\mathcal{w}_1\rangle_v\end{aligned}\ ] ] with the help of eq.([aa ] ) and eq.([first ] ) we obtain the following self - consistent expression for @xmath60 , @xmath76ds } { \frac{a^2}{n}\int_{0}^n ds \
\langle{\bf r}^2(s)\rangle_v}.\ ] ] calculating the averages in fourier space , where @xmath77 , @xmath78 , and @xmath79 ) , we obtain @xmath80}{n^4 } } { \left(\sum\limits_{n=1}^{n}\frac{1-\cos[n \pi(s - s')/n]}{n^2}+\frac{3 \pi^2a_0 ^ 2 } { 2a^2 n}\right)^{5/2}}\\ \nonumber & - & \kappa\frac { ( \frac{3}{2})^{5/2}(\frac{\pi}{2})^{3/2}}{(a^2)^{5/2}n^{3/2}\left(\sum\limits_{n=1}^n \frac{1}{n^2}\right ) } \sum\limits_{\{s_i , s_j\ } } \frac{\sum\limits_{n=1}^{n}\frac{1-\cos[n \pi(s_i - s_j)/n]}{n^4 } } { \left(\sum\limits_{n=1}^{n}\frac{1-\cos[n \pi(s_i - s_j)/n]}{n^2}+\frac{3 \pi^2\sigma^2 } { 2a^2 n}\right)^{5/2}}.\\ \nonumber\end{aligned}\ ] ] the best estimate of the effective monomer size @xmath60 can be obtained by numerically solving eq.([self ] ) provided the contact map is known .
a bound for the actual size of the chain is @xmath81 . because we are interested only in the collapsibility of proteins we use the definition of the @xmath13-point to assess the condition for protein compaction instead of solving the complicated eq.([self ] ) numerically .
the volume interactions are on the right hand side of eq.([self ] ) . at the @xmath13-point ,
the @xmath82-term should exactly balance the @xmath48-term .
since at the @xmath13-point the chain is ideal with @xmath83 , we can substitute this value for @xmath60 in the sums in the denominators of the @xmath82- and @xmath48-terms . by equating the two
, we obtain an expression for @xmath84 .
thus , from eq.([self ] ) , the specific interaction strength at which two - body repulsion ( @xmath82-term ) equals two - body attraction ( @xmath48-term ) is : @xmath85}{n^4 } } { \left(\sum\limits_{n=1}^{n}\frac{1-\cos[n \pi(s - s')/n]}{n^2}+\frac{3 \pi^2 } { 2 n}\right)^{5/2 } } } { \sum\limits_{\{s_i , s_j\ } } \frac{\sum\limits_{n=1}^{n}\frac{1-\cos[n \pi(s_i - s_j)/n]}{n^4 } } { \left(\sum\limits_{n=1}^{n}\frac{1-\cos[n \pi(s_i - s_j)/n]}{n^2}+\frac{3 \pi^2\sigma^2 } { 2a_0 ^ 2 n}\right)^{5/2}}}. \label{ktheta}\ ] ] the numerator in eq.([ktheta ] ) is a consequence of chain connectivity and the denominator encodes protein topology through the contact map , determining the extent to which the sizes in @xmath21 and @xmath22 states change as @xmath14 becomes less than @xmath15 .
the numerical value of @xmath84 is a measure of collapsibility .
* native topology determines collapsibility : * the central result in eq.([ktheta ] ) can be used to quantitatively predict the extent to which a given protein has a propensity to collapse .
we used a list of proteins with low mutual sequence identity selected from the protein data bank pdbselect @xcite , and calculated @xmath84 using eq.([ktheta ] ) for these proteins . in all we considered 2306 proteins . for each contact @xmath86
, the energetic contribution due to interaction between @xmath87 and @xmath88 is @xmath89 according to eq.([hamiltv ] ) .
thus , @xmath90 is the average strength ( in units of @xmath91 ) of a contact at the @xmath13-point .
if @xmath84 , calculated using eq.([ktheta ] ) , is too large then the extent of polypeptide chain collapse is expected to be small .
it is worth noting that the theory can not be used to determine the stability of the folded state , because in the hamiltonian there are only two states , @xmath21 ( @xmath48=0 in eq.([hamiltv ] ) ) and @xmath22 ( @xmath92 ) .
the strength of contacts in real proteins ( excluding possibly salt bridges ) is typically on the order of few @xmath93 in the absence of denaturants .
this is the upper bound for the contact strength any theory should predict , as adding denaturant only decreases the strength .
if @xmath94 is unrealistically high ( tens of @xmath93 ) then the attractive interactions of the protein would be too weak to counteract the excluded volume repulsion even at zero denaturant concentration , resulting in negligible difference in @xmath0 between the @xmath21 and @xmath22 states .
fig.([fig : pdbhist ] ) shows a two - dimensional histogram of the pdbselect proteins in the @xmath95 plane . for the majority of small proteins ( less than 150 residues )
the value of @xmath84 is less than 3 @xmath93 , indicating that the unfolded states of all of these proteins should become compact at @xmath96 .
that collapse must occur , as predicted by our theory and established previously in lattice @xcite , and off - lattice models of proteins@xcite , does not necessarily imply that it can be easily detected in standard scattering experiments , because the changes could be small requiring high precision experiments ( see below ) .
* weight function of a contact : * for a given @xmath2 , the criterion for collapsibility in eq.([ktheta ] ) depends on the architecture of the proteins explicitly represented in the denominator through the contact map .
analysis of the weight function of a contact , defined below , provides a quantitative measure of how a specific contact influences protein compaction .
some contacts may facilitate collapse to a greater extent than others , depending on the location of the pair of residues in the polypeptide chain . in this case , the same number of native contacts @xmath40 in the protein of the same length @xmath2 might yield a lower ( easier collapse ) or higher ( harder collapse ) value of @xmath94 . in order to determine the relative importance of the contacts with respect to collapse , we consider the contribution of the contact between residues @xmath87 and @xmath88 in the denominator of eq.([ktheta ] ) , @xmath97}{n^4 } } { \left(\sum\limits_{n=1}^{n}\frac{1-\cos[n \pi(i - j)/n]}{n^2}+\frac{3 \pi^2\sigma^2 } { 2a_0 ^ 2 n}\right)^{5/2}}. \label{weighting}\ ] ] a plot of @xmath98 in fig.([fig : weighting ] ) for different values of the chain length @xmath2 shows that the weight depends on the distance between the residues along the chain .
contacts between neighboring residues have negligible weight , and there is a maximum in @xmath98 at @xmath99 ( for @xmath100 ) , almost independent of the protein length .
the maximum is at a higher value for proteins with @xmath101 residues .
the figure further shows that longer range contacts make greater contribution to chain compaction than short range contacts .
the results in fig.([fig : weighting ] ) imply that proteins with a large fraction of non - local contacts are more easily collapsible than those dominated by short range contacts , which we elaborate further below . * maximum and minimum collapsibility boundaries : * using @xmath98 in eq.([weighting ] )
, we can design protein sequences to optimize for `` collapsibility '' . to design a `` maximally collapsible '' protein , for fixed @xmath2 and number of native contacts @xmath40
, we assign each of the @xmath40 contacts one by one to the pair @xmath102 with a maximal @xmath103 among the available pairs with the criterion that @xmath104 .
such an assignment necessarily implies that the artificially designed contact map will not correspond to any known protein .
similarly , we can design an artificial contact map by selecting @xmath102 pairs with minimal @xmath103 till all the @xmath40 are fully assigned .
such a map , which will be dominated by local contacts , are minimally collapsible structures .
the white lines in fig.([fig : pdbhist ] ) show @xmath94 of chains of length @xmath2 with @xmath105 contacts distributed in ways to maximize or minimize collapsibility .
we estimated @xmath106 from the fit of the whole pdbselect set .
since the lines are calculated for @xmath40 from the fit over the entire set , and not from @xmath40 for every protein , there are proteins below the minimal and above the maximal curve in fig.([fig : pdbhist ] ) . for a given protein , with @xmath2 and @xmath40 defined by its pdb structure , @xmath94 for all possible arrangements of native contacts
is largely in between the maximally and minimally collapsible lines in fig.([fig : pdbhist ] ) .
the majority of proteins in our set are closer to the maximal collapsible curves , suggesting that the unfolded proteins have evolved to be compact under native folding conditions .
this theoretical prediction is in accord with our earlier studies which suggested that foldability is determined by both collapse and folding transitions @xcite , and more recently supported by experiments @xcite .
* @xmath6-sheet rather than @xmath7-helical proteins undergo larger compaction : * the weight function @xmath107 ( eq.([weighting ] ) and fig.([fig : weighting ] ) ) suggests that contacts in @xmath7-helices ( @xmath108 ) only make a small contribution to collapse .
contacts corresponding to the maximum of @xmath107 at @xmath109 are typically found in loops and long antiparallel @xmath6-sheets .
fig.([fig : alphabeta ] ) shows a set of proteins with high @xmath7-helix ( @xmath110 ) and a set with high content of @xmath6-sheets ( @xmath111 ) .
the values of @xmath94 for the two sets are very distinct , so they barely overlap .
we find that many of the @xmath7-helical proteins lie on or above the curve of minimal collapsibility while the rest are closer to the maximal collapsibility .
the smaller @xmath6-rich proteins lie on the curve of maximal collapsibility slightly diverging from it as the chain length grows .
these results show that the extent of collapse of proteins that are mostly @xmath112-helical is much less than those with predominantly @xmath6-sheet structures .
a note of caution is in order .
the minimal collapsibility of most @xmath7-helical proteins in the set may be a consequence of some of them being transmembrane proteins , which do not fold in the same manner as globular proteins .
instead , the transmembrane @xmath7-helices are inserted into the membrane by the translocon , one by one , as they are synthesized .
such proteins would not have the evolutionary pressure to be compact .
* extent of compaction varies depending on the native structure : * our conclusions , summarized in figs.(1 - 3 ) , are based on an approximate theory . to validate the approximations and illustrate the impact of the location of contacts on chain compaction we performed monte carlo simulations of chains of lengths @xmath113 , 72 and 153 using the energy function given by eq.([hamiltv ] ) with harmonic spring chain connectivity .
the proteins are villin headpiece ( @xmath114 36 ) , protein - l ( @xmath114 72 ) , and myoglobin ( @xmath114 153 ) .
the contact maps and cartoon representation of these proteins are in fig.([fig : cmap_pro ] ) .
fig.([fig : ratio ] ) shows that the ratio between end - to - end distance ( @xmath115 ) and radius of gyration ( @xmath0 ) decreases continuously with the interaction strength , indicative of chain compaction .
fig.([fig : ratio ] ) also shows that @xmath116 , which is well known for gaussian chains ( @xmath117 ) and self - avoiding walks ( @xmath118 limit ) for which the ratio @xmath119 .
based on @xmath53 estimates we conclude that protein - l ( @xmath120 ) is more collapsible than villin headpiece and myoglobin ( helical proteins ) .
fig.([fig : rg]a ) shows the variations in the probability distribution of @xmath0 , @xmath121 for protein - l as a function of @xmath48 .
the broader distribution with zero interaction strength corresponds to the extended chain .
we find that @xmath121 becomes narrower as the attractive strength increases
. the continuous shift in to the compact state with gradual increase in the attractive strength shows the compaction of unfolded proteins as the denaturant concentration decreases .
thus , generally @xmath0 of the @xmath22 state is less than that of the @xmath21 state .
the end - to - end distribution , @xmath122 , for different values of interaction strength has been shown in fig.([fig : rg]b ) .
the broad distribution at @xmath118 corresponds to the unfolded protein .
average @xmath115 decreases as attractive strength increases and the distribution becomes narrower .
the results in fig.([fig : rg ] ) show that both @xmath115 , which can be inferred using fret , and @xmath0 ( measurable using saxs ) , are smaller in the @xmath22 state than the @xmath21 state . however , the extent of decrease is greater in @xmath115 than @xmath0 .
* rnas are compact : * there are major differences between how rna and proteins fold @xcite . in contrast to the apparent controversy in proteins , it is well established that rna molecules are compact@xcite at high ion concentrations or at low temperatures .
because our theory relies only on the knowledge of contact map that we used , it assess collapsibility in _ azoarcus _ ribozyme and mmtv pseudoknot to merely illustrate collapsibility of rna .
the @xmath123 values ( green stars in fig.[fig : alphabeta ] ) are close to the lower @xmath6-sheet line , indicating that these molecules must undergo compaction .
this prediction from the theory is fully supported by both equilibrium and time - resolved saxs experiments@xcite on _ azoarcus _ ribozyme . in this case ( @xmath124 ) the changes are so large that even using low resolution experiments collapse is readily observed@xcite .
we should emphasize that the size of different rnas ( for example viral , coding , non - coding ) vary greatly .
for a fixed length , single - stranded viral rnas have evolved to be maximally compact , which is rationalized in terms of the density of branching .
although the sizes of the viral rnas considered in @xcite are much longer than the _ azoarcus _ ribozyme the notion that compaction is determined by the density of branching might be valid even when @xmath125 .
we have shown that polymer chains with specific interactions , like proteins ( but ones without a unique native state ) , become compact as the strength of the specific interaction changes .
a clear implication is that the size of the @xmath126 state should decrease continuously as @xmath14 decreases . in other words ,
the unfolded state under folding conditions is more compact than it is at high denaturant concentrations .
compaction is driven roughly by the same mechanism as the collapse transition in homopolymers in the sense that when the solvent quality is poor ( below @xmath15 ) the size of the unfolded state decreases continuously .
when the set of specific interactions is taken from protein native contacts in the pdb , our theory shows that the values of @xmath94 are in the range expected for interaction between amino acids in proteins .
this implies that collapsibility should be a universal feature of foldable proteins but the extent of compaction varies greatly depending on the architecture in the folded state .
this is manifested in our finding that proteins dominated by @xmath6-sheets are more collapsible compared to those with @xmath7-helical structures . *
magnitude of @xmath123 and plausible route to multi - domain formation : * the scaling of @xmath123 with @xmath2 allows us to provide arguments for the emergence of multi - domain proteins . in eqs.([self ] ) or ( [ ktheta ] ) attractive ( @xmath48- ) and repulsive ( @xmath82- ) terms have the same structure .
the only difference in their scaling with @xmath2 is due to the difference in the sums ( over all the monomers in the repulsive term and over native contacts in the attractive term ) .
double summation over all the monomers gives a factor of @xmath127 to the repulsive term .
the summation over native contacts in the attractive term scales as @xmath40 .
therefore , to compensate for the repulsion , @xmath40 should scale as @xmath127 .
however , for a given protein with a certain length @xmath2 and certain numbers of contacts , it is not clear how the denominator in eq.([ktheta ] ) scales with @xmath2 .
empirically we find @xmath105 dependence across a representative set of sequences scales as @xmath128 with @xmath129 less than 2 ( likely on the order of 1.5 ) .
thus , it follows from eq.([ktheta ] ) that @xmath94 increases without bound as @xmath2 continues to increase . because this is unphysical
, it would imply that proteins whose lengths exceeds a threshold length @xmath130 can not become maximally compact even at @xmath131 .
an instability must ensue when @xmath2 exceeds @xmath130 .
this argument in part explains why single domain proteins are relatively small@xcite .
scaling of @xmath40 as a power law in @xmath128 with @xmath132 means that as protein size grows , the value of @xmath94 will deviate more and more from those found in globular proteins , implying such proteins can not be globally compact under physiologically relevant conditions .
however , such an instability is not a problem because larger proteins typically consist of multiple domains .
thus , if the protein does not show collapse as a whole , the individual domains could fold independently , having lower values of @xmath94 for each domain of the multi - domain protein .
it would be interesting to know if the predicted onset of instability at @xmath130 provides a quantitative way to assess the mechanism of formation of multi - domain proteins .
extension of the theory might yield interesting patterns in the assembly of multi - domain proteins .
for instance , one can quantitatively ascertain if the n - terminal domains of large proteins , which emerge from the ribosome first , have higher collapsibility ( lower @xmath84 ) than c - terminal domains .
* saxs - smfret controversy resolved : * our theory resolves , at least theoretically , the contradictory results using saxs and fret experiments on compaction of small globular proteins
. it has been argued , based predominantly on protein - l ( @xmath133 ) saxs experiments that @xmath0 of @xmath21 and @xmath22 states are virtually the same at denaturant concentrations that are less than @xmath15 @xcite .
this conclusion is not only at variance with saxs experiments on other proteins but also with interpretation of smfret data on a number of proteins .
the present work , surveying over 2300 proteins , shows that the compact state has to exist , engendered by mechanisms that have much in common with homopolymer collapse . for protein - l , the @xmath134 , a very typical value , right on the peak of the heat map in fig.([fig : pdbhist ] ) .
we have previously argued that because the change in @xmath0 between the @xmath21 and @xmath22 states for small proteins is not large , high precision experiments are needed to measure the predicted changes in @xmath0 between @xmath22 and @xmath21 . for protein -
l the change is less than 10%@xcite , making its detection in ensemble experiments very difficult .
similar conclusions were reached in recent experiments @xcite . a clear message from our theory is that , tempting as it may be , one can not draw universal conclusions about polypeptide compaction by performing experiments on just a few proteins .
one has to survey a large number of proteins with varying @xmath2 and native topology to quantitatively assess the extent of compaction .
our theory provides a framework for interpreting the results of such experiments . * random contact maps , local and non - local contacts : * in order to differentiate collapsibility between evolved and random proteins , we created twelve random contact maps keeping the total number of contacts the same as in protein - l ( see fig.([fig : random72 ] ) for examples ) . for each of these pseudo - proteins we calculated @xmath123 using eq.([ktheta ] ) .
we find that for all the random contact maps the @xmath123 values are less than for protein - l , implying that the propensity of the pseudo - proteins to become compact is greater than for the wild type .
this finding is in accord with studies based on homopolymer and heteropolymer collapse with random crosslinks .
these studies showed that the polymer undergoes a collapse transition as the density of crosslinks is increased @xcite .
of particular note is the demonstration by camacho and schanke @xcite , who showed using exact enumeration of random heteropolymers and scaling arrangements that the collapse can be either a first or second order transition depending on the function of hydrophobic residues @xcite .
some time ago abkevich et al .
@xcite showed , using monte carlo simulations of protein - like lattice polymers , that the folding transition in proteins with predominantly non - local contacts was first order like , which is not the case for proteins in which local contacts dominate . in light of this finding , it is interesting to examine how compaction is affected by local and non - local contacts .
we created for @xmath2=72 ( protein - l ) a contact map with 185 ( same number as with wt
protein - l ) , predominantly local contacts ( fig.(8b ) ) .
the values of @xmath94 for these pseudo - proteins is considerably larger than for the wt , implying that proteins dominated by local contacts are minimally collapsible .
we repeated the exercise by creating contact maps with predominantly non - local contacts ( fig.(8c ) ) .
interestingly , @xmath94 values in this case are significantly less than for the wt .
this finding explains why in proteins with varied @xmath135 topology there is a balance between the number of local and non - local contacts .
such a balance is needed to achieve native state stability and speed of folding@xcite with polypeptide compaction playing an integral part @xcite . based on these findings
we conclude that @xmath0 of the unfolded states of proteins dominated by non - local contacts must undergo greater compaction compared to those with that have mostly local contacts . the results in fig .
3 also show that proteins rich in @xmath6-sheet are more collapsible than predominantly @xmath7-helical proteins .
it follows that @xmath6-sheet proteins must have a larger fraction of non - local contacts than proteins rich in @xmath7-helices . in fig .
( 8d ) we plot the distribution of the fraction of non - local contacts for the 2306 proteins .
interestingly , there is a clear separation in the distribution of non - local contacts between @xmath7-helical rich and @xmath6-sheet rich proteins .
the latter have substantial fraction of non - local contacts which readily explains the findings in fig .
( 8c ) and the predictions in fig .
we have created a theory to assess collapsibility of proteins .
the major implications of the theory are the following .
( i ) because single domain proteins are small , the changes in the radius of gyration of the unfolded states as the denaturant concentration is lowered are often small .
thus , it has been difficult to detect the changes in saxs experiments in a couple of proteins raising the question if unfolded polypeptide chains become compact below @xmath15 . here
, we have solved this long - standing problem showing that the unfolded states of single - domain proteins do become compact as the denaturant concentration decreases , sharing much in common with the physical mechanisms governing homopolymer collapse . by adopting concepts from polymer physics , and using the contact maps that reflect the topology of the native states , we established that proteins are collapsible .
( ii ) based on a survey of over two thousand proteins we surmise that there is evolutionary pressure for collapsibility is universal although the extent of collapse can vary greatly , because this ensures that the propensity to aggregate is minimized even if environmental fluctuations under cellular conditions transiently populate unfolded states .
two factors contribute to aggregation .
first , the rate of dimer formation by diffusion controlled reaction would be enhanced if a pair of @xmath21 rather than @xmath22 molecules collided due cellular stress because the contact radius in the former would be greater than in the latter .
second , the fraction of exposed hydrophobic resides in @xmath21 is much greater than in @xmath22 , thus greatly increasing the probability of aggregation .
the second factor is likely to be more important than the first .
consequently , transient population of @xmath22 due to cellular stress minimizes the probability of aggregation .
( iii ) we have also shown that the position of the residues forming the native contact greatly influences the collapsibility of @xmath6 sheet proteins ( containing a number of non - local contacts showing greater compaction than @xmath7 helical proteins , which are typically stabilized by local contacts .
our theory also shows that most rnas may have evolved to be compact in their natural environments .
although the evolutionary pressure to be compact is likely to be substantial for viral rnas @xcite , it is apparent that even non - coding rnas are also likely to be almost maximally compact in their natural environments .
our theory suggests that , to a large extent , collapsibility of rna is similar to proteins with @xmath6-sheet structures .
both classes of biological macromolecules are stabilized by non - local contacts .
interestingly , it has been argued that the need to be compact ( `` compaction selection hypothesis '' @xcite ) could be a major determinant for evolved biopolymers to have minimum energy compact structures as their ground states .
( in units of @xmath93 ) for a set of 2,306 pdb structures as a function of the length @xmath2 of the proteins .
white lines show the @xmath94 at the boundaries for maximally and minimally collapsible proteins ( lower and upper lines respectively ) .
colors give a rough estimate of the number of proteins , which decreases from red to violet.,scaledwidth=90.0% ] ( eq.([weighting ] ) ) of a contact , showing how much a contact between residues @xmath87 and @xmath88 contributes to the compaction of a protein .
the colors are for different @xmath2 values ( shown in the inset ) .
interestingly , the location of the maximum is roughly independent of @xmath2.,scaledwidth=90.0% ] on the secondary structure content of proteins .
we display @xmath94 for @xmath7-rich ( @xmath136 ) and @xmath6-rich ( @xmath111 ) proteins .
proteins that are predominantly @xmath7-helical tend to be close to minimally collapsible ( upper line ) , while @xmath6-rich proteins are close to maximally collapsible curve ( lower curve ) .
the green stars are for rna with the left one at small
n corresponding to the__mouse mammary tumor virus _
_ ( mmtv ) pseudoknot ( n=34 ) and the other is _ azoarcus _ ribozyme ( n=196).,scaledwidth=90.0% ] ) and radius of gyration ( @xmath0 ) for three different chain lengths ( n=36 , 72 and 153 ) .
it decreases continuously with increasing interaction strength .
the presence of contacts between residues decreases @xmath137 from the maximum value attainable for a self - avoiding walk .
, scaledwidth=80.0% ] for different values of interaction strength @xmath138 for protein - l . as @xmath138 increases
, the distribution becomes narrower .
( b ) same as ( a ) except this panel shows end - to - end distribution @xmath122 for different values of attractive strength @xmath138 for protein - l .
the similarity between @xmath122 and @xmath121 shows that @xmath115 also is a reasonable measure of compaction .
, scaledwidth=80.0% ] and @xmath88 ( @xmath104 ) satisfying @xmath139 , where @xmath140 is the distance between centers of mass of the nucleotides @xcite .
mmtv has two stem basepairs ( cyan and green in a - c ) , which contribute to non - local contacts ( cyan and green in c ) .
_ azoarcus _ ribozyme has several hairpin basepairs ( p2 , p5 , p6 , p8 and p9 ) which can be seen in the vicinity of the diagonal in the contact map ( f ) .
there are also basepairs between nucleotides far along the sequence such as p3 , p4 and p7 , as well as tertiary interactions such as tl2-tr8 and tl9-tr5 .
these non - local contacts contribute to the collapsibility of the ribozyme.,scaledwidth=80.0% ] and total number of contacts @xmath141 ( same as protein - l ) , residue pairs @xmath86 are randomly chosen to satisfy the following conditions : ( a ) uniformly distributed , @xmath142 ; ( b ) local contacts only , @xmath142 and @xmath143 ; , and ( c ) non - local contacts only , @xmath144 . the calculated values of @xmath94 are explicitly shown .
the @xmath94 value for protein - l is @xmath145.(d ) distribution of the fraction of non - local contacts in the 2306 proteins . for each protein ,
the fraction is calculated as the number of non - local contacts ( @xmath146 ) divided by the total number of contacts ( @xmath147 ) . a contact between residues @xmath87 and @xmath88 is non - local ( nl ) " if @xmath144 .
there is a clear separation in this distribution for proteins rich in @xmath7 helices compared to those that are rich in @xmath6-sheets implying that the latter are more collapsible than the former . ] | the propensity of single domain globular proteins , the workhorses in cells , to be compact is the key reason that their folded states achieve high packing density . it is known that the radius of gyration , @xmath0 , of both the folded and unfolded ( created by adding denaturants ) states increase as @xmath1 where @xmath2 is the number of amino acids in the protein .
the values of the celebrated flory exponent @xmath3 are , respectively , @xmath4 and @xmath5 in the folded and unfolded states , which coincide with those found in homopolymers in poor and good solvents .
however , the extent of compaction of the unfolded state of a protein under low denaturant concentration , conditions favoring the formation of the folded state , is unknown .
this problem which goes to the heart of how proteins fold and has implications for the evolution of foldable sequences is unsolved .
we develop a theory based on polymer physics concepts that uses the contact map of proteins as input to quantitatively assess collapsibility of proteins .
the model , which includes only two - body excluded volume interaction and interactions reflecting the strength of the contact map , has only expanded and compact states .
surprisingly , we find that although protein collapsibility is universal , the propensity to be compact depends on the protein architecture .
application of the theory to over two thousand proteins shows that the extent of collapsibility depends not only on @xmath2 but also on the contact map reflecting the native fold structure .
a major prediction of the theory is that @xmath6-sheet proteins are far more collapsible than structures dominated by @xmath7-helices .
the theory fully resolves the apparent controversy between conclusions reached using different experimental probes assessing the extent of compaction of a couple proteins .
in addition , it reveals that there are considerable similarities between the physical mechanisms of homopolymer and protein collapse .
the theory provides quantitative insights into the reasons why single domain proteins are small and the physical reasons for the origin of multi - domain proteins .
we also show that non - coding rna molecules , whose collapsibility is similar to proteins with @xmath6-sheet structures , must undergo collapse prior to folding , adding support to `` compaction selection hypothesis '' proposed in the context of rna compaction . |
in particle induced x - ray emission ( pixe ) measurements an energetic ion beam , normally a 14 mev hydrogen or helium ion - beam , is directed to the sample to be analyzed , and sample composition for elements typically heavier than na or mg can be determined by means of characteristic x - ray emission @xcite . over the last four decades pixe
has developed to become a standard tool for elemental analysis in many fields of science . in geology ,
art restoration and medical diagnostic , to name a few , pixe has been used to determine the elemental composition of samples @xcite .
the development of micro - pixe , where the particle beam is focused to about 1 @xmath1 m diameter spot , has opened the possibility to raster the target in order to get the positional elemental information of the sample .
one interesting special feature of pixe is that it is possible to irradiate samples with ion beams in ambient conditions , thus making the analysis of biological and large specimens , like paintings , possible .
in addition , compared to electron beam excitement , the bremsstrahlung from protons is negligible because of their higher mass @xcite .
this means that the continuum x - ray background is orders of magnitude smaller compared to scanning electron microscope energy - dispersive x - ray spectroscopy ( sem - edx ) .
traditionally , reverse biased silicon or germanium detectors have been the work horse for pixe because of their wide dynamic range in energy , ease of use and reasonable cost .
these properties have so far compensated the limited energy resolution . on the other hand ,
wavelength dispersive detectors with their great energy resolution have been successfully used with pixe @xcite , but their intrinsically low throughput makes the measurements cumbersome and time consuming . by using transition - edge sensor ( tes )
microcalorimeters operating at cryogenic temperatures , one combines the benefits of energy dispersive detectors ( efficiency , wide energy range ) and wavelength dispersive detectors ( resolution ) @xcite .
tes detectors have matured to the state where they are used in number of applications , thanks to their superior energy resolution and sensitivity .
a tes is a device that operates between the superconducting and the normal state of a metallic thin film with a coupled absorber @xcite . within the superconducting transition region ,
the resistance of a tes is very sensitive to changes in temperature .
the device is connected to a heat bath via a weak thermal link so that when a photon hits the absorber and is converted into heat , due to the small heat capacity of the detector , a relatively large temperature excursion and change in the resistance of the detector is produced .
the change in the current through the tes during the photon event is read out with a superconducting quantum interference device ( squid ) , which acts as a highly sensitive current sensor that can be coupled to the low impedance of the tes @xcite .
multiplexing is needed , when the number of detectors increases and when the number of measurement wires needs to be minimized in order to limit the thermal load to the cryostat .
demonstrations of combining a single tes detector pixel with pixe have been published before @xcite , but the energy resolution achieved ( @xmath0 18 ev at 1.7 kev ) has been much worse than in this work , and quite far from the expected theoretical limits of @xmath0 1.3 ev for a typical tes detector at that energy @xcite .
a pelletron accelerator with 1.7 mv maximum terminal voltage was used to produce 2.015 mev incident @xmath2h@xmath3 ion beam that was used in the measurements .
the tes x - ray detector array was positioned to an angle of 90@xmath4 with respect to the incident ion beam .
the sample was tilted 45@xmath4 towards the detector ( see fig .
[ acc ] ) .
the backscattered incident ions were stopped before reaching the detector snout by means of 80 @xmath1 m polyethylene terephthalate ( pet ) film inside the sample chamber , which also filtered out low - energy x - rays .
an amptek x-123sdd silicon drift detector ( sdd ) ( 130 ev resolution for 5.9 kev x - rays from @xmath5fe source ) positioned to an angle of 135@xmath4 with respect to beam line was used as a reference detector .
the microcalorimeter tes detector array consisting of 160 pixels was fabricated at nist boulder .
the superconducting thin film is a molybdenum - copper bilayer film , where the proximity effect @xcite is used to achieve a critical temperature of @xmath0 100 mk .
the absorber is made of bismuth with horizontal dimensions of 350 @xmath1m@xmath6350 @xmath1 m and with 2.5 @xmath1 m thickness .
a collimator with a 320 @xmath1 m @xmath6 305 @xmath1 m aperture size for each pixel is placed on top of the detector chip to stop photons from hitting the substrate .
the separation between the detector and the collimator is 20 @xmath1 m .
the x - rays from the sample chamber enter through a vacuum window ( ap3.3 ultra - thin polymer / al silicon grid x - ray window by moxtek inc . ) into the cryostat .
inside the detector snout , the x - rays still have to penetrate through three layers of ir radiation filters ( each a 1 @xmath1 m pet film with 1 @xmath1 m aluminum coating ) at three different temperature stages of the snout ( 60 k , 1 k , 50 mk ) , before impinging on the tes detector array . in these first measurements only 12 pixels were connected out of which 9 gave good data .
one pixel had major problems with squid locking and did not produce good enough data and two pixels had over an order of magnitude smaller count rates and their data was removed from the final spectra due to energy calibration problems of insufficient number of events .
the read - out of the detectors was realized with nist time - division - multiplexing ( tdm ) squid electronics @xcite . in tdm , many microcalorimeters are read out in a single set of wires by turning the squids on sequentially .
the cryogenic cooling was achieved with an adiabatic demagnetization refrigerator ( adr ) ( denali model 102 , high precision devices , inc ) .
the cryostat has a special snout designed to couple the x - rays into the detectors , with two layers of cylindrical magnetic shielding from a4k material ( amuneal corp . ) at 60 k and 1 k , and one cylindrical superconducting al shield at 50 mk .
the pre - cooling of the adr s two stages ( 60 k and 3 k ) is done with cryomech pulse - tube refrigerator model no .
pt407 rm , making the whole system cryogen - free .
vibrations from the pulse tube are minimized by using a remote valve system and a flexible bellows between the pulse tube and cryostat body . the control software and hardware for the adr s 4 t magnet was made by star cryoelectronics llc .
the adr has a base temperature of @xmath0 30 mk and can be regulated at the 65 mk operation point with below 15 @xmath1k rms accuracy for about 12 hours with the read - out electronics on . in this first proof of principle
measurement , the detector array was in a location with a solid angle of only @xmath7=1.3@xmath8 of 4@xmath9 sr ( for the 12 pixels and only the absorber area visible through the collimator ) at a distance of 30 cm from the sample .
the solid angle will be increased by a factor of 36 in the future by moving the detector array closer to the sample from 30 cm to 5 cm .
also , by connecting all of the 160 pixels the solid angle will increase by another factor of 13 , leading to a total increase by a factor 468 .
this increase will translate straight to increase in the x - ray count rates and it gives the possibility to study thin film samples in contrast to the bulk samples discussed here .
a spectrum of mn k@xmath10 lines produced by a @xmath5fe - source was measured .
after the subtraction of the natural line shape and width of the emission , the instrumental energy resolution of the best pixel was 3.06 ev full width at half maximum ( fwhm ) , which is within a few percent of the expected resolution based on the measured average signal and noise .
the arithmetic mean resolution of 9 pixels was ( 3.8 @xmath11 0.6 ) ev each having approximately the same number of events .
no external magnetic field optimization was done during the measurements for any of the results . in fig .
[ mnfe ] , we present the comparison between the @xmath5fe produced mn spectrum and spectrum from a pixe measurement with 2 mev protons and a mn sample , with the combined data from the 9 pixels .
the fwhm broadens by less than 1 ev with the pixe data compared to the @xmath5fe measurement ( 4.45 ev vs. 3.75 ev respectively ) .
the origin of this broadening is not yet fully understood , but there was some electrical interference present during the mn pixe measurement unlike in the @xmath5fe measurement , that could affect the energy resolution .
complex spectrum of pixe and @xmath5fe source .
spectra are combined from the 9 pixels .
inset shows a satellite peak in the pixe spectrum discussed later .
( color figure online ) ] a bulk copper sample was also measured with 2 mev proton excitation , using both tes and sdd detectors to compare the two detector types , with the results shown in fig .
the ni impurities in the sample can not easily be resolved with the sdd detector but they are clear in the tes data .
the satellite peak on the high energy side of the cu k@xmath12 and cu k@xmath13 emission line visible in the tes spectrum is possibly a multivacancy satellite line , where two ( or more ) inner shell electron vacancies are created simultaneously @xcite .
the same feature is also seen in the mn spectrum in fig .
[ mnfe ] . ,
k@xmath13 and ni k@xmath10 impurity lines of the sample are more visible .
the theoretical energy values @xcite for the impurity lines are represented with the vertical arrows .
( color figure online ) ]
in this paper , we have demonstrated particle induced x - ray emission ( pixe ) spectroscopy using a multiplexed microcalorimeter tes detector array with an unprecedented energy resolution for pixe applications . in the future
, the beam line and the target chamber will be optimized to maximize the x - ray flux for thin film analysis .
this will make it possible to measure not only the elemental composition of thin films in greater detail , but also the chemical environment of the different elements in cases where the chemical shifts are large .
another very interesting future application is the study of the formation of embedded metallic clusters in insulators after implantation and annealing .
this work was supported by the finnish funding agency for technology and innovation tekes , academy of finland project no .
260880 and academy of finland center of excellence in nuclear and accelerator based physics ( ref .
. m. palosaari would personally like to thank the national graduate school in materials science for funding .
99 s. a. e. johansson in _ particle - induced x - ray emission spectrometry ( pixe ) , volume 133 in chemical analysis : a series of monographs on analytical chemistry and its applications _ edited by .
s. a. e. johansson , j. l. campbell k. g. malmqvist , john wiley & sons , inc .
, new york , ( 1995 ) . | in this paper we present a new measurement setup , where a transition - edge sensor detector array is used to detect x - rays in particle induced x - ray emission measurements with a 2 mev proton beam .
transition - edge sensors offer orders of magnitude improvement in energy resolution compared to conventional silicon or germanium detectors , making it possible to recognize spectral lines in materials analysis that have previously been impossible to resolve , and to get chemical information from the elements .
our sensors are cooled to the operation temperature ( @xmath0 65 mk ) with a cryogen - free adiabatic demagnetization refrigerator , which houses a specially designed x - ray snout that has a vacuum tight window to couple in the radiation . for the best pixel ,
the measured instrumental energy resolution was 3.06 ev full width at half maximum at 5.9 kev .
we discuss the current status of the project , benefits of transition - edge sensors when used in particle induced x - ray emission spectroscopy , and the results from the first measurements .
pacs numbers : 82.80.ej , 85.25.oj , 87.64.kd , |
the unitary fermi gas ( ufg ) is expected to be one of the most strongly correlated systems in nature as it saturates the unitarity bound for the @xmath1-wave cross section @xmath2 , where @xmath3 is the relative wave vector of scattering particles .
the strong correlations are responsible for a multitude of interesting phenomena .
the most surprising ones include the existence of the pseudogap regime between the superfluid state and the normal state @xcite and the nearly ideal hydrodynamic behavior @xcite .
therefore the properties of the ufg attract enormous attention from several communities including atomic physics , nuclear physics , relativistic heavy - ion collisions , and high-@xmath4 superconductivity ( see review papers @xcite ) .
experimentally the ufg has been realized with trapped fermionic atoms by means of feshbach resonances @xcite and currently represents one of the most controllable quantum systems .
however , a theoretical description poses a real challenge due to the lack of any small , dimensionless expansion parameter .
fortunately , for unpolarized ufg the genuine nonperturbative quantum monte carlo ( qmc ) techniques are free from the infamous fermionic sign problem , and can be used to produce reliable predictions , providing a solid benchmark for experimental as well as other theoretical studies . in particular , the energy of the ground state agrees with experimental results within @xmath5@xcite and the same accuracy was obtained for the equation of state for the system being in the normal state @xcite . for temperatures around the critical temperature
the discrepancy between experiment and the simulations does not exceed @xmath6 @xcite .
other properties well established by qmc include the evolution of the pairing gap @xcite , the critical temperature for the superfluid - to - normal phase transition @xcite , and the calculation of transport coefficients of the ufg , namely , the shear viscosity @xcite and the spin - diffusion coefficient @xcite .
determination of transport coefficients is currently of great interest as these provide a verification of various hypothetical bounds having their roots directly in quantum mechanics .
the best known example is a conjecture formulated by kovtun , son , and starinets ( kss ) of the existence of a lower bound @xmath7 on the ratio of the shear viscosity @xmath8 to the entropy density @xmath1 , for all fluids @xcite . due to strong correlations which imply short
mean free path , the ufg appears as one of the best candidates for being a `` perfect fluid '' ( besides the quark - gluon plasma created in heavy ion collisions ) , defined as the one with the lowest ratio @xmath9 .
experimental measurements for trapped systems yield @xmath10 @xcite , while the recent qmc simulations set the minimum value to be @xmath11 @xcite .
still there are contradicting scenarios of how the shear viscosity behaves in the low temperature regime .
whereas some authors expect the pairing correlations to reduce the viscosity significantly and even to decrease it to zero at @xmath12 @xcite , others favor the upturn originating from the phonon contribution @xcite .
recently it was shown by chafin and schfer that classical hydrodynamic fluctuations in a non - relativistic fluid lead to the appearance of the minimum in the shear viscosity to density ratio @xmath13 as a function of the temperature @xcite . in the vicinity of the critical temperature
the bound was established to be @xmath14 .
this result was confirmed by romatschke and young @xcite .
while the experimental results for trapped systems are consistent with the hydrodynamic bound , the qmc results for uniform systems violate this bound as they predict values @xmath15 in the superfluid phase @xcite . however , to extract the shear viscosity within qmc approach one has to perform an analytic continuation of the imaginary time correlator to real frequencies , which is a highly nontrivial procedure .
it is known that this step represents an ill - posed numerical problem , where statistical fluctuations which are tied up with the qmc are greatly enhanced during the continuation process .
indeed , it was pointed out in the supplement of @xcite , and also in @xcite , that the statistical accuracy of the quantum monte carlo signal allows to produce the ratio @xmath13 consistent with the hydrodynamic bound . here we present qmc results for a temperature evolution of the shear viscosity of higher accuracy than in our previous studies @xcite and discuss the errors associated with the analytic continuation . under the assumption of the smoothness of frequency dependent shear viscosity @xmath0 ,
we show that the new results violate the classical hydrodynamic bound @xcite for @xmath13 in a low temperature regime , while the ratio @xmath9 stays above the kss bound .
moreover the hydrodynamic bound is violated for the temperatures at which the cooper pairs are formed in the system .
we show that the agreement with hydrodynamic bound can be achieved only if there exists a low frequency sharp peak in the @xmath0 which is overlooked by analytic continuation procedure . in order to determine the shear viscosity of the ufg we employ the qmc technique on the lattice , which provides numerical results with controllable accuracy , up to quantifiable systematic uncertainties ( for details see ref .
these simulations are very similar to those of ref .
@xcite and we therefore only briefly describe the main stages of the computational process , focusing mainly on improvements .
henceforth we define units : @xmath16 .
we performed simulations using three lattice sizes @xmath17 with corresponding average densities @xmath18 and @xmath19 , respectively .
for these lattices the systematic errors were estimated to be less than @xmath20 and are the most severe in the superfluid phase ( for detailed discussion in the context of the transport coefficients see the supplemental material of refs .
these errors are related mainly to corrections coming from the nonzero effective range @xmath21 and exclude the universal high momenta tail in the occupation probability by finite momentum cut - off @xmath22 .
we obtain rather moderate values of @xmath23 and @xmath24 , respectively for lattices @xmath25 and @xmath26 [ @xmath27 stands for fermi momentum ] .
the statistical uncertainties are typically of the order of @xmath5 as an ensemble of @xmath28 uncorrelated samples were collected at each temperature .
the fact that we have used three different lattices allows us to check the stability of the results as we approach the thermodynamic @xmath29 and continuum @xmath30 limits , where @xmath31 ( the lattice constant is set to unity ) .
these new results were obtained with a new ( parallelized ) code and for densities significantly lower than those reported in ref .
@xcite , and thus are closer to the desired limit @xmath32 .
the frequency dependent shear viscosity @xmath0 is obtained from the imaginary - time ( euclidean ) stress tensor - stress tensor correlator @xmath33 evaluated at zero momentum @xmath34 , by inversion of the relation @xmath35 } { \sinh\left ( \omega\beta/2\right ) } d\omega , \label{eqn : inversionproblem}\ ] ] where @xmath36 is the inverse of the temperature .
taking the limit of zero frequency we obtain the static shear viscosity @xmath37 . the average in eq .
( [ eqn : gpi ] ) is performed over the grand canonical ensemble , @xmath38 , @xmath39 is the hamiltonian of the system , @xmath40 is the chemical potential , and @xmath41 is the particle number operator . for evaluation of the shear viscosity
it is sufficient to use the kinetic part of the stress tensor only ( see , e.g. , ref .
@xcite ) : @xmath42 the correlator @xmath43 was evaluated for at least 61 points in imaginary time @xmath44 , uniformly distributed in the interval @xmath45 $ ] .
since the correlator is symmetric with respect to @xmath46 , we restrict the inversion procedure to the interval @xmath47 $ ] , and apply the symmetrization : @xmath48/2 $ ] in order to decrease statistical fluctuations . the inversion of eq .
( [ eqn : inversionproblem ] ) represents numerically an ill - posed problem .
however , by using known theoretical constraints the solution of this problem becomes easier to determine . besides the non - negativity of the shear viscosity @xmath49 , the sum rule @xmath50 = \dfrac{\varepsilon}{3 } , \label{eqn : sumrule}\ ] ] and the asymptotic tail behavior
@xmath51 have been used as _ a priori _ information .
the tan contact density @xmath52 and the energy density @xmath53 are obtained consistently within the same simulation , and are in agreement with results reported in previous studies @xcite . to perform the inversion we applied the methodology which combines two complementary methods : singular value decomposition ( svd ) and maximum entropy method ( mem ) ,
both described in ref .
the inversion procedure consists of two steps : ( i ) the svd method provides us with the projection of the solution onto the subspace , where the inverse problem is well - posed @xmath54 , ( ii ) the inversion is performed using the self - consistent mem and we impose the svd solution as additional external constraint , i.e. , @xmath55=\tilde{\eta}$ ] , where @xmath56 stands for the projection operator onto svd subspace . in order to estimate the stability of the method with respect to the algorithm parameters , the bootstrap " strategy was applied . for more details
see the supplemental material of @xcite .
an important ingredient of the self - consistent mem is an appropriately chosen class of _ a priori _ models for the solution .
we assume that the frequency dependent shear viscosity @xmath0 has lorentzian - like structure at low frequencies , smoothly evolving into the asymptotic tail behavior : @xmath57\ , \mathcal{l}(\omega,\{\mu,\gamma , c\}),\end{aligned}\ ] ] where @xmath58 and @xmath59 the parameters @xmath60 describe admissible degrees of freedom of the model and are adjusted automatically in the self - consistent manner .
this choice of model space is reasonable as it is compatible with the low frequency behavior of the @xmath0 predicted by hydrodynamic fluctuations formalism @xcite and with results obtained within luttinger - ward theory @xcite .
finally , we assume that there is no sharp structure in the shear viscosity @xmath0 in the low frequency limit .
this assumption is of crucial importance , as it is well known , that sharp structures can be overlooked by numerical analytic continuation @xcite .
clearly , note that adding @xmath61 to the shear viscosity @xmath0 corresponds to increasing the imaginary time correlator by a constant value @xmath62 .
lattice for two different temperatures : ( upper panel ) @xmath63 ( close to the critical temperature ) and ( lower panel ) @xmath64 ( normal phase ) .
the gray bands show uncertainty of the reconstruction . in the insets , the corresponding qmc correlators ( blue points )
are presented .
the lines show the quality of the reproduction of the qmc data . [
fig : eta_dynamic ] ] in fig .
[ fig : eta_dynamic ] , the dimensionless frequency dependent shear viscosity @xmath65 is shown for two selected temperatures obtained for a @xmath66 lattice : close to the critical temperature of superfluid - to - normal phase transition @xmath67 and in the normal phase @xmath64 , where @xmath68 is the fermi energy of the non - interacting gas . in the normal phase @xmath69
, one may observe that @xmath0 possesses a lorentzian - like structure at low frequencies with a maximum in the vicinity of zero frequency .
for temperatures below @xmath70 the suppression of @xmath0 develops for the zero frequency .
we associate this depletion with the onset of pairing correlations , which has been identified to exist in ufg above the critical temperature up to temperatures @xmath71 , commonly referred to as pseudogap regime @xcite . as a function of temperature for an @xmath72 lattice [ solid ( red ) circles ] , a @xmath66 lattice [ ( blue ) squares ] , and a @xmath73 lattice [ ( green ) diamonds ] .
vertical black dotted lines indicate the critical temperature of superfluid - to - normal phase transition for @xmath67 and the onset of cooper pair formation , @xmath74 , respectively .
the black line indicates smooth approximation of the qmc results and the ( yellow ) band reflects uncertainty of the viscosity computation .
the ( purple ) asterisk shows the shear viscosity from the qmc calculations of ref .
the phonon contribution to the viscosity , evaluated according to ref .
@xcite is shown as a dot - dashed ( brown ) line and it is accounted for in the ( yellow ) band .
the hydrodynamic bound of ref .
@xcite is plotted by the black dashed line .
the inset shows the ratio of the shear viscosity to entropy density ( @xmath9 ) as a function of @xmath75 , obtained from the smooth approximation of the qmc results .
the black dashed line indicates kss bound : @xmath76 .
[ fig : viscosity ] ] in fig .
[ fig : viscosity ] , the dimensionless static shear viscosity @xmath13 is shown as a function of dimensionless temperature @xmath77 .
results obtained for different lattices exhibit reasonable agreement , even though they are affected by significant uncertainty of the order of @xmath78 .
for the @xmath73 lattice we provide results only for temperatures greater than @xmath79 as we were not able to produce the correlator of required accuracy to perform reliable inversion for lower temperatures
. the temperature dependence of the shear viscosity can be resolved , which is clearly visible when a smooth approximation of qmc results is performed , while the amount of scattering of the qmc points can be used to estimate final uncertainty of the viscosity computation ( yellow band in fig .
[ fig : viscosity ] ) .
moreover , the new results are consistent with previous estimate of the viscosity coefficient of ref .
we have checked the stability of our results with respect to the variation of the density , e.g. , for the lattice size @xmath66 and @xmath80 and obtained @xmath81 and @xmath82 for densities @xmath83 and @xmath84 , respectively . in the low temperature regime
no upturn originating from phonon contribution is observed , however , this does not imply that there is no upturn at all . precisely , the boxes used in the simulations are not capable of properly capturing the phonon - related physics .
it becomes clear if we note that the phonon occupation probability @xmath85 $ ] ( where @xmath86 is the sound velocity ) has a peak at momenta of the order of @xmath87 , which for @xmath88 is smaller than the smallest lattice momentum @xmath89 .
thus infrared momentum cutoff @xmath89 removes almost entirely the pure phonon contribution . to take into account this source of the systematic error we modified the estimated error of the shear viscosity computation in such a way as to encapsulate the pure phonon contribution of ref .
phonons appear to contribute significantly only well below the critical temperature @xmath4 . above the critical temperature , when the system is in a normal state and the superfluid order parameter is vanishing the calculation approach adopted in ref .
@xcite is not valid anymore .
this approach is based on an effective theory for the phonon field alone , valid at very low temperatures only , where cooper pair breaking and pseudogap physics can be ignored .
comparing our _ ab initio _ results with the hydrodynamic bound we find that as long as the system is in the normal phase ( for @xmath90 ) the static shear viscosity is located above the bound .
the results obtained for all three lattices indicate that the hydrodynamic bound is violated when the system enters the pseudogap regime , which occurs above the critical temperature , at @xmath91 .
interestingly , the corresponding ratio @xmath9 stays above the kss bound @xmath92 for all considered temperatures , with the minimal value @xmath93 , located in the vicinity of the critical temperature .
for the temperature evolution of the entropy we have used the mit results of ref .
@xcite , as our results reproduce experimental data with reasonable accuracy ( see @xcite ) .
the presented methodology , which produces the most probable results , together with the smoothness assumption of @xmath0 clearly allows for the hydrodynamic bound to be violated in the low temperature regime for @xmath94 .
the hydrodynamics bound was obtained by adding the contribution due to thermal fluctuations to an undefined classical value of the viscosity @xcite and with an amplitude depending on a cutoff parameter , which can be estimated only by order of magnitude , and thus a factor of order unity can not be refuted .
these estimates of a possible lower bound for viscosity are purely classical in character and they show that the role of classical thermal fluctuations could become important if the bare viscosity , evaluated in the absence of such thermal fluctuations , is small . the bare viscosity can be evaluated in a classical kinetic approach , but to what extent such an estimate is valid in the case of a unitary fermi gas remains an open question . even though a unitary fermi gas is dilute in the sense that @xmath95 , two - body collisions do not dominate because @xmath96 , where @xmath97 is the interaction cross section between fermions at the fermi surface . except the role of very low momentum phonons at very low temperatures which can not be properly included in qmc within the presently tractable lattice sizes , all other effects ( pairing , pseudogap , fermion collisions )
are included . as seen from fig .
[ fig : viscosity ] phonons show a significant contribution only for temperatures below the critical temperature @xmath4 .
it implies that even if the phonons were included , the viscosity would have remained below the classical hydrodynamic bound @xcite up to temperatures smaller than @xmath98 . as another possibility
, one can assume that hydrodynamic bound originates from long - wavelength fluctuations , which are not included properly into our simulations due to finite size of the box .
however , in this case we would expect to see a systematic trend in the results as we increase the box size , while we do not detect it . even though we have no reason to expect that any alleged unaccounted for thermal fluctuations of the kind advocated in refs .
@xcite might appear in the pseudogap regime below the temperature @xmath99 , we included such a contribution into the spectral function to investigate whether our qmc data are consistent with this classical hydrodynamical bound .
three tests were performed using data obtained for a @xmath66 lattice at the critical temperature .
hydrodynamic fluctuations @xcite suggest nonanalytic contributions for small frequencies of the form @xmath100 where the momentum diffusion constant @xmath101 is positive .
thus , @xmath0 is predicted to be a monotonically decreasing function for small frequencies . in the first test
, the mem solution is forced to be consistent with this assumption by replacing the lorentzian - like structure at low frequencies ( [ eqn : lorentzian ] ) by the expression @xmath102 where @xmath103 .
this modification implies that the space for default models contains only solutions which are monotonically decreasing and explicitly consistent with predictions taking into account hydrodynamic fluctuations .
in addition the constraint provided by svd is removed .
the resulting frequency dependent viscosity is shown with a ( red ) solid line in fig .
[ fig : tests ] . in the second test ,
the modified model is fit to the imaginary time correlator , in a similar way as has been done in ref .
this method typically violates the sum rule ( [ eqn : sumrule ] ) and the resulting @xmath0 is shown with a ( blue ) dot - dashed line in fig . [
fig : tests ] .
it is noted that our fit reveals a significantly different result than reported in ref .
@xcite , which may be related to the fact that we use different models for the fitting procedure . finally , the qmc correlator was fit to the model proposed by enss _
_ in the paper @xcite @xmath104 where the only free parameter is the viscous transport scattering time @xmath105 , as the total drude weight @xmath106 is fixed by the sum rule ( [ eqn : sumrule ] ) [ see ( red ) dotted line in fig .
[ fig : tests ] ] .
all three approaches are able to reproduce the imaginary time correlator within its error bars .
these simple tests result in a static shear viscosity below the hydrodynamic bound . for mem with the model ( [ eqn : lorentzianm ] ) , we find that @xmath13 is typically located below the bound
. however , we also find instances of the mem algorithm that produce results above the bound as reflected in the error band depicted in fig .
[ fig : tests ] .
similar results were determined for other temperatures .
obtained by three different methodologies of the reconstruction as described in the text .
the ( purple ) asterisk indicates the hydrodynamic bound obtained in refs .
[ fig : tests ] ] based on these tests we conclude that the qmc results are compatible with the hydrodynamic bound only if we assume the existence of a very narrow peak located around @xmath107 .
the classical hydrodynamic fluctuation formalism predicts this type of peak , which has its origin in nonanalytic contributions to the shear viscosity . in ref . @xcite it was estimated that the width of the nonanalytic structure can be of the order of @xmath108 and at the critical temperature it gives @xmath109 , which is in agreement with our first test .
however , we have to point out several important aspects of our results .
( i ) we do not observe any abrupt change in the frequency dependent shear viscosity profile as we decrease the temperature , which is a typical signal of losing the resolution .
instead , the profile we construct evolves smoothly between the two cases presented in fig .
[ fig : eta_dynamic ] .
( ii ) our qmc simulations admit the existence of the pseudogap regime for the ufg @xcite .
other theoretical approaches that admit the presence of the pseudogap predict the shear viscosity to be significantly suppressed , as in systems where a fermionic gap opens at @xmath99 , and exhibit a clear violation of the hydrodynamic bound @xcite .
similar behavior was predicted for the spin - diffusion coefficient @xcite .
our _ ab initio _ results are qualitatively consistent with these observations and support scenarios provided by these theories .
( iii ) we find it difficult to rationalize the existence of a nonanalytical classical hydrodynamical correction to the spectral function deep in the quantum regime .
in summary , we have presented the temperature evolution of a shear viscosity obtained within fully _ ab initio _ simulations . assuming the smoothness of the @xmath0 the results violate the hydrodynamic bound for @xmath13 in a low temperature regime
, while the ratio @xmath9 stays above the kss bound for all studied temperatures .
violation of the hydrodynamic bound occurs simultaneously with the occurrence of the cooper pairs in the system .
the accuracy of the qmc signal allows a result with the static viscosity being above the hydrodynamic bound only if we assume an existence of a sharp structure in the @xmath0 located in the vicinity of zero frequency .
we thank p. romatschke for making the results of ref .
@xcite available to us .
we acknowledge support under u.s .
doe grants no .
de - fg02 - 97er41014 and no .
de - fc02 - 07er41457 , and contract no .
n n202 128439 of the polish ministry of science
. one of the authors ( g.w . ) acknowledges the polish ministry of science for support within the program `` mobility plus - i edition '' under contract no .
calculations reported here have been performed at the interdisciplinary centre for mathematical and computational modelling ( icm ) at warsaw university and on the university of washington hyak cluster funded by the nsf mri grant no .
this research also used resources of the national center for computational sciences at oak ridge national laboratory , which is supported by the office of science of the department of energy under contract no .
de - ac05 - 00or22725 .
stewart , j.p .
gaebler , and d.s .
jin , nature * 454 * , 744 ( 2010 ) .
j. p. gaebler , j. t. stewart , t. e. drake , d. s. jin , a. perali , p. pieri , and g. c. strinati , nature phys . * 6 * , 569 ( 2010 )
. a. perali , _ et al .
106 * , 060402 ( 2011 ) .
a. turlapov , j. kinast , b. clancy , l. luo , j. joseph , j.e .
thomas , j. low temp .
* 150 * , 567 ( 2008 ) . c. cao , e. elliott , j. joseph , h. wu , j. petricka , t. schaefer , j. e. thomas , science * 331 * , 58 ( 2011 ) . c. cao , e. elliott , h. wu and j.e .
thomas , new j. phys . * 13 * , 075007 ( 2011 ) .
s giorgini , l.p .
pitaevskii , s. stringari , rev .
mod . phys . * 80 * , 1215 ( 2008 ) .
i. bloch , j. dalibard , w. zwerger , rev .
phys . * 80 * , 885 ( 2008 ) .
t. schfer and d. teaney , rep .
. phys . * 72 * , 126001 ( 2009 ) . _ the bcs - bec crossover and the unitary fermi gas _ lecture notes in physics , edited by w. zwerger ( springer - verlag , berlin , 2012 ) , vol . 836 . _ ultracold fermi gases _ , proceedings of the international school of physics `` enrico fermi , '' courseclxiv , varenna , june 20 - 30,2006 , edited by m. inguscio , w. ketterle , and c. salomon ( ios press , amsterdam , 2008 )
. j. carlson _
et al . _ ,
a * 84 * , 061602(r ) ( 2011 ) .
k. van houcke , f. werner , e. kozik , n. prokofev , b. svistunov , m. j. h. ku , a. t. sommer , l. w. cheuk , a. schirotzek , and m. w. zwierlein , nature phys . * 8 * , 366 ( 2012 ) .
drut , t.a .
lhde , g. wlazowski , p. magierski , phys .
a * 85 * , 051601(r ) ( 2012 ) .
j. carlson and s. reddy , phys .
* 95 * , 060401 ( 2005 ) .
a. gezerlis and j. carlson , phys .
c * 77 * , 032801(r ) ( 2008 ) .
p. magierski , g. wlazowski , a. bulgac , and j.e .
drut , phys . rev
. lett . * 103 * , 210403 ( 2009 ) .
p. magierski , g. wlazowski , and a. bulgac , phys .
lett . * 107 * , 145304 ( 2011 ) .
e. burovski , n. prokofev , b. svistunov , and m. troyer , phys . rev
* 96 * , 160402 ( 2006 ) .
e. burovski , e. kozik , n. prokofev , b. svistunov , and m. troyer , phys .
* 101 * , 090402 ( 2008 ) .
a. bulgac , j.e .
drut , and p. magierski , phys .
a * 78 * 023625 ( 2008 ) .
g. wlazowski , p. magierski , and j.e .
drut , phys .
lett . * 109 * , 020406 ( 2012 ) .
g. wlazowski , p. magierski , j.e .
drut , a. bulgac , and k.j .
roche , phys .
* 110 * , 090401 ( 2013 ) .
kovtun , d.t .
son , and a.o .
starinets , phys .
lett . * 94 * , 111601 ( 2005 ) .
bruun and h. smith , phys .
a * 75 * , 043612 ( 2007 ) .
h. guo , d. wulin , c .- c .
chien , and k. levin , phys .
* 107 * , 020403 ( 2011 ) ; new j. phys .
* 13 * , 075011 ( 2011 ) .
g. rupak and t. schfer , phys .
a * 76 * , 053607 ( 2007 ) . c. chafin , t. schfer , phys .
a * 87 * , 023629 ( 2013 ) .
p. romatschke , r.e .
young , phys .
a * 87 * , 053606 ( 2013 ) .
t. enss , r. haussmann , w. zwerger , ann . phys . * 326 * , 770 ( 2011 ) .
p. magierski and g. wlazowski , comput .
phys . commun . * 183 * , 2264 ( 2012 ) .
ku , a.t .
sommer , l.w .
cheuk , m.w .
zwierlein , science * 335 * , 563 ( 2012 ) .
d. wulin , h. guo , c .- c .
chien , and k. levin , phys .
a * 83 * , 061601(r ) ( 2011 ) . | we present an _ ab initio _ determination of the shear viscosity for the unitary fermi gas based on finite temperature quantum monte carlo ( qmc ) calculations and the kubo linear - response formalism .
the results are confronted with the bound for the shear viscosity originating from hydrodynamic fluctuations .
assuming smoothness of the frequency dependent shear viscosity @xmath0 , we show that the bound is violated in the low temperature regime and the violation occurs simultaneously with the onset of the cooper paring in the system . in order to preserve the hydrodynamic bound in qmc @xmath0
has to possess a sharp structure located in the vicinity of zero frequency which is not resolved by an analytic continuation procedure . |
in recent years simulations of babcock - leighton type flux - transport ( hereafter blft ) solar dynamos in both 2d and 3d @xcite demonstrated the crucial role the sun s global meridional circulation plays in determining solar cycle properties .
time variations in speed and profile of meridional circulation have profound influence on solar cycle length and amplitude .
the recent unusually long minimum between cycles 23 and 24 has been explained by implementing two plausible changes in meridional circulation , ( i ) by implementing the change from a two - celled profile in latitude in cycles 22 to a one - celled profile in cycle 23 @xcite , and ( ii ) by performing a vast number of simulations by introducing a flow - speed change with time during the declining phase of each cycle @xcite . accurately knowing the speed and profile variations of the meridional circulation would greatly improve prediction of solar cycle features .
the meridional circulation has been observed in the photosphere and inside the upper convection zone in the latitude range from the equator to @xmath1 in each hemisphere @xcite .
however , the speed , pattern and time variations of the circulation at high latitudes and in the deeper convection zone are not known from observations yet . theoretical models of meridional circulation @xcite provide some knowledge , but the flow patterns derived from model outputs vary from model to model , primarily because of our lack of knowledge of viscosity and density profiles and thermodynamics in the solar interior , which are essential ingredients in such models . as differential rotation does not change much with time compared to meridional circulation , in this first study
we focus on time variation of meridional flow - speed , using a set - up similar to that used previously @xcite .
since the meridional circulation is a specified parameter in kinematic blft dynamos and the dynamo solutions depend sensitively on the spatio - temporal patterns of this circulation , we ask the question : can we infer the meridional circulation ( in both space and time ) from observations of the magnetic field ?
the purpose of this paper is to describe an ensemble kalman filter ( enkf ) data assimilation in a 2d blft solar dynamo model for reconstructing meridional flow - speed as a function of time for several solar cycles .
a subsequent paper will investigate the reconstruction of spatio - temporal patterns of meridional circulation in the solar convection zone .
data assimilation approaches have been in use for several decades in atmospheric and oceanic models , but such approaches have been implemented in solar and geodynamo models only recently .
@xcite introduced a variational data assimilation system into an @xmath2-@xmath3 type solar dynamo model to reconstruct the @xmath2-effect using synthetic data .
very recently @xcite applied a variational data assimilation method to estimate errors of the reconstructed system states in a stratified convection model .
a detailed discussion of data assimilation in the context of the geodynamo can be found in @xcite . in a sequential data assimilation framework ,
a set of dynamical variables at a particular time defines a `` model state '' , which is the time - varying flow speed in the context of the present paper .
scalar functions of these state variables that can also be observed using certain instruments are called `` observation variables '' , which are magnetic fields here .
more detailed terminology for identifying data assimilation components with solar physics variables is given in 2 . in brief , the goal of sequential data assimilation is to constrain the model state at each time - step to obtain model - generated observation variables that are as close to the real observations as possible .
the basic framework is based on statistical multidimensional regression analysis , a well - developed method that has been applied in atmospheric and oceanic studies ( see @xcite for details ) .
the enkf sequential data assimilation framework also allows adding model parameters to the set of model states and estimating values of these parameters that are most consistent with the observations .
it is a common practice to perform an `` observation system simulation experiment '' ( osse ) in order to validate and calibrate the assimilation framework for a particular model .
an osse generates synthetic observations from the same numerical model that is used in the assimilation . in this case
the numerical model is a simple blft dynamo model containing only a weak nonlinearity in the @xmath2-quenching term ; thus adding gaussian noise to model - outputs for producing synthetic observations works well . in a more realistic situation for a large system with highly nonlinear processes , such as in numerical weather prediction models
, it may be necessary to use a non - gaussian ensemble filter ( see , e.g. @xcite ) .
a few examples of predicting model parameters using sequential data assimilation techniques have been presented by @xcite and @xcite in the context of estimating neutral thermospheric composition , and most recently by @xcite for estimating thermospheric mass density .
an enkf data assimilation framework has recently been applied to a 3d , convection - driven geodynamo model for full state estimation using surface poloidal magnetic fields as observations @xcite .
we implement enkf sequential data assimilation to reconstruct time - variations in meridional flow - speed for several solar cycles , using poloidal and toroidal magnetic fields as observations .
we note certain differences in our case compared to the cases described above , namely , unlike neutral thermospheric composition and thermospheric mass density , the meridional flow - speed is not governed by a deterministic equation .
in order to describe the enkf data assimilation methodology , we first identify the data assimilation components with solar physics variables .
a physical blft dynamo model , which generates magnetic field data for a given the meridional flow - speed , is called the `` forward operator '' .
the time - varying meridional flow - speed at a given time is the `` model state '' and will be estimated by the enkf system .
the enkf requires a prediction model that generates a forecast of the meridional flow - speed at a later time given the value at the current time . here
, that prediction model simply adds a random draw from a gaussian distribution to the current meridional flow - speed , because the blft dynamo model we are using here is a kinematic dynamo model . in the future ,
the results can be further refined by imposing additional physical conditions using a dynamical dynamo model .
it is important not to confuse the prediction model with the blft dynamo model that acts as a forward operator for the data assimilation process .
figure 1 schematically depicts the data assimilation framework considering an `` ensemble of three members '' of the prediction model state ( meridional flow - speeds ) .
sequential assimilation can usually be described as a two - step procedure , a forecast followed by an analysis each time an observation is available . in figure 1 , @xmath4 near the label ( a ) denote three different realizations of initial flow - speeds , which are input to the prediction model .
we generate `` prior '' estimates of the model state ( denoted by @xmath5 near the label ( b ) ) by using the prediction model to advance the estimates of the flow - speed from the initial time ( @xmath6 ) to the time ( @xmath7 ) at which the first observations of magnetic field are available . in this case , the prediction model just adds a different draw from the gaussian noise distribution to each initial ensemble estimate of the flow speed ( @xmath8 along the solid green arrows in figure 1 ) .
the resulting ensemble of flow speeds is referred to as a prior ensemble estimate .
the central equation for estimating the prior state is a stochastic equation given by , @xmath9 in which , @xmath10 is a function for generating normalized gaussian random numbers with unit amplitude and unit standard deviation , and the amplitude of the prediction model noise is governed by @xmath11
. thus the evolution of the system from label ( a ) to ( b ) through equation ( 1 ) , denoted by eq1 in figure 1 , constitutes the first step of the two - step procedure in sequential assimilation .
the second step includes ( i ) producing observations from outputs of the forward operator and ( ii ) estimating posterior flow - speeds by employing regression among these observations , real observations ( synthetic in osse ) and prior flow - speeds .
the forward operator ( blft dynamo in this case ) is denoted by @xmath12 along the solid black arrows , which uses the prior estimates of flow speed to produce a prior `` ensemble of observation estimates '' which are magnetic field outputs .
three realizations of magnetic fields from the forward operator are denoted by @xmath13 , @xmath14 , @xmath15 in figure 1 .
note that the statement , `` forward operator operating on three model states generates three prior observation estimates '' , is equivalent to the statement , `` blft dynamo running with three meridional flow speeds produces three sets of model outputs of magnetic fields '' . to elucidate the second step of assimilation
, we describe the function of the `` filter '' . for real prediction using enkf data assimilation
, we would have real data ( observations ) from instruments ; in our osse case it is synthetic , denoted by @xmath16 in figure 1 .
synthetic observations are generated by applying the forward operator to a specified time series of meridional flow - speeds ( as shown in figure 2a ) .
our goal is to apply an enkf to obtain an improved distribution of estimated flow - speeds ( i.e. `` posterior states '' , @xmath17 ) using the prior ensemble and the observation of magnetic field .
the enkf ( black - dotted box in the diagram ) first compares the prior ensemble of observation - estimates to the actual observation and computes increments to the prior observation - estimates .
these observation - increments are then regressed using the joint prior - ensemble distribution of flow - speed and magnetic field observations to compute increments for the prior - ensemble of flow - speeds .
the enkf can also produce a posterior - ensemble of magnetic field observations ( @xmath18 ) , which can be used for diagnostic purposes .
the posterior - ensemble distribution of flow - speeds is the best estimate of the flow - speed distribution at time @xmath7 given the available observations .
mean flow - speed at time @xmath7 can be calculated by taking the average over all ensemble members . to proceed with the reconstruction of flow - speeds at the time of the next observation , @xmath19 , the reconstructed flow - speeds at time @xmath7 are used as the input to the prediction model ( equation ( 1 ) ) , and the same procedure described in the previous paragraphs
is repeated .
random gaussian noise through equation ( 1 ) prevents degeneration of ensemble .
thus after many time steps , the entire time series of the ensemble distribution of flow - speeds can be constructed .
time series of the mean flow - speeds can be calculated by taking the average over all ensemble members at each time .
however , it may produce a better reconstruction in some cases if one ensemble member , which produced observation ( @xmath20 ) closest to real observation ( @xmath16 ) , is chosen . in order to perform an osse ,
it is now necessary to define the `` true state '' flow - speed as a function of time .
synthetic observations are generated at selected times by applying the forward operator ( blft dynamo model ) to the true state flow - speed at the appropriate time and adding on a random draw from a specified observational error distribution to simulate instrumental and other errors . as noted above
, the forward operator is a kinematic blft dynamo model , described in detail in @xcite .
the dynamo ingredients are a solar - like differential rotation , a single - celled meridional circulation , a babcock - leighton type surface @xmath2-effect and a depth - dependent diffusivity ; the mathematical forms are prescribed in @xcite .
dynamo equations , computation domain and boundary conditions are used as given in @xcite . at @xmath21
we start the integration of the dynamo - dart system over the first assimilation - step by initializing the forward operator ( ftd model ) with a converged solution for a flow - speed of @xmath22 . in the subsequent assimilation - steps , the solution at the end of previous assimilation - step for each ensemble - member is used as initial condition .
we construct a time - varying flow - speed for a span of 35 years ( see figure 2(a ) ) , guided by observations @xcite , which has a natural variation of 20 - 40% with respect to the mean flow considered here , i.e. 14 @xmath23 , as shown by the thin black line in figure 2(a ) .
this specified time - varying flow - speed is referred to as the true model state .
thus , keeping the spatial pattern of meridional circulation fixed , we consider here reconstructing the time - series of the scalar flow - speed . to generate synthetic magnetic field data , we incorporate the time - varying true meridional flow - speed in our blft dynamo , and simulate the time series of idealized magnetic fields in the entire computation domain .
then we construct the time series of synthetic magnetic observations by adding synthetic observational error to the simulated idealized magnetic data . figure 2(b ) shows a single observation , which is created from the simulated poloidal field by extracting from the location , @xmath24 and @xmath25 .
however , note that the dynamo simulation in @xmath26 grids can give us as many as 20202 synthetic magnetic observations of poloidal and toroidal fields .
considering only one observation , as shown in figure 2(b ) , we perform assimilation runs with 16 ensemble members with an observational error of @xmath27 , which means an error of @xmath27 about the ideal magnetic field generated using the true meridional flow speed .
to estimate the prior states of flow - speed , we use equation ( 1 ) , in which we set @xmath28 .
if the meridional flow - speed varies up to @xmath29 ( i.e. @xmath30 for a mean flow - speed of @xmath31 ) during six months , the variation in 15 days can be @xmath32 .
thus we chose @xmath28 so that it is large enough to capture the variation in flow - speed within our selected updating time - step of 15 days and also large enough to avoid ensemble collapse , but not so large as to produce unusual departures from cyclic behavior in a flux - transport dynamo .
we show in figure 3 the reconstructed meridional flow - speed as a function of time ( panel ( a ) ) and the estimates of the observation computed from the flow speed ensemble after data assimilation ( panel ( b ) ) .
we see in figure 3(a ) that the reconstruction is reasonably good except for the time windows between @xmath33 to 18 years and 33 to 35 years during which we find @xmath34 error in the reconstructed flow speed .
observations indicate @xmath35 error in the measurement of meridional flow speed @xcite .
the inset of figure 3(a ) reveals , for an initial guess , far - off from the true - state , the reconstructed states asymptotically converges towards the true - state .
even though they oscillate around the truth with large amplitude , the oscillations damp with time .
figure 3(b ) shows histograms for the normalized distribution of flow - speeds before ( in cyan ) and after ( in magenta ) the analysis stage , along with true - state ( blue ) , for the time instances of 5 , 6.9 , 10.1 and 27.5 years ( marked by vertical lines in figure 3(a ) ) during assimilation . for this case with 16 ensemble members we chose 10 bins for an optimal display .
four time instances are chosen in such a way as to present the following four different phases of reconstructions : ( i ) the distribution of prior and posterior states has small overlap ( top left frame of figure 3(b ) ) , ( ii ) distribution is sharply peaked in one bin each for prior and posterior ( top right frame ) , ( iii ) distribution is broad and has significant overlap ( bottom left frame ) and ( iv ) distribution has no overlap at all ( bottom right frame ) .
however , we can clearly see the successful performance of the enkf which reveals that in all cases the analysis phase brings the posterior distribution ( magenta bars ) closer to true - state ( blue ) .
figure 3(c ) reveals that the assimilated magnetic observation ( blue solid curve ) is well - reproduced when compared with real observation ( red - dashed curve ) .
this is not surprising , because it has already been noted that short - term small fluctuations in flow speed do not significantly influence the overall evolution of global magnetic fields generated by a babcock - leighton dynamo ( see , e.g. @xcite ) .
we also plot the innovation ( @xmath36 ) for this observation ( black curve ) and the cumulative innovation ( orange curve ) .
the innovation at the @xmath37 analysis - step ( @xmath36 ) is defined as the signed difference between real and reconstructed observations , whereas the cumulative innovation ( @xmath38 ) at @xmath37 analysis - step is the normalized sum of norm of innovation vectors over all the previous analysis - steps . @xmath36 and @xmath38 being small ,
both have been ten - fold magnified to superimpose on observations .
we see that , at three different time instances ( @xmath39 , 17 and 27 years ) , the innovation is relatively large ; this is because at @xmath21 the initial guess is far - off from the truth , and at @xmath40 and 27 years there are relatively sharp changes in flow - speed .
the cumulative innovation asymptotes to zero as expected , implying no bias of the system . to investigate the possibility of further improvement in the quality of the enkf reconstruction , we examine the consequences of three important aspects of the enkf : variation in observational error , size of ensemble and number of observations . we perform convergence tests by estimating the error in the reconstructed state as functions of observational errors , size of ensemble and number of observations .
we define the error as the root mean square of the difference between the reconstructed state and the true - state .
the assimilation interval is chosen to be 15 days in the present case ; denoting every 15 days by the index @xmath41 , the true state and the reconstructed state at the @xmath37 assimilation step by @xmath42 and @xmath43 respectively , we define the error as , @xmath44 , in which @xmath45 is the total number of indices for the 15-days assimilation intervals during the entire time - span of 35 years .
figure 4 shows the rms errors in reconstructed flow speed as functions of observational error ( figure 4(a ) ) , size of ensemble ( figure 4(b ) ) and number of observations ( figure 4(c ) ) .
we see in figure 4 that the error decreases systematically and asymptotes for certain values of the observational errors ( 1% ) , size of ensemble ( 192 members ) and number of observations ( 180 ) . in the case of more than one observation
, we include more poloidal field observations ( synthetic ) from various locations at and near the surface , and more toroidal field observations at and near the bottom of the convection zone .
while we vary observational errors in panel ( a ) , we use @xmath46 observational error in panels ( b ) and ( c ) .
panels ( a , b ) show convergence to typical hyperbolic patterns , as is often seen in numerical convergence tests ( see , e.g. figure 3 of dikpati ( 2012 ) ) . but panel ( c ) shows that the reconstruction can be improved systematically only when there are more than a certain number of observations ( four in our case ) . in all these experiments
, we used the same enkf scheme .
bias or systematic error in an osse reconstruction may arise primarily because of the following assumptions made in the assimilation system : ( i ) the evolution of the ensemble spread is linear , ( ii ) the ensemble is sufficiently large , ( iii ) the forward operators are linear . though these assumptions are roughly valid , they are not strictly true .
but in general , the resulting systematic errors are small unless the assimilation is applied in a large , highly nonlinear system like a numerical weather prediction model . with the knowledge gained from the convergence experiments shown in figure 4
, we consider a case with 192 ensemble members and 180 magnetic observations , consisting of equal numbers of poloidal and toroidal magnetic fields at various latitude and depth locations .
each of these observations has an observational error of @xmath46 .
we present the reconstructed flow - speed from this assimilation in figure 5 . the reconstructed flow - speed ( red curve )
matches very well with the true state , and thus the true state plotted in blue is essentially hidden behind red and green curves . it is not realistic to expect an observational error of as small as 1% ; figure 5 presents here an illustrative example of one of the best possible reconstructions .
in fact , several additional assimilation runs indicate that the reconstruction is still good if the observational error does not exceed 40% , and reasonably good when 90 out of 180 observations have up to @xmath47 errors .
but the reconstruction fails when all observations have more than 40% error .
what we have demonstrated so far is that the time - dependent amplitude of meridional circulation , having one flow - cell per hemisphere , can be reconstructed by implementing enkf data assimilation with synthetic data . in reality
we do not know from observational data whether they were produced by a dynamo operating with a single - celled flow in each hemisphere , or with a more complex flow profile , or with a combination of complex time - variations in all possible dynamo ingredients . in order to investigate the outcome from this method
when the assumption made about the flow profile is wrong , we carry out an experiment to reconstruct flow - speed assuming a one - cell flow , while using observations of magnetic fields produced from a flow pattern that has two cells in latitude in each hemisphere ( see @xcite for prescription of a two - celled flow ) .
we obtain synthetic data for a case with two flow cells in latitude .
using 192 ensemble members and 180 magnetic observations with 1% error in each of these observations , we estimate the time variation in flow - speed by assuming a single - celled flow , and plot in figure 6(a ) .
we find that the reconstruction is relatively poor , as expected .
however , with a closer look we can see that the reconstructed speed is trying to approach the true - state from a lower value for the first 12 years and from a higher value for the next 13 years .
when the trend in the true - state reverses near the year 27 , the osse has greater difficulty in converging on it .
figure 6(b ) shows the innovation for two typical observations of poloidal fields near the surface , at @xmath25 ( i.e. at low latitude near the equator ) and at @xmath48 ( at high latitude near the pole ) .
much larger innovation at high latitude than at low latitude reflects the fact that the spatial pattern is more erroneous at polar latitudes .
the cummulative innovation ( orange curve ) does no longer asymptote to zero ; this implies bias in the system , due to incorrect assumption of flow - pattern . in the future
, we will extensively explore the reconstruction of spatial pattern of meridional circulation .
we have demonstrated through osses that enkf data assimilation into a babcock - leighton flux - transport dynamo can successfully reconstruct time - varying meridional circulation speed for several cycles from observations of the magnetic field . to obtain the best reconstruction , we have fed the assimilation system of 192 ensemble members with 180 observations with 1% observational error .
however , a reasonably good reconstruction can be obtained when all observations have up to @xmath49 error , or half of the observations have up to @xmath50 errors , but the rest of them have much smaller errors .
@xcite noted that the response time of a babcock - leighton dynamo model to changes in meridional flow is @xmath51 months , so the relatively poor reconstruction with only one observation with 33% observational error and 16 ensemble members can be improved if this information about the dynamo model s response time to flow changes can be exploited .
a forthcoming paper will investigate how to use this response time during assimilation . throughout this paper , we used an assimilation interval of 15-days .
the reason for this choice is as follows .
recently @xcite have done a very thorough assessment of the predictability of solar flux - transport dynamos .
predictability of a model refers to the time it takes for two solutions that start from slightly different values of either initial conditions or input parameters to diverge from each other to the point that they forecast substantially different outcomes .
@xcite found an @xmath52-folding time of about 30 years .
the flux - transport dynamo model we use here is physically very similar to theirs .
therefore we judge that the @xmath52-folding time for our model will be similar .
it is clear that the time interval for updating the data in our osse s should be much less than the predictability limit of the model , but it should be long compared to the integration time step ( a few hours ) , and consistent with the time scale for changes in axisymmetric solar observations , which is one solar rotation .
it should also be shorter than the `` response time '' of our dynamo model ( @xmath51 months , see @xcite ) to a sudden change in inputs .
we have therefore chosen our updating interval to be 15 days ; we plan to test the sensitivity of assimilation to changes in the updating interval from 15 days to longer ( for instance , 30 days ) and shorter ( @xmath53 days ) .
in this study we have demonstrated what it takes to reconstruct the amplitude variations with time of a one - celled meridional circulation of fixed profile with latitude and radius . in reality
the sun s meridional circulation may not be a one - celled pattern it may undergo changes both in profile and speed with time .
we have demonstrated that the innovation in observation - forecast can be very large when the assumption about the spatial structure of the flow - pattern is incorrect , and it can be even larger where the departure in assumed spatial pattern from actual pattern in flow is larger .
thus an obvious next step with our data assimilation system would be to attempt to reconstruct the spatio - temporal pattern of meridional flow .
our ultimate goal is to perform assimilation runs from actual observations instead of synthetic data , in cases of reconstruction as well as future predictions . from this study
we can build confidence about the power of enkf data assimilation for reconstructing not only the flow speed but also the profile of meridional circulation in the entire convection zone of the sun in the future .
we thank nancy collins and tim hoar for their invaluable help with assimilation tools and software .
we extend our thanks to two reviewers for many helpful comments and constructive suggestions , which helped significantly improve the paper .
the dart / dynamo assimilation runs have been performed on the yellowstone supercomputer of nwsc / ncar under project number p22104000 , and all assimilation tools used in this work are available to the public from @xmath54 .
this work is partially supported by nasa grant nnx08aq34 g .
dhrubaditya mitra was supported by the european research council under the astrodyn research project no .
227952 and the swedish research council under grant 2011 - 542 and the hao visitor program .
national center for atmospheric research is sponsored by the national science foundation . , j. l. , wyman , b. , zhang , s. & hoar , t. , assimilation of surface pressure observations using an ensemble filter in an idealized global atmospheric prediction system , j. atmos .
, 62 , 2925 - 2938 , 2005 , a. , hulot , g. , jault , d. , kuang , w. , tangborn , a. , gillet , n. , canet , e. , aubert , j. & lhuillier , f. , an introduction to data assimilation and predictability in geomagnetism , space sci rev , 155 , 247 - 291 , 2010 , j. , bogart , r. s. , kosovichev , a. g. , duvall , t. l. , jr . ,
hartlep , h. , detection of equatorward meridional flow and evidence of double - cell meridional circulation inside the sun , apj lett . , 774 , l29 , 1 - 6 , 2013 | accurate knowledge of time - variation in meridional flow - speed and profile is crucial for estimating a solar cycle s features , which are ultimately responsible for causing space climate variations .
however , no consensus has been reached yet about the sun s meridional circulation pattern observations and theories . by implementing an ensemble kalman filter ( enkf ) data assimilation in a babcock - leighton solar dynamo model using data assimilation research testbed ( dart ) framework
, we find that the best reconstruction of time - variation in meridional flow - speed can be obtained when ten or more observations are used with an updating time of 15 days and a @xmath0 observational error .
increasing ensemble - size from 16 to 160 improves reconstruction .
comparison of reconstructed flow - speed with `` true - state '' reveals that enkf data assimilation is very powerful for reconstructing meridional flow - speeds and suggests that it can be implemented for reconstructing spatio - temporal patterns of meridional circulation . |
after loading the mot for @xmath46 s we shut off the mot beams and wait for @xmath47s before probing the atoms for @xmath47s .
the probe laser beam is split up after the cavity and sent to two detectors which allow for a low bandwidth detection ( @xmath48 mhz ) of the transmitted power and a high bandwidth detection ( @xmath49 mhz ) for the phase measurement .
the nice - ohms phase signal is obtained by generating fourier sidebands in the probe light at exactly the free spectral range ( fsr ) of our cavity at @xmath50 mhz using an eom .
the carrier is tuned close to the atomic resonance and will experience a phase shift due to the atoms .
the sidebands , on the other hand , are far off resonance and will not be influenced by the atoms .
they will , however , still be transmitted through the cavity since they are displaced by exactly one fsr , and experience similar technical noise and low - frequency amplitude noise as the carrier .
demodulating the heterodyne beat signal between the carrier and the sidebands after the high bandwidth detector will result in a signal proportional to the phase shift induced by the atoms at a significantly reduced noise level .
the transmission is measured by normalising the signal from the slow detector obtained just after the mot is shut off with a signal obtained after waiting an additional 50 ms for all the atoms to have left the cavity . to minimize frequency fluctuations of the 689 nm laser we lock the laser to a high finesse ule cavity ( @xmath51 ) using the pound - drever - hall method @xcite . at 1
ms of integration time the laser has a linewidth of about @xmath52 hz .
we monitor the number of atoms though the atomic absorption ( low bandwidth detector ) and indirectly through the mot fluorescence measured by a photomultiplier .
typically , we trap about 5@xmath53 @xmath54 atoms in the mot while only @xmath55 atoms overlap with the cavity volume .
the number of atoms is tuned by changing the loading time of the mot . by optical alignment of the mot
beams the trap center is off - set from the quadrupole magnetic field zero such that atoms experience a constant magnetic field of about 1 mt aligned parallel to the probe light polarization .
this provides a local quantization axis and ensures that we only probe the @xmath56-transition ; the @xmath57 transitions are shifted by tens of mhz .
the mirrors for the 689 nm cavity are attached outside the view ports of the vacuum chamber , limiting the finesse of the cavity to @xmath58 and the cavity length to @xmath59 cm .
our atom - cavity parameters are characterized by the single atom cooperativity @xmath60 , the atom - light coupling @xmath61 590.6 hz , the cavity decay time constant @xmath3 and the atomic decay time constant @xmath62 . here
@xmath63 is the cavity waist and @xmath64 the transition wavelength .
the collective coupling parameter is given by @xmath65 where @xmath66 refers to the total number of atoms in the cavity ( @xmath67 ) .
our set of parameters is given by @xmath68 ) .
the collective cooperativity @xmath69 is for our system about @xmath70 , thus placing our system in the the regime of high collective cooperativity ( which requires @xmath71 ) in the so - called bad cavity limit ( the cavity linewidth @xmath15 mhz is about a factor of 800 larger than the atomic linewidth @xmath10 khz ) , but not in the cqed strong collective coupling regime which requires @xmath72 and @xmath73 @xcite .
we model our system as a collection of 2-level atoms inside a single mode optical cavity , using a born - markov master equation to describe the open quantum system , @xmath74 + \hat{\mathcal{l}}\left [ \hat{\rho } \right ] , \label{me1}\end{aligned}\ ] ] where , @xmath75 the hamiltonian @xmath76 describes the coherent evolution of the coupled atom cavity system in an interaction picture which rotates at the frequency of the cavity , and @xmath77 is the atom - cavity detuning .
the pauli spin matrices are @xmath78 , @xmath79 is the classical drive amplitude @xmath80 , with @xmath81 being the input optical power , and @xmath82 is the annihilation operator of the cavity mode .
the atom - cavity coupling rate is given by : @xmath83 where @xmath84 is the wave number of the cavity , @xmath85 and @xmath86 are the longitudinal and axial positions , @xmath87 is the doppler shift in terms of the atom velocity @xmath21 , @xmath63 is the waist of the gaussian cavity mode , @xmath88 is the vacuum rabi frequency with @xmath89 the effective mode volume of the cavity , @xmath90 is the dipole moment of the atomic transition , and @xmath91 is the vacuum permittivity the incoherent evolution is described by the liouvillian @xmath92 $ ] , @xmath93 & = & -\frac{\kappa}{2 } \left\ { \hat{a}^{\dagger } \hat{a } \hat{\rho } + \hat{\rho } \hat{a}^{\dagger } \hat{a } - 2 \hat{a } \hat{\rho } \hat{a}^{\dagger } \right\ } \nonumber \\ & & -\frac{\gamma}{2 } \sum_{j=1}^n \left\ { \hat{\sigma}_{j}^{+ } \hat{\sigma}_{j}^{- } \hat{\rho } + \hat{\rho } \hat{\sigma}_{j}^{+ } \hat{\sigma}_{j}^{- } - 2 \hat{\sigma}_{j}^{- } \hat{\rho } \hat{\sigma}_{j}^{+ } \right\ } \nonumber \\ & & + \frac{1}{2t_2 } \sum_{j=1}^n \left\ { \ \hat{\sigma}_{j}^{z } \hat{\rho } \hat{\sigma}_{j}^{z } - \hat{\rho } \right\},\end{aligned}\ ] ] where @xmath3 is the decay rate of the cavity , @xmath94 is the spontaneous emission rate for the atoms , and @xmath95 is the inhomogeneous dephasing .
we derive @xmath25-number langevin equations corresponding to equation ( [ me1 ] ) . assuming that the classical drive @xmath79 is strong enough a mean - field description provides an accurate representation .
we therefore define the mean values @xmath96 which evolve according to @xmath97 in the moving frame of reference of each atom , the cavity field appears as a travelling wave , containing two frequencies shifted above and below the cavity frequency by the doppler shift . to solve this problem that intrinsically contains a bi - chromatic drive we proceed in two ways .
we first numerically integrate equations ( [ a1 ] ) - ( [ sz1 ] ) , approximating the sum in equation ( [ a1 ] ) as an integral , with the integrand weighted by the thermal velocity distribution .
the integral is then partitioned into finite segments .
the positions are chosen in an analogous manner , with the atoms being distributed in a gaussian distribution of width @xmath98 by the mot .
the velocity partition must be chosen with care , since the system exhibits doppleron resonances .
specifically , at lower velocity , more resolution in the partition is required .
we adopt a floquet analysis @xcite , in which we define for each atom , @xmath99 where @xmath100 represents the contribution from doppleron resonance of order @xmath30 , and @xmath29 is the number of photons absorbed from one direction and @xmath30 photons emitted into the opposite direction as stated in the article . upon substitution into equations ( [ a1 ] -[sz1 ] ) , equations for the amplitudes @xmath101,@xmath102 , and @xmath103 are found : as a second method , we consider the steady state solution where equation ( [ a1 ] ) becomes , @xmath108 where the sum over atoms has again been approximated as in integral , and @xmath109 is the thermal velocity distribution .
the @xmath106 that solves equation ( [ anm ] ) is found by applying newton s method for root finding .
we see excellent agreement between the two methods .
it is interesting to compare the situation observed in fig .
2(b ) and fig .
2(c ) in the article against lower temperature results . in fig .
[ fig : temp_dependence ] we plot theoretical transmission and phase curves for @xmath110 mk , @xmath111k and @xmath112k , where @xmath113 mk corresponds to the typical experimental situation .
the effect of going to lower temperature is evident on the transmission curves , where the dopplerons become less and less dominant .
the phase is less affected by dopplerons , although the slope on resonance tends to increase slightly with decreasing temperature .
the slope around zero - frequency determines the potential stability of this system in application as a frequency - lock .
however , considering the additional experimental complexity of decreasing the temperature by orders of magnitude and the possible reduction of atom number in the process , the gain in slope is minimal .
based on our experimental parameters we have estimated the shot noise limited locking performance of our setup .
we proceed as in @xcite to determine the shot noise limited linewidth with one difference : we do not assume the local oscillator power , _
i.e. _ the sum of the power in each sideband @xmath114 , is large compared to signal power @xmath115 .
@xmath116 is the carrier power leaking out of the cavity .
generally , when using the nice - ohms detection scheme , the local oscillator power may be comparable to the sideband power .
lifting the high local oscillator power assumption modifies the result by a factor of @xmath117 , resulting in @xmath118 our calculation assumes unit quantum detector efficiency , and additional quantum noise due to fluctuations in intra - cavity photon number is expected to be unimportant ( a @xmath119 effect where @xmath120 is the number of photons in the cavity ) . the intra - cavity power and the phase slope @xmath121
are calculated using the floquet simulation , and found to be in excellent agreement with experimental values as shown in the article .
for our present numbers , @xmath122 , and we estimate a shot noise limited laser linewidth of @xmath123 mhz .
this linewidth estimate could realistically be improved by at least an order of 10 - 20 by improving the experimental parameters , thus reaching state - of - the - art performance .
first of all , optimizing the eom modulation index to allow half of the optical power in the sidebands , such that @xmath124 , would immediately decrease the linewidth by a factor 15 .
secondly , the phase slope scales directly with atom number , and an increase in the number of atoms by a factor of 3 - 5 is realistic ( _ i.e. _ , a decrease of the linewidth by a factor 9 - 25 , see equation ) .
possible atom numbers exceeding even @xmath125 have been reported in literature @xcite by operating the strontium oven at somewhat higher temperatures compared to our setup and employing more efficient zeeman slower design in combination with strong transverse cooling section .
efficient transverse cooling alone may increase number of trapped atoms number by a factor of 4 .
additionally , increasing the finesse of the cavity a factor of 10 , from the current @xmath58 to @xmath126 , would reduce the laser linewidth by a factor of 100 .
however , to avoid undesirable power broadening effects the input power here must also be lowered .
this makes the experimental realization of this method less attractive as it adds more technical challenges on the detection system . a realistic linewidth reduction due to an increase in the finesse we estimate as a conservative factor of 10 .
from the floquet theory we may further investigate the role of the doppleron resonances on the shot noise limited linewidth .
this is done by decomposing the slope @xmath127 according to the doppleron order @xmath30 . in the floquet decomposition
the order @xmath128 corresponds to the @xmath129/2 order doppleron resonance @xcite , where @xmath130 . to investigate the importance of dopplerons
, we compare the slope of the phase around resonance with increasing number of doppleron orders included . to make the comparison
, we use the parameters of fig . 2 from the article , i.e. , an input power of 975 nw , a total number of atoms in the mot of @xmath33 and temperature @xmath34 mk .
the slope for varying orders of dopplerons is shown in fig .
[ fig : slopedopplerons ] , where we sum all contributions to the slope @xmath127 with orders up to @xmath30 .
we find that the solution is sufficiently converged by third order , i.e. , the sum of contributions from @xmath131 .
the difference between third order and sixth order solutions is less than a few percent .
the slope change is found to be @xmath132 at resonance for our conditions . as the shot noise limited linewidth scales as @xmath133 the predicted linewidth calculated with dopplerons will be larger than the linewidth calculated without dopplerons by a factor of 4 . on resonance ( zero detuning ) as a function of doppleron order @xmath30 .
here we sum all contributions to the slope with orders up to @xmath30 . for @xmath30 = 0 , only the zeroth order slope is included , while for @xmath30 = 1 both the 0 and the 1 order contribution to the slope are included and so forth .
[ fig : slopedopplerons],scaledwidth=45.0% ] | as an alternative to state - of - the - art laser frequency stabilisation using ultra - stable cavities , it has been proposed to exploit the non - linear effects from coupling of atoms with a narrow transition to an optical cavity .
here we have constructed such a system and observed non - linear phase shifts of a narrow optical line by strong coupling of a sample of strontium-88 atoms to an optical cavity .
the sample temperature of a few mk provides a domain where the doppler energy scale is several orders of magnitude larger than the narrow linewidth of the optical transition .
this makes the system sensitive to velocity dependent multi - photon scattering events ( dopplerons ) that affect the cavity field transmission and phase . by varying the number of atoms and the intra - cavity power we systematically study this non - linear phase signature which displays roughly the same features as for much lower temperature samples
this demonstration in a relatively simple system opens new possibilities for alternative routes to laser stabilization at the sub 100 mhz level and superradiant laser sources involving narrow line atoms .
the understanding of relevant motional effects obtained here has direct implications for other atomic clocks when used in relation with ultranarrow clock transitions .
state - of - the - art atomic clocks rely on highly coherent light sources to probe narrow optical transitions @xcite . however , these clocks are limited by the frequency noise of the interrogation oscillator through the dick effect @xcite .
only recently multi - atom optical clocks have surpassed single ion clocks in stability owing to the enhanced laser stability @xcite .
achieving a better stability has so far been hampered by thermal noise in the reference cavity used for laser stabilisation @xcite .
recent proposals suggest an alternative approach to laser stabilisation @xcite where atoms in an optical lattice are probed on the narrow clock transition inside an optical cavity .
this brings non - linear effects into the system dynamics that could considerably enhance the spectral sensitivity and potentially lead to laser stability comparable to or better than the current state - of - the - art .
however , for finite temperature samples of atoms the principal mechanisms that are relevant to this physical domain have not been investigated in detail . in such systems with highly non - linear phase response ,
_ a priori _ unpredictable effects such as bi - stability @xcite and the finite temperature of the atomic ensemble can change the phase response in an undesirable way , which could reduce the performance of the stabilisation scheme for all practical implementations . to achieve a better understanding of cavity - mediated effects with a narrow optical transition
we have constructed a system with @xmath0sr atoms probed on the @xmath1transition at 689 nm inside an optical cavity ( see fig .
[ fig : cavity ] ) . to capture the basic physics of the strong non - linear phenomena one can consider @xmath2 atomic dipoles strongly coupled to a single mode of the cavity field .
the dipole moment associated with this narrow transition is around five orders of magnitude smaller than that for a typical dipole - allowed transition in an alkaline element .
also , at finite temperature only a small fraction of the atomic sample is probed due to doppler broadening . here
, the role of the cavity is to enhance the weak interaction by order of the finesse of the cavity .
experimentally we operate in the so - called `` bad cavity '' regime , where the atomic dipole decay rate is a factor of 1000 smaller than the cavity decay rate @xmath3 . in our experiment
we use the @xmath0sr @xmath4 transition at @xmath5 nm to cool and trap atoms in a magneto - optical trap ( mot ) .
we load about @xmath6 atoms in the mot at a temperature of @xmath7 mk inside an optical cavity prepared for light at 689 nm .
the cavity waist of @xmath8 m ensures a good overlap with the mot and negligible transit time broadening ( @xmath9 khz ) compared to the natural line width ( @xmath10 khz ) of the probe transition . the dimensionless number @xmath11 , where @xmath12 depends on the single - atom / cavity coupling constant @xmath13 , is known as the collective cooperativity and is a measure of how strong the coherent atom - cavity coupling is with respect to the dissipation channels . in our configuration ( @xmath14 hz , @xmath15 mhz ) we are able to generate a collective cooperativity of about @xmath16 , thus placing our system in the regime of high collective cooperativity in the bad cavity limit , but outside the more restrictive cqed strongly coupled regime @xcite
. our experiment is operated in a cyclic fashion .
we start each cycle preparing the atomic sample by loading a mot inside the optical cavity . after loading we shut off the mot beams and probe the atoms at 689 nm while recording both the intensity and phase shift of the transmitted probe light via two detectors ( see fig . [
fig : cavity ] ) .
the total cycle time is typically around @xmath17 s. for the phase measurement we employ cavity - enhanced fm spectroscopy by using the so - called noise - immune cavity - enhanced optical - heterodyne molecular spectroscopy ( nice - ohms ) technique @xcite ( see supplemental information ) .
this technique has a clear advantage over heterodyne signals generated , for example , from interferometric methods in terms of superior noise reduction and simplicity . during experiments we lock the cavity resonance to the 689 nm laser frequency using a hnsch - couillaud scheme @xcite .
the standing wave generated in the cavity will thus be present at all times while the 689 nm laser frequency is scanned . in the limit of @xmath18 and for very low cavity field
intensities several solutions exist for the steady - state intra - cavity field @xcite .
this is known as optical bi - stability , which would render the system unsuited for frequency stabilization . however , at finite temperatures when motional effects are included this picture changes . in this case
, there is a critical temperature @xmath19 above which only one solution for the steady - state intra - cavity field exists . for our parameters
@xmath19 is of the order of a few hundred nk while experiments are typically performed at mk temperatures .
the non - zero velocity of the atoms brings additional photon resonance phenomena into play , which changes the complex amplitude of the cavity field around the atomic resonance @xmath20 . in the rest frame of an atom moving with velocity @xmath21 the atom experiences a bi - chromatic light field given by @xmath22 and @xmath23 , where @xmath24 is the laser frequency and @xmath25 is the speed of light .
resonant scattering events will take place if the atom is doppler - tuned into resonance at @xmath20 , e.g. , @xmath26 , such that the atom may absorb a photon from a given direction of the cavity field .
higher order resonances are also possible where , e.g. , the atom absorbs two photons from one direction at @xmath27 and emits one photon in the other direction at @xmath28 .
generally , the resonance condition for @xmath29 absorbed and @xmath30 emitted photons becomes @xmath31 for @xmath32 @xcite .
the process is illustrated in fig .
[ fig : power transmitted](a ) .
these non - linear multi - photon scattering effects are known as dopplerons and give rise to a series of velocity dependent resonances @xcite , which change the transmitted field amplitude around resonance . in the direction of the cavity axis .
the first resonance condition ( top equation ) involves only one photon and corresponds to the usual doppler effect .
the next resonance involves two photons absorbed and one photon emitted , and so forth .
( b , c ) typical frequency scan without any averaging across the atomic resonance for an input power of 975 nw and a total number of atoms in the mot of @xmath33 . the data in ( b ) displays the transmission of the probe light through the cavity normalized to a signal with no atoms in the cavity .
the data in ( c ) is the phase shift of the cavity - transmitted field obtained using the nice - ohms method .
the solid lines are theoretical predictions based on our theoretical model which includes the doppler effect and the spatial overlap of the thermal cloud ( here with temperature @xmath34 mk ) with the cavity field . at maximum phase shift ( around detunings of @xmath35 mhz )
our detection system starts to saturate giving a slightly flatter appearance of the phase data .
inset : zoom on central phase feature with similar experimental parameters ( data identical to fig .
5(a ) ) . here
, we have included a theoretical plot that does not take the dopplerons into account ( black , dashed curve ) .
the effect of the dopplerons is readily apparent .
units on axes are the same as in ( c ) .
[ fig : power transmitted ] ] in fig .
[ fig : power transmitted](b , c ) we show typical results for a frequency scan across the @xmath1line resonance ( red circles ) .
the input power was @xmath36 nw corresponding to an average saturation parameter of @xmath37 .
it is clear that the phase signal in fig .
[ fig : power transmitted](c ) has a significantly higher signal - to - noise ratio ( snr = 70 ) than the transmitted power signal in fig .
[ fig : power transmitted](b ) ( snr @xmath38 ) , demonstrating the effectiveness of the nice - ohms technique . currently , the factor limiting the signal - to - noise ratio of the phase signal is the shot - to - shot atom number fluctuations and residual amplitude modulation from the eom .
we model the dynamics of the system by a hamiltonian describing the coherent time evolution of an ensemble of atoms , where each atom with a given velocity is coupled to a single mode of the optical cavity . solving the corresponding optical bloch equations
yields the cavity - transmitted intensity and phase as a function of detuning , number of atoms , and temperature .
our model is also adapted to take into account the spatial extent of the cavity field and atomic density profile .
the blue solid curves in fig . [
fig : power transmitted ] ( b , c ) are the theoretical prediction based on the hamiltonian presented in supplemental equation 2 @xcite . in our theoretical model
we fix the number of atoms , laser input power , laser line width , cavity waist , and cavity finesse based on experimental values , but allow a scaling factor for the absolute phase .
the temperature is allowed to vary in the range of 2 - 4 mk , in accordance with the experimental condition .
considering the transmission in fig .
[ fig : power transmitted](b ) we can identify three spectral features : ( 1 ) the broad ( @xmath39 mhz wide ) doppler absorption feature consistent with the sample temperature of a few mk ; ( 2 ) a central region ( @xmath40 mhz wide ) with enhanced transmission due to saturation , affected by the doppleron resonances which lead to enhanced back - scattering ( or reduced forward transmission ) , limiting the height of the saturated absorption peak ;
( 3 ) finally , in the central region around zero velocity ( i.e. , on resonance ) , the doppleron mechanism breaks down and the saturated absorption takes place again with increased transmission as a result .
the dopplerons also have an effect in the phase signal ( fig .
[ fig : power transmitted](c ) ) , although the effect is negligible for large laser detunings corresponding to larger atom velocities . in the inset of fig .
[ fig : power transmitted](c ) we zoom in on the phase of the central saturated absorption feature were we have plotted experimental data ( with parameters corresponding to fig . [
fig : number dependence](a ) ) and theoretical curves without dopplerons ( black , dashed ) and with dopplerons ( blue ) . here
, the effect of dopplerons becomes clear and there is an observable effect on the phase signature which is a decrease in slope around resonance , showing consistency between our theoretical model and the experimental data .
this decrease in slope is important in the determination of the frequency stability that is achievable using this system , since the stability depends inversely on square of the slope around resonance , and reducing the temperature further does not significantly improve this slope @xcite .
is progressively decreased from 1950 nw ( a ) to 650 nw ( d ) .
the number of atoms is about @xmath41 .
each point is an average of 3 data points .
the solid lines are theoretical predictions based on our theoretical model .
[ fig : power dependence],scaledwidth=50.0% ] to evaluate and characterize our physical system experimentally and test it against the theoretical model we have mapped out the central phase feature as a function of probe input power with fixed atom number . in addition to a validation of the theoretical model this will provide an understanding of the behaviour and sensitivity of the phase signal to typical experimental variables relevant to , _
e.g. _ , laser stabilisation . in fig .
[ fig : power dependence ] we show the phase signal for a fixed number of atoms as a function of laser detuning for different input powers in the range 650 - 1950 nw . for high input powers we strongly saturate the dipole and power
broaden the central saturated absorption peak .
as we gradually lower the input power , the power broadening is reduced leaving the central phase feature with a larger slope without reducing the signal - to - noise ratio .
( a ) to @xmath42 ( d ) .
the input power used for all plots was 650 nw . each point is an average of 3 data points .
the solid lines are theoretical predictions based on our theoretical model .
the central slope scales linearly with atom number .
[ fig : number dependence],scaledwidth=50.0% ] fig .
[ fig : number dependence ] ( a ) - ( d ) shows the evolution of the phase signal for fixed probe power as the number of atoms inside the cavity mode is changed from @xmath43 in ( a ) to @xmath44 in ( d ) .
we observe a strong dependence on atom number with increasing phase response and increasing slope on resonance for increasing atom numbers as expected , and the slope can straight forwardly be improved by increasing the number of atoms .
however , our system is strongly non - linear and other optimal parameters , such as input power , for a given number of atoms , may not be trivially assigned to our experiment , but must be found numerically or experimentally . using the central phase slope for laser frequency locking
, we estimate a shot noise limited line width of @xmath45mhz based on our experimental parameters . this number can be improved by at least a factor of 20 with realistic improvements of the experimental parameters , _
e.g. _ by optimizing the eom modulation index ( a factor 15 ) , and increasing the atom number and the cavity finesse ( both a factor 10 ) , which would render the system comparable to state - of - the - art frequency stabilisation references @xcite ( see supplemental information for details ) . in conclusion ,
we have constructed a system dominated by highly saturated multi - photon absorption with laser - cooled strontium atoms coupled to a low - finesse optical cavity .
the transmission through the cavity is altered by thermal effects , but apart from a small decrease in slope , the central phase response of the atoms remains relatively immune to these effects while displaying a high signal - to - noise ratio ( snr ) owing to the cavity and detection technique .
the atomic phase signature was observed via cavity enhanced fm spectroscopy ( nice - ohms ) on the narrow optical @xmath1inter - combination line of @xmath0sr providing snr of exceeding 7000 for one second of integration .
the understanding obtained here of the `` bad cavity '' physics lends promise to further development in this area , such as a new generation of frequency stabilisation @xcite or superradiant laser sources @xcite .
specifically , the physical understanding of a `` warm '' system ( mot temperature ) obtained in his work will prove valuable when future atomic clocks and/or stable lasers will be operated under more noisy and compact environments , e.g. in vehicles and space crafts , where the size , ruggedness and convenience of the setup might dictate higher atomic temperatures than what is currently used for state - of - the - art systems .
in this situation , this work will serve as an important piece of technical understanding for out - of - lab clocks employing warm atoms .
we would like to acknowledge support from the danish research council and esa contract no .
4000108303/13/nl / pa - npi272 - 2012 .
dt , mh , and jy also wish to thank the darpa quasar program , the nist and the nsf for financial support .
1 b. j. bloom _
et al _ , _ nature _ * 506 * , 71 - 75 ( 2014 ) .
n. hinkley _
et al _ , _ science _ * 341 * , 1215 - 1218 ( 2013 ) .
r. le targat _ et al _ , _ nat .
commun . _ * 4 * , 2109 ( 2013 ) .
m. takamoto , f .-
l .
hong , r. higashi , and h. katori , _ nature _ * 435 * , 321 - 324 ( 2005 ) .
c. w. chou , d. b. hume , j. c. j. koelemeij , d. j. wineland , and t. rosenband , _ phys .
rev .
lett . _ * 104 * , 070802 ( 2010 ) .
g. santarelli , c. audoin , a. makdissi , p. laurent , g. j. dick , and a. clairon , _ ieee trans .
ultrason .
ferroelectr .
freq .
control _ * 45 * , pp .
887 - 894 ( 1998 ) .
t. l. nicholson , m. j. martin , j. r. williams , b. j. bloom , m. bishof , m. d. swallows , s. l. campbell , and j. ye , _ phys .
rev .
lett . _ * 109 * , 230801 ( 2012 ) .
t. kessler , t. legero , and u. sterr , _ j. opt
. soc . am .
b _ * 29 * , 178 - 184 ( 2012 ) .
t. kessler _
et al _ , _ nature photonics _ * 6 * , 687 - 692 ( 2012 ) .
m. j. martin _
et al _ , _ science _ * 341 * , 632 - 636 ( 2013 ) .
d. meiser , j. ye , d. r. carlson , and m. j. holland , _ phys .
rev .
lett . _ * 102 * , 163601 ( 2009 ) .
d. meiser and m. j. holland , _ phys .
rev .
a. _ * 81 * , 033847 ( 2010 ) .
m. j. martin , d. meiser , j. w. thomsen , j. ye , and m. j. holland , _ phys .
rev . a _ * 84 * , 063813 ( 2011 ) . h. tanji - suzuki , ian d. leroux , m. h. schleier - smith , m. cetina , a. grier , j. simon , and v. vuletic . , _ adv . at .
mol .
opt .
phys . _ * 60 * , 201 ( 2011 ) .
see supplemental material [ url ] , which includes refs .
@xcite .
j. ye , l .- s .
ma , and j. l. hall , _ j. opt .
soc . am .
b _ * 15 * , 6 - 15 ( 1998 ) .
l .- s .
ma , j. ye , p. dub , and j. l. hall , _ j. opt . soc .
am .
b _ * 16 * , 2255 - 2268 ( 1999 )
. t. w. hnsch and b. couillaud , _ opt .
commun .
_ * 35 * , 441 - 444 ( 1980 ) .
a. tallet , _ j. opt .
soc . am .
b _ * 11 * , 1336 - 1349 ( 1994 ) .
e. kyrl and s. stenholm , _ opt .
commun . _
* 22 * , 123 ( 1977 ) .
m. bishof , x. zhang , m. j. martin , and jun ye , _ phys .
rev .
lett . _ * 111 * , 093604 ( 2013 ) .
y. y. jiang , a. d. ludlow , n. d. lemke , r. w. fox , j. a. sherman , l .- s .
ma , and c. w. oates , _ nature photonics _ , * 5 * , 158161 ( 2011 ) .
j. millo , et al .
, _ phys .
rev . a _ * 79 * , 053829 ( 2009 )
. j. g. bohnet , z. chen , j. m. weiner , d. meiser , m. j. holland , and j. k. thompson , _ nature _ * 484 * , 78 - 81 ( 2012 ) .
t. maier , s. kraemer , l. ostermann , and h. ritsch , _ opt .
express _ * 22 * , 13269 ( 2014 ) .
r. w. p. drever _
et al _ , _ appl .
phys .
b _ * 31 * , issue 2 , 97 - 105 ( 1983 ) .
z. ficek and h.s .
freedhoff , _ phys .
rev . a _ * 48 * , 3092 - 3104 ( 1993 )
. s. stellmer , r. grimm , and f. schreck , _ phys .
rev .
a _ , * 87 * , 013611 ( 2013 ) . |
supersymmetry is expected to be one of the key ingredients to describe physics beyond the standard model ( sm ) .
while tree - level supersymmetry breaking within the sm sector leads to light sfermions , the breaking sector is separated from the sm one and is mediated by some effective operators or quantum effects . among various mechanisms for realizing this scenario ,
the gauge mediation , relevant to this paper , is one of the most promising candidates with a strong prophetic power ( for a review , see @xcite ) .
it is known that pseudomoduli directions are present in the supersymmetry - breaking vacuum of oraifeartaigh - like models with the canonical khler potential @xcite .
an important implication of this result is that , if such models are used as the hidden sector of gauge mediation , gaugino masses are generally suppressed or the vacuum is unstable somewhere along the pseudomoduli @xcite .
there have been various ways in the literature to avoid such a phenomenologically unfavorable situation , such as including nonminimal terms in the potential @xcite , quantum effects from specific scalar and/or vector multiplets @xcite , or accepting metastable vacua @xcite .
another way , as we discussed before , is to introduce gauge multiplets and take non - negligible d term into account for making the vacuum stable without pseudomoduli . in our previous paper
, we classified supersymmetry - breaking models with nonvanishing f and d terms @xcite .
first , the models are divided into two categories based on whether the f - term potential has a supersymmetric minimum ( at finite field configuration ) .
we then add the d term by gauging flavor symmetry and analyze the vacuum of the full scalar potential in each category . for models that do not satisfy the f - flatness conditions , we found that the full potential generally shows runaway behavior . on the other hand , when the f - flatness conditions are satisfied , supersymmetry can be broken without pseudomoduli only in the presence of the fayet - iliopoulos ( fi ) term . by using the latter class of models
, we constructed a model of gauge mediation where gaugino masses are generated at the one - loop order . in this paper
, we discuss another possibility for the classification : the f - term potential is minimized at some infinite field configuration , i.e. , it shows a runaway behavior .
it is found that the runaway direction of the f - term potential can be uplifted by the d term , and a supersymmetry - breaking vacuum emerges at finite field configuration .
there are several reasons to explore this class of models in detail .
first of all , contrary to our previous result , there is no need to add the fi term for supersymmetry breaking .
secondly , the vacuum automatically suppresses pseudomoduli directions associated with the f - term potential , since it has a runaway behavior and is stabilized by the d - term potential .
we propose a minimal model with such properties and couple it with an appropriate messenger sector . in this model
including the messenger sector , r symmetry is spontaneously broken at the tree level even though the model contains chiral superfields with u(1)@xmath0 charge 0 or 2 only .
we notice that r symmetry breaking does not occur for oraifeartaigh - like models with such a u(1)@xmath0 charge assignment @xcite .
this class of supersymmetry breaking can therefore provide a realistic model for gauge mediation , where leading - order gaugino masses are obtained at the stable vacuum .
the outline of this paper is as follows . in sec .
[ sec : review ] , we briefly review our classification of f- and d - term supersymmetry breaking . in sec .
[ sec : runaway ] , we discuss the case in which the f - term potential shows runaway behaviors .
after some general arguments , a minimal model is presented to realize the vacuum property listed above .
further , appropriate messenger sectors are discussed and shown to be viable for generating gaugino masses .
section [ sec : conclusion ] is devoted to summarizing our results and discussions on future directions . in the appendix
, we show the potential analysis of the model given in sec .
[ sec : runaway ] .
we first review our previous result of the classification of supersymmetry breaking with both f and ( abelian ) d terms @xcite . throughout this paper
, we assume the khler potential is canonical .
the superpotential @xmath1 has a polynomial form of chiral superfields @xmath2 with u(1 ) charges @xmath3 ( the latin indices label their species ) .
the scalar potential @xmath4 is then given by @xmath5 where @xmath6 and @xmath7 are the contributions from f and d terms : @xmath8 here @xmath9 is the u(1 ) gauge coupling constant and @xmath10 is the coefficient of the possible fayet - iliopoulos term @xcite . in the following , we abbreviate field derivatives of the superpotential as @xmath11 ( @xmath12 ) .
we first divide models into two categories .
the criterion for the classification is whether the f - flatness condition , @xmath13 , is satisfied or not at its minimum defined by @xmath14 .
if the f - flatness condition is ( not ) satisfied , a model is called in the second ( first ) class .
we then add the u(1 ) d term and analyze the tree - level behavior of the full scalar potential .
for the first class of models , the scalar potential shows supersymmetric runaway behaviors when the d - term contribution is included .
for the second class of models , on the other hand , supersymmetry can be broken without any pseudomoduli .
that is , however , realized only in the presence of the fi term .
( see table [ table : class ] for the classification . )
.classification of supersymmetry breaking with f and d terms . [ cols="<,^,^,^",options="header " , ]
+ the superpotential is @xmath15 this form can be the most generic , renormalizable one if @xmath16 and @xmath17 have the charges @xmath18 and @xmath19 under an additional @xmath20 symmetry ( @xmath21 ) .
we here comment on the role of each term in the superpotential : the first and second terms make the origin unstable along the meson @xmath22 .
the third term lifts up the @xmath23 direction , and naively supersymmetry is broken with the f - term potential coming from the first three terms .
however , the potential minimum is found to run away to infinity of the moduli space along the @xmath24 direction ( with a finite value of @xmath22 ) .
it is further noticed that anomaly cancellation requires the existence of a negatively charged field ( @xmath17 ) . without the last fourth term ,
the value of @xmath17 is free and the direction @xmath25 becomes d flat , along which the potential minimum goes to infinity and supersymmetry is recovered . in this way , the superpotential is regarded as minimal one for the present purpose . the scalar potential @xmath26 is explicitly given by @xmath27 & & v_d \,=\ , \frac{g^2}{2}d^2 \,=\ , \frac{g^2}{2}\big(|x_+|^2+|\varphi_+|^2-|x_-|^2-|\varphi_-|^2\big)^2 \ , .
\label{dterm_pot}\end{aligned}\ ] ] the f - term potential @xmath6 has the u(1 ) runaway directions discussed in the previous section .
we find along the following direction @xmath28 and all the f terms vanish except for the positively charged field @xmath29 , @xmath30 the runaway direction is parametrized by @xmath31 .
as @xmath32 , the positively charged field @xmath24 goes to infinity and the f - term potential ( @xmath29 ) approaches to zero .
furthermore , no u(1)@xmath0 runaway is expected since all the scalar fields in the present model have r charges 0 or 2 @xcite .
it is easily seen that the f - term runaway direction is stabilized by the d - term contribution because , along this direction , the d term increases as @xmath33 ( see also fig .
[ fig : uplift ] . )
we again note that even when the runaway is lifted by the d term , it does not necessarily mean that the vacuum of the scalar potential is in the direction .
( dashed line ) and @xmath4 ( solid line ) along the runaway direction . the runaway direction of @xmath6
is uplifted by the d - term contribution @xmath7.,width=245 ] we then analyze the scalar potential in detail to confirm that a stable supersymmetry - breaking vacuum is obtained in some parameter region .
the parameters appearing in the superpotential are assumed to be real and positive without loss of generality .
the vacuum is identified by solving the stationary conditions for the full scalar potential @xmath4 .
a trivial configuration satisfying the stationary conditions is the origin at which all the scalar fields vanish .
this point however is unstable and we do not consider it in the following . by some calculation
( the detail is summarized in the appendix ) , we can show that the vacuum satisfies @xmath34 where several f terms vanish ; @xmath35 then the stationary conditions for @xmath36 , @xmath37 , and @xmath16 automatically hold , and the scalar potential simplifies to @xmath38 with @xmath39 . with this reduced potential ,
the stationary condition for @xmath17 reads @xmath40 indicating @xmath41 or @xmath42 .
we discuss these two cases separately below , solving the remaining stationary conditions for @xmath43 . : in this case , for a large mass ( @xmath44 ) or a small one ( @xmath45 ) , we can approximately write down the analytic solution to the stationary conditions for @xmath43 . for the large mass regime ,
the solution is given by @xmath46 where the f and d components other than become @xmath47 at the leading order .
therefore , the scalar potential is dominated by @xmath48 , i.e. , @xmath49 . on the other hand , for the small mass regime , we have @xmath50 where the f and d components are @xmath51 and the scalar potential is found to be dominated by @xmath29 , i.e. , @xmath52 .
notice that , in both regimes of @xmath53 , the r symmetry is unbroken as @xmath54 .
the stability of these vacua is read off from the eigenvalues of the squared mass matrix for the scalar fields .
we find that all eigenvalues except for @xmath17 are positive semidefinite . along the @xmath17 direction ,
the eigenvalue is given by @xmath55 this eigenvalue is positive when @xmath56 for the large mass regime and @xmath57 for the small mass regime .
therefore the supersymmetry - breaking vacuum realized is stable for these parameters .
otherwise , @xmath58 is a saddle point and the true vacuum is given by the following second case : : in this case , the stationary conditions are complicated and we only present a typical numerical solution .
for example , when the model parameters are set as @xmath59 , which do not satisfy the above condition for the stability of @xmath58 , the vacuum is located at @xmath60 where nonvanishing f and d components are @xmath61 all the f and d terms become comparable to each other and contribute to the scalar potential @xmath62 .
the stability of this vacuum is confirmed numerically .
we have also checked that , for a wider parameter region , the scalar potential has a similar behavior . in fig .
[ fig : vac - height ] , we show the normalized potential @xmath63 characterizing supersymmetry breaking at the vacuum as a function of @xmath64 for the large mass regime ( @xmath65 . without the @xmath64 term , supersymmetry recovers at infinity of the moduli space , as we mentioned before .
( for the @xmath66 case ) .
the other parameters are fixed to @xmath67 . in the limit @xmath68 ,
supersymmetry is recovered , but some expectation values run away to infinity.,width=245 ] as @xmath64 becomes larger , @xmath17 is stabilized for @xmath56 and the vacuum is shifted up to , where @xmath69 .
we then discuss the gauge - mediation scenario by employing the above model as a supersymmetry breaking sector .
appropriate messenger fields and their superpotential are identified for generating one - loop - order gaugino masses for the two cases , @xmath58 and @xmath66 , separately . :
as we have shown , the f and d terms have different behaviors depending of whether the mass parameter @xmath53 is large or small . for the large @xmath53 case , ,
the supersymmetry - breaking scale is governed by @xmath48 from which sfermions are expected to receive soft masses . for generating a similar size of gaugino masses ,
a simple way is to introduce the messenger fields @xmath70 and @xmath71 which are vectorlike under the sm gauge symmetry and have the superpotential , @xmath72 for @xmath73 , the standard one - loop diagram of messenger fields generates a gaugino mass @xmath74 ; @xmath75 where @xmath76 is the sm gauge coupling , and @xmath77 is the dynkin index of the sm gauge symmetry for @xmath70 and @xmath71 .
the gaugino mass is given at the vacuum discussed in the previous section .
it is , however , noticed that the r symmetry is softly broken by the parameter @xmath78 that makes @xmath79 a local minimum .
another way of obtaining gaugino masses would be to consider the direct gauge mediation , i.e. , to generalize u(1 ) to non - abelian symmetry containing the sm one . by adding a small supersymmetric mass for @xmath43 , they behave as messengers and
would induce the sm gaugino masses without introducing additional multiplets . for the small mass case ,
, the charged f term @xmath29 dominates the supersymmetry - breaking scale . to split messenger masses with @xmath29
, we must introduce two pairs of vectorlike messengers with u(1 ) charges @xmath80 and @xmath81 and couple them with @xmath36 in the superpotential . for further details of the messenger sector , the stability of the vacuum , and gaugino mass generation , see ref .
@xcite . : in the vacuum , the u(1)@xmath0 symmetry given in table [ table : assign ] is broken by the f term of @xmath16 which is neutral under both u(1 ) and u(1)@xmath0 .
we consider the following messenger superpotential @xmath82 where @xmath70 and @xmath71 are vectorlike multiplets , charged under the sm gauge symmetry .
it is noticed that the @xmath20 symmetry is softly broken in this messenger sector .
note that the superpotential ( [ example_superpotential ] ) has an anomalous u(1) symmetry , under which @xmath16 and @xmath17 are charged , but it is explicitly broken in ( [ messenger ] ) .
the r symmetry is spontaneously broken in the full potential with both ( [ example_superpotential ] ) and ( [ messenger ] ) , if the vacuum in the previous section is stable .
the vacuum stability is ensured by taking the parameter @xmath78 sufficiently large .
the gaugino mass is evaluated for @xmath83 as @xmath84 which comes from a one - loop diagram , where @xmath70 , @xmath71 circulate in the loop . from these observations , we conclude that this class of supersymmetry - breaking models is useful to build a simple , realistic hidden sector of gauge mediation .
we have studied supersymmetry - breaking models with both f and d terms being nonvanishing . in particular ,
we have focused on the case that the f - term potential shows runaway behaviors that originate from symmetries of theory considered .
the runaway directions are uplifted by the d term and a supersymmetry - breaking vacuum is realized at finite field configuration .
an interesting property of this approach is that ( phenomenologically disfavored ) pseudomoduli are absent in the vacuum since they are related to the minimization of the f - term potential .
moreover , there is no need to add the fi term for supersymmetry breaking . along this line
, a minimal renormalizable model has been presented where supersymmetry is broken . for an application to gauge mediation ,
we have introduced appropriate messenger sectors and confirmed that r symmetry is spontaneously broken , and gaugino masses in the visible sector are generated at the comparable order of sfermion masses .
this class of models may open up a new way to build realistic models of gauge mediation , circumventing the lemma proved by komargodski and shih @xcite .
as remarked , the d - term lifted runaway might be destabilized along other orthogonal directions such as the u(1)@xmath0 runaway .
we do not have any criteria to ensure that the d - term uplifting of runaway directions can lead to the stable and global minimum of the scalar potential .
it would be interesting if one could carry out a general argument on this issue .
we would like to thank l.b .
anderson for useful advice on the numerical analysis .
thanks the yitp workshop on string theory and field theory and the 2012 simons workshop on mathematics and physics , where this work was partly done , for hospitality .
is in part supported by the jsps postdoctoral fellowship for research abroad and is grateful to the center for the fundamental laws of nature at harvard university for support
. k.y . is supported in part by the grant - in - aid for scientific research no .
23740187 and also in part by keio gijuku academic development funds .
this work is supported by
the next generation of physics , spun from universality and emergence , " the gcoe program from the ministry of education , culture , sports , science and technology of japan .
in this appendix , we give some details of the potential analysis for the model given in sec . [
subsec : example ] .
in particular , we explain the derivation of the vacuum expectation values .
the field derivatives of the superpotential are given by @xmath85 w_{\varphi_+ } = \lambda x_0\varphi_- + \lambda'x_-\varphi_0 \ , , \qquad w_{\varphi_- } = \lambda x_0\varphi_+ + mx_+ \ , , \qquad w_{\varphi_0 } = \lambda'x_-\varphi_+ \,.\end{gathered}\ ] ] the scalar potential @xmath4 is the sum of the contributions from f and d terms , @xmath6 and @xmath7 , and these explicit forms are written down in and .
the stationary conditions for @xmath4 are then given by @xmath86 first , by using and , we express @xmath37 and @xmath16 in terms of the other fields : @xmath87 \varphi_0 & \,=\ ; \frac{m}{\lambda'}\ , \frac{x_+x_-^*\varphi_-}{\varphi_+\big(|x_-|^2 + |\varphi_+|^2+|\varphi_-|^2\big ) } \ , .
\label{ph02}\end{aligned}\ ] ] if @xmath88 , we find from and , @xmath89 this implies that @xmath90 is negative at the stationary point ( if @xmath88 ) . on the other hand , from and , we have @xmath91 while the equation @xmath92 gives @xmath93 & & \qquad -\varphi_-\lambda x_0 ( \lambda x_0\varphi_- + \lambda'x_-\varphi_0)^ * + \lambda x_0(\lambda x_0\varphi_+ + mx_+)^ * \,=\ , 0 \ , .
\label{ph+ph-}\end{aligned}\ ] ] by using the relations , @xmath94 following from and , we find another expression for @xmath90 from and : @xmath95 \ , .
\label{dpos}\ ] ] that implies @xmath90 is positive at the stationary point . for both conditions and to be true , @xmath96 is the only solution , but it can not be satisfied because the origin of meson direction @xmath22 is destabilized by the f - term potential .
therefore @xmath88 , which is the assumption for , should not be realized . in the end , from and ,
we find the following vacuum expectation values : @xmath97 l. g. aldrovandi and d. marques , jhep * 0805 * ( 2008 ) 022 [ arxiv:0803.4163 [ hep - th ] ] ; y. nakai and y. ookouchi , jhep * 1101 * ( 2011 ) 093 [ arxiv:1010.5540 [ hep - th ] ] ; t. s. ray , phys . rev .
d * 85 * ( 2012 ) 035003 [ arxiv:1111.4266 [ hep - ph ] ] .
m. dine and j. mason , phys .
d * 77 * ( 2008 ) 016005 [ hep - ph/0611312 ] ; k. intriligator and m. sudano , jhep * 1006 * ( 2010 ) 047 [ arxiv:1001.5443 [ hep - ph ] ] ; e. dudas , s. lavignac and j. parmentier , phys .
b * 698 * ( 2011 ) 162 [ arxiv:1011.4001 [ hep - th ] ] .
r. kitano , h. ooguri and y. ookouchi , phys .
d * 75 * ( 2007 ) 045022 [ arxiv : hep - ph/0612139 ] ; b. k. zur , l. mazzucato and y. oz , jhep * 0810 * ( 2008 ) 099 [ arxiv:0807.4543 [ hep - ph ] ] ; a. giveon , a. katz and z. komargodski , jhep * 0907 * ( 2009 ) 099 [ arxiv:0905.3387 [ hep - th ] ] ; s. a. abel , j. jaeckel and v. v. khoze , phys .
b * 682 * ( 2010 ) 441 [ arxiv:0907.0658 [ hep - ph ] ] ; m. bertolini , l. di pietro and f. porri , jhep * 1201 * , 158 ( 2012 ) [ arxiv:1111.2307 [ hep - th ] ] .
t. azeyanagi , t. kobayashi , a. ogasahara and k. yoshioka , jhep * 1109 * ( 2011 ) 112 [ arxiv:1106.2956 [ hep - ph ] ] . | we study the d - term effect on runaway directions of the f - term scalar potential .
a minimal renormalizable model is presented where supersymmetry is broken without any pseudomoduli .
the model is applied to the hidden sector of gauge mediation for spontaneously breaking r symmetry and generating nonvanishing gaugino masses at the one - loop order . |
it has recently become possible to study the emission regions of a handful of black holes on sub - horizon scales with millimeter wave very long baseline interferometry ( mm - vlbi ) .
already , this technique has produced horizon - scale information on sagittarius a * ( sgr a * ) , the @xmath6 black hole located at the center of the milky way @xcite , using the _
arizona radio observatory sub - millimeter telescope _ ( smt ) on mount graham , arizona , _ james clerk maxwell telescope _ ( jcmt ) and _ sub - millimeter array _ ( sma ) atop mauna kea in hawaii , and the _ combined array for research in millimeter - wave astronomy _ ( carma ) in cedar flat , california @xcite . due to the limited signal - to - noise of these early experiments ,
they have produced primarily visibility amplitudes , related to the modulus of the fourier transform of the intensity distribution of the source .
nevertheless , when analyzed in the context of physically motivated accretion flow models , they have resulted in dramatic constraints upon the spin orientation and magnitude @xcite .
however , the absence of phase information introduces fundamental degeneracies in the orientation of the modeled image , and systematic uncertainties in the image structure generally .
the importance of visibility phase information has been appreciated since the beginning of radio interferometry . in the context of vlbi
, it has only been widely possible since the introduction of `` closure phases '' , and the associated development of self - calibration techniques ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the closure phases , which are equivalent to the argument of the bispectrum , are combinations of the visibility phases on baseline triangles ( and thus vlbi station triples ) that are insensitive to individual station - based phase errors , which otherwise typically dominate the phase uncertainties , e.g. , due to pathlength variations from atmospheric turbulence .
antennas there are only @xmath7 independent closure phases , and thus @xmath8 additional phases must be supplied to produce the full compliment of phase information .
frequently , these are obtained via self - calibration techniques , a combination of non - linear algorithms in which the unknown phases are chosen such that the resulting image satisfies various physical constraints ( see , e.g. , * ? ? ? * for a detailed summary of self - calibration techniques ) . ] specifically , if @xmath9 is the visibility phase on the baseline between stations @xmath10 and @xmath11 , the closure phase associated with three stations is @xmath12 even a handful of closure phases are diagnostic of the underlying image structure , e.g. , the closure phases of a point source or gaussian flux distribution ( including asymmetric cases ) are identically @xmath13 , while that of an annulus may be either @xmath13 or @xmath3 , depending upon the particular baselines considered .
more general flux distributions produce non - vanishing closure phases , indicative of the symmetry of the image .
recently , @xcite reported the measurement of a closure phase at @xmath14 for sgr a * of @xmath0 on the smt - jcmt - carma triangle .
note that this represents a new constraint upon models for the structure of the emitting region surrounding sgr a * , independent of those associated with the visibility amplitudes alone .
the _ event horizon telescope _ ( eht ) is a project underway that will extend current mm - vlbi arrays to shorter wavelengths ( @xmath15 ) , increased sensitivity , and greater baseline coverage , substantially improving the ability of mm - vlbi to study black holes on schwarzschild radius scales .
@xcite . anticipated and potential future stations sites include chile ( _ atacama pathfinder experiment _ , _ atacama submillimeter telescope _ and _ atacama large millimeter array _ ; apex , aste , and alma , respectively ) , mexico ( _ large millimeter telescope _ ; lmt ) , the south pole ( _ south pole telescope _ ; spt ) , and the iram telescopes in spain ( _ pico veleta _ ; pv ) and france ( _ plateau de bure _ ; pdb ) . among these the longest baselines are @xmath16 , corresponding to a maximum angular resolution of @xmath17 at @xmath18 ( @xmath14 ) and @xmath19 at @xmath20 ( @xmath21 ) . expected increases in bandwidth and the phasing together of elements within hawaii , chile and carma ( e.g. , * ? ? ?
* ) , will lead to substantial improvements in sensitivity . as a consequence
, it will become possible in the near future to measure mm - vlbi closure phases on a variety of additional triangles with uncertainties considerably smaller than that of the @xcite result .
here , motivated by previous efforts to model the @xmath14-vlbi visibilities using physically motivated accretion models for sgr a * , we compute the closure phases implied by the radiatively inefficient accretion flow models and @xmath14-vlbi visibility amplitude fits presented in @xcite . by doing so we address three immediate questions : 1 .
is the new closure phase estimate consistent with the accretion flow models we have considered in particular and radiatively inefficient accretion flow models generally ? 2 .
what is the strength of the constraint placed upon physically motivated accretion flow models and the estimates of black hole spin by the measured closure phase ?
what are the strength of the constraints that will possible in the near future as the eht develops ? in section [ sec :
cafcp ] we briefly describe the accretion models and how we compute the closure phases . in section [ sec : cwmv ] we compare the predicted closure phases with the measured values . in section
[ sec : cpft ] we predict the closure phases for the eht and compare these with the estimated uncertainties of the eht . finally , we summarize our conclusions in section [ sec : c ] .
we model sgr a * s accretion flow as a radiatively inefficient accretion flow , the details of which may be found in @xcite , and references therein , and thus we only summarize the model here . sgr a * transitions from an inverted , presumably optically thick spectrum to an optically thin spectrum near millimeter wavelengths .
this implies that near 1.3 mm sgr a * is only becoming optically thin , and thus absorption in the surrounding medium is likely to be important .
this transition does not occur isotropically , happening at longer wavelengths for gas that is receding and at shorter wavelengths for gas that is approaching .
therefore , properly modeling the structure and relativistic radiative transfer is crucial to producing high fidelity images . for concreteness , as in @xcite
, we follow @xcite and employ a model in which the accretion flow has a keplerian velocity distribution , a population of thermal electrons with density and temperature @xmath22 and @xmath23 respectively , and a toroidal magnetic field in approximate ( @xmath24 ) equipartition with the ions ( which are responsible for the majority of the pressure ) , i.e. , @xmath25 in all of these , @xmath26 is the schwarzschild radius , @xmath27 is the cylindrical radius and @xmath28 is the vertical coordinate . inside of the innermost - stable circular orbit ( isco ) we assume the gas is plunging upon ballistic trajectories . in principle
the plunging gas can still radiate , though in practice it contributes little to the overall emission due to the large radial velocities it develops . in the case of the thermal quantities
the radial structure was taken from @xcite , and the vertical structure was determined by assuming that the disk height is comparable to @xmath27 .
note that all of the models we employ necessarily have the spin aligned with the orbital angular momentum of the accretion flow .
for the regions that dominate the @xmath29 emission , this assumption is well justified due to disk precession and viscous torques , though it may be violated at large distances .
thermal electrons alone are incapable of reproducing the nearly - flat spectrum of sgr a * below @xmath30 .
thus it is necessary to also include a nonthermal component . as with the thermal components ,
we adopt a self - similar model for a population of nonthermal electrons , @xmath31 with a power - law distribution corresponding to a spectral index of @xmath32 and cut off below lorentz factors of @xmath33 ( consistent with * ? ? ?
the radial power - law index was chosen to reproduce the low frequency spectrum of sgr a * , and is insensitive to the black hole properties due to the distant location of the long - wavelength emission .
the primary emission mechanism at the wavelengths of interest is synchrotron radiation , arising from both the thermal and nonthermal electrons .
we model the emission from the thermal electrons using the emissivity described in @xcite , appropriately altered to account for relativistic effects ( see , e.g. , * ? ? ?
since we perform polarized radiative transfer via the entire complement of stokes parameters , we employ the polarization fraction for thermal synchrotron as derived in @xcite . in doing so
, we have implicitly assumed that the emission due to thermal electrons is isotropic , which while generally not the case is unlikely to change our results significantly .
for the nonthermal electrons , we follow @xcite for a power - law electron distribution , with an additional spectral break associated with the minimum electron lorentz factor . for both emission components the absorption coefficients are determined directly via kirchhoff s law .
images are then produced using the fully relativistic ray - tracing and radiative transfer schemes described in @xcite and @xcite . because @xcite neglected relativistic effects and assumed spherical symmetry , it is not directly applicable here .
for these reasons , as in @xcite , the coefficients @xmath34 were adjusted to fit the radio spectral energy distribution ( sed ) of sgr a * ( see * ? ? ? * for details on the fitting procedure ) .
we repeated this procedure for a large number of positions in the dimensionless spin - inclination ( @xmath35-@xmath36 ) parameter space , with @xmath37 corresponding to a maximally rotating kerr spacetime .
we define @xmath36 such that models with @xmath38 are viewed along the spin axis , and thus the thick accretion disk viewed face - on .
conversely , models with @xmath39 are viewed perpendicular to the spin axis , and thus the thick accretion disk is viewed edge - on . ] , producing a tabulated set of the coefficients @xmath34 at a large number of points throughout the @xmath35-@xmath36 parameter space . in all cases it was possible to fit the sed with extraordinary precision . from the tabulated values , the coefficients
are then obtained at arbitrary @xmath35 and @xmath36 using high - order polynomial interpolation . during the @xmath1-vlbi observations
sgr a * s flux varied by roughly @xmath40 .
we model this as a variable accretion rate , moving the electron density normalization up and down . in practice
, we reduced the electron density normalization by an amount sufficient to produce a total flux of @xmath41 , and then multiplied the resulting images by a correction faction during the mm - vlbi data analysis . because the source is not uniformly optically thin ,
this is not strictly correct , though this makes a small change to the images themselves .
we produced 9090 images , with flux normalized as described above , at @xmath42 for each @xmath43 .
we then produce models with arbitrary position angles , @xmath44 , by rotating the images on the sky . for this purpose
we define @xmath44 as the position angle ( east of north ) of the projected spin vector .
the intrinsic complex visibilities are obtained in the standard fashion : @xmath45 where @xmath46 is the intensity at a given set of angular coordinates .
these are subsequently modified to account for the observed interstellar electron scattering ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) . the closure phase for a given triplet of observatories is then given by @xmath47 + \arg\left[v(u_{jk } , v_{jk})\right ] + \arg\left[v(u_{ki } , v_{ki})\right]\,.\ ] ] we compute the @xmath48 for all spin magnitude , inclination and position angles considered in @xcite . from these we determine the probability density of a given @xmath48 : @xmath49\,,\ ] ] in which @xmath50 is the probability of a given vector spin . for our purposes
here we consider two forms of @xmath50 : an isotropically distributed spin direction with a flat prior upon the spin magnitude , the same prior as that adopted in @xcite , and the posterior probability of a given spin , defined in @xcite , after fitting the @xmath1-vlbi amplitudes . we will refer to these as the `` isotropic '' and `` amplitude - fitted '' priors , respectively .
interpolation errors associated with the finite density of points in the spin parameter space and the finite size of the imaged region , resulted in some small - amplitude saw - tooth features in the @xmath51 .
as a consequence , we smoothed the @xmath51 shown on scales of @xmath52@xmath53 , depending upon the particular case , with @xmath54 being typical .
where large spike features appeared we did no smoothing , and in all cases the wings of the probability distribution remained unchanged .
note that fitting the visibility amplitudes necessarily imparts a @xmath3 degeneracy in the position angle , corresponding to a sign change in the resulting closure phase . as a result , in principle ,
all closure phase estimates we present are unique only up to a sign .
nevertheless , we show the closure phases associated with one particular choice for the position angle , specifically that for which the spin of the most likely configuration is oriented @xmath55 east of north . ) and inclination ( @xmath36 ) for a number of position angles ( @xmath44 ) .
black contour lines show the closure phase limit of @xmath0 reported by @xcite . for reference ,
the posterior probability contours ( white ) associated with the @xmath56@xmath57 ( solid ) , @xmath58@xmath57 ( dashed ) , and @xmath59@xmath57 ( dotted ) cumulative probabilities as determined in @xcite are overlaid . note that these are degenerate to @xmath60.,scaledwidth=50.0% ] @xmath57 and @xmath58@xmath57 regions , defined by cumulative probability , and the light grey hatched region shows the measured value ( @xmath0 ) .
the vertical red line indicates the closure phase associated with the most likely accretion flow model . for reference , the distribution of closure phases when the @xmath1-vlbi prior is not applied is shown in the upper inset . note that these are degenerate to @xmath60.,scaledwidth=50.0% ]
lrcc hawaii & 8 & 23 & 4900 + carma & 8 & 27 & 6500 + smt & 1 & 10 & 11,900 + lmt & 1 & 32 & 10,000 + apex & 1 & 12 & 6500 + alma & 10 & 38 & 500 + the precision with which a given closure phase may be measured depends upon the signal - to - noise ratio ( snr ) of the visibilities on the individual baselines , @xmath61 , where @xmath62 is the visibility amplitude , @xmath63 is the bandwidth , @xmath64 is the atmospheric coherence time , and @xmath65 is the system equivalent flux density for a given station . here
we take @xmath66 , typical of the atmospheric coherence times at the wavelengths and sites of interest . however , because closure phases are insensitive to atmospheric phase fluctuations , they may be coherently averaged over time scales much larger than @xmath64 , with the measurement uncertainties decreasing as @xmath67 .
this is limited by the time scales on which the baseline orientations and the intrinsic structure of sgr a * change significantly . in quiescence ,
the former is the limiting factor , and thus we choose @xmath68 .
however , during periods of flaring activity , or if sgr a * exhibits some dynamical structure , this could be considerably shorter . finally , unless otherwise stated , we assume @xmath69 , the anticipated near - term bandwidth target of the eht , and adopt the @xmath65 reported in table 1 of @xcite , partially reproduced in table [ tab : sefds ] here .
results in baseline snrs approximately an order of magnitude larger than those obtained during the @xmath1-vlbi observations reported in @xcite . ]
specifically we assume that the sma , cso and jcmt are phased together in hawaii , six of the @xmath70 and two of the @xmath71 carma antennas are phased together , and ten of the @xmath72 alma antennas are phased together in chile . in terms of the @xmath73 ,
the closure phase uncertainty is given by , @xmath74\,,\ ] ] where @xmath75\ ] ] in which @xmath76 and @xmath77 are the hyperbolic bessel functions of the first kind @xcite .
in the high snr limit ( @xmath78 ) this reduces to @xmath79^{-1/2}\,,\ ] ] though we will make use of the exact expression . note that since @xmath80 depends upon the individual snr values , it is model dependent ; the values we quote are associated with the most likely accretion model , as determined via the @xmath14-vlbi visibility amplitudes alone .
there are two @xmath81 periods over which sgr a * s closure phase was measured on the smt - jcmt - carma triangle on the night of april 6 , 2009 ( corresponding to day 96 ) , beginning on 12:20 utc and 12:40 utc @xcite .
the distribution of the closure phases at 12:20 utc in spin magnitude ( @xmath35 ) , spin inclination ( @xmath36 ) , and spin position angle ( @xmath44 ) are shown in figure [ fig:1220cps ] , with the closure phases at 12:40 utc differing by only a small amount . for reference
the @xmath56@xmath57 , @xmath58@xmath57 , and @xmath59@xmath57 posterior probability contours from the @xmath14-vlbi visibility fits obtained in @xcite are overlaid , shown by the solid , dashed , and dotted white contours , respectively .
it is immediately apparent that small closure phases are not generic . assuming the isotropic prior
, fewer than 49% of the accretion flow models are consistent with the closure phase measurement .
the inset in figure [ fig : day96cps ] bears this out ; while small closure phases are marginally preferred , a large closure phase island exists with a substantial likelihood of any phase being measured .
the excluded models are not confined to a particular region in the spin parameter space , extending over a wide range of spin magnitudes , inclinations and position angles .
thus , despite the large uncertainties , the measured closure phase has considerable diagnostic ability . however , this changes dramatically when we restrict our attention to those models which are consistent with the @xmath1-vlbi amplitudes .
the white contours in figure [ fig:1220cps ] show the posterior probability of a given spin vector based upon the @xmath1-vlbi visibility amplitudes alone , i.e. , the amplitude - fitted prior .
the regions of high probability are confined to a narrow valley of low closure phases .
hence , while accretion models generally are not necessarily consistent with the observed closure phases , the vast majority of those that were found to reproduce the visibility amplitudes _ are _ broadly consistent with the recent measurement . once the amplitude - fitted prior is adopted , the probability of finding a closure phase that is larger than @xmath82 falls to less than 3% .
the amplitude - fitted models that violate the recent closure phase measurement are confined to the @xmath58@xmath57 significance regions , primarily at large spin magnitude and large negative position angles , where the closure phase is @xmath83 .
this behavior is similar to that exhibited by annular models , in which the closure phase varies between @xmath13 and @xmath3 , depending upon the angular size of the annulus . as a result , improvements of factors of a few in sensitivity will allow closure phase measurements that can compete with the visibility amplitudes in their power to constrain the structure of sgr a*. these conclusions
are borne out by the closure phase probability distributions shown in figure [ fig : day96cps ] , computed with the amplitude - fitted prior .
for these we show both times , which exhibit slightly different distributions , though in both cases the closure phases are dominated by peaks near @xmath84 and @xmath85 . the probability of measuring a closure phase within the allowed range exceeds 97% at both times .
the closure phase of the model with the highest amplitude - fitted prior probability ( based upon the @xmath14-vlbi visibility amplitude ) is located at @xmath86
. however , significant probabilities extend out to @xmath87 , and thus factor of few improvements in sensitivity necessarily result in a significant likelihood of detecting non - trivial closure phases .
the development of the eht will be characterized by the inclusion of additional baselines and improvements in station sensitivity .
thus , we also compute the closure phases along triangles including intermediate and large north - south baselines .
figure [ fig : allcps ] show the probability distribution for closure phases on all triangles made with the smt , carma , hawaii , the lmt , and chile , as seen at 12:20 utc on day 96 of 2009 .
closure phase distributions on other days and times are qualitatively similar . while the closure phase on the smt - carma - hawaii triangle is small , this is not generic .
closure phases with magnitudes of @xmath88 are likely on five of ten triangles , including the smt - hawaii - chile and hawaii - chile - lmt triangles , for which the most probable model has closure phases comparable to @xmath89 ( shown by the red vertical lines in each panel)-@xmath57 region .
this is not unexpected given the definition of the @xmath56-@xmath57 and @xmath58-@xmath57 regions , and does not indicate that the model with the highest posterior probability is inherently unlikely .
these deviations are generally largest for large open triangles , associated with the intrinsically larger value and distribution of closure phases in these cases . ] . more importantly , with improvements in the station sensitivity , arising from increases in bandwidth and the phasing together of multiple elements at individual stations , substantial improvements in the precision with which closure phases can be measured are possible .
despite the small visibilities associated with the most probable model ( see * ? ? ?
* ) , on all baselines constructed from the five stations we consider , sgr a * is detectable on the typical atmospheric coherence time ( @xmath90 ) with snr exceeding unity . over the dynamical time
scale of sgr a * , roughly @xmath91 , the uncertainty in the closure phase is reduced by an additional order of magnitude , resulting in the red , horizontal error bars in the individual panels of figure [ fig : allcps ] , ranging from @xmath92@xmath93 . in all cases
this is much smaller than the presently allowed range of values , and provides sufficient precision to distinguish likely values from zero .
the high precision with which closure phases may be measured in the near future , warrants a brief examination of their ability to help constrain the spin of sgr a*. to estimate this we consider the effects of including a single closure phase measurement on each of the various station - triangles discussed in section [ sec : cpft ] . in practice , where many closure phase measurements for each triangle are anticipated ,
these are expected to be even more constraining .
figure [ fig : cpspinlimits]a shows the posterior probability distribution obtained from analysing the 1.3mm - vlbi amplitudes alone , as a function of spin magnitude ( @xmath35 ) and inclination ( @xmath36 ) , after marginalized over the spin position angle ( @xmath44 ) due to this marginalization . ] . including the closure phase measurement reported in @xcite
eliminates some of the high - spin , high - inclination models , producing the probability distribution seen in figure [ fig : cpspinlimits]b . nevertheless , the @xmath56-@xmath57 and @xmath58-@xmath57 contours are not significantly affected .
the represents the present constraints upon the spin of sgr a*. to illustrate the improvement associated with higher - precision measurements and additional triangles , we will assume the most likely model , as defined by the 1.3mm - vlbi visibility amplitudes alone , is the `` true '' model for sgr a * s emission region .
this allows us to compute the visibilities throughout the @xmath94-@xmath95 plane , obtain the snrs at each point , and thus infer both the expected closure phase , @xmath96 and the precision with which it may be measured , @xmath97 . based upon this
, we can construct a likelihood associated with these observations , @xmath98 ^ 2}{2\sigma_{\phi_{ijk,0}}^2 } \right]\,,\ ] ] where @xmath99 is some constant and @xmath100 is the likelihood of observing the measured visibility amplitudes alone .
this may then be used to produce a posterior probability density , @xmath50 ) , as described in @xcite , the particulars of which depend upon which closure phase measurements we choose to include and their associated measurement uncertainties . in what follows
we consider a handful of possible measurements that are relevant for upcoming observations . improving the precision of the closure phase measurement on the smt - carma - jcmt triangle to @xmath101
substantially improves the spin estimation in two respects .
firstly , it would conclusively eliminate the @xmath3 degeneracy in the position angle resulting from fitting visibility amplitudes alone . secondly ,
as seen in figure [ fig : cpspinlimits]c , it begins to significantly modify the 1-@xmath57 and 2-@xmath57 contours .
thus , even with modest precisions , closure phase measurements can substantial improve black hole spin estimation in sgr a*. for comparison with ongoing observations , we show in figure [ fig : cpspinlimits]d the constraints associated with closure phases measured with a phased hawaii , phased carma , smt , and apex , all with a @xmath102 bandwidth ( in contrast to the @xmath103 bandwidth assumed elsewhere ) . even without the substantial sensitivity improvements associated with large bandwidths and phasing up multiple alma antennas ,
the resulting constraint is strengthened dramatically , in principle allowing a spin magnitude estimate with precision @xmath104 .
the spin constraints associated with observations with a phased alma are shown in figure [ fig : cpspinlimits]e , at which point the allowed model parameter space becomes too small to resolve with the parameter sampling we employed .
when the lmt is included ( figure [ fig : cpspinlimits]f ) , the allowed region is reduced to a single grid point in our model space ( @xmath105 , @xmath106 ) .
whether such high precision spin estimation is possible in practice is almost certainly going to be limited by the systematic uncertainties in the modeling of the emission region .
nevertheless , the constraining power of even a handful of closure phase measurements is clear .
the measurement of mm - vlbi closure phases in sgr a * provide an independent , significant constraint upon the structure of the emission region .
models based upon radiatively inefficient accretion flows do not generally have small closure phases , and thus are not generally consistent with the recently measured value on the smt - carma - jcmt triangle . as a consequence , even in the presence of large uncertainties
, closure phases can place significant constraints upon the structure of sgr a*. however , when only accretion flow models that fit the @xmath1-vlbi visibility amplitudes are considered , the probability of finding an excluded closure phase falls to less than 3% .
hence , the vast majority of the acceptable models in @xcite are broadly consistent with all current @xmath1-vlbi constraints .
this is surprising since the @xmath14-vlbi amplitudes and closure phase measurements provide independent constraints , and thus having fit the accretion models to the former there is no a priori reason to expect that they would be consistent with the latter .
based upon the models that are consistent with the @xmath1-vlbi visibility amplitudes , we predict that the closure phase on the smt - carma - jcmt triangle during the time it was measured was between @xmath2 , with @xmath107 most preferred .
improving the snr of mm - vlbi observations by a factor of a few , achievable by phasing up multiple carma antennas and employing the recently phased sites in hawaii , should allow detection of a non - zero closure phase on the smt - carma - jcmt triangle .
note that this would eliminate the existing @xmath3 ambiguity in the orientation of sgr a*. if sgr a * is well described by radiatively inefficient accretion flow models , currently acceptable closure phases on large triangles ( e.g. , smt - hawaii - chile and hawaii - chile - lmt ) can be as large as @xmath5 , and typical values on triangles which include long baselines of @xmath4 .
future improvements in sensitivity associated with the near - term development of the eht will substantially increase the precision with which closure phases can be measured . on all triangles we considered ( constructed from the smt , carma , hawaii , the lmt , and chile ) the resulting precision should be sufficient to measure the deviations from @xmath13 . improved closure phase measurements and baseline coverage result in dramatically improved constraints upon the spin of sgr a*. including a single antenna in chile ( e.g. , apex or a single alma dish ) produces spin estimates with precisions of roughly @xmath104 . with a partially phased alma station ,
this improves to @xmath108 , and with the lmt is better than @xmath109 , at which point we fail to resolve the likely region . in the latter cases ,
the spin estimation will almost certainly be limited by systematic uncertainties in the modeling of the emission region motivating the consideration of more sophisticated accretion and outflow models . | millimeter wave very long baseline interferometry ( mm - vlbi ) provides access to the emission region surrounding sagittarius a * , the supermassive black hole at the center of the milky way , on sub - horizon scales .
recently , a closure phase of @xmath0 was reported on a triangle of earth - sized baselines ( smt - carma - jcmt ) representing a new constraint upon the structure and orientation of the emission region , independent from those provided by the previously measured @xmath1-vlbi visibility amplitudes alone . here
, we compare this to the closure phases associated with a class of physically motivated , radiatively inefficient accretion flow models , and present predictions for future mm - vlbi experiments with the developing _ event horizon telescope _
( eht ) .
we find that the accretion flow models are capable of producing a wide variety of closure phases on the smt - carma - jcmt triangle , and thus not all models are consistent with the recent observations .
however , those models that reproduce the @xmath1-vlbi visibility amplitudes overwhelmingly have smt - carma - jcmt closure phases between @xmath2 , and are therefore broadly consistent with all current mm - vlbi observations .
improving station sensitivity by factors of a few , achievable by increases in bandwidth and phasing together multiple antennas at individual sites , should result in physically relevant additional constraints upon the model parameters and eliminate the current @xmath3 ambiguity on the source orientation .
when additional stations are included , closure phases of order @xmath4@xmath5 are typical . in all cases
the eht will be able to measure these with sufficient precision to produce dramatic improvements in the constraints upon the spin of sgr a*. |
we consider the grassmannian @xmath0 , the set of two - planes inside @xmath1 .
a plane @xmath2 is given by two linearly independent vectors or by any two linear combinations of them that are independent , so @xmath3 there is a transitive action of @xmath4 ( or @xmath5 ) on @xmath0 .
@xmath6 if we consider a particular point @xmath7 @xmath8 @xmath9 we notice that the conformal group of space time , @xmath10 , has spin group @xmath11 .
its complexification , @xmath12 , has spin group @xmath13 . how to extract the minkowski space from @xmath0 ?
notice that since the two vectors are independent , @xmath14 at least one of the @xmath15 determinants in this matrix is @xmath16 .
the space is covered by the atlas @xmath17 @xmath18 is the _ big cell _ , and using the @xmath19 freedom , a plane in @xmath18 can be represented by @xmath20 with the entries of @xmath21 totally arbitrary .
so @xmath22 , and it is a good candidate for the minkowski space .
what about the group action ?
@xmath18 is left invariant by the _
lower parabolic _ subgroup of @xmath23 , @xmath24 and it acts on @xmath21 as @xmath25 the group is @xmath26 so it is the poincar group where instead of the lorentz group we have put its double cover .
@xmath21 belongs to the _ twistor space _ associated to spacetime . using the pauli matrices
, we can revert to the spacetime notation and obtain the standard action of the poincar group on minkowski spacetime @xmath27 also the spacetime metric has an interpretation in the twistor formalism , @xmath28
quantization of spacetime means to deform the commutative algebra of functions ( can be polynomials or smooth functions ) to a non commutative algebra .
other properties that we want to consider in the quantum setting have to be first defined in the algebraic formalism and then quantized. this is the case of the group actions . the respective algebras are @xmath29/ ( \det g-1),\qquad a , b=1,\dots , 4.\ ] ] @xmath30/ ( \det x\cdot \det y -1 ) , \qquad i , j=1,2,\quad a , b=3,4.\ ] ] @xmath31.\ ] ] the group law is expressed in terms of a _ coproduct _ @xmath32with the property @xmath33 the action on the minkowski space is a _ coaction _ @xmath34
in refs . @xcite one substitutes the group @xmath13 by @xmath35 in the twistor construction .
all the scheme of coaction and big cell can be repeated in the quantum case , which gives a quantization for the minkowski space as a big cell inside a quantum conformal space ( a quantum grassmannian ) .
we just state the result : the quantum minkowski space is a quantum matrix algebra with the rows interchanged .
the correspondence @xmath36 is given in terms of the respective generators : @xmath37 this means that the commutation relations among the quantum generators are the following @xmath38 what happened to the groups ? they have become _ quantum groups _ with a non commutative product and a coproduct that is _ the same _ than the one we had before .
this means that the group law has not changed , nor the coaction on the grassmannian and the minkowski space .
the only change is that all these varieties have become non commutative .
it is a remarkable property of matrix quantum groups that the coproduct is compatible with both , the commutative product and the non commutative one . in the quantum version @xmath39 ,
the sets of generators @xmath40 , @xmath41 and @xmath42 are separately isomorphic to @xmath15 matrix algebras , but while @xmath40 and @xmath41 commute among them , @xmath43 does not commute with the rest of the generators .
a quantum matrix algebra is an algebra over @xmath44 $ ] , where @xmath45 is a parameter .
moreover , as a module over @xmath46 , it is a free module , which means that it has a basis .
it is well known that there is at least one _ ordering _ among the generators such that the standard monomials associated to this ordering are a basis of the quantum matrix algebra .
( this is a non trivial property ) .
the ordering is the following @xmath47 and then there is an isomorphism ( _ ordering rule _ or _ quantization map _ ) between @xmath48 and @xmath49 $ ] : @xmath50@ > q_{\mathrm{m}}>>{\mathcal{o}}_q({\mathrm{m}})\\t_{41}^a t_{42}^b t_{31}^c t_{32}^d @ > > > { \hat{t}}_{41}^a { \hat{t}}_{42}^b { \hat{t}}_{31}^c { \hat{t}}_{32}^d\end{cd}.\ ] ] with the quantization map we can pull back the non commutative product to @xmath49 $ ] .
this defines a _ star product _ , @xmath51,\ ] ] which can be computed _ explicitly _
@xmath52 @xmath53 are numerical factors defined recursively .
we recover the semiclassical interpretation of the algebra being an algebra of functions , but with a star product .
the previous formula for the star product is nice and compact , but can only be computed on polynomials .
can we extend it to smooth functions ? not obvious .
we prove that there exists a ( unique ) differential star product that coincides with the one given above on polynomials .
change of the parameter : @xmath54 .
we expand in powers of @xmath55 so we obtain a star product of the form @xmath56 with @xmath57 at each order , we have contributions from each of the terms with different @xmath58 @xmath59 we want to write @xmath60 as a bidifferential operator . but this is not trivial because all the dependence in the exponents should cancel .
@xmath61 antisymmetrizing and changing variables we obtain the poisson bracket @xmath62 notice that if the @xmath40 s are real , then the poisson bracket is pure imaginary . also , it is quadratic .
we have computed explicitly up to the order @xmath63 , but the expression is already too big to display it here .
we looked for an argument to show it at arbitrary order .
this can be done by careful inspection .
the proof that it is differential at each order is rather technical and we do not reproduce it here @xcite but having the explicit formula for the polynomials is essential to apply the argument .
an example of the possible difficulty : suppose that we want to reproduce @xmath64 as the result of applying a differential operator to @xmath65 .
we have several choices , @xmath66 but none of them is both , independent on the exponent @xmath67 and polynomial in the variable .
so the right combination of coefficients should appear in order to cancel the factors that appear when differentiating .
for example , if the result were @xmath68 , then we have a differential operator @xmath69 since we have recovered the interpretation of functions for the non commutative algebra , we can try to express the coaction as an action of this space of functions . remember that , formally , for the generators the coaction is the same than in the commutative algebra .
we just need to pull it back to the star product algebra .
we define the transformed variables ( no translations are considered here ) @xmath70 @xmath71 one just has to expand the star products in the right hand side .
up to order @xmath55 we have computed it in terms of a differential operator , @xmath72 and the coefficients are polynomials of order 6 in the variables @xmath73 .
let @xmath74 be a commutative algebra over @xmath75 .
an involution @xmath76 of @xmath74 is an antilinear map satisfying , for @xmath77 and @xmath78 @xmath79 let us consider the set of fixed points of @xmath76 , @xmath80 it is easy to see that this is a real algebra whose complexification is @xmath74 .
@xmath81 is a _
real form _ of @xmath74 .
_ classical minkowski space _ : @xmath82 the combinations @xmath83 are fixed points of the involution .
one has @xmath84.\ ] ] _ classical euclidean space _ : @xmath85 the commbinations @xmath86 are fixed points of @xmath87 , and as before , @xmath88.\ ] ] formally , the same expressions on the generators as in the classical case are involutions in the quantum algebra .
a few things change : * checking with the commutation relations they are antiautomorphisms , this is @xmath89 * this discards the interpretation of the real form of the non commutative algebra as the set of fixed points of the involution . *
when pulling back to the star product algebra , the poisson bracket is purely imaginary .
finally one finds also the corresponding involutions in the group , @xmath90 it is not difficult to realize that in the minkowskian case the real form of the lorentz group ( corresponding to the generators @xmath40 and @xmath41 ) is @xmath91 and in the euclidean case is @xmath92 .
d. cervantes wants to thank the departament de fsica terica , universitat de valncia for the hospitality during the elaboration of this work . | we present a deformation of the minkowski space as embedded into the conformal space ( in the formalism of twistors ) based in the quantum versions of the corresponding kinematic groups .
we compute explicitly the star product , whose poisson bracket is quadratic .
we show that the star product although defined on the polynomials can be extended differentiably .
finally we compute the eucliden and minkowskian real forms of the deformation . |
non - equilibrium current noise in mesoscopic structures is a consequence of the discreteness of the charge carriers ( for reviews , see refs .
@xcite ) . for conductors with open channels the fermionic statistics of electrons results in a suppression of shot noise below the classical schottky limit @xcite .
this was first noted by khlus @xcite and lesovik @xcite for single channel conductors .
subsequently , bttiker generalized this suppression for many - channel conductors @xcite .
mesoscopic conductors are often probed by two or more leads .
the quantum statistics induces cross - correlations between the currents in different terminals .
since these cross - correlations vanish in the classical limit , even their sign is not obvious a priori . using only the unitarity of the scattering matrix
, bttiker proved that cross - correlations for non - interacting fermions are _ always negative _ for circuits with leads maintained at constant potentials @xcite .
note that this also holds in the presence of a magnetic field .
it has also been found that an interacting paramagnetic dot shows negative cross - correlations in the absence of a magnetic field @xcite .
spin - dependent cross - correlations in a non - interacting 4-terminal spin valve were studied @xcite and found to be negative . on the experimental side
negative cross - correlations were measured by henny _
. _ @xcite and oliver _ et al . _
@xcite in mesoscopic beam splitters .
several ways to produce positive cross - correlations in fermionic systems have been proposed ( see e.g. @xcite for a recent review ) . among these possibilities
are sources which inject correlated electrons @xcite and finite - frequency voltage noise @xcite .
the question of the existence of intrinsic mechanisms , i. e. due to interactions occuring in the beam - splitter device itself , has been answered positively by us @xcite .
surprisingly , a simple quantum dot connected to ferromagnetic contacts can lead to positive cross - correlations due the so - called _ dynamical spin - blockade_. simply speaking , up- and down - spins tunnel through the dot with different rates . in the limit where the coulomb interaction prevents a double occupancy of the dot ,
the spins which tunnel with a lower rate modulate the tunneling of the other spin - direction , which leads to an effective _ bunching _ of tunneling events . in a three terminal geometry with one input and two outputs , this results in positive cross - correlation between the two output currents .
independently , sauret and feinberg proposed a slightly different setup of a ferromagnetic quantum dot , which also produces positive cross - correlations @xcite .
experimentally , it is more difficult to fabricate quantum dots with ferromagnetic leads .
however , quantum dots with paramagnetic leads have shown to exhibit spin - dependent transport .
a magnetic field lifts the spin - degeneracy and a spin - polarized current with nearly 100% efficiency can be created @xcite . in this letter
, we will address the current correlations in a few - electron quantum dot connected to three paramagnetic leads .
we will show below that positive cross - correlations can be produced in this device simply by applying a _
magnetic field_. furthermore , this system also shows a super - poissonian shot noise . to arrive at these conclusions we consider a quantum dot with one orbital energy level @xmath0 connected to three terminals by tunnel contacts .
the junctions are characterized by bare tunneling rates @xmath1 ( @xmath2 ) and capacitances @xmath3 .
we assume that a magnetic field @xmath4 is applied to the dot , which leads to a zeeman splitting of the level according to @xmath5 , where @xmath6 is the bohr magneton .
the double occupancy of the dot costs the charging energy @xmath7 , with @xmath8 .
the energy spacing to the next orbital is @xmath9 .
we will assume @xmath10 according to these inequalities , the dot can be only singly occupied and we have to take into account only one orbital level . in the sequential - tunneling limit @xmath11 , the time evolution of the occupation probabilities @xmath12 of states @xmath13
is described by the master equation : @xmath14 where @xmath15{ccc}% -\gamma_{\uparrow}^{-}-\gamma_{\downarrow\uparrow } & \gamma_{\uparrow \downarrow } & \gamma_{\uparrow}^{+}\\ \gamma_{\downarrow\uparrow } & -\gamma_{\downarrow}^{-}-\gamma_{\uparrow \downarrow } & \gamma_{\downarrow}^{+}\\ \gamma_{\uparrow}^{- } & \gamma_{\downarrow}^{- } & -\gamma_{\uparrow}% ^{+}-\gamma_{\downarrow}^{+}% \end{array } \right ] \,.\label{matrixm}%\ ] ] the rate for an electron to tunnel on / off the dot ( @xmath16 ) through junction @xmath17 is given by @xmath18)$ ] , where @xmath19 and @xmath20 . here
, we took the fermi energy @xmath21 for lead 2 as a reference .
the total tunneling rates are @xmath22 and @xmath23 .
spin flips on the dot are described by rates @xmath24 , which obey the detailed balance rule @xmath25 . from eq .
( [ masterequation ] ) the stationary occupation probabilities @xmath26 are @xmath27 and @xmath28 .
these probabilities can be used to calculate the average value @xmath29 of the tunneling current @xmath30 through junction @xmath17 as @xmath31 where @xmath32 is the state of the dot after the tunneling of an electron with spin @xmath33 in the direction @xmath34 , i. e. , @xmath35 and @xmath36 .
the frequency spectrum of the noise correlations can be defined as @xmath37 where @xmath38 is the deviation from the average current in terminal @xmath39 .
it can be calculated using the method developed in refs .
@xcite as:@xmath40 where the first term is the schottky noise produced by tunneling through junction @xmath17 , and @xmath41 with @xmath42 . in the following
we study the dot in a beam - splitter configuration , in which a bias voltage @xmath43 is applied between terminal 2 and terminals 1 and 3 .
we consider the case @xmath44 , so that it is energetically more favorable for electrons to go from lead 2 to leads 1 and 3 .
we will limit our discussion to the case in which the two zeeman sublevels are below the fermi energy at equilibrium ( _ i . e. _ @xmath45 ) .
the opposite case was discussed in ref .
@xcite for a two - terminal dot .
we are mostly interested in the total zero - frequency current noise @xmath46 and the cross - correlations @xmath47 between the two output leads .
it is useful to define the fano factor @xmath48 and , correspondingly , @xmath49 . , @xmath50 , @xmath51 , @xmath52 , @xmath53 , and different values of @xmath54 .
the average current @xmath55 through lead @xmath56 is plotted in units of @xmath57 ; the voltage is in units of @xmath0 .
the positions of @xmath58 and @xmath59 are indicated in dotted lines .
, width=377 ] in the following , we will first assume that @xmath60 .
transport through the down level is energetically allowed for @xmath61 . however , for @xmath62 , the dot is blocked by an up spin , thus down spins can not cross the dot .
around @xmath63 , the lower zeeman level is close to the fermi level of leads 1 and 3 , as represented by the level diagram in the lower right inset of fig .
[ fig : current ] .
the blockade of the dot is then partially lifted and transport through both levels is allowed . in this regime , we can write the tunneling rates as @xmath64 , @xmath65 , @xmath66 , @xmath67 , @xmath68 , and @xmath69 , where @xmath70)$ ] ranges from 0 to 1 with increasing voltage . furthermore , taking @xmath71 and @xmath72 , we find for the current @xmath73 for the fano factor @xmath74 and for the cross correlations @xmath75 we observe that the current increases with voltage ( i.e. with @xmath76 ) around the voltage step @xmath58 . note that this current is _ not spin - polarized _ because up and down spin have the same probability to enter the dot , regardless of what happens at the output . the fano factor
@xmath77 and the cross - correlations @xmath78 deviate from their poissonian values depending on the applied voltage .
our main results are that @xmath77 can be super - poissonian and @xmath78 positive for @xmath79 , as can be cleary seen from ( [ f2para ] ) or ( [ eq : c2para ] ) in the limit @xmath80 .
these features are a consequence of dynamical spin blockade : up spins leave the dot with a rate smaller than down spins , leading to a bunching of tunneling events @xcite . in the limit @xmath81 ,
the fano factor is always sub - poissonian and the cross - correlations always negative .
this is due to the fact that the tunnel rates of up and down spins are equal , thus the zeeman splitting plays no role and the dot is equivalent to a simple quantum dot with a spin - degenerate level . in the limit @xmath82 , one could also expect the super - poissonian nature of @xmath77 and the positivity of @xmath78 to be lost since the transport is enabled only by thermally activated processes .
however , below the voltage threshold @xmath58 , the fano factor tends to : @xmath83 which is always super - poissonian .
if the coupling to terminal 2 dominates , i. e. @xmath80 , the fano factor takes a maximal value of 3 . in the opposite limit @xmath84 ,
@xmath77 approaches the poisson limit of uncorrelated single charge transfer .
it is interesting to note that a symmetric junction @xmath85 produces twice the poisson noise level .
the cross - correlations in the same limit have the form @xmath86 in the three cases discussed above , the cross - correlations thus take the limiting values @xmath87 for @xmath84 , @xmath88 for @xmath85 and @xmath89 for @xmath90 .
hence , both the super - poissonian nature of @xmath77 and the positivity of @xmath78 can persist for @xmath82 .
even if the transport is enabled only by thermally activated processes , dynamical spin blockade already results in a correlated transfer of electrons . of the total current as a function of voltage , for the same circuit parameters as in fig .
[ fig : current ] . left panel : data for different values of @xmath91 and @xmath71 .
right panel : effect of spin flip scattering for @xmath92 and different values of @xmath93 .
the curves displayed in both panels are independent of the asymmetry between the output leads . ]
we now turn to the discussion of the general results displayed in figs .
( [ fig : current])-([fig : cross ] ) , obtained from an exact treatment of the full master equation . fig .
[ fig : current ] shows the full voltage dependence of the average current . as expected , the current shows a single step at @xmath94 @xcite .
the step width is about @xmath95 , whereas its position varies only slightly with the asymmetry of the junctions ( the maximal variation is about @xmath96 ) .
the left panel of figure [ fig : fano ] shows the voltage dependence of the fano factor in the absence of spin - flip scattering , for some values of @xmath97 .
the divergence @xmath98 of the fano factor at zero voltage is simply a result of the dominating thermal noise in the limit @xmath99 .
note that similarly to @xmath55 , the fano factor @xmath77 shows one single step at @xmath100 .
the right panel of fig .
[ fig : fano ] shows the effect of spin - flip scattering on @xmath101 for the case @xmath102 . for @xmath103 spin
flips become effective when @xmath104 , see eq .
( [ matrixm ] ) .
the sensitivity to @xmath93 thus increases with b. below @xmath59 , even smaller spin - flip rates suppress the super - poissonian noise because the dwell time of electrons on the dot is very long . far above @xmath58 , spin - flip scattering has no effect on the sub - poissonian noise . between leads 1 and 3 as a function of voltage , for the same circuit parameters as in fig .
[ fig : current ] and different values of @xmath54 . in all curves of the main frame @xmath105
the inset shows the effect of spin flip scattering for @xmath92 and different values of @xmath93 .
right : @xmath78 as a function of voltage for different values of @xmath4 .
the curves are shown for @xmath106 .
a cross indicates the position of @xmath59 for each case.,title="fig:",width=272 ] between leads 1 and 3 as a function of voltage , for the same circuit parameters as in fig .
[ fig : current ] and different values of @xmath54 .
in all curves of the main frame @xmath105 the inset shows the effect of spin flip scattering for @xmath92 and different values of @xmath93 .
right : @xmath78 as a function of voltage for different values of @xmath4 .
the curves are shown for @xmath106 .
a cross indicates the position of @xmath59 for each case.,title="fig:",width=257 ] the left panel of fig .
[ fig : cross ] shows the voltage dependence of the cross - correlations factor @xmath78 between the two output terminals , for the same parameters as in fig .
[ fig : fano ] .
first , around the voltage threshold @xmath58 , we observe the features discussed above .
the cross - correlations develop from a positive or negative level below @xmath58 depending on the ratio @xmath97 to the usual negative cross - correlations above @xmath58 , where the spin - splitting plays no role anymore .
remarkably , in contrast to @xmath77 , the cross - correlations also show a step around the lower voltage threshold @xmath59 .
this illustrates clearly that @xmath78 and
@xmath77 are qualitatively different . the absence of the lower step for @xmath77 can be interpreted as a consequence of the unidirectionnality of tunneling through junction 2 . indeed
, @xmath107 means that @xmath77 depends only on @xmath108 and @xmath109 [ see ( [ eq : current ] ) and ( [ eq : noise2 ] ) ] .
now , for @xmath110 @xmath111 implies that the contribution of these terms is independent of @xmath43 . on the contrary
, @xmath78 also depends on @xmath112 and @xmath113 with @xmath114 . for @xmath111 ,
these last terms depend strongly on @xmath115 which varies itself significantly with @xmath43 around @xmath59 .
note that the absence of a step in @xmath77 implies a redistribution of the noise between @xmath116 , @xmath117 and @xmath118 when the threshold @xmath59 is crossed ( due to charge conservation , @xmath119 ) .
the extra step of @xmath78 disappears for @xmath80 . in this limit , the cross - correlations display a single low voltage plateau @xmath89 , which is an upper bound for the two low voltages plateaux found in the general case . the inset in the left panel of fig .
[ fig : cross ] shows the effect of spin - flip scattering on the cross - correlations .
as expected , they suppress all spin - effects and the positive cross - correlations become finally negative .
like for the fano factor , very small spin - flip scattering rates @xmath93 are already sufficient to modify @xmath78 for @xmath120 .
since the positive cross - correlations found in this work are intimately related to the dynamical spin - blockade , we expect a strong dependence on the magnetic field . the right panel of fig .
[ fig : cross ] shows the voltage dependence of @xmath78 around the step @xmath58 , for a fixed temperature and various magnetic fields . just below @xmath58
the limiting value of @xmath78 is determined by formula ( [ eq : cross_below_step ] ) .
thus , for a constant voltage @xmath121 we predict a cross - over from negative to positive cross - correlations with increasing magnetic field .
one can see a qualitative change in the curves , which can be understood by a gradual splitting of the voltage steps @xmath59 and @xmath58 .
the lower step is at @xmath122 . as long as @xmath123 , the two voltage steps are indistinguishable
however , positive cross - correlations are already expected for @xmath124 . for @xmath125
the two steps still overlap , resulting in a broad peak , whereas for the higher magnetic field @xmath126 the lower threshold at @xmath59 is outside the plotting region .
the regime @xmath100 has the advantage that the current is not exponentially small ( c.f .
[ fig : current ] ) and thus observable more easily in an experiment . for @xmath127 ,
the maximum value obtained for the cross - correlations is @xmath128 at @xmath129 , i. e. @xmath130 in fig .
[ fig : cross ] . with @xmath131 ghz
, this corresponds to @xmath132a@xmath133s , a noise level accessible experimentally @xcite . in conclusion
we have studied current correlations for a three terminal quantum dot with unpolarized leads , placed in a magnetic field . below the voltage threshold @xmath58 , as a result of dynamical spin - blockade , the fano factor of the input current shows an interesting super - poissonian behavior and the cross - correlations in the two output leads can be positive . at higher voltages the fano factor becomes sub - poissonian and the cross - correlations negative , as usual .
the effect we predict should be observable in semiconductor quantum dots of ref . | we investigate current fluctuations in a three - terminal quantum dot in the sequential tunneling regime .
dynamical spin blockade can be induced when the spin - degeneracy of the dot states is lifted by a magnetic field .
this results in super - poissonian shot noise and positive zero - frequency cross - correlations .
our proposed setup can be realized with semiconductor quantum dots . |
face centered cubic ( fcc ) iron - based alloys are widely used for developing of stainless austenitic steels especially for using in critical temperature ranges , aggressive environment and other severe external conditions .
doping of these steels with light interstitial impurities ( c , n ) influence mechanics and kinetics of structure phase transitions in fe - based alloys .
distribution of carbon and nitrogen atoms in solid solutions influence electrical and mechanical properties of alloys .
nitrogen doping enables to solve the problem of the strengthening of stainless steels .
investigation of the influence of nitrogen on physical properties of multicomponent systems is a complicated problem .
the solution of this problem should be made in several stages .
on the first stage it seems important to study electronic structure of iron nitride fe@xmath1n , because binary @xmath3-fen@xmath4 alloy can be modeled by non - stoichiometric nitride @xmath3-fe@xmath1n@xmath5 .
there are a lot of experimental data about atomic and electronic structure of fe@xmath1n .
we believe that mssbauer spectroscopy gives most interesting information about impurity distribution , electronic structure and magnetic interactions in alloys .
studying of hyperfine structure of the energy spectra of nuclei is a powerful tool for investigation of interactions of atomic nucleus with local electric and magnetic fields .
these interactions cause shifts and splittings of nuclear energy levels and enable us to obtain information about symmetry of charge distribution around the nucleus , about electronic configurations of atoms and ions , as well as about peculiarities of atomic structure of solids .
a number of experimental papers show substantial differences in mssbauer spectra of binary fe - n and fe - c alloys .
these differences are believed to be connected with different c and n distribution in fe fcc alloys @xcite . in this paper
we present calculation results of hyperfine interaction parameters for iron nitride fe@xmath1n as well as for hypothetical fe@xmath1n structure with another distribution of nitrogen atoms .
this allows us to determine changes in mssbauer spectra caused by redistribution on nitrogen atoms .
wien97 programme package @xcite , employing full - potential linearized augmented plane wave ( flapw ) method was used for calculations . as far as flapw
is an all - electron method ( unlike pseudopotential methods ) , it allows to perform calculations of hyperfine interaction parameters from first principles .
obtained theoretical parameters of interaction of a nucleus with electric and magnetic fields can be successfully compared with parameters of experimental mssbauer spectra .
generalized gradient approximation ( gga ) according to perdew - burke - ernzerhof @xcite model was used for exchange - correlation potential .
the radii of atomic spheres were chosen as 1.9 a.u .
and 1.68 a.u for fe and n atoms respectively .
the accuracy of calculation results depends on several basic parameters : number of @xmath6-points in brillouin zone , number of @xmath7-components and fourier coefficients in charge density and potential decomposition and number of plane waves in interstitial region .
the choice of the values of these parameters was based on convergence condition .
convergence tests gave the value @xmath8 , which corresponds to 205 plane waves per atom in the basis set .
inside atomic spheres the wave function was decomposed up to @xmath9 .
charge density and potential was decomposed inside atomic spheres using lattice harmonics basis up to @xmath10 . in the interstitial region
fourier expansion was used with 850 coefficients .
calculations were performed for 3000 k - points in the brillouin zone ( 84 k - points in the irreducible part ) .
the values of all parameters ensure accuracy of 0.1 mry in total energy of the system . due to ferromagnetic nature of iron
nitride all calculations were performed using spin - polarized approximation .
the unit cell of iron nitride fe@xmath0n is a unit cell of @xmath3-fe with a nitrogen atom introduced in the centre of the cube ( figure [ cell]a ) .
the structure has two symmetry types of fe atoms : fe@xmath11 type forms octahedron around impurity atom and has two impurity atoms in the first coordination sphere located at @xmath12 ; fe@xmath13 type is located in the corners of the cell and has no impurity atoms in the first coordination sphere . in order to find the value of lattice parameter , corresponding to the minimum total energy of the system we performed calculations for five different values of lattice parameter .
the results were approximated by second - order polynomial using least square fit method .
then the value of lattice parameter was obtained analytically . calculated value of lattice parameter c=7.164 a.u .
is in good agreement with experimental value c=7.17 a.u .
calculations of electronic structure and hyperfine interaction parameters were performed using obtained optimized value of lattice parameter .
optimization procedure was also performed for another two systems : @xmath3-fe ( fe - fcc ) and hypothetical fe@xmath0n structure ( figure [ cell]b ) .
both calculations were also performed in spin - polarized approximation .
calculation of @xmath3-fe seems interesting because it allows us to determine the influence of introducing of nitrogen atom on electronic structure and hyperfine interaction parameters of fe - fcc .
optimization of hypothetical fe@xmath0n structure ( figure [ cell]b ) was performed including variation of interatomic fe - n distance .
this structure has two types of iron atoms : fe@xmath14 and fe@xmath11 , having one and two n atoms in the nearest neighbourhood respectively .
we found out that with this type of ordering the total energy of the structure is 18 mry larger , than the total energy of iron nitride .
this shows that the formation of fe
- n - fe - n chains ( fe@xmath11 configuration ) without fe - n pairs ( fe@xmath14 configuration ) is energetically favourable . from the point of view of atomic structure of fe - n alloys
iron nitride fe@xmath1n is the simplest structure where fe - n chains are combined with fe - fe chains .
this makes possible to perform detailed analysis of the influence of nitrogen on electronic structure and hyperfine interaction parameters and to use obtained information later for analyzing more complicated fe - n alloys .
this step - by - step approach , based on application of up - to - date _ ab initio _ calculation method , appears to be most accurate for solving the problem of the influence of impurity atoms on the properties of fe - based alloys . * .
[ cols= " < , < , < ,
< , < " , ] first of all let us analyze the electric field gradient , caused by the whole valent band ( valence efg ) .
table [ tab3 ] shows that @xmath15 contribution is negligibly small for all energy regions .
efg is practically formed by @xmath16 and @xmath17 contributions , which have different signs .
we see that negative @xmath16 contribution is dominating , although it is greatly canceled by positive @xmath17 contribution .
the analysis of table [ tab3 ] data allows us to conclude that main contribution to efg formation is made by @xmath18 and @xmath19 nitrogen states , and contribution from @xmath18 states is dominant .
it should be noted that we are now talking about electrons , which have energy of nitrogen @xmath18 states , but spatially are located in fe@xmath11 sphere .
substantial asymmetry of spatial distribution of these electrons ( figure [ den1]a ) gives main negative contribution to total efg .
the interesting fact is that more than 70% of charge , formed by these electrons , have @xmath20symmetry , and 30% - @xmath21symmetry relative to the centre of fe@xmath11 sphere .
this follows from analysis of @xmath16 and @xmath17 contributions for n-@xmath18 interval ( table [ tab3 ] ) .
the same situation is observed for @xmath22electrons of nitrogen , but their total efg contribution is 2.5 times smaller than @xmath23electrons contribution .
the efg contribution of n-@xmath22electrons is practically caused only by electrons of @xmath20symmetry relative to the centre of fe@xmath11 sphere .
substantial negative efg , caused by described asymmetry of nitrogen electrons , is greatly canceled by positive gradient , caused by own fe electrons ( figure [ den1]b ) .
these are first of all fe-@xmath24electrons , which density is distributed in such a way , that greatly cancels asymmetry of @xmath19 and @xmath18 electrons of nitrogen .
comparative analysis of @xmath16 and @xmath17 contributions shows that more than 60% of asymmetrically distributed electrons of fe-@xmath25 energy region have @xmath21symmetry , and less than 40% - @xmath20symmetry ( table [ tab3 ] ) relative to the centre of fe@xmath11 sphere .
redistribution of rather localized fe-@xmath26 states also gives positive efg contribution , but this contribution is almost 5 times smaller than contribution of @xmath24states . it should be noted that quadrupole splitting was calculated using total efg value , given in table [ tab2 ] , and for calculation of valence efg contribution summation in ( [ rho ] ) was done up to @xmath9 . on the other hand , for analysis of partial efg contributions we limited the summation only up to @xmath27 ( due to the absence of `` chemical '' @xmath28-states )
this is the reason for small differences of valence efg values in tables [ tab2 ] and [ tab3 ] .
the values of local magnetic moments play an important role in efg formation in iron nitride fe@xmath1n .
the calculations show that the upper edge of fe-@xmath25 spin down states is located 2 ev above the fermi level .
substantial difference in population of states with different spin orientation greatly influence efg formation .
table [ tab3 ] gives efg contributions from states with different spin orientation for n-@xmath18 and fe-@xmath25 energy regions .
we notice considerable difference in efg values , caused by spin up and down electrons of fe-@xmath25 interval .
this effect mostly influence the value of @xmath17 contribution of this energy region .
this contribution is mainly formed by density distribution of own @xmath24electrons of iron .
spin up contribution is more than 3 times larger than spin down contribution for @xmath17 component of fe-@xmath24region .
the fact that spin down states are partially `` pulled '' above the fermi level considerably decreases compensating positive component , and leads to increasing of absolute value of total ( negative ) efg . in the absence of this effect
we would have considerably larger compensating component , which would lead to decreasing of total efg value and , as a consequence , to decreasing of quadrupole splitting in mssbauer spectra .
we performed _ ab initio _ band structure calculations of iron nitride fe@xmath0n .
all our results are in very good agreement with experimental data .
structure characteristics as well as magnetic properties are well reproduced .
we also performed calculations of hypothetical fe@xmath0n disordered structure .
comparative analysis of these two calculations showed that the formation of fe - n pairs in fe@xmath0n compound is energetically unfavourable .
we also calculated hyperfine interaction parameters for both fe@xmath0n structures taking into account the presence of magnetic field . for iron nitride structure we obtained perfect agreement of quadrupole splitting values with experimental data .
for disordered fe@xmath0n structure we observe considerable decreasing of quadrupole splitting for fe - n - fe - n chains , and very small quadrupole splitting for fe - n pairs .
detailed analysis of efg formation in iron nitride fe@xmath0n allows us to conclude that efg is mainly formed by space distribution of nitrogen @xmath19 and @xmath18 electrons , although it is greatly canceled by fe@xmath24electrons .
magnetization reduces this cancellation effect and increases the value of total efg in iron nitride structure .
99 v.g.gavriljuk _ et al _ , acta mater 45 , 225 ( 1997 ) , acta mater 47 , 927 ( 1999 ) p.blaha , k.schwarz , and j.luitz , wien97 , a full potential linearized augmented plane wave package for calculating crystal properties ( karlheinz schwarz , techn .
universitt wien , austria ) , 1999 .
isbn 3 - 9501031 - 0 - 4 perdew j.p.,burke s. and ernzerhof m. 1996,phys.rev.let.77,3865 j.hglund , phys .
b 47 , 566 ( 1993 ) d.singh , phys . rev .
b 43 , 6388 ( 1994 ) p.dufek , p.blaha and k.schwarz , phys .
75 , 3545 ( 1995 ) m.acet , h.zhres , e.f.wasserman and w.pepperhoff , phys . rev .
b 49 , 6012 ( 1994 ) h.jacobs , d.rechenbach and u.zachwieja , journal of alloys and compounds 227 , 10 ( 1995 ) h.c.herper , e.hoffmann and p.entel , phys .
b 60 , 3839 ( 1999 ) p.blaha , k.schwarz and p.herzig , phys . rev .
lett . , 54 , 1192 ( 1985 ) k.schwarz , c.ambrosch-draxl and p.blaha , phys .
b , 42 , 2051 ( 1990 ) p.rochegude and j.foct , phys
( a ) 98 , 51 ( 1986 ) r.coehoorn , g.h.o.daalderop and h.j.f.jansen , phys . rev .
b , 48 , 3830 ( 1993 ) | iron nitride fe@xmath0n is studied by full - potential lapw method .
structure parameters , electronic and magnetic properties as well as hyperfine interaction parameters are obtained .
we observe perfect agreement with experimental results .
hypothetical fe@xmath0n structure was also calculated to study the influence of disordering effects on parameters of mssbauer spectra .
we performed detailed analysis of efg formation on fe nuclei including magnetization effects .
we show that the formation of n - fe - n local configuration is energetically favourable in nitrogen austenites .
* electronic structure , hyperfine interactions and disordering effects in iron nitride fe@xmath1n * + + + _ @xmath2 rostov state university of transport , narodnogo opolcheniya 2 , rostov - on - don , 344038 russia _ + |
consider a massless dirac fermion interacting with a background gauge field .
our present goal is to formulate a nonperturbatively regularized quantum field theory which at least satisfies the following physical constraints : \(a ) in the classical limit , it reproduces the classical physics of the action , @xmath0 \psi(x ) $ ] .
\(b ) for topologically trivial gauge backgrounds , and in the weak coupling limit , it agrees with the predictions of weak coupling perturbation theory of the action .
\(c ) for topologically nontrivial gauge backgrounds , it possesses exact zero modes satisfying the atiyah - singer index theorem . although wilson s idea @xcite of formulating gauge theories on the spacetime lattice is the most successful nonperturbative regularization for pure gauge fields , putting massless dirac fermions @xcite on the lattice has been a notorious problem for more than twenty years .
the resolution of the lattice fermion problem first appeared in the context of the domain - wall fermion @xcite , and it motivated the overlap formalism @xcite which led to the construction of overlap - dirac operator @xcite in 1997 .
we refer to ref .
@xcite for a recent review of the domain - wall fermions , and to ref .
@xcite for a recent review of the overlap .
however , if we consider a weyl fermion interacting with a background gauge field , then a completely satisfactory nonperturbative regularization for chiral guage theories ( e.g. , the standard model ) has not yet been presented up to this moment . in the following
, we will concentrate our discussions on the general principles to construct chiral dirac fermion operators on a finite lattice , in particular , for vector - like gauge theories such as qcd . with
the constraints imposed by the nielson - ninomiya no - go theorem@xcite , one can construct a gauge covariant dirac operator @xmath1 on a finite lattice such that : ( i ) : : @xmath1 breaks the chiral symmetry ( i.e. , @xmath2 ) at finite lattice spacing but recovers the chiral symmetry in the continuum limit @xmath3 . (
ii ) : : @xmath1 is local .
( @xmath4 with @xmath5 ; or @xmath6 for @xmath7 , where @xmath8 is much less than the size of the lattice . ) ( iii ) : : in the free fermion limit , @xmath1 is free of species doublings . ( the free fermion propagator @xmath9 has only one simple pole at the origin @xmath10 in the brillouin zone . ) ( iv ) : : in the free fermion limit , @xmath1 has correct continuum behavior .
( in the limit @xmath3 , @xmath11 around @xmath10 . )
however , one can not push the property * ( i ) * any further , while maintaining properties * ( ii)-(iv)*. for example , if @xmath1 is chirally symmetric at finite lattice spacing , then it must violate at least one of the three properties * ( ii)-(iv)*. we note that these four properties * ( i)-(iv ) * form the necessary conditions to meet the requirements * ( a)-(c ) * , however , they are not sufficient to guarantee that * ( c ) * will be satisfied .
an example satisfying * ( i)-(iv ) * is the standard wilson - dirac fermion operator@xcite [ eq : dw ] d_w = _ t_+ w where [ eq : tmu ] t_(x , y ) = [ u_(x ) _ x+,y - u_^(y ) _ x-,y ] , [ eq : sigma ] _ _ ^ + _ _ ^ = 2 _ , [ eq : gamma ] _ & = & ( cc 0 & _ + _ ^ & 0 ) and @xmath12 is the wilson term [ eq : wilson ] w(x , y ) = _ . the color , flavor and spinor indices have been suppressed in ( [ eq : dw ] ) .
the first term on the r.h.s . of ( [ eq : dw ] ) is the naive fermion operator which is chirally symmetric at any lattice spacing and satisfies properties * ( ii ) * and * ( iv ) * but violates * ( iii ) * since it has @xmath13 fermion doubled modes .
the purpose of the wilson term @xmath12 is to give each doubled mode a mass of @xmath14 such that in the continuum limit ( @xmath3 ) , each doubled mode becomes infinitely heavy and decouples from the fermion propagator .
however , the introduction of the wilson term has serious drawbacks .
it causes @xmath15 artifacts and also leads to the notorious problems such as vector current renormalization , additive fermion mass renormalization , and mixings between operators in different chiral representations . during the last two years
, it has become clear that the proper way to break the chiral symmetry of @xmath1 at finite lattice spacing is to conform with the ginsparg - wilson relation @xcite [ eq : gwr ] d 5 + 5 d = 2 a d r _ 5 d where @xmath16 is a positive definite hermitian operator which is local in the position space and trivial in the dirac space .
then the generalized chiral symmetry ( [ eq : gwr ] ) can ensure that the theory is free of above mentioned problems of the wilson - dirac operator @xcite . the general solution to the ginsparg - wilson relation
can be written as @xcite [ eq : gen_sol ] d = d_c ( + a r d_c ) ^-1 = ( + a d_c r ) ^-1 d_c where @xmath17 is any chirally symmetric ( @xmath18 ) dirac operator which must violate at least one of the three properties * ( ii)*-*(iv ) * above .
now we must require @xmath17 to satisfy * ( iii ) * and * ( iv ) * , but violate * ( ii ) * ( i.e , @xmath17 is nonlocal ) , since ( [ eq : gen_sol ] ) can transform the nonlocal @xmath17 into a local @xmath1 on a finite lattice for @xmath19 with @xmath20 in the proper range @xcite , while the properties * ( iii ) * and * ( iv ) * are preserved .
moreover , the zero modes and the index of @xmath17 are invariant under the transformation @xcite .
that is , a zero mode of @xmath17 is also a zero mode of @xmath1 and vice versa , hence , [ eq : npm ] n_+ ( d_c ) = n_+ ( d ) , n_- ( d_c ) = n_- ( d ) , [ eq : index ] ( d_c ) = n_-(d_c ) - n_+(d_c ) = n_-(d ) - n_+(d ) = ( d ) .
since the massless dirac fermion operator in continuum is antihermitan , we also require that @xmath17 is antihermitian ( @xmath21 ) even at finite lattice spacing .
then the chiral symmetry of @xmath17 together with its antihermiticity implies that @xmath17 satisfies the @xmath22-hermiticity [ eq : hermit ] d_c^ = 5 d_c 5 .
this implies that @xmath1 in the general solution ( [ eq : gen_sol ] ) also satisfies the @xmath23-hermiticity [ eq : hermit_d ] d^ = 5 d 5 .
then the eigenvalues of @xmath1 are either real or come in complex conjugate pairs . furthermore , from ( [ eq : gen_sol ] ) , since @xmath16 is a positive definite hermitian operator , the lower bound of real eigenvalues of @xmath1 is zero , thus @xmath24 is real and nonnegative , and is amenable to hybrid monte carlo simulation for dynamical fermions with _ any _ number of fermion flavors . in particular , for @xmath19 with @xmath25 , the analysis [ eqs .
( 24)-(41 ) ] in ref .
@xcite goes through with trivial modifications .
the main results are : ( @xmath26 ) : : the eigenvalues of @xmath1 fall on a circle with center at @xmath27 , and radius @xmath27 , and have the reflection symmetry with respect to the real axis .
( @xmath28 ) : : the real eigenmodes ( if any ) at @xmath29 and @xmath30 have definite chirality @xmath31 or @xmath32 . ( @xmath33 ) : : the chirality of any complex eigenmodes is zero . ( @xmath34 ) : : total chirality of all eigenmodes must vanish .
@xmath35 where @xmath36 denotes the number of zero modes of positive ( negative ) chirality , and @xmath37 the number of @xmath30 modes of positive ( negative ) chirality . from ( @xmath34 ) , we immediately see that _ any zero mode must be accompanied by a real @xmath30 mode with opposite chirality _ , and the index of @xmath1 is [ eq : index_d ] ( d ) n_- - n_+ = - ( n_- - n_+ ) .
now the central problem is to construct the chirally symmetric @xmath17 which is nonlocal , and satisfies * ( iii ) * , * ( iv ) * , and ( [ eq : hermit ] ) .
furthermore we also require that @xmath17 is topologically proper ( i.e. , satisfying the atiyah - singer index theorem ) for any prescribed smooth gauge background .
these constitute the necessary requirements @xcite for @xmath17 to enter ( [ eq : gen_sol ] ) such that @xmath1 could provide a nonperturbative regularization for a massless dirac fermion interacting with a background gauge field .
explicitly , these necessary requirements are : ( a ) : : @xmath17 is antihermitian ( hence @xmath22-hermitian ) and it agrees with @xmath38 in the classical continuum limit .
( b ) : : @xmath17 is free of species doubling .
( c ) : : @xmath17 is nonlocal .
( d ) : : @xmath17 is well defined in topologically trivial background gauge field .
( e ) : : @xmath17 has zero modes as well as simple poles in topologically non - trivial background gauge fields ( each zero mode of @xmath17 must be accompanied by a simple pole of @xmath17 ) .
furthermore , the zero modes of @xmath17 satisfy the atiyah - singer index theorem for any prescribed smooth gauge background .
the general solution of @xmath17 satisfying these requirements had been investigated in ref .
@xcite and was reported by the author at chiral99 . however , in general , given any lattice dirac operator @xmath1 , there exists a transformation @xmath39 for @xmath1 such that the transformed dirac operator @xmath40 $ ] is chirally symmetric @xcite .
therefore , for the sake of completeness , the transformation @xmath39 is outlined in section 2 . then the construction in ref .
@xcite is reviewed in section 3 .
concluding remarks are briefly outlined in section 4 .
given any lattice dirac operator @xmath1 , in general , there are many different ways to construct a chirally symmetric @xmath17 out of @xmath1 . for example , we can construct [ eq : ds ] d_s = ( d - _ 5 d _ 5 ) which is chirally symmetric ( @xmath41 ) .
however , this transformation does not necessarily preserve the property * ( iii)*. for example , if we apply this transformation to the wilson - dirac operator ( [ eq : dw ] ) , we obtain @xmath42 , the naive fermion operator which suffers from the species doublings . although @xmath43 is free of species doublings in the continuum limit , the transformation ( [ eq : ds ] ) can not preserve this property since @xmath44 is local .
therefore , we need a transformaton which preserves the properties * ( iii ) * , * ( iv ) * , ( [ eq : hermit_d ] ) and ( [ eq : npm ] ) , but exchanges the locality of @xmath1 for its chiral symmetry at finite lattice spacing .
the transformation @xcite [ eq : tit ] ( r ) : d d= ( r)[d ] d ( + r d ) ^-1 = ( + d r ) ^-1 d which generalizes ( [ eq : gen_sol ] ) can serve our purposes .
evidently , the set of these transformations , @xmath45 , form an abelian group with group parameter space @xmath46 @xcite . for any @xmath1 satisfying @xmath22-hermiticity , ( [ eq : hermit_d ] ) , there exists a hermitian @xmath47 , [ eq : rc ] r_c = - ( d^-1 + _ 5 d^-1 _ 5 ) such that [ eq : d5d ] d_c = ( r_c ) [ d ] = 2 5 d ( 5 d - d 5 ) ^-1 d is chirally symmetric ( @xmath48 ) and antihermitian ( @xmath21 ) .
it is evident that for any two lattice dirac operators @xmath49 and @xmath50 satisfying ( [ eq : hermit_d ] ) , their corresponding chiral limits obtained from ( [ eq : d5d ] ) , say , @xmath51 and @xmath52 , are in general different .
however , they are related by the transformation [ eq : dct ] d_c^(1 ) = _ i t_i d_c^(2 ) t_i^ where each @xmath53 commutes with @xmath54 . in general , given any two @xmath51 and @xmath52 , it is a nontrivial task to obtain all @xmath53 in ( [ eq : dct ] ) . for the wilson - dirac operator ( [ eq : dw ] ) , the transformation ( [ eq : d5d ] ) gives [ eq : dtwc ] d_c = _ t_- w ( _ t_)^-1 w
. it is antihermitian , nonlocal ( due to the second term ) , and satisfies * ( i ) , ( iii ) * and * ( iv)*. then we can substitute ( [ eq : dtwc ] ) into ( [ eq : gen_sol ] ) with @xmath19 to obtain a gw dirac operator [ eq : dw_gw ] d = where [ eq : c ] c = ( ^_t _ ) - w ( _ t_)^-1 w . on a finite lattice , the gw dirac operator ( [ eq : dw_gw ] ) can be constructed to be local but not highly peaked in the diagonal elements if the value of @xmath20 is within a proper range @xcite . in other words ,
if @xmath20 is too small , then @xmath1 must be nonlocal since @xmath1 is close to @xmath17 .
on the other hand , if @xmath20 is too large , then @xmath1 becomes highly peaked in the diagonal elements .
in these two extreme cases , @xmath1 does not respond properly to the background gauge field ( e.g. , the anomaly function of @xmath1 does not agree with the chern - pontryagin density of the prescribed gauge background ) . in the limit
the number of sites in each dimension goes to infinity , the proper range of @xmath20 extends to @xmath55 . the anomaly function of @xmath1 ( [ eq : dw_gw ] )
can be written [ eq : anxc ] = 2 ( x , x ) where @xmath56 denotes the trace over the color , flavor and spinor space .
it is instructive to compare the anomaly function of the wilson - dirac operator @xmath43 ( [ eq : dw ] ) to that of the gw wilson - dirac operator @xmath1 ( [ eq : dw_gw ] ) , in a topologically trivial background gauge field , on a finite lattice , as shown in fig . 1 in ref .
@xcite and fig . 1 in ref
@xcite , respectively .
the anomaly function of @xmath43 is very different from the chern - pontryagin density @xmath57 on a @xmath58 torus , while that of the gw dirac - wilson operator @xmath1 is in good agreement with @xmath57 at each site .
this demonstrates that the transformation @xmath59 plays the important role in converting @xmath43 into a gw dirac operator @xmath60 $ ] which is free of @xmath15 lattice artifacts , thus @xmath1 can reproduce the correct anomaly function even on a _ finite _ lattice provided that the local fluctuations of the background gauge field are not too violent .
however , a lattice dirac operator satisfying properties * ( i)-(iv ) * , ( [ eq : hermit_d ] ) and the gw relation ( [ eq : gwr ] ) does not guarantee that it has the correct anomaly function in topologically nontrivial gauge backgrounds .
this can be seen as follows .
it is well known that the sum of the axial anomaly over all sites of a finite lattice is equal to the index of the lattice dirac operator .
therefore _ the necessary condition for a lattice dirac operator to have the correct anomaly function is that it possesses exact zero modes satisfying the atiyah - singer index theorem _
i.e. , @xmath61 . however , not every lattice dirac operator has exact zero modes in topologically nontrivial gauge backgrounds .
for example , the wilson - dirac operator , @xmath43 in ( [ eq : dw ] ) , it does not have exact zero modes in topologically nontrivial sectors .
then , according to ( [ eq : npm ] ) , @xmath60 $ ] also does not have any exact zero modes , even though @xmath1 satisfies the gw relation ( [ eq : gwr ] ) , and properties * ( i)-(iv ) * and ( [ eq : hermit_d ] ) .
this implies that the anomaly function of @xmath43 or @xmath60 $ ] does not agree with the chern - pontryagin density for topologically nontrivial gauge fields on _ any _ finite lattices .
consequently , _ the disagreement must persist in the continuum limit @xmath3_.
this provides an example to illustrate that any lattice dirac operator @xmath1 must possess a nonperturbative attribute , the topological characteristics , @xmath62 $ ] @xcite . in general , @xmath62 $ ] is a rational number , a functional of @xmath1 and the gauge configuration . in a gauge background with nonzero topological charge @xmath63 ,
the index of a lattice dirac operator is [ eq : cd ] n_- - n_+ = c[d ] q . in the case of @xmath43 and
@xmath60 $ ] ( [ eq : dw_gw ] ) , @xmath64 , thus the topological characteristics @xmath65 = c[d ] = 0 $ ] . for the vacuum sector with nonzero topological charge density ( i.e. , @xmath66 but @xmath67 ) , @xmath62 $ ] can not be defined by ( [ eq : cd ] ) , however , it can be defined unambiguously @xcite in the context of the anomaly function .
we note in passing that for the overlap - dirac operator @xmath68 @xcite , the transformation ( [ eq : d5d ] ) gives @xmath69 , as expected . substituting @xmath17 into ( [ eq : gen_sol ] ) with @xmath19
, we obtain the generalized overlap - dirac operator [ eq : do_gw ] d = 2 ( + v ) [ - v + 2 r ( + v ) ] ^-1 .
suppose that we do not have any lattice dirac operator to begin with .
it is still possible for us to construct @xmath17 @xcite according to the necessary requirements * ( a)-(e ) * listed in section 1 .
since @xmath17 is antihermitian , there exists one to one correspondence between @xmath17 and a unitary operator @xmath70 , d_c = ( + v ) ( - v ) ^-1 , v = ( d_c - ) ( d_c + ) ^-1 . [ eq : vdc ] where @xmath70 also satisfies the @xmath22-hermiticity , @xmath71 .
then the unitary operator @xmath70 can be expressed in terms of a hermitian operator @xmath72 , [ eq : v5h ] v = _ 5 h = ( cc & 0 + 0 & - ) ( cc h_1 & h_2 + h_2^ & h_3 ) = ( cc h_1 & h_2 + -h_2^ & -h_3 ) where @xmath73 and @xmath74 . using the unitarity condition @xmath75 , we have @xmath76 , [ eq : h^2 ] h^2 = ( cc h_1 ^ 2 + h_2 h_2^ & h_1 h_2 + h_2
h_3 + h_2^ h_1 + h_3 h_2^ & h_2^ h_2 + h_3 ^ 2 ) = ( cc & 0 + 0 & ) then we obtain d_c [ eq : dc ] = where @xmath77 . the general solution to eq .
( [ eq : h^2 ] ) can be written as [ eq : h1 ] h_1 & = & + [ eq : h2 ] h_2 & = & b e^ i where @xmath78 is an arbitrary phase , and @xmath79 is any operator . in the following
we will restrict @xmath80 to zero , and also pick the minus sign for @xmath81 .
then the general solution for @xmath17 can be written in the following form [ eq : dl ] d_l & = & b^-1 + [ eq : dr ] d_r & = & ^-1 b due to the presence of the square root in ( [ eq : dl ] ) and ( [ eq : dr ] ) , @xmath17 is nonlocal for nontrivial @xmath82 , thus the requirement * ( b ) * is satisfied . in the following ,
we outline a construction of @xmath82 which satisfies constraints * ( a)-(d ) * but not * ( e)*. in order to have @xmath17 satisfy the constraint * ( a ) * , we first try [ eq : b ] b = w^-1 _ _ t_where @xmath83 and @xmath84 are defined in ( [ eq : tmu ] ) and ( [ eq : sigma ] ) respectively , and @xmath12 is a non - singular hermitian operator which is trivial in the dirac space and goes to a constant in the classical continuum limit operator in this section is different from the wilson term in eq .
( [ eq : wilson ] ) . ] .
( [ eq : dl ] ) and ( [ eq : dr ] ) become [ eq : dlw ] d_l & = & ( t ) ^-1 + [ eq : drw ] d_r & = & - d_l^ where [ eq : sigma_t ] t & = & _ _ t _ + [ eq : t2 ] t^2 & = & - ( t ) ( ^ t ) however , @xmath83 suffers from species doublings . now
our task is to construct a hermitian operator @xmath12 such that the doubled modes are completely decoupled from the fermion propagator @xmath85 , even at finite lattice spacing . as discussed in ref .
@xcite , we want to construct a hermitian @xmath12 such that in the free fermion limit , it satisfies the following condition , [ eq : wp ] w(p ) = \ { ll > 0 & + < 0 & + .
a simplest solution to ( [ eq : wp ] ) is [ eq : wpc ] w(p ) = c - 2 _ ^2 ( p_a / 2 ) , c ( 0 , 2 ) note that the role of @xmath12 in the general solution of @xmath17 is quite different from the wilson term ( [ eq : wilson ] ) in the wilson - dirac operator ( [ eq : dw ] ) . in the general solution of @xmath17 , the chiral symmetry is always preserved , and _ the role of @xmath12 is to suppress the doubled modes completely at finite lattice spacing _ , while in the wilson - dirac operator , the wilson term breaks the chiral symmetry explicitly and gives a mass of order @xmath86 to the doubled modes such that they can be decoupled in the continuum limit ( @xmath3 ) .
after the gauge links are restored , @xmath12 in the position space becomes [ eq : wxy ] w(x , y ) = c _ x , y - _ this is one of the simplest solution of @xmath12 satisfying the requirement ( [ eq : wp ] ) in the free fermion limit .
certainly , there exists other solutions to ( [ eq : wp ] ) . substituting the @xmath17 [ eqs .
( [ eq : dlw])-([eq : drw ] ) with @xmath83 in ( [ eq : tmu ] ) , and @xmath12 in ( [ eq : wxy ] ) with @xmath87 , and the normalization constant @xmath88 ] into ( [ eq : gen_sol ] ) with @xmath19 , we obtain the gw dirac operator [ eq : twc ] d = where [ eq : bwt ] b & = & ( w - ) ^-1 ( t ) the anomaly function of @xmath1 is [ eq : anx ] = 2 r^2 ( x , x ) where @xmath56 stands for the trace over the color , flavor and spinor space . since the @xmath17 of ( [ eq : twc ] ) in the free fermion limit is free of species doubling and has the correct continuum behavior , the perturbation calculation in ref .
@xcite showed that @xmath89 has the correct chiral anomaly in topologically trivial gauge backgrounds .
this has been verified explicitly on a finite lattice .
an example is shown in fig . 2 of ref .
@xcite .
however , for topologically nontrivial sectors , the anomaly function of @xmath1 also depends on the topological characteristics , @xmath62 $ ] , which is a non - perturbative attribute of @xmath1 . since @xmath1 does not have any exact zero modes in nontrivial sectors , the anomaly function of @xmath1 does not agree with the chern - pontryagin density for nontrivial gauge backgrounds .
now we have two examples of gw dirac operators , ( [ eq : dw_gw ] ) and ( [ eq : twc ] ) , which have the correct axial anomaly in the vacuum sector but not in the nontrivial sectors .
in general , if @xmath17 is well defined in nontrivial sectors ( i.e. , not satisfying * ( e ) * ) , then @xmath90 = \mbox{index}[d ] = 0 $ ] @xcite ( i.e. , @xmath62 = c[d_c ] = 0 $ ] , @xmath1 and @xmath17 are _ topologically trivial _ ) . on the other hand ,
if @xmath17 has zeros and poles , then @xmath62 \ne 0 $ ] , however , it does not necessarily imply that @xmath62 = 1 $ ] .
sometimes @xmath62 $ ] may even become a fraction ( see table 4 in ref .
the important point is that in general we can not predict the value of @xmath62 $ ] in the topologically nontrivial sectors solely based on the properties of @xmath1 in the free fermion limit or in the vacuum sector .
the necessary conditions for a lattice dirac operator @xmath1 to provide a nonperturbative regularization for massless fermion interacting with a background gauge field are * ( a)-(c ) * listed in section 1 . the conditions * ( a ) * and * ( b ) * can be satisfied if @xmath1 fulfils * ( i)-(iv ) * and ( [ eq : hermit_d ] ) , and the chiral symmetry is broken according to the ginsparg - wilson relation ( [ eq : gwr ] ) .
since any gw dirac operator can be represented by the transformation ( [ eq : gen_sol ] ) [ preserving * ( iii ) , ( iv ) * , ( [ eq : hermit_d ] ) and ( [ eq : npm ] ) ] on a chirally symmetric @xmath17 , then the conditions for @xmath1 become the necessary requirements * ( a)-(e ) * for @xmath17 .
it should be emphasized that the requirement * ( c ) * for @xmath1 ( or the requirement * ( e ) * for @xmath17 ) can not be expressed in terms of any conditions in the free fermion limit or in the vacuum sector .
this naturally leads to the concept of topological characteristics @xcite associated with each lattice dirac operator @xmath1 .
although we have no problems to construct @xmath1 to satisfy * ( i)-(iv ) * , ( [ eq : hermit_d ] ) and ( [ eq : gwr ] ) , it remains an interesting question whether one can construct a topologically proper @xmath1 on a four dimensional lattice without square root operations .
if one has a lattice dirac operator @xmath1 satisfying * ( i)-(iv ) * , ( [ eq : hermit_d ] ) and ( c ) , then one can transform @xmath1 into a gw dirac operator through the transformation @xmath91 $ ] which not only preserves all essential physics of @xmath1 but also can restrict the quantum corrections to behave properly such that @xmath1 is free of @xmath15 lattice artifacts , additive mass renormalization , etc .
recently it has been shown that the effective four - dimensional action ( @xmath92 ) of the domain wall fermion is local and satisfying the gw relation , for gauge fields with small field strength @xcite . however , it is possible to transform the domain - wall fermion operator into a local gw dirac operator even at finite @xmath93 @xcite , then the anomalous effects ( e.g. , the residual pion mass in domain - wall qcd ) due to chiral symmetry violations at finite @xmath93 could be suppressed even at moderate @xmath93 .
further investigations are required before one can tell whether this scenario can be realized or not .
i would like to thank all participants of chiral99 , for their inspiring talks , interesting questions , and enlightening discussions .
this work was supported by the national science council , r.o.c . under the grant number nsc89 - 2112-m002 - 017 . | we discuss how to formulate dirac fermion operator on a finite lattice such that it can provide a nonperturbative regularization for massless fermion interacting with a background gauge field . 11.15.ha , 11.30.rd , 11.30.fs |
the continuous search for the physical mechanism which sets the cosmic acceleration of the universe and the corresponding conditions for stability has stimulated interesting and sometimes fascinating discussions on cosmological models @xcite .
the dynamical mass attributed to neutrinos or dark matter @xcite can , for instance , regulate the time evolution of the dynamical dark energy providing the setup of the cosmic acceleration followed by the cosmological stability . in this context ,
the coupling of mass varying dark matter with neutrinos yields interesting relations between the present mass of neutrinos and the dark energy equation of state . in a previous issue
@xcite , it was demonstrated that an effective generalized chaplygin gas ( gcg ) scenario @xcite can be reproduced in terms of a dynamical dark energy component @xmath0 with equation of state given by @xmath1 and a cold dark matter ( cdm ) component with a dynamical mass driven by the scalar field @xmath0 .
dark matter is , most often , not considered in the mass varying neutrino ( mavan ) models .
the treatment of dark energy and dark matter in the gcg unified scheme naturally offers this possibility . identifying sterile neutrinos as dark matter coupled with dark energy provides the conditions to implement such unified picture in the mavan formulation since active and sterile neutrino states are connected through the _ seesaw _ mechanism for mass generation .
the constraints imposed by the _
seesaw _ mechanism allows one to establish an analytical connection to the gcg in terms of a real scalar field .
the dynamics of the coupled fluid composed by neutrinos , dark matter and dark energy is driven by one single degree of freedom , the scalar field , @xmath2 .
the simplest realization of mavan mechanisms @xcite consists in writing down an effective potential which , in addition to a scalar field dependent term , contains a term related to the neutrino energy density .
it results in the so - called adiabatic condition , which sometimes leads to the stationary regime for the scalar field in respect with an effective potential @xcite .
one indeed expect a tiny contribution from cosmological neutrinos to the energy density of the universe .
mavan scenarios essentially predict massless neutrinos until recent times .
when their mass eventually grows close to its present value , they form a non - relativistic ( nr ) fluid and the interaction with the scalar field stops its evolution .
the relic particle mass is generated from the vacuum expectation value of the scalar field and becomes linked to its dynamics : @xmath3 .
it is presumed that the neutrino mass has its origin on the vacuum expectation value ( vev ) of the scalar field and its behavior is governed by the dependence of the scalar field on the scale factor .
in fact , it is well - known that the active neutrino masses are tiny as compared to the masses of the charged fermions .
this can be understood through the symmetry of the standard model ( sm ) of electroweak interactions .
it involves only left - handed neutrinos such that no renormalizable mass term for the neutrinos is compatible with the sm gauge symmetry @xmath4 .
once one has assumed that baryon number and lepton number is conserved for renormalizable interactions , neutrino masses can only arise from an effective dimension five operator .
it involves two powers of the vacuum expectation value of the higgs doublet .
they are suppressed by the inverse power of a large mass scale @xmath5 of a sterile right - handed majorana neutrino , since it has hypercharge null in the sm .
this super - massive majorana neutrino should be characteristic for lepton number violating effects within possible extensions beyond the sm . at our approach
, the mass scale @xmath5 has its dynamical behavior driven by @xmath0 . in some other words
, the sterile neutrino mass is characteristic of the aforementioned mass varying dark matter which indirectly results in mavan s , i. e. active neutrinos with mass computed through _ seesaw _ mechanism . in section
ii , we report about the main properties of a unified treatment of dark matter and dark energy prescribed by the mass varying mechanism .
one sees that model dependent choices of dynamical masses of dark matter allows for reproducing the conditions for the present cosmic acceleration in an effective gcg scenario .
the stability conditions resulted from a positive squared speed of sound , @xmath6 , is recovered . in section iii , we discuss the neutrino mass generation mechanism in the context of the gcg model . following the simplest formulation of the _ seesaw _ mechanism , dirac neutrino masses with analytical dependencies on @xmath0 , @xmath7 and @xmath8 are considered . in section iv
, we discuss the conditions for stability and the perturbative modifications on the accelerated expansion of the universe in the framework here proposed .
one can state the conditions for reproducing the gcg scenario .
we draw our conclusions in section v.
to understand how the mass varying mechanism takes place for different particle species , it is convenient to describe the relevant physical variables as functionals of a statistical distribution @xmath9 .
this counts the number of particles in a given region around a point of the phase space defined by the conjugate coordinates : momentum , @xmath10 , and position , @xmath11 .
the statistical distribution @xmath9 can be defined in terms of the comoving momentum , @xmath12 , and for the case where @xmath9 is a fermi - dirac distribution , it can be written as @xmath13 } + 1\right\}^{\mi\1},\ ] ] where @xmath14 is the relic particle background temperature at present . in the flat frw scenario , the corresponding particle density ,
energy density and pressure can thus be depicted from the einstein s energy - momentum tensor @xcite as @xmath15 where @xmath16 is the scale factor ( cosmological radius ) for the flat frw universe , for which the metrics is given by @xmath17 . from the dependence of @xmath18 on @xmath16 one
can obtain the energy - momentum conservation equation , @xmath19 that translates the dependence of @xmath5 on @xmath0 into a dynamical behavior , and where @xmath20 is the expansion rate of the universe and the _ overdot _ denotes differentiation with respect to time ( @xmath21 ) .
simple mathematical manipulations allow one to easily demonstrate that @xmath22 from which , one can see that the strength of the coupling between relic particles and the scalar field is suppressed by the relativistic increase of pressure ( @xmath23 ) .
as long as particles become ultra - relativistic ( @xmath24 ) the matter fluid and the scalar field fluid tend to decouple and evolve adiabatically separated . adding eq .
( [ gcg03 ] ) to the equation of motion for a cosmon field @xmath0 like @xmath25 rewritten as @xmath26 results in the equation of energy conservation for a unified fluid with a dark energy component and a mass varying dark matter component , @xmath27 for which we identify @xmath28 and @xmath29 . from this point
, it is suggested that such a unified fluid corresponds to an effective description of a gcg universe .
the gcg model is characterized by an exotic equation of state @xcite given by @xmath30 which can be obtained from a generalized born - infeld action @xcite . inserting the above equation of state into the unperturbed energy conservation eq .
( [ gcg06 ] ) and following a straightforward integration @xcite , one obtains @xmath31^{\1/(\1 \pl \al ) } , \label{gcg21}\ ] ] and @xmath32^{-\al/(\1 \pl \al)}. \label{gcg22}\ ] ] one of the most striking features of the gcg fluid is that its energy density interpolates between a dust dominated phase , @xmath33 , in the past , and a de - sitter phase , @xmath34 , at late times .
this property makes the gcg model an interesting candidate for the unification of dark matter and dark energy .
notice that for @xmath35 , gcg behaves always as matter whereas for @xmath36 , it behaves always as a cosmological constant .
hence to use it as a unified candidate for dark matter and dark energy one has to exclude these two possibilities so that @xmath37 must lie in the range @xmath38 .
furthermore , this evolution is controlled by the model parameter @xmath39 .
assuming the canonical parametrization of @xmath18 and @xmath40 in terms of a scalar field @xmath0 , @xmath41 and following ref .
@xcite , one can obtain through eq .
( [ gcg21]-[gcg22 ] ) the effective dependence of @xmath0 on @xmath16 implicitly given by @xmath42^{-\al/(\al \pl \1 ) } , \label{pap02}\ ] ] and explicit expressions for @xmath18 , @xmath40 and @xmath43 in terms of @xmath0 . assuming a flat evolving universe described by the friedmann equation @xmath44 ( with @xmath45 in units of @xmath46 and @xmath18 in units of @xmath47 , one obtains @xmath48 } , \label{pap03}\ ] ] where it is assumed that @xmath49}. \label{pap04}\ ] ] one then readily finds the scalar field potential , @xmath50^{\frac{\2}{\al \pl \1 } } + \left[\cosh{\left(3\bb{\alpha + 1 } \phi/2\right)}\right]^{-\frac{\2\al}{\al \pl \1 } } \right\}. \label{pap05}\ ] ] if one supposes that the energy density , @xmath18 , may be decomposed into a mass varying cdm component , @xmath51 , and a dark energy component , @xmath52 , connected by the scalar field equations ( [ gcg04])-([gcg05 ] ) , the equation of state ( [ gcg20 ] ) is just assumed as an effective description of the cosmological background fluid of the universe . since the cdm pressure , @xmath53 , is null , the dark energy component of pressure , @xmath54 , results in the gcg pressure , @xmath55 .
assuming that dark energy behaves like a cosmological constant , that is , its equation of state is given by @xmath56 , the dark energy density can be parameterized by a generic quintessence potential , @xmath57 , since its kinetic component has to be null for a canonical formulation .
it results in @xmath58 , where @xmath40 is the gcg pressure given by eq .
( [ gcg22 ] ) . by substituting the result of eq.([pap03 ] ) into the eq.([gcg22 ] ) , and observing that @xmath44 , with @xmath18 given by eq.([gcg21 ] ) , it is possible to rewrite the gcg pressure , @xmath40 , in terms of @xmath0 .
it results in the following analytical expression for @xmath59 , @xmath60^{-\frac{\2 \al}{1 \pl \al } } , \label{pap08}\ ] ] which is consistent with the result for @xmath61 from eq .
( [ pap05 ] ) . since @xmath62 , the eq .
( [ gcg05 ] ) is thus reduced to @xmath63 and the problem is then reduced to finding a relation between the scalar potential @xmath59 and the variable mass @xmath64 . from the above equation , the effective potential governing the evolution of the scalar field is naturally decomposed into a sum of two terms , one arising from the original quintessence potential @xmath59 , and other from the dynamical mass @xmath64 . for appropriate choices of potentials and coupling functions satisfying eq .
( [ pap09 ] ) , the competition between these terms leads to a minimum of the effective potential . in the adiabatic regime ,
the matter and the scalar field are tightly coupled together and evolve as one effective fluid . at our approach ,
once one assumes a @xmath65-_like _ equation of state @xmath66 , without any additional constraint on cosmon field equations , eq .
( [ pap09 ] ) is naturally obtained . in the gcg
cosmological scenario , the effective fluid description is valid for the background cosmology and for linear perturbations .
the equation of state of perturbations is the same as that of the background cosmology where all the effective results of the gcg paradigm are maintained .
( [ pap08 ] ) leads to @xmath67 which , in the cdm limit , gives @xmath68 since the dependence of @xmath5 on @xmath16 is exclusively intermediated by @xmath2 , i. e. @xmath69 , from eqs .
( [ gcg21 ] ) , ( [ gcg22 ] ) and ( [ pap03 ] ) , after simple mathematical manipulations , one obtains @xmath70^{\frac { \2 \al}{1 \pl \al } } \label{pap11}\ ] ] which is consistent with eq .
( [ pap09 ] ) . from the above result
, one can infers that the adequacy to the adiabatic regime is left to the mass varying mechanism which drives the cosmological evolution of the dark matter component .
for mass varying cdm coupled with @xmath65-_like _ dark energy , with @xmath66 , the gcc leads to similar predictions for the equation of state , @xmath71 , independently of the scale parameter @xmath16 .
the same is not true for hot dark matter ( hdm ) @xcite .
unlike photons and baryons , cosmological neutrinos have not been observed , so arguments about their contribution to the total energy density of the universe are necessarily theoretical .
otherwise , neutrinos coupled to dark energy can lead to a number of significant phenomenological consequences .
the neutrino mass should be a dynamical quantity and neutrinos remain essentially massless until recent times .
when their mass eventually increases close to its present value , their interaction with the background scalar field almost ceases @xcite .
the energy of the scalar field becomes the dominant contribution to the energy density of the universe and cosmic acceleration ensues . without loss of generality , such a behavior
can be easily implemented to a degenerate fermion gas ( dfg ) of neutrinos . in the limit where @xmath14 tends to @xmath72 in eq .
( [ gcg01 ] ) , the fermi distribution @xmath9 becomes a step function that yields an elementary integral for the above equations , with the upper limit equal to the fermi momentum here written as @xmath73 .
it results in the equations for a dfg @xcite , which can be useful for parameterizing the transition between ultra - relativistic ( ur ) and non - relativistic ( nr ) thermodynamic regimes .
the equation of state can be expressed in terms of elementary functions of @xmath74 and @xmath75 , @xmath76\\ p_m\bb{a } & = & \frac{1}{8 \pi^{\2 } } \left[\beta ( \frac{2}{3 } \beta^{\2 } - m^{\2})\sqrt{\beta^{\2 } + m^{\2 } } + \mbox{arc}\sinh{\left(\beta / m\right)}\right]\nonumber \label{gcg01d}\end{aligned}\ ] ] the relation given by eq .
( [ gcg02 ] ) can be verified for the above definition . besides the mass dependence
, it is necessary to determine for which values of the scale factor the neutrino - scalar field coupling becomes important .
the dfg approach take care of this .
let us then turn back to the explanation for the smallness of the neutrino masses . according to the _ seesaw _ mechanism , the tiny masses , @xmath77 , of the usual left - handed neutrinos are obtained through a very massive , @xmath5 , sterile right - handed neutrino . the lagrangian density that describes the simplest version of the _ seesaw _ mechanism via yukawa coupling between a light scalar field and a single neutrino flavor
is given by @xmath78 where it is shown that at scales well below the right - handed neutrino mass , one has the effective lagrangian density @xcite @xmath79 phenomenological consistency with the sm implies that logarithm corrections to the above terms are small , while it is well - know from the results of solar , atmospheric , reactor and accelerator neutrino oscillation experiments that neutrino masses given by @xmath80 lie in the sub-@xmath81 range .
it is also clear that promoting the scalar field @xmath0 into a dynamical quantity leads to a mechanism in the context of which neutrino masses are time - dependent . associating the scalar field to the dark energy field allows linking nr neutrino energy densities to late cosmological times @xcite .
this scenario can be implemented through the perturbative approach via eq .
( [ gcg06 ] ) @xcite .
it is evident that this approach is fairly general , as well as independent of the choice of the equation of state and of the dependence of the neutrino mass on the scalar field .
the form of @xmath64 and of the equation of state can indeed lead to quite different scenarios .
in particular , we consider three cases of neutrino mass dependence on @xmath0 . in case 01
we set @xmath82 , in case 02 we have set @xmath83 , and in case 03 we have set @xmath84 .
energy densities , @xmath18 , as function of the scale factor for the gcg fluid and for the decomposed components of the unified effective gcg background fluid , namely , the mass varying dark matter , @xmath85 , the cosmon-_like _ dark energy , @xmath52 , and the perturbative dfg of neutrinos , @xmath86 , are obtained in the fig .
[ gcg-0 ] .
the neutrino densities are computed for neutrinos masses correlated to the dark matter mass , @xmath64 , in correspondence with the gcg scenario with @xmath87 , that represents a realistic fraction of dark energy , and with @xmath88 and @xmath89 . in the fig .
[ gcg - b ] we have taken into account the energy densities for photons ( @xmath90 ) and baryons ( @xmath91 ) in complement to our effective gcg scenario .
then we have computed the relative modifications on the spectrum of the cosmic evolution of the dark sector components . conveniently in agreement with the phenomenological predictions ,
we have set the present values of the energy densities as : @xmath92 , @xmath93 , @xmath94 , @xmath95 , and @xmath96 .
assuming that neutrinos are nr at present , and observing the behavior of eqs .
( [ gcg01d ] ) , we have set @xmath97 .
this condition allows one to establish the correspondence between the values of @xmath16 for which the coupling transition takes place , i. e. the transition between relativistic and nr regimes , so that the neutrino masses achieve the present - day values . for our model
such a transition is governed by the dfg behavior of eq .
( [ gcg01d ] ) .
these scenarios are by no means the only possibilities .
in particular , we choose them as they correspond to the simplest feasible possibility for active neutrino mass generation , given that the scalar field has mass dimension one @xcite .
the essential information of the mass dependence on the scale factor is that , for case 03 , the neutrino mass increases with decreasing @xmath0 , in opposition to what happens for cases 01 and 02 . through the stability analysis performed in the next section
we shall elucidate some important consequences due to such a distinct behavior . in the fig .
[ gcg - c ] we verify how the energy density @xmath18 and the corresponding equations of state @xmath71 for the composed fluid deviate from the effective gcg scenario . for mass varying cdm coupled with dark energy with @xmath66 , the effective gcc leads to similar predictions for @xmath98 , independently of the scale factor @xmath16 .
the same is not true for hdm which , in the dfg approach , while weakly coupling with dark energy , leads to the same behavior of the gcg just at late times ( @xmath99 ) .
the possibility of adiabatic instabilities in cosmological scenarios was previously pointed out @xcite in a context of a mass varying neutrino model of dark energy . in opposition , in the usual treatment where dark matter are just coupled to dark energy , cosmic expansion together with the gravitational drag due to cold dark matter have a major impact on the stability of the cosmological background fluid .
usually , for a general fluid for which we know the equation of state , the dominant effect on the sound speed squared @xmath6 arises from the dark sector component and not from the neutrino component . for the models where the adiabatic regime ( cf .
( [ pap09 ] ) ) implies an equation of state reproducing cosmological constant , @xmath65 , effects , @xmath100 , one obtains @xmath101 from the very start of the analysis .
the effective gcg is free from this inconsistency .
a detailed quantitative analysis of the stability conditions for the gcg in terms of the squared speed of sound is discussed in ref .
positive @xmath6 implies that @xmath102 in the gcg equation of state .
the coupling of the dark energy component with dynamical dark matter is responsible for removing such inconsistency by setting @xmath103 . in the fig .
[ gcg - d ] we show the results for @xmath104 for a cosmological background scenario corresponding to the sum of the energy components of an effective gcg fluid : mass varying dark matter ( dm ) , cosmon-_like _ dark energy ( de ) and perturbative dfg of neutrinos ( @xmath105 ) .
we have considered the gcg phenomenological given by @xmath87 and @xmath106 ( solid line ) , @xmath107 ( dot line ) , @xmath108 ( dash dot line ) , and @xmath89 ( dash line ) .
the perturbative influence of neutrinos on the positiveness of @xmath104 has its magnitude measured by comparing the results with those for the gcg scenario with and and without neutrino perturbation for each one of the three cases in correspondence with fig .
[ gcg-0 ] . in this scenario ,
the most important result illustrated by fig .
[ gcg - d ] is that growing neutrino masses parameterized by @xmath109 ( case 03 ) results in instabilities at present times .
this instability is characterized by a negative squared speed of sound for an effective coupled neutrino / dark energy fluid , and results in the exponential growth of small scale modes @xcite . a natural interpretation for this
is that the universe becomes inhomogeneous with the neutrinos forming denser structures or lumps @xcite .
effectively , the scalar field could mediate an attractive force between neutrinos leading to the formation of neutrino nuggets .
anyway , these results are consistent with the accelerated expansion of the universe ruled by the dynamical masses of cdm and neutrinos since positive values for @xmath110 are observed , as we can notice in the fig .
[ gcg - e ] .
growing neutrinos coupled to mass varying dark matter plus cosmon-_like _ dark energy were studied assuming that the cosmological background unified fluid presents an effective behavior similar to that of the gcg .
the essential ingredient of this class of models is a neutrino mass that , through the _ seesaw _ mechanism , depends on the cosmon field and grows in the course of the cosmological evolution .
our setup is an effective gcg decomposed into two interacting components @xcite .
the first one behaves like mass varying cdm since it is pressure - less .
the second one corresponds to a cosmon-_like _ dark energy component @xmath0 with equation of state given by @xmath1 .
apparently the model does not look different from the interacting quintessence models where one has two different interacting fluids .
the present fraction of dark energy is set by a dynamical mechanism .
as soon as dark matter become nr , their coupling to the cosmon triggers an effective stop ( or substantial slowing ) of the evolution of the cosmon . before this event
, the quintessence field follows the tracking of the minimum of an effective potential , in a kind of stationary condition which drives the adiabatic regime , and for which the coupling strength is strong compared to gravitational strength . in the scope of finding a natural explanation for the cosmic acceleration and the corresponding adequation to stability conditions ,
we have proposed a systematic procedure to treat variations of fundamental parameters of neutrino cosmology ( @xmath0 , @xmath77 and @xmath5 ) independent of any particular theoretical model which enforces restrictive relations among these parameters .
we have quantified the perturbative influence of such variations on the calculation of the squared speed of sound and on the cosmic acceleration of the universe , noticing that the former quantity is much more sensible to variations with respect to the exact gcg background cosmology .
an interesting feature is however that the gcg has its cosmological evolution reproduced by a cdm mass varying mechanism consistent with the mass generation mechanism for neutrinos .
most of dark sector models predict an entangled mixture of interacting dark matter and dark energy .
we have consistently introduced the possibility of such a mixed coupling with active and sterile neutrinos , for which we have tested three hypothesis for _ seesaw _ masses . at our approach , the mass scale @xmath5 and scalar field @xmath0
are tightly coupled together and evolve as a unique effective fluid .
the effective potential driving the evolution of the scalar field is decomposed into a sum of two terms , one arising from the original quintessence potential @xmath59 , and the other from the dynamical mass of the dark matter .
the sterile neutrino mass produces the mass varying dark matter effects which indirectly result in mavan s , and the active neutrino has its mass perturbatively computed via _ seesaw _ relations . since the results for the cosmic acceleration followed by cosmological stability for gcg scenarios are well - defined , one can assert that the increase of neutrino mass acts as a cosmological clock or trigger for the crossover to the effective scenario here studied .
in fact , a scalar field associated to dark energy in connection with the sm neutrinos and the electroweak interactions may bring important insights on the physics beyond the sm @xcite . the case of cosmological neutrinos , in particular , is a fascinating example where salient questions concerning sm particle phenomenology can be addressed and hopefully better understood .
99 i. zlatev , l. m. wang and p. j. steinhardt , phys .
lett . * 82 * , 896 ( 1999 ) .
l. m. wang , r. r. caldwell , j. p. ostriker and p. j. steinhardt , astrophys .
j. * 530 * , 17 ( 2000 ) .
p. j. steinhardt , l. m. wang and i. zlatev , phys . rev . *
d59 * , 123504 ( 1999 ) .
t. barreiro , e. j. copeland and n. j. nunes , phys . rev . * d61
* , 127301 ( 2000 ) . o. bertolami and p. j. martins , phys . rev . * d61 * , 064007 ( 2000 ) .
r. bean , e. e. flanagan and m. trodden , arxiv:0709.1128 [ astro - ph ] .
kari enqvist , steen hannestad , and martin s. sloth , phys .
lett . * 99 * , 031301 ( 2007 ) ; m. lattanzi and j. w. f. valle , phys .
lett . * 99 * , 121301 ( 2007 ) ; c. arina , f. bazzocchi , n. fornengo , j. c. romao , and j. w. f. valle , phys .
* 101 * , 161802 ( 2008 ) .
a. e. bernardini , phys . rev .
* d80 * , 103522 ( 2009 ) .
a. kamenshchik , u. moschella and v. pasquier , phys . lett . *
b511 * , 265 ( 2001 ) ; n. bili , g. b. tupper and r. d. viollier , phys . lett . * b535 * , 17 ( 2002 ) .
m. c. bento , o. bertolami and a. a. sen , phys
. rev . * d66 * , 043507 ( 2002 ) ; m. c. bento , o. bertolami and a. a. sen , phys . rev .
* d67 * , 063003 ( 2003 ) ; phys .
b575 * , 172 ( 2003 ) ; l. amendola , f. finelli , c. burigana and d. carturan , jcap * 0307 * , 005 ( 2003 ) .
o. bertolami , a. a. sen , s. sen and p. t. silva , mon . not .
. soc . * 353 * , 329 ( 2004 ) .
t. krger , m. neubert and c. wetterich , phys . lett .
* b663 * , 21 ( 2008 ) .
p. q. hung , arxiv : hep - ph/0010126 .
p. gu , x. wang and x. zhang , phys . rev .
* d68 * , 087301 ( 2003 ) .
r. fardon , a. e. nelson and n. weiner , jcap * 0410 * 005 ( 2004 ) .
r. d. peccei , phys . rev . *
d71 * , 023527 ( 2005 ) .
a. e. bernardini and o. bertolami ; phys .
b662 * , 97 ( 2008 ) .
a. w. brookfield , c. van de bruck , d. f. mota and d. tocchini - valentini , phys .
lett . * 96 * , 061301 ( 2006 ) ; a. w. brookfield , c. van de bruck , d. f. mota and d. tocchini - valentini , phys
. rev . * d73 * , 083515 ( 2006 )
. o. e. bjaelde _
et al . _ , jcap * 0801 * , 026 ( 2008 ) .
a. e. bernardini and o. bertolami ; phys . rev . * d77 * , 083506 ( 2008 ) .
b. patt and f. wilczek , arxiv : hep - ph/0605188 ; s. dodelson , _ modern cosmology : anisotropies and inhomogeneities in the universe _ , ( academic press , new york , 2003 ) .
b. zeldovich and i. d. novikov , _ relativistic astrophysics - vol.i - stars and relativity _ ( university of chicago press , chicago , 1974 ) . m. gell - mann , p. ramond and r. slansky , in _ complex spinors and unified theories _ in supergravity , ( north holland , amsterdan , 1979 ) ; t. yanagida , _ proceedings of the workshop on unified theory and the baryon number of the universe _ , ( kek , japan , 1979 ) ; r. n. mohapatra , _ unification and supersymmetry _ ( springer - verlag , berlin , 1986 )
. n. arkani - hamed , s. dimopoulos , g. r. dvali , and j. march - russell , phys . rev . *
d65 * 024032 ( 2002 ) .
n. afshordi , m. zaldarriaga and k. kohri , phys .
* d72 * , 065024 ( 2005 ) . n. brouzakis and n. tetradis , jcap * 0601 * , 004 ( 2006 ) .
a. e. bernardini and o. bertolami , phys . lett . *
b684 * , 96 ( 2010 ) .
b. patt and f. wilczek , arxiv : hep - ph/0605188 ; o. bertolami and r. rosenfeld , int .
a23 * , 4817 ( 2008 ) .
m. c. bento , a. e. bernardini and o. bertolami , j. phys .
: conf . ser . *
174 * , 012060 ( 2009 ) .
a. e. bernardini , phys . lett . *
b684 * , 162 ( 2010 ) ; a. e. bernardini and o. bertolami , phys .
* d81 * , 123013 ( 2010 ) . | neutrinos coupled to an underlying scalar field in the scenario for unification of mass varying dark matter and cosmon-_like _ dark energy is examined . in the presence of a tiny component of mass varying neutrinos ,
the conditions for the present cosmic acceleration and for the stability issue are reproduced .
it is assumed that _
sterile _ neutrinos behave like mass varying dark matter coupled to mass varying _ active _ neutrinos through the _ seesaw _ mechanism , in a kind of _ mixed _ dark matter sector .
the crucial point is that the dark matter mass may also exhibit a dynamical behavior driven by the scalar field .
the scalar field mediates the nontrivial coupling between the mixed dark matter and the dark energy responsible for the accelerated expansion of the universe .
the equation of state of perturbations reproduce the generalized chaplygin gas ( gcg ) cosmology so that all the effective results from the gcg paradigm are maintained , being perturbatively modified by neutrinos . |
optical lattices have become a powerful tool to enhance the effects of interaction in ultracold atomic systems to create strong correlations and probe many - body physics beyond the mean - field theory @xcite . simply through varying the depth of the lattice potential ,
one changes the tunneling rate as well as the on - site interaction energy by changing the confinement of the atoms .
the strength of the atomic interaction can be directly tuned with a magnetic feshbach resonance @xcite . in comparison to @xmath7rb , which has been used in almost all experiments on optical lattices , @xmath0na has stronger and wider feshbach resonances that are experimentally accessible @xcite .
one such resonance has been used to produce quantum degenerate na@xmath5 molecules @xcite .
therefore , a sodium condensate loaded into an optical lattice would be a rich and flexible system for studying strong correlations .
so far , most optical lattice experiments have been performed with relatively heavy atomic species ( e.g. rubidium and potassium ) for which the recoil frequencies are lower and lasers are readily available to achieve trap depths of several tens of recoil frequencies at a few tens of milliwatts . for @xmath0na , high power single - mode lasers are necessary for similar experiments . in this work
, we chose to use a dye laser red - detuned by @xmath1 nanometers from the d lines of sodium ( 589 nm ) .
the spontaneous scattering rate limited the time window of the experiment to less than 50 ms , but was still sufficient to satisfy the adiabaticity condition to explore the quantum phase transition from a superfluid to a mott insulator .
we also observed strong atom losses at various lattice laser detunings , which were interpreted as photoassociation transitions .
the particular molecular states responsible for these transitions were identified through theoretical calculations and previous experimental data .
a @xmath0na bose - einstein condensate containing up to @xmath8 atoms in the @xmath9 state was first produced in a magnetic trap and subsequently loaded into a crossed optical dipole trap .
the optical trap was derived from a single - mode 1064 nm infrared laser , with the horizontal and vertical beams detuned by 60 mhz through acousto - optic modulators .
the number of condensed atoms was varied through three - body decay in a tight trap ( @xmath10 hz ) , after which the trap was decompressed ( @xmath11 hz ) to allow further evaporation and re - thermalization . a vertical magnetic field gradient was applied to compensate for gravity and avoid sagging in the weaker trap .
a dye laser operated at 594.710 nm was used to set up a three dimensional optical lattice .
the three beams were focused to @xmath12-waist of @xmath13 @xmath14 m at the condensate , and retro - reflected to form standing waves .
the two horizontal beams were orthogonal to each other , while the third beam was slanted at @xmath15 with respect to the vertical axis due to limited optical access .
the three beams were frequency - shifted by @xmath16 mhz and 80 mhz to eliminate cross interference between different beams .
the gaussian profile of the lattice beams added an additional harmonic trapping potential , while the localization of atoms at the lattice sites increased the repulsive mean field interaction . at the maximum lattice depth , the trap frequencies due to the combined potentials of the optical dipole trap and the lattice beams were @xmath17 hz for all three dimensions .
the trap parameters were chosen such that during the ramping of the optical lattice potential , the overall size of the cloud ( parametrized by thomas - fermi radii ) remained approximately constant in order to minimize intra - band excitations ( the mean thomas - fermi radius is @xmath18 m for @xmath8 atoms ) .
the peak per - lattice - site occupancy numbers achieved in our experiment were between 3 to 5 .
atoms held in a shallow optical lattice can tunnel freely from site to site and form a superfluid phase . as the lattice is made deeper , the atomic interaction is increased while the tunneling rate between lattice sites is exponentially suppressed .
the system then undergoes a phase transition to an insulating phase the mott - insulator in which each lattice site contains a definite fixed number of atoms . according to the mean - field theory for the homogenous systems of atoms in the lowest band of an optical lattice , the critical point for the phase transition from a superfluid to a mott - insulator state with @xmath19 atoms per lattice site
is determined by @xcite : @xmath20 where : @xmath21 is the on - site interaction energy ; @xmath22 is the tunneling rate between adjacent lattice sites ; @xmath23 is the number of nearest neighbors in the lattice ( 6 for a cubic lattice ) ; @xmath24 is the atomic mass ; @xmath25 is the @xmath26-wave scattering length ( 2.75 nm for @xmath0na ) ) ; @xmath27 is the wannier function ; @xmath28 is the lattice wavelength ; @xmath29 is the lattice potential .
figure [ uandj ] shows @xmath30 and @xmath31 for a cubic lattice as a function of the lattice depth , obtained through a band - structure calculation .
all energies are expressed in units of the recoil energy @xmath32 , where @xmath33 is the lattice wavenumber . with this scaling @xmath34
is independent of @xmath28 .
the peak occupancy number in our experiment was @xmath35 . from fig .
[ uandj ] , we find that the the critical points are at a lattice depth of 14.2 , 16.2 , 17.6 , 18.7 , and 19.5 ( all in units of @xmath36 ) for @xmath37 and 5 respectively .
the inset of fig .
[ uandj ] shows that the ratio of @xmath38 increases rapidly with increasing lattice depth . when a weak harmonic trap is present in addition to the lattice potential , as is the case for the experiment , the atomic density is not uniform
. nevertheless , eqs .
( [ eq1 ] [ eq3 ] ) can be used to estimate the lattice depth needed to observe the mott - insulator phase transition at any point in the harmonic trap .
given the local density of the initial condensate , a local value of @xmath19 can be estimated and thus the local critical lattice depth can be read off from fig .
[ uandj ] .
since the critical depth increases with @xmath19 , one expects that as the lattice depth is increased , shells of different occupancies will undergo the transition to the mott - insulator phase starting from the edge of the density profile and moving in towards the center .
occurs when @xmath31 as given by equation ( [ eq1 ] ) equals @xmath30 .
the horizontal location of the crossing point where @xmath39 gives the critical lattice depth .
the inset shows the ratio of @xmath30 over @xmath34 as a function of the lattice depth.,width=302 ] in our experiment , the optical lattice was linearly ramped up to a maximum potential of @xmath40 in a variable time @xmath41 ( @xmath42 khz for our system ) .
the lattice depth was calibrated by probing the energy difference between the first and the third band at zero quasi - momentum with small amplitude modulation of the lattice beams ( see , e.g. , @xcite ) . after reaching the peak value ,
the lattice was ramped back down again in @xmath41 .
the ramp sequence was stopped at different times when both the trap and the lattice were suddenly switched off ( in @xmath43s ) .
absorption images were then taken after some time - of - flight as shown in figure [ supmott ] .
the disappearance of the interference pattern as the lattice depth was increased indicated the loss of phase coherence and a transition from the superfluid state to the mott insulator state @xcite .
the subsequent revival of the interference patterns as the lattice depth was reduced ensured that the system remained essentially in the ground state during the ramping process .
different @xmath41 s were used to check the adiabaticity condition .
the peak spontaneous light scattering rate was about 21 s@xmath44 at the maximum intensity .
therefore for @xmath45 ms , less than 20% of the atoms spontaneously scattered a photon . after the lattice was fully ramped down , most of the atoms ( @xmath46 ) remained in the condensed fraction while the rest were heated and distributed across the first brillouin zone .
based on the number of atoms that remained in the condensate after the lattice was fully ramped down , we conclude that @xmath47 ms satisfies the intra - band adiabaticity condition . in the following discussion
, all measurements were performed for @xmath48 ms , but only the data for @xmath49 ms are shown as representative of similar results unless otherwise noted . to characterize the lifetime of the mott - insulator phase , we held the lattice depth at the maximum level for various amounts of time before ramping the lattice down to @xmath50 ( below the mott - insulator transition point ) and taking the time - of - flight image . if the system remains in the ground state , the contrast of the interference pattern should be recovered , whereas additional heating populates the brillouin zone and reduces the interference contrast .
a cross - section of the density profile was taken along the horizontal direction showing the interference peaks on top of a broad background ( see figure [ lifetime ] ) .
the 5 interference peaks and the broad background were fit by 6 gaussians .
the ratio between the total integrated area of the peaks and the background was used as the contrast to quantify the heating of the system .
the contrast gives a more sensitive measure of the heating compared to simply counting the recovered condensate atoms .
as the atoms in the interference peaks quickly move apart , they are not as broadened by the mean - field expansion as a single condensate .
we performed the same measurement for two different peak occupancy numbers @xmath51 and 5 .
figure [ lifetime ] shows the decay of the contrast and the lifetime @xmath52 was determined using an exponential fit .
the fitting error on the lifetime was less than 17% .
the lifetime was about 50% longer for @xmath53 , implying that inelastic collision processes significantly contributed to the heating of the system .
the three - body decay rate at the maximum lattice potential for the peak on - site atomic density ( @xmath54 @xmath55 for @xmath56 ) is about 100 s@xmath44 @xcite , consistent with the observed lifetimes of @xmath57 ms .
the peak density of a condensate in a harmonic trap and therefore the peak occupancy number scales with @xmath58 power of the total number of atoms , and our method for varying the number of atoms ( through three body decay ) was unable to consistently produce low enough atom numbers for peak occupancy @xmath59 .
the signal - to - noise ratio of our current imaging system also became marginal for such low atom numbers .
in this experiment , in addition to losses due to three body recombination , we observed large losses of atoms for certain specific tunings of the lattice laser in the range 592 nm to 595 nm .
a sample of such a loss feature is shown in fig .
[ photoassoc ] . for this measurement ,
the same ramp sequence was used as before with @xmath60 ms .
the peak intensity is about 280 w/@xmath61 in each lattice beam . due to the intentional frequency shifts between the three lattice beams the effective bandwidth of the lattice light as seen by the atoms
is @xmath62 mhz . for the narrow frequency scan range of fig .
[ photoassoc ] , the relative frequency scale was determined to better than 25 mhz using a fabry - perot cavity with a 2 ghz free spectral range .
axis corresponds to a laser wavelength of 594.490 nm as determined with a wavemeter with accuracy @xmath63 ghz .
the error bars are one standard deviation uncertainty .
the line is only a guide and does not have a theoretical underpinning.,width=302 ] single - photon photoassociations proved to have caused these losses .
the lattice laser is tuned by @xmath64 cm@xmath44 to the red of the atomic @xmath2s@xmath3p@xmath66 transition and thus is in a spectral region where it might drive photoassociation transitions @xcite to rovibrational levels in molecular states dissociating to either the @xmath2s@xmath67p@xmath4 or @xmath2s@xmath67p@xmath66 limits .
such a photoassociation transition , followed by the spontaneous radiative decay of the excited molecule into either a bound ground electronic - state molecule or into `` hot '' atoms , results in significant losses of atoms from the lattice .
it is therefore important to identify the locations and strengths of these resonances and choose the appropriate lattice wavelength to avoid such losses .
there is an extensive body of knowledge on the photoassociation of ultracold alkali - metal atoms and the behavior of the molecular potentials dissociating to the @xmath2s@xmath67p@xmath4 or @xmath2s@xmath67p@xmath66 limits @xcite .
figure [ pots ] shows the relevant excited molecular potentials as a function of internuclear separation @xmath68 .
the ground electronic states of na@xmath5 are the x@xmath69 and a@xmath70 states and two colliding ground state atoms will be some mixture of these symmetries . to the extent that the excited states are well described as @xmath71 or @xmath72 states , the @xmath73 and @xmath74 selection rules
imply that photoassociation transitions are allowed only to the two @xmath71 states ( a@xmath75 and ( 1)@xmath76 ) .
@xmath2s@xmath77p separated atom limit ( 1 @xmath78 = 0.0529 nm ) : the zero of the vertical scale is the @xmath2s@xmath77p@xmath66 limit , the @xmath2s@xmath77p@xmath4 limit lies 17.2 cm@xmath44 lower .
the hatched region indicates the energy range explored in the experiment .
the potentials are labeled by both hund s case ( a ) state labels of the form @xmath79 and by hund s case ( c ) labels @xmath80 given in parenthesis after the case ( a ) label . for small @xmath68 the curves labeled by a given case ( a ) label become degenerate , while for larger @xmath68 , as the effect of spin - orbit coupling becomes increasingly important ,
the degeneracy is lifted .
, width=302 ] previous experiments have identified the locations of the strong transitions to the a@xmath75 state @xcite and the weak transitions to the @xmath81 state @xcite .
these are shown in fig .
[ compare ] .
we looked for but failed to find any significant losses attributable to the weak @xmath81 state resonances .
we were able to confirm one of the a - state resonances , indicated by the dot in fig .
[ compare ] .
since the primary goal of the present work was to avoid photoassociation losses we did not investigate known a - state locations further .
our search focused on the strong resonances due to the ( 1)@xmath76 state which had not been previously observed in this spectral region . in 0.5 ghz steps , we scanned through a 30 ghz range around the theoretically predicted locations based on the model and auxiliary experimental data discussed below . in all but one such scans
, we were able to observe between 1 to 3 dips in the remaining atom numbers within a range of @xmath82 ghz , including the loss feature shown in fig .
[ photoassoc ] . as shown in fig .
[ compare ] the agreement of the observed locations with the predictions and auxiliary measurements confirms that these loss features are due to photoassociation to the ( 1)@xmath76 state . the locations of the rovibrational levels of the b@xmath83 state are not known in the current tuning range .
their spacing , however , should be equal to that of the @xmath84 levels and our observations are not consistent with such spacings .
the potential curves used in the calculation of the ( 1)@xmath76 vibrational levels were generated from an extended version of the model developed by movre and pichler for calculating the combined effects of the @xmath85 resonant dipole interaction and the atomic spin - orbit interaction .
such models have been extensively used to interpret photoassociation experiments @xcite .
to the long range potentials generated by the movre - pichler type calculation we append the results of _ ab initio _
calculations on the short range molecular potentials ( in the chemical bonding region ) .
these short range potentials are not sufficiently accurate to allow predictions of absolute vibrational positions .
it is necessary to make slight adjustments to these potentials to match experimentally measured vibrational positions , which were obtained through a separate photoassociation spectroscopy experiment in a dark - spot magnetooptical trap ( mot ) containing na @xmath2s@xmath86 atoms at about 300 @xmath14k , using a two color ionization scheme and an ion detector @xcite .
the photoassociation spectra taken in the mot have higher resolution than the loss features in the lattice experiment . in the spectral region of interest for the lattice experiment , three vibrational levels of the ( 1)@xmath76 were identified by measurements in the mot .
the spectra show a @xmath87 component with a complicated hyperfine / rotational pattern , and , slightly higher ( @xmath88 cm@xmath44 ) in energy , a @xmath89 component with a simpler nearly rotational pattern .
this ordering of the @xmath87 and @xmath90 components is in agreement with the movre - pichler model . for these levels ,
the @xmath91 feature of the @xmath89 component were found to be at 16845.155 cm@xmath44 , 16852.585 cm@xmath44 , and 168908.201 cm@xmath44 , calibrated to iodine lines with an estimated uncertainty of about 0.004 cm@xmath44 and are shown in fig .
[ compare ] .
this new data on the ( 1)@xmath76 was sufficient to calibrate our extended movre - pichler model without any further adjustable parameters .
the predicted positions shown in fig .
[ compare ] agree well with those of the loss features observed in the lattice experiment , thus identifying the loss features as photoassociation transitions to levels in the ( 1)@xmath76 state . based on this insight
, we chose a wavelength of 594.710 nm for our lattice experiment ( corresponding to -158.5 cm@xmath44 in fig .
[ compare ] ) .
this tuning lies @xmath92 ghz from the closest molecular resonance . given the observed on - resonance photoassociation rate of 1 ms@xmath44 in the lattice ,
photoassociative decay can be ignored at such detunings as the rate scales as the square of the ratio of the natural linewidth to the detuning .
in this paper , we explored the possibility of using a dye laser detuned @xmath1 nm from the na d lines to study many - body physics of a sodium bec in an optical lattice , which could allow for an independent control of interaction using magnetic feshbach resonances .
the superfluid to mott - insulator transition was observed in a na bose - einstein condensate for the first time .
the main technical difficulties are due to the heating from the spontaneous light scattering and three - body decay processes .
in addition , several photoassociation resonances were observed and identified by means of auxiliary spectroscopy measurements combined with theoretical modeling .
these resonances were avoided by choosing an appropriate lattice wavelength . in future experiments , we plan to use a high - power single - mode infrared laser at 1064 nm to eliminate atomic light scattering ( figure [ uandj ] also shows the transition points for a 1064 nm lattice ) .
moreover , heating from three body recombination can be avoided by using occupancy numbers less than three .
the authors would like to thank fredrik fatemi and paul lett for their help in obtaining the mot photoassociation data used here , and widagdo setiawan for experimental assistance .
we also thank carl williams for initiating the communications regarding the photoassociation resonances .
this research is supported by nsf , onr , aro , and nasa .
e. tiemann , h. knckel , and h. richling , z. phys .
d * 37 * , 323 ( 1996 ) , ( given our understanding of the implied dissociation limit for the rkr curve , the numbers in @xcite , and the x@xmath69 ( @xmath93 ) energy in @xcite , we infer a dissociation limit which gives a binding energy @xmath940.07 cm@xmath44 too low for the @xmath95 level when compared to the binding energy obtained directly from the photoassociation measurement in @xcite .
even though @xcite reported only the one line position as an incidental bit of information , other consistency tests suggest to us that it is better to use this line to fix the energy scale .
this small difference is not important for the present work . ) .
l. ratliff _
et al . _ , j. chem
* 101 * , 2638 ( 1994 ) , ( while transitions to the two @xmath72 states are nominally forbidden , the residual effects of the fine structure interaction provide sufficient mixing that transitions are possible even at these detunings . in particular , transitions to the @xmath81 state played a prominent role in early photoassociation experiments , due largely to the fact that a single color laser can not only produce molecules in this state but also readily ionizes them , making transitions to this state particularly easy to detect . ) . | the phase transition from a superfluid to a mott insulator has been observed in a @xmath0na bose - einstein condensate .
a dye laser detuned @xmath1 nm red of the na @xmath2s@xmath3p@xmath4 transition was used to form the three dimensional optical lattice .
the heating effects of the small detuning as well as the three - body decay processes constrained the timescale of the experiment .
certain lattice detunings were found to induce a large loss of atoms .
these loss features were shown to be due to photoassociation of atoms to vibrational levels in the na@xmath5 @xmath6 state . |
the fact that two parts of a polymer chain can not occupy the same place at the same time due to their finite volume has been recognised in the polymer literature for many years now as being an extremely important microscopic phenomenon that governs the macroscopic behavior of polymer solutions @xcite . like hydrodynamic interaction ,
the _ excluded volume _
effect influences the properties of polymer solutions even in the limit of extremely long chains because it is responsible for segments remote from each other along the polymer chain interacting with each other .
while the effect of excluded volume on static properties of polymer solutions has been widely studied , there have been very few attempts at examining its influence on properties far from equilibrium .
excluded volume effects can be incorporated into bead - spring chain models for polymer solutions in a relatively straightforward manner by adding the excluded volume interaction force between a particular bead and all the other beads in the chain ( pairwise ) to the other potential forces that are acting on the bead .
an noteworthy aspect of this approach is the kind of repulsive potential that is adopted to represent the excluded volume interactions . in static theories of polymer solutions ,
the excluded volume interaction is typically assumed to be a very short range @xmath2-function potential .
fixman @xcite and more recently ahn et al .
@xcite have attempted to predict the rheological properties of dilute polymer solutions by approximately incorporating the effects of both hydrodynamic interaction and excluded volume in a self - consistent manner into a bead - spring chain model .
( ahn et al
. also include finitely extensible springs in place of hookean springs ) . in order to obtain a solvable model ,
fixman @xcite used a repulsive quadratic excluded volume potential in place of a @xmath2-function potential .
this leads to a tractable model since the bead - connector vectors are then described by a gaussian non - equilibrium distribution function .
results obtained with the quadratic excluded volume potential have , however , not been compared so far with the results of other models for the excluded volume potential .
andrews et al .
@xcite have recently carried out a numerical study of the influence of excluded volume interactions on rheological and rheooptical properties of dilute solutions , with the help of brownian dynamics and configuration biased monte carlo simulations .
a bead - spring chain model , with `` fraenkel '' springs between beads and a morse potential to represent excluded volume interactions , was used to to model the flexible polymer molecule .
attention was largely confined to the prediction of properties in elongational flow and transient shear flow .
the predictions of their theories in the limit of long chains have not been considered by fixman @xcite , ahn et al .
@xcite and andrews et al . @xcite . on the other hand ,
the universal character of excluded volume effects have been studied using renormalisation group theory methods based on kinetic theory models ( with a @xmath2-function excluded volume potential ) by ttinger and coworkers @xcite . while the work of andrews et al .
@xcite is based on brownian dynamics simulations , the accuracy of the other _ approximate _ treatments of excluded volume cited above has not been tested by comparison with brownian dynamics simulations ( which are an ideal tool for testing approximations for nonlinear effects ) .
this is in contrast to the situation that exists for kinetic theory models that only incorporate hydrodynamic interaction effects , where extensive comparisons between the exact results of brownian dynamics simulations and various approximations have been made @xcite .
it is the purpose of this paper to examine the influence of the excluded volume effect on the rheological properties of a dilute polymer solution by using a _ narrow gaussian potential _ to describe the excluded volume interactions .
since the narrow gaussian potential tends to the @xmath2-function potential in the limit of a parameter @xmath0 ( that describes the width of the potential ) going to zero , it provides a means of evaluating results obtained with a singular @xmath2-function potential .
compared to the @xmath2-function potential , analytical calculations are not significantly harder with the narrow gaussian potential ; quite often , upon setting @xmath3 at the end of a calculation , the predictions of a @xmath2-function potential can be obtained .
furthermore , since brownian dynamics simulations can not be performed with a @xmath2-function potential , simulations carried out with the narrow gaussian potential for small values of the parameter @xmath0 provide a means of asymptotically obtaining the predictions of a @xmath2-function potential model .
any molecular theory that seeks to describe the dynamics of polymers in good solvents must simultaneously incorporate both the microscopic phenomena of hydrodynamic interaction and excluded volume , since hydrodynamic interaction effects have been shown to have an unavoidable influence on the dynamic behavior of polymer solutions
. however , it would be difficult to explore the consequences of such a theory for two reasons .
firstly , the incorporation of hydrodynamic interaction would lead to the complication of multiplicative noise . secondly , since brownian dynamics simulations for long chains would be extremely computationally intensive , any approximations that are developed can only be tested for very short chains . for these reasons , and being in the nature of a preliminary investigation , we examine excluded volume effects independently from hydrodynamic interaction effects , and confine attention here to a hookean dumbbell model for the polymer .
this enables the careful evaluation of various approximations .
it is hoped that , in the future , the best approximation can be used for chains with both hydrodynamic interaction and excluded volume , with the ultimate objective of calculating universal properties predicted in the long chain limit .
it must be noted that unlike in the case of the quadratic potential , the use of the narrow gaussian potential does not lead to exact solvability .
indeed , as has been observed in earlier treatments of other non - linear microscopic phenomenon such as hydrodynamic interaction and internal viscosity , it is found that it is not possible to obtain an analytical solution valid at all shear rates . as a result
, retarded motion expansions , brownian dynamics simulations , and perturbative and non - perturbative approximation procedures are used in this work to obtain the material functions predicted by the narrow gaussian potential .
an important consequence of using the narrow gaussian potential is that the nature of fixman s quadratic excluded volume potential can be explored .
predictions of fixman s quadratic excluded volume potential in simple shear flow will be compared with the predictions of the narrow gaussian potential .
the outline of this paper is as follows . in section 2 , the basic equations required to discuss the dynamics of hookean dumbbells in good solvents are derived . in section 3 , the various material functions considered in this paper are defined . a retarded motion expansion for the stress tensor , for arbitrary excluded volume potentials , is derived in section 4 , and power series expansions for the material functions in simple shear flow are obtained .
section 5 is devoted to examining the consequences of describing the excluded volume interaction with a narrow gaussian potential . in section 5.1 ,
expressions for the material functions predicted with this potential at zero shear rate are obtained by using the retarded motion expansion .
the formulation of a brownian dynamics simulation algorithm is discussed in section 5.2 , and a gaussian approximation for the configurational distribution function is introduced in section 5.3 . in section 5.4 ,
a first order perturbation expansion in the strength of the excluded volume interaction is derived .
fixman s theory for dumbbells is presented in section 6 , in terms of the framework adopted for the rest of the discussion in this paper .
the results of the various exact and approximate treatments are compared and discussed in section 7 , and the principal conclusions of the paper are summarised in section 8 .
the hookean dumbbell model represents a macromolecule by a mechanical model that consists of two identical beads connected by a spring .
the solvent in which the beads are suspended is assumed to be newtonian , and attention is restricted to flows which have a homogeneous velocity field , _ ie .
_ of the form @xmath4 , where @xmath5 is a constant vector , @xmath6 is a traceless tensor , and @xmath7 is the position vector with respect to a laboratory - fixed frame of reference . the instantaneous position of the beads
are specified by bead position vectors @xmath8 and @xmath9 .
this paper examines the consequence of introducing an excluded volume interaction between the beads of the dumbbell , so that the total potential experienced by the beads , @xmath10 , is the sum of the spring potential @xmath11 , and the excluded volume potential @xmath12 .
the force on bead @xmath13 due to this potential , @xmath14 , is then given by @xmath15 . for hookean springs ,
the spring potential is given by @xmath16 , where @xmath17 is the spring constant .
two forms of the excluded volume potential are considered in this work , namely , the narrow gaussian potential , and fixman s quadratic potential .
these are discussed in greater detail subsequently . here , we summarise the governing equations that are valid for an arbitrary choice of the excluded volume potential @xmath12 . for homogeneous flows ,
the configurational distribution function @xmath18 depends only on the internal configuration of the dumbbell , specified by the bead - connector vector @xmath19 , and not on the center of mass .
the quantity @xmath20 is then the probability that at time @xmath21 the dumbbell has a configuration in the range @xmath22 to @xmath23 . using the framework of polymer kinetic theory @xcite
one can show that the distribution function @xmath24 , in the presence of excluded volume , satisfies the following _ diffusion equation _
, @xmath25 where , @xmath26 is the bead friction coefficient ( _ ie . , _
@xmath27 for spherical beads with radius @xmath28 , in a solvent with viscosity @xmath29 ) , @xmath30 is boltzmann s constant , and @xmath31 is the absolute temperature .
the stress tensor , @xmath32 , in a polymer solution is considered to be given by the sum of two contributions , @xmath33 , where @xmath34 is the contribution from the solvent , and @xmath35 is the polymer contribution . since the solvent is assumed to be newtonian , @xmath36 , where @xmath37 is the rate of strain tensor , @xmath38 .
the rheological properties of a dilute polymer solution may thus be obtained by calculating the polymer contribution to the stress tensor , @xmath35 . for a dumbbell model in the presence of excluded volume
, it is given by the kramer s expression @xcite , @xmath39 here , @xmath40 is the number density of polymers , and angular brackets represent averaging with respect to the configurational distribution function @xmath41 . for both the excluded volume potentials considered here , it will turn out that calculation of the second moment @xmath42 is necessary in order to evaluate the average in equation ( [ kram ] ) . a time evolution equation for the second moment can be obtained by multiplying the diffusion equation ( [ diff ] ) by @xmath43 and integrating over all configurations , @xmath44 \label{secmom}\ ] ] it proves convenient for subsequent calculations to introduce the following dimensionless variables , @xmath45 where , @xmath46 is the familiar time constant , @xmath47 and @xmath48 are the non - dimensional hookean spring potential and the non - dimensional excluded volume potential , respectively .
note that @xmath49-solvent values of a typical time scale ( @xmath50 ) and a typical length scale ( @xmath51 ) are used for the purpose of non - dimensionalisation , regardless of the form of the excluded volume potential .
the intramolecular force is expected to be in the direction of the bead - connector vector .
therefore , it can be written in terms of non - dimensional variables as , @xmath52 where @xmath53 is an arbitrary function of the magnitude of @xmath54 . as a result , the diffusion equation ( [ diff ] ) , in terms of non - dimensional variables , is given by , @xmath55 while , the kramers expression ( [ kram ] ) assumes the form , @xmath56 and the second moment equation ( [ secmom ] ) becomes , @xmath57 equation ( [ nondimsecmom ] ) is in general not a closed equation for the second moments since it depends on the form of the function @xmath53 ( _ ie .
_ on the choice of excluded volume potential ) .
on examining equations ( [ nondimkram ] ) and ( [ nondimsecmom ] ) , it is straightforward to see that the second moment equation ( [ nondimsecmom ] ) is nothing but the giesekus expression for the stress tensor , @xmath58 while the equations derived in this section are valid for arbitrary homogeneous flows , in this paper we confine attention to the prediction of rheological properties in simple shear flows , defined in the section below .
steady simple shear flows are described by a tensor @xmath59 which has the following matrix representation in the laboratory - fixed coordinate system , @xmath60 where @xmath61 is the constant shear rate .
the three independent material functions used to characterize such flows are the viscosity , @xmath62 , and the first and second normal stress difference coefficients , @xmath63 , respectively .
these functions are defined by the following relations @xcite , @xmath64 a transient experiment that is used often to characterise polymer solutions is small amplitude oscillatory shear flow , where the tensor @xmath6 is given by , @xmath65 here , @xmath66 is the amplitude , and @xmath67 is the frequency of oscillations in the plane of flow .
the polymer contribution to the shear stress , @xmath68 , depends on time through the relation @xcite , @xmath69 where @xmath70 and @xmath71 are the material functions characterising oscillatory shear flow .
they are represented in a combined form as the complex viscosity , @xmath72 . in the linear viscoelastic flow regime ,
the stress tensor is described by the linear constitutive relation , @xmath73 where @xmath74 is the relaxation modulus . as a result , for oscillatory
shear flows with a small amplitude @xmath66 , expressions for the real and imaginary parts of the complex viscosity can be found in terms of the relaxation modulus from the expression , @xmath75 note that the zero shear rate viscosity @xmath76 and the zero shear rate first normal stress difference @xmath77 , which are linear viscoelastic properties , can be obtained from the complex viscosity in the limit of vanishing frequency , @xmath78
a retarded motion expansion for the stress tensor , derived previously for the fene dumbbell model @xcite , can be adapted to the present instance by recognising that the fene spring potential ( as indeed any choice of excluded volume potential ) is but a particular example of a connector force potential between the beads of the dumbbell . in this section ,
we briefly summarise the development of a retarded motion expansion for an arbitrary choice of the excluded volume potential .
details of the derivation may be found in bird et al .
@xcite .
we seek a solution of the diffusion equation ( [ nondimdiff ] ) whereby the configurational distribution function , @xmath41 , can be written as a product of an equilibrium contribution and a flow contribution , @xmath79 the governing equation for the flow contribution @xmath80 , @xmath81 can be obtained by substituting equation ( [ prod ] ) into the diffusion equation ( [ nondimdiff ] ) , and exploiting the fact that the equilibrium distribution function is given by , @xmath82 where @xmath83 is the normalisation constant .
regardless of the form of the excluded volume potential , at steady state , an exact solution to equation ( [ flowdiff ] ) can be found for all homogeneous _ potential _ flows @xcite . for more general homogeneous flows , however , it is necessary to seek a perturbative solution .
the flow contribution , @xmath80 , is assumed to be expandable in a power series in the velocity gradients , @xmath84 where @xmath85 is of order @xmath86 in the velocity gradient .
partial differential equations governing each of the @xmath85 may be obtained by substituting equation ( [ flowdistexp ] ) into equation ( [ flowdiff ] ) and equating terms of like order . following the procedure suggested in bird et al .
@xcite , one can judiciously guess the specific forms for the functions @xmath85 by noting that each of these functions must have certain properties .
the form of the function @xmath87 which satisfies these requirements is , @xmath88 while the form of @xmath89 can be guessed to be , @xmath90 where , @xmath91 denotes an average with @xmath92 , @xmath93 is the vorticity tensor , defined here as @xmath94 , and the scalar function @xmath95 obeys the following second order differential equation , @xmath96 it is difficult to suggest boundary conditions for equation ( [ aode ] ) other than to say that the solution must be such that @xmath97 is bounded . in the case of fene springs ,
it is possible to explicitly obtain the particular solution of a similar second order differential equation .
it is clear from equation ( [ giesekus ] ) that at steady state , the stress tensor can be found once @xmath98 is known .
the second moment @xmath98 can be found correct to second order in velocity gradients by using the power series expansion ( [ flowdistexp ] ) for @xmath80 , and the specific forms for @xmath87 and @xmath89 in equations ( [ phi1 ] ) and ( [ phi2 ] ) , respectively .
this leads to the following expression for the stress tensor , correct to third order in velocity gradients , @xmath99 the cayley - hamilton theorem has been used to eliminate the term @xmath100 in equation ( [ retmot ] ) , and an isotropic term that does not affect the rheological properties has been dropped .
the stress tensor in simple shear flow , for small values of the non - dimensional shear rate @xmath101 , can be found by substituting equation ( [ ssf1 ] ) for the rate of strain tensor , in equation ( [ retmot ] ) . using the definitions of the viscometric functions in equation ( [ sfvis ] ) ,
the following power series expansions are obtained , @xmath102 clearly , equation ( [ psi1 ] ) indicates that one must expand to higher orders in velocity gradients before the shear rate dependence of the first normal stress difference can be obtained .
zero shear rate properties , however , can be obtained from equations ( [ etap ] ) and ( [ psi1 ] ) .
in the static theory of polymer solutions it is common to represent the dimensionless excluded volume potential , between two points on the polymer chain separated by a non - dimensional distance @xmath54 , with the dirac delta function , @xmath103 where , @xmath104 is a non - dimensional parameter which represents the strength of the excluded volume interaction , and in which @xmath105which has the dimensions of volume is called the ` excluded volume parameter ' @xcite .
the parameter @xmath1 is frequently used in theories that incorporate excluded volume , as it is considered to be the appropriate parameter to be used in perturbation expansions . as mentioned earlier , excluded volume interactions are taken into account in this work by means of a narrow gaussian potential @xcite . the narrow gaussian potential has the following form in terms of non - dimensional variables , @xmath106 it is clear from equation ( [ nondimnagpot ] ) that the non - dimensional parameter @xmath0 controls the extent of the excluded volume interaction , and as @xmath107 , the narrow gaussian potential tends to the @xmath2potential . the narrow gaussian potential , as mentioned earlier , serves as a means of regularising the singular @xmath2potential and consequently , permits the evaluation of results obtained with a @xmath2potential . with excluded volume interactions described by the narrow gaussian potential , the function @xmath108 , which appears in equation ( [ evforce ] ) for the non - dimensional force between the beads of the dumbbell , is given by , @xmath109 \label{nondimevforce}\ ] ]
the complex form of this function implies that the diffusion equation ( [ nondimdiff ] ) can not be solved exactly analytically to obtain the non - equilibrium configurational distribution function @xmath110 .
furthermore , the time evolution equation for the second moments ( [ nondimsecmom ] ) is not a closed equation for the second moments since it involves the higher order moment @xmath111 on the right hand side . as a result ,
perturbative methods , non - perturbative approximation procedures or numerical schemes must be used in order to obtain the material functions predicted by the narrow gaussian potential .
the viscosity and the first normal stress difference predicted by the narrow gaussian potential at low shear rates can be obtained from the equations ( [ etap ] ) and ( [ psi1 ] ) , respectively , once the equilibrium averages that occur in these expressions are evaluated . for the narrow gaussian potential ,
the equilibrium distribution function is given by equation ( [ eqdist ] ) , with the excluded volume contribution to the non - dimensional potential @xmath112 given by equation ( [ nondimnagpot ] ) .
we denote the various non - dimensional moments of the equilibrium distribution for a narrow gaussian potential by , @xmath113 in order to obtain the viscosity at non - zero shear rates , it is necessary to find the function @xmath114 that satisfies the second order differential equation ( [ aode ] ) [ with @xmath115 . unlike in the case of an fene dumbbell
, it has not been possible to obtain the particular solution to equation ( [ aode ] ) . as a result
, attention is confined here to obtaining the zero shear rate predictions of the narrow gaussian potential , @xmath116 for which only the moments @xmath117 and @xmath118 are required .
alternative methods will be used in sections 5.2 to 5.4 to obtain the shear rate dependence of the viscometric functions .
the non - dimensional equilibrium moments @xmath117 and @xmath118 can be obtained exactly , as will be shown in section 5.1.1 below .
the need to calculate equilibrium moments is also frequently encountered in static theories of polymer solutions . in these theories , as mentioned earlier , it is common to represent excluded volume interactions with a delta - function excluded volume potential , and furthermore , to obtain universal predictions by considering the limit of long chains .
the hamiltonian then typically has two singular objects making it impossible to evaluate equilibrium moments exactly .
the most succesful approach so far towards approximately evaluating these moments has been to develop a perturbation expansion in the parameter @xmath1 , and to use renormalisation group methods to refine the results of the perturbation calculation @xcite .
an alternative and simpler non - perturbative route is the uniform expansion model @xcite .
the use of the narrow gaussian potential albeit in the simple context of hookean dumbbells provides an opportunity to compare the results of these approximate models with the exact solution .
the rigorous solution is discussed in section 5.1.1 below , while the perturbation expansion and the uniform expansion models are discussed in sections 5.1.2 and 5.1.3 , respectively .
it is straight forward to show that the moments @xmath119 are given by the ratio of two integrals , @xmath120 , where , @xmath121 in the limit of @xmath107 and @xmath122 , these integrals can be evaluated analytically .
consider the quantities , @xmath123 which are the integrands for the integrals @xmath124 , and @xmath125 now , @xmath126 , for all values of @xmath0 . at any non - zero value of @xmath127 , it is clear from equation ( [ nondimnagpot ] ) that , @xmath128 in other words , for all values of @xmath127 , the quantities @xmath129 tend _ pointwise _ to @xmath130 as @xmath0 tends to zero or to infinity .
furthermore , for all values of @xmath0 , it can be shown that @xmath129 are bounded functions of @xmath127 on @xmath131 .
it then follows from a theorem of the calculus @xcite that , @xmath132 as a result , the asymptotic values of @xmath119 for @xmath133 and @xmath122 are found to be _ independent _ of @xmath1 , and are equal to the @xmath49-solvent values , @xmath134 this implies from equations ( [ etap02 ] ) and ( [ psi102])that the use of a delta - function potential to represent excluded volume interactions leads to the prediction of zero shear rate properties in good solvents which are identical to those in @xmath49-solvents . away from these limiting values of @xmath0 , _
i.e_. at non - zero finite values of @xmath0 , the integrals @xmath124 can be found by numerical quadrature .
here they have been evaluated using a routine given in numerical recipes @xcite for the integration of an exponentially decaying integrand .
discussion of zero shear rate property predictions in this case is taken up in section 7 .
static theories for polymer solutions indicate that accounting for excluded volume interactions with a delta function potential leads to the prediction of a _ swelling _ ( _ i.e_. an increase in the mean square end - to - end distance ) of the polymer chain , which is in close agreement with experimental observations @xcite . in the case of hookean dumbbell , however , we have seen above that the use of a delta function potential to account for the presence of excluded volume ( which corresponds to the limit @xmath107 ) , does not lead to any change in the prediction of equilibrium moments when compared to the @xmath49-solvent case .
it is worthwhile therefore to examine the nature of the perturbation expansion in @xmath1 , and to compare it with the results of the exact calculation . upon expanding @xmath135 in a power series ,
the integral @xmath124 has the form , @xmath136 where , @xmath137 in order to carry out a term by term integration of the functional series @xmath138 in equation ( [ sjexp ] ) , it is necessary for the series to be _ uniformly convergent _ on @xmath131 . for all values of @xmath1 , and @xmath139 , uniform convergence can be established with the help of the weierstrass @xmath140 test @xcite .
therefore , a term by term integration in equation ( [ sjexp ] ) can be caried out , _ except _ when @xmath141 . assuming that @xmath139 , and performing the integration in equation ( [ sjexp ] ) , one obtains , @xmath142 consider the moment @xmath117 , which is undoubtedly the most interesting physical moment . using the perturbation expansion for @xmath124 , @xmath117
is given by the ratio , @xmath143 , where , @xmath144 and @xmath145 are defined by , @xmath146 if only the first order perturbation term is retained , one obtains , @xmath147 curiously , the limit @xmath107 can be carried out in this case .
it leads to a result which is in line with static theories , which indicate a finite non - vanishing effect due to the presence of @xmath2-potential excluded volume interactions .
however , such a limit can not be carried out if higher order terms are retained , since both the sums @xmath144 and @xmath145 diverge as @xmath107 . in static theories ,
higher order perturbation expansions are obtained by _ dropping _ divergent terms , as they are postulated not to matter .
these divergent terms arise from products of the @xmath2-potential for the _ same _ pair of interacting beads . in dumbbells , these are the only kind of beads present . as a result , there is no meaningful or mathematically consistent way of going beyond first order perturbation theory for dumbells in the limit @xmath107 . note that both the series , @xmath144 and @xmath145 , are alternating series . furthermore ,
for given values of @xmath1 and @xmath0 , the terms decrease monotonically for large enough values of @xmath40 .
it follows then , from the leibnitz criterion for alternating series @xcite , that both @xmath144 and @xmath145 converge for all values of @xmath1 and @xmath139 .
this suggests that , even though it is not possible to switch the integral and summation in equation ( [ sjexp ] ) for @xmath148 , the value of @xmath117 at @xmath148 can be found by setting it equal to the limit of @xmath149 as @xmath107 .
we shall see in section 7 that such a limiting process is infeasible since it becomes numerically impossible to evaluate the sums ( [ s0s1 ] ) for small enough values of @xmath0 .
the uniform expansion model seeks to approximate the average @xmath150 of any quantity @xmath151 , with @xmath152 , where @xmath153 denotes an average with the gaussian equilibrium distribution function , @xmath154 with @xmath155^{3 / 2}$ ] .
the aim is to find the parameter @xmath156 that leads to the best possible approximation .
as may be expected , this depends on the quantity @xmath151 that is averaged .
the motivation behind the uniform expansion model , and details regarding the calculation of @xmath156 are given in appendix a. since the equilibrium distribution function in the absence of excluded volume is gaussian , this assumption expects the equilibrium distribution to remain gaussian upon the incorporation of excluded volume , albeit with an increased end - to - end vector .
if we define @xmath157 , such that @xmath158 , then clearly @xmath157 is the uniform expansion model approximation for @xmath159 .
the uniform expansion model predictions of the zero shear rate properties are given by ( [ etap02 ] ) and ( [ psi102 ] ) , with @xmath159 replaced by @xmath157 .
results of material properties obtained by numerical quadrature , and by the uniform expansion model , are discussed in section 7 .
development of the retarded motion expansion has proved useful in obtaining exact expressions for the zero shear rate properties . at non - zero shear rates ,
exact predictions of the viscometric functions can be obtained by solving the ito stochastic differential equation , @xmath160 \cdot \bq^ * \ , dt + d \bw \label{ito}\ ] ] which corresponds to the non - dimensional diffusion equation ( [ nondimdiff ] ) , and in which @xmath161 is a three - dimensional wiener process . for the narrow gaussian potential , since @xmath162 , equation ( [ ito ] ) is non - linear . as a result , it can not be solved analytically .
two different brownian dynamics simulation algorithms have been adopted here for the numerical solution of equation ( [ ito ] ) .
both schemes use a second order predictor - corrector algorithm with time - step extrapolation @xcite .
the first scheme obtains steady - state expectations by the simulation of a single long trajectory , and is based on the assumption of ergodicity @xcite .
it has been used to obtain results at equilibrium , and for large values of the shear rate .
a second algorithm which employs a variance reduction procedure has been used at low values of the shear rate , since the variance for the viscometric functions is found to be relatively large at these shear rates .
reduction in the variance is obtained by following a scheme suggested by wagner and ttinger @xcite . the scheme which constructs an ensemble of trajectories over several relaxation times , from start - up of flow to steady - state essentially consists of subtracting the rheological properties obtained from a parallel equilibrium simulation .
while this does nt change the average values of the properties , it significantly reduces the fluctuations , since the fluctuations are virtually the same at zero and small shear rates .
the results of these simulation algorithms are discussed in section 7 .
the main obstacle ( in the configuration space of the dumbbell ) to obtaining the rheological properties predicted by a narrow gaussian potential is that the second moment equation is not a closed equation .
a closure problem has also been encountered earlier in treatments of the phenomenon of hydrodynamic interaction and internal viscosity , where it has been shown that an accurate approximation can be obtained by assuming that the non - equilibrium configurational distribution function is a gaussian distribution @xcite . in this section , a similar systematic approximation procedure for the treatment of excluded volume interactions described by a narrow gaussian potential
is introduced .
the assumption that @xmath163 is a gaussian distribution , @xmath164 makes the second moment equation ( [ nondimsecmom ] ) a closed equation , since the higher order moment @xmath165 can be expressed in terms of the second moment . on performing this reduction
, it can be shown that the gaussian approximation leads to the following closed second moment equation , @xmath166 where , @xmath167 it is also straight forward to show that on introducing the gaussian approximation , the giesekus expression for the stress tensor has the form , @xmath168 the steady state viscometric functions [ defined by equations ( [ sfvis ] ) ] can therefore be found once equation ( [ gasecmom ] ) is solved for @xmath169 .
it is worth examining the nature of the polymer contribution to the stress tensor . in the limit @xmath107
, @xmath170 reduces to @xmath171 , and as a result , the presence of excluded volume only makes an indirect contribution through its influence on the second moment @xmath169 .
this follows from the fact that an isotropic contribution to the stress tensor makes no difference to the rheological properties of the polymer solution .
on the other hand , for non - zero values of @xmath0 , the rheological properties are also affected directly by excluded volume . under the gausssian approximation ,
linear viscoelastic properties can be obtained by deriving a first order codeformational memory integral expansion for the stress tensor .
the tensor @xmath169 is expanded , in terms of deviations from its isotropic equilibrium solution , upto first order in velocity gradient , @xmath172 where , the parameter @xmath173 ( commonly called the swelling ratio ) is defined by , @xmath174 here , @xmath175 , is the mean square end - to - end distance in the absence of excluded volume .
clearly , @xmath173 represents the _ equilibrium _ swelling of the polymer chain caused by the presence of excluded volume .
@xmath173 is not an independent parameter since at equilibrium the second moment equation ( [ gasecmom ] ) reduces to the following consistency relationship between @xmath1 , @xmath0 and @xmath173 , @xmath176 the well known scaling relation for the end - to - end distance with the number of monomers @xmath177 , namely , @xmath178 , may be obtained from equations ( [ alpha ] ) and ( [ equirel ] ) in the limit of large @xmath177 by noting that since excluded volume is a pairwise interaction@xmath1 must scale as @xmath179 for dumbbells @xcite . substituting the expansion ( [ linvisexp ] ) into the evolution equation ( [ gasecmom ] ) leads to , @xmath180 where , @xmath181^{\ , -1 } \label{lvtau}\ ] ]
furthermore , the stress tensor ( [ gakram ] ) upto first order in the velocity gradient ( without the rheologically unimportant isotropic contribution ) is given by , @xmath182 where , @xmath183 upon integrating equation ( [ lvdif ] ) , which is a first order ordinary differential equation for @xmath184 , and substituting the result into equation ( [ lvkram ] ) , the following codeformational integral expansion is obtained , @xmath185}(t , s ) \label{lvmemexp}\ ] ] where @xmath186}$ ] is the codeformational rate - of - strain tensor @xcite and the memory function @xmath187 is given by @xmath188 the product @xmath189 is usually interpreted as a relaxation time , and @xmath190 as a relaxation weight .
clearly , the incorporation of excluded volume effects increases the relaxation time in a good solvent relative to a theta solvent by a factor of @xmath191 . the memory function @xmath187 can now be used to derive the linear viscoelastic material properties .
substituting @xmath192 , into equation ( [ usf6 ] ) , leads to , @xmath193 upon taking the limit of @xmath194 we obtain , @xmath195 at moderate to large values of the shear rate , it is not possible to obtain analytical expressions for the shear rate dependence of the viscometric functions , and consequently a numerical procedure is required . since the second moment @xmath169 shares the symmetry of the flow field in simple shear flow , its cartesian components can be denoted by ,
@xmath196 upon substituting equation ( [ momscompsf ] ) into equation ( [ gasecmom ] ) , a system of four first order ordinary differential equations for the quantities @xmath197 is obtained .
steady state viscometric functions , as functions of shear rate , can then be found by numerically integrating these equations with respect to time ( using a simple euler scheme ) until steady state is reached .
results obtained by this procedure are discussed in section 7 .
the influence of excluded volume effects on the universal shear rate dependence of viscometric functions has been studied , as mentioned earlier , by using renormalisation group methods @xcite .
the renormalisation group theory approach is essentially a method for refining the results of a low order perturbation expansion in @xmath1 , by introducing higher order interactions effects so as to remove the ambiguous definition of the bead size .
results of this approach , based on a @xmath2-function excluded volume potential , indicate that the presence of excluded volume has a non - trivial influence on the shear rate dependence of the viscometric functions .
we have seen earlier in this work that as far as equlibrium swelling and zero shear rate properties are concerned , the use of a @xmath2-function excluded volume potential leads to trivial results . in this section ,
a first order perturbation expansion in @xmath1with a narrow gaussian excluded volume potential is constructed , in order to compare its predictions of shear rate dependence with those obtained with brownian dynamics simulations , and with the gaussian approximation .
the dependence of the predictions on the width parameter @xmath0 is of particular interest .
a first order perturbation expansion can be constructed , following the procedure suggested in references @xcite , from the second moment equation ( [ nondimsecmom ] ) .
we assume that the configurational distribution function @xmath198 can be written as @xmath199 , where @xmath200 is the distribution function in the absence of excluded volume , i.e. in a @xmath49-solvent , and @xmath201 is the correction to first order in the strength of the excluded volume interaction .
the averages performed with these contributions will be denoted by @xmath202 and @xmath203 , respectively . on equating terms of equal order , equation ( [ nondimsecmom ] )
can be rewritten as two equations , namely , a zeroth order second moment equation and a first order second moment equation .
the zeroth order equation which is linear in the moment @xmath204is the well known second moment equation for hookean dumbbells in a @xmath49-solvent @xcite .
it has the analytical solution @xcite , @xmath205}(t^ * , s^ * ) \label{qqtheta}\ ] ] where , @xmath206}$ ] is the codeformational relative strain tensor @xcite . the first order second moment equation has the form , @xmath207 the @xmath49-solvent distribution function @xmath200 is a gaussian @xcite , and consequently , the complex moment on the right hand side of equation ( [ secmomz ] )
can be reduced to a function of @xmath208 .
the following equation is obtained on performing this reduction , @xmath209 where , @xmath210 clearly , equation ( [ redsecmomz ] ) could have also been derived by expanding the second moment equation for the gaussian approximation ( [ gasecmom ] ) to first order in @xmath1 .
it follows , therefore , that the gaussian approximation is exact to first order in @xmath1 .
this is also the situation in the case of the gaussian approximation introduced for the treatment of hydrodynamic interaction effects , where it was found to be exact to first order in the strength of hydrodynamic interaction , @xmath211 @xcite .
equation ( [ redsecmomz ] ) is a system of linear inhomogeneous ordinary differential equations , whose solution is , @xmath212 where , @xmath213 is the displacement gradient tensor @xcite . the expression for the stress tensor ( [ nondimkram ] ) can also be expanded to first order in @xmath1 .
after reduction of complex moments to second moments , the stress tensor depends only on the second moments @xmath214 and @xmath215 .
equations ( [ qqtheta ] ) and ( [ qqz ] ) may then be used to derive the following first order perturbation theory expression for the stress tensor in arbitrary homogeneous flows , @xmath216}(t^ * , s^ * ) - \be(t^ * , s^ * ) \cdot \by \cdot \be{^t}(t^ * , s^ * ) \right ) \label{pertau}\ ] ] note that @xmath217 , the direct contribution to the stress tensor , is isotropic only in the limit @xmath107 .
the form in steady shear flow , of the tensors @xmath206}$ ] and @xmath213 , has been tabulated in reference @xcite . using the expression for the stress tensor
( [ pertau ] ) , and the definition of the viscometric functions ( [ sfvis ] ) , the following first order perturbation theory results for the viscometric functions are obtained , @xmath218 where , @xmath219 - \lambda_h^2 { \dot \gamma}^2 $ ] . in the limit of @xmath220
going to zero , the expression for the viscosity ( [ etaper ] ) reduces to the expression derived earlier in section 5.1using the retarded motion expansion and the equilibrium perturbation expansion for the zero shear rate viscosity .
one can also show that @xmath221 reduces to the equilibrium moment @xmath117 [ see equation ( [ equimom ] ) ] , in the limit @xmath222 .
the first order perturbation results are compared with simulation results , and with results of the guassian approximation , in section 7 .
many years ago , in path - breaking seminal work , fixman @xcite considered the simultaneous inclusion of hydrodynamic interaction and excluded volume in bead - spring chain models for dilute polymer solutions . in order to render the problem solvable ,
fixman introduced a number of approximations .
since we are only concerned with excluded volume in the context of hookean dumbbells in this work , we shall only consider those approximations which are relevant in this context .
the introduction of the quadratic potential , the governing equations of fixman s theory , and the calculation of material functions predicted by the theory in simple shear flow are considered in this section . with regard to excluded volume , the most crucial approximation of fixman @xcite is the replacement of the delta function potential with a quadratic potential . by adopting a boson operator formulation of the governing equations
, fixman has shown that the delta function potential ( [ delpot ] ) may be represented by , @xmath223 where @xmath224 is a symmetric function of various configuration dependent quantities introduced in the boson operator formalism . from this expression
it is clear that a quadratic potential for the excluded volume may be obtained by replacing the fluctuating quantity @xmath225 with an average .
fixman obtains a quadratic potential by replacing @xmath225 with a configuration dependent average , as described below . as a result of replacing @xmath225 with its average
, one can show from equation ( [ bosone ] ) that , @xmath226 in other words , for any given potential @xmath227 , one can find @xmath228 provided that the non - equilibrium distribution function @xmath229 with which to carry out the average on the right hand side of equation ( [ avgg ] ) is known .
it turns out that @xmath229 , in the presence of a quadratic excluded volume potential and with consistently averaged hydrodynamic interaction @xcite , is a gaussian distribution .
evaluation of the average for a @xmath2potential ( [ delpot ] ) , with a gaussian distribution ( [ gausdist ] ) , leads to the following non - dimensional quadratic potential , @xmath230 it is straightforward to show that the potential ( [ modquadpot ] ) leads to an unphysical _ non - central _ excluded volume force between the beads .
fixman , perhaps for this reason , introduces a further approximation which consists of replacing the above potential with the following simpler form , @xmath231 where , @xmath173 is defined as before by equation ( [ alpha ] ) . however , in this case , @xmath173 obeys the consistency relation , @xmath232 interestingly enough , equation ( [ equirel ] ) reduces to equation ( [ fixalp ] ) in the limit @xmath107 ( which corresponds to a @xmath2potential ) .
it is appropriate here to note that while fixman has presented all his arguments for bead - spring chain models , the form of fixman s potential for dumbbells given above can be found in the book by larson @xcite .
the consequences of adopting the quadratic excluded volume potential ( [ quadpot ] ) are briefly discussed in the following section .
it is worthwhile to point out here that fixman s original formulation of the problem was not in terms of the gaussian distributions and second moments discussed below .
rather , his attempt was to directly solve the diffusion equation ( [ diff ] ) for the configurational distribution function , and then carry out the average in equation ( [ kram ] ) to obtain the rheological properties .
it is , however , possible to discuss his approach within the framework developed subsequently by " ottinger @xcite and this is the procedure that is adopted here . a graphic exposition of fixman s algorithm is given in reference @xcite . with excluded volume interactions described by the quadratic potential ( [ quadpot ] )
, the diffusion equation ( [ diff ] ) becomes linear in the bead - connector vector . as a result , the diffusion equation is exactly satisfied by a gaussian distribution ( [ gausdist ] ) . in fixman
s theory therefore , a tractable model is obtained not by approximating the distribution function , as in the case of the gaussian approximation , but by introducing the quadratic potential ( [ quadpot ] ) . while @xmath229 is a gaussian both in the gaussian approximation and in fixman s theory , the second moment which completely determines these distributions is different in the two cases . in fixman s theory , the non - dimensional second moment @xmath169 is governed by the equation , @xmath233 \ , \avel \bq^ * \bq^ * \aver + \bu \label{fixsecmom}\ ] ] note that the second moment equation reduces to equation ( [ fixalp ] ) at equilibrium .
the giesekus expression for the stress tensor has the form , @xmath234 \
, \moms + \bu \label{fixkram}\ ] ] as a result , the stress tensor in fixman s theory , for any flow situation , may be obtained once equation ( [ fixsecmom ] ) is solved for @xmath169 .
it is clear from equation ( [ fixkram ] ) that in fixman s theory , as for @xmath235 in the gaussian approximation , rheological properties are affected both directly and indirectly by the presence of excluded volume .
as will be shown in the section below , at _ steady state _ in simple shear flow , the problem of solving the governing equation for the second moments ( [ fixsecmom ] ) , reduces to one of solving a single nonlinear algebraic equation . in the linear viscoelastic limit , however , analytical expressions for the various properties can be derived in the same manner as described earlier for the gaussian approximation .
indeed , it can be shown that the linear viscoelastic properties are given by equations ( [ etaprime ] ) and ( [ lvetap0 ] ) , where the quantities @xmath190 and @xmath236 are now given by , @xmath237 in steady simple shear flow , substituting equation ( [ momscompsf ] ) for @xmath169 and equation ( [ ssf1 ] ) for @xmath59 , into the second moment equation ( [ fixsecmom ] ) , leads to the following equations for the components of @xmath169 , @xmath238 where , @xmath239 must satisfy the nonlinear algebraic equation , @xmath240 the normalised viscometric functions can be found by using the definitions ( [ sfvis ] ) , and the results above for the zero shear rate properties , @xmath241 the nonlinear algebraic equation ( [ sfs2 ] ) is solved here with a newton - raphson scheme .
the material functions predicted by fixman s theory are compared with the predictions of the narrow gaussian potential in the section below .
the predictions of rheological properties in simple shear flow , by the various theories for the excluded volume effect , are compared in this section .
predictions in the limit of zero shear rate are first considered below , and those at finite non - zero shear rates subsequently .
figure [ fig1 ] is a plot of @xmath242 versus @xmath0 for @xmath243 and @xmath244 .
the continuous curves are exact predictions obtained by numerical quadrature , while the triangles and circles are exact results of brownian dynamics simulations carried out at equilibrium ( without variance reduction ) [ see equation ( [ etap02 ] ) ] .
the dashed lines are the predictions of the gaussian approximation for the narrow gaussian potential [ _ i.e_. equation ( [ lvetap0 ] ) , with @xmath191 and @xmath190 given by equations ( [ lvtau ] ) and ( [ lvh ] ) , respectively ] , the dot - dashed curve is the prediction of the first order perturbation theory [ _ i.e_. equation ( [ etaper ] ) in the limit @xmath245 , and the filled squares are the predictions of fixman s theory [ _ i.e_. equation ( [ lvetap0 ] ) , with @xmath191 and @xmath190 given by equations ( [ fixlvspk ] ) ] .
the parameter @xmath0 does not enter into fixman s theory , however , these values are plotted corresponding to @xmath3 , since the quadratic potential is used in fixman s theory as an approximation for the @xmath2-potential .
the first feature to be noticed in figure [ fig1 ] is the reassuring closeness of the exact results obtained by using the retarded motion expansion and by brownian dynamics simulations .
the retarded motion expansion provides a means of validating the results of brownian dynamics simulations . in the limit @xmath107 , and for large values of @xmath0 ,
the continuous curves and brownian dynamics simulations reveal that , as expected , the exact predictions of the narrow gaussian potential tend to the @xmath1-independent @xmath49-solvent value , @xmath246 .
this implies , as pointed out earlier , that the use of a @xmath2-function potential to represent excluded volume interactions does not lead to any change in the zero shear rate viscosity prediction .
on the other hand , figure [ fig1 ] seems to suggest that a finite range of excluded volume interaction is required to cause a change from the @xmath49-solvent value .
away from these limits , at non - zero values of @xmath0 , the narrow gaussian potential predicts an increase in the value of zero shear rate viscosity .
the existence of shear thinning in good solvents can be attributed to this increase .
this follows from the fact that at high shear rates , as the effect of the excluded volume interaction diminishes , the viscosity is expected to return to its @xmath49-solvent value .
we shall see later that this expectation is indeed justified .
the dashed lines in figure [ fig1 ] indicate that in the limit of zero shear rate , for a given value of @xmath1 , the gaussian approximation is reasonably accurate above a certain value of @xmath0 .
this limiting value of @xmath0 appears to be smaller for smaller values of @xmath1 .
a similar behavior is also observed with regard to the prediction of the zero shear rate first normal stress difference ( see figure [ fig2 ] ) .
thus it appears that the exact configurational distribution function @xmath41 becomes increasingly non - gaussian as the narrow gaussian potential becomes narrower , and as the strength of the excluded volume interaction becomes larger .
for the large values of @xmath1 considered in figure [ fig1 ] , results obtained with the first order perturbation expansion in @xmath1 can not be expected to be accurate .
it is clear from the dot - dashed lines that the perturbation expansion results deviate significantly from exact results for small values of @xmath0 .
however , they become increasingly accurate as @xmath0 increases , for a given value of @xmath1 . this can be understood by considering equation ( [ etaper ] ) , which indicates that in the limit @xmath247 , the first order correction to the @xmath49-solvent value increases as @xmath1 increases , but decreases as @xmath0 increases .
as @xmath0 increases from zero , values of the zero shear rate viscosity predicted by the gaussian approximation , approach the exact values more rapidly than the predictions of the first order perturbation theory , for a given value of @xmath1 .
the gaussian approximation is a non - perturbative approximation
however , it was shown in section 5.4 , to be exact to first order in @xmath1 .
one way to understand this is to consider the gaussian approximation to consist of an infinite number of higher order terms , whose nature is unknown . in this sense
, it is an _ uncontrolled _ approximation , which remains accurate at values of @xmath1 and @xmath0 , where the first order perturbation expansion becomes inaccurate . as will be seen
shortly , these remarks apply also to the results obtained at finite shear rate .
the difference in the prediction of the zero shear rate viscosity by fixman s theory and by the narrow gaussian potential is evident in figure [ fig1 ] .
it is also worth noting that , although the relaxation weight and the relaxation time are different in fixman s theory and in the gaussian approximation for @xmath3 , they lead to the same prediction of the zero shear rate viscosity .
this is , however , not true for the zero shear rate first normal stress difference .
the ratio @xmath248 , defined by @xcite , @xmath249 is equal , in the zero shear rate limit , to @xmath250 in the gaussian approximation , while it has a constant value of 2 in fixman s theory . both the uniform expansion model and the gaussian approximation use gaussian distributions in order to evaluate averages .
however , the uniform expansion model uses different gaussian distributions for different equilibrium averages , such that the best approximation is obtained .
while this does not lead to any difference in the prediction of the zero shear rate viscosity by the two approximations , figure [ fig2 ] reveals that , at small enough values of @xmath0 , there is a significant difference in the prediction of the zero shear rate first normal stress difference .
clearly , the uniform expansion model continues to be a reasonable approximation for values of @xmath0 at which the gaussian approximation is no longer accurate .
however , even the uniform expansion model leads to a poor approximation at sufficiently small values of @xmath0 . in the discussion of the equilibrium perturbation expansion in section 5.1.2 ,
it was pointed out that , in principle , the equilibrium moment @xmath117 can be obtained for any non - zero value of @xmath0 , provided that enough numbers of terms in the series for the quantities @xmath144 and @xmath145 , are summed [ see equations ( [ s0s1 ] ) ] .
figure 3 displays the sum @xmath144 , for different numbers of summed terms in the perturbation expansion , as a function of @xmath0 , for @xmath251 .
exact results are obtained by noting that @xmath252 , and evaluating @xmath253 by numerical quadrature .
clearly , more terms of the expansion are required for convergence as @xmath0 decreases .
the terms of the series , which are alternating in sign , keep increasing rapidly in magnitude , until @xmath40 is approximately @xmath254 , before they begin to decrease .
therefore , as @xmath1 increases , or as @xmath0 decreases , more and more terms are required for the sum to converge .
above a threshold value of @xmath255 however , it becomes impossible to evaluate @xmath144 since the round - off errors due to the summation of large numbers makes the perturbation expansion numerically useless .
a similar problem is also encountered while evaluating the sum @xmath145 . in short ,
the hope of extrapolating finite @xmath0 results to the limit @xmath3 , can not be realised . therefore ,
in the case of hookean dumbbells , one can not obtain equilibrium moments for a delta function excluded volume potential by using a narrow gaussian potential and considering a perturbation expansion in the limit @xmath107 .
the results of brownian dynamics simulations ( without variance reduction ) displayed in figure [ fig4 ] reveal that the dependence of the viscosity and the first normal stress difference on @xmath0 , at a value of the non - dimensional shear rate @xmath256 , is similar in shape to the dependence observed in the limit of zero shear rate . at small and large values of @xmath0
the material functions tend to the @xmath49-solvent value , and exhibit a maximum at some value in between . since , even at this non - zero value of shear rate , it appears that the viscosity and the first normal stress difference remain at their @xmath49-solvent values for @xmath141 , it implies that the use of a @xmath2-potential to represent excluded volume interactions would not predict any shear thinning . on the other hand , as we shall see subsequently , a quadratic potential does predict substantial shear thinning . at @xmath256 ,
the gaussian approximation seems to be accurate above roughly the same values of @xmath0 as were observed in the limit of zero shear rate .
figures [ fig5 ] and [ fig6 ] are plots of non - dimensional viscosity and first normal stress difference versus the non - dimensional shear rate @xmath220 .
figure [ fig5 ] displays the dependence of these viscometric functions on the parameter @xmath0 , for a fixed value of @xmath257 , while figure [ fig6 ] displays the dependence on the parameter @xmath1 , for a fixed value of @xmath258 .
the prediction of shear thinning for non - zero values of @xmath0 is apparent , and our earlier expectation in this direction is justified . in particular , the predictions of brownian dynamics simulations , the gaussian approximation , and the first order perturbation theory tend to @xmath49-solvent values at high shear rates .
shear thinning , which is seemingly physically meaningful , is also predicted , as can be seen from figure [ fig5 ] , by both the gaussian approximation and the first order perturbation theory , for @xmath3 .
this corresponds to a @xmath2-function excluded volume potential , and is clearly an artifact of the perturbation expansion , since rigorous calculations indicate a trivial result .
it remains to be seen if the situation is different in the limit of long chains . for small enough values of @xmath1 , and large enough values of @xmath0 ,
the results of the gaussian approximation , and the first order perturbation expansion , agree exceedingly well with the exact results of brownian dynamics simulations .
indeed , as @xmath220 increases for fixed values of @xmath1 and @xmath0 , both the gaussian approximation , and the first order perturbation expansion become increasingly accurate . in particular ,
if both the approximations are accurate at zero shear rate , they continue to remain accurate at non - zero shear rates .
this can be understood in the case of the first order perturbation expansion by considering equations ( [ etaper ] ) and ( [ psi1per ] ) .
clearly , the departure from the @xmath49-solvent values decreases as @xmath220 increases .
this is inline with the intuitive expectation of decreasing excluded volume interactions with increasing shear rate . for a fixed value of the shear rate @xmath220 , as @xmath1 increases , or @xmath0 decreases , the predictions of the gaussian approximation and the first order perturbation expansion become increasingly inaccurate , with the first order perturbation expansion breaking down before the gaussian approximation
this can be expected , since the gaussian approximation being exact to first order in @xmath1is at least as accurate as the first order perturbation expansion .
the two different brownian dynamics simulation algorithms mentioned in section 5.2 were used to obtain the data in figures [ fig5 ] and [ fig6 ] .
while the algorithm with variance reduction was used for shear rates upto @xmath259 , the algorithm without variance reduction was used at higher shear rates . with regard to the results obtained by variance reduction
, it was found that the variance was typically reduced by a factor of five to ten by the parallel equilibrium simulation subtraction procedure . while the magnitude of the reduced variance was relatively independent of shear rate for the viscosity , it decreased with increasing shear rate for the first normal stress difference .
the time to reach steady - state , from start - up of flow , was roughly ten relaxation times for @xmath260 , @xmath243 and @xmath261 , and roughly fifteen relaxation times for @xmath244 .
rheological properties in the two parallel simulations remained correlated during the time required to reach steady - state , and as a result the present technique proved adequate for the purpose of variance reduction .
we have seen earlier in figure [ fig1]that both fixman s theory , and the gaussian approximation for @xmath3 , lead to identical values for the zero shear rate viscosity .
this coincidence is , however , restricted to the limit of zero shear rate . at non - zero shear rates , as can be seen from figure [ fig7 ] , there is considerable divergence between the predictions of the two theories .
fixman s theory for dumbbells , though differing considerably from the narrow gaussian potential in terms of its predictions of rheological properties , has the appealing aspect that it captures some of the _ universal _ features of the behavior of good solvents which can only be expected from bead - spring chain theories in the limit of a large number of beads .
we have seen this universal behavior earlier in the correct prediction of the end - to - end distance scaling , and in the parameter free nature of the ratio @xmath248 .
figure [ fig8 ] displays the prediction by fixman s theory of the reduced variable @xmath262 versus the non - dimensional shear rate @xmath263 where , @xmath264 , is a characteristic relaxation time . here , @xmath265 is the zero shear rate intrinsic viscosity , @xmath140 is the molecular weight and @xmath266 is avagadro s number . for dilute solutions one can show that , @xmath267 .
the figure clearly reveals that as @xmath268 , the curves for different values of @xmath1 overlap .
therefore , in this respect also , fixman s theory mimics the universal behavior expected of long chains .
the source of this behavior can be understood by examining equation ( [ sfs2 ] ) . in the limit of @xmath269
, one can show that @xmath270^{1/5}$ ] . as a result , @xmath271^{-2/5}$ ] leading to the observed scaling . at small values of @xmath220 ,
one can therefore construct the analytical expression , @xmath273 and expect it to be accurate at very small and large values of @xmath274 . the dot - dashed curve in figure [ fig8 ]
shows that equation ( [ anal ] ) is accurate over a fairly wide range of @xmath274 . the universal behavior discussed above is not exhibited by the gaussian approximation .
the reason for this can be easily understood in the limit @xmath107 , where , the second moment equation ( [ gasecmom ] ) reduces at steady - state to a non - linear algebraic equation .
indeed , one can show that for @xmath269 , @xmath275^{-1/5}$ ] . as a result
, the normalised material functions do not collapse onto a single curve when plotted versus @xmath274 .
it is , however , not realistic to expect a dumbbell model to exhibit universal features , and the real verification of universal behavior requires the development of a theory for long bead - spring chains .
the non - dimensional ratio @xmath248 [ see equation ( [ upsieta ] ) ] has a constant value of two , independent of shear rate , for hookean dumbbells in @xmath49-solvents , and in fixman s theory for good solvents .
figure 9 displays the predictions of @xmath248 by the narrow gaussian potential , obtained by brownian dynamics simulations , the gaussian approximation , and the following first order perturbation expansion ( which can be derived from equations ( [ etaper ] ) and ( [ psi1per ] ) ) , @xmath276 where , @xmath277 has been defined below equation ( [ psi1per ] ) . since
a logarithmic scale has been chosen for the shear rate axis , it is difficult to represent the zero shear rate value of @xmath248 . however , since it is very nearly constant at low values of shear rate , the zero shear rate value is represented in figure 9 by the filled circles on the @xmath278-axis .
all the data in figure 9 have the same trend of remaining nearly constant at low shear rates , and approaching asymptotically the value of two at high shear rates . in the case of the first order perturbation expansion ,
this can be understood by considering equation ( [ upsietaper ] ) in the limit @xmath279 .
the dependence of the mean squared end - to - end distance @xmath280 on @xmath220 is revealed in figure 10 .
the circles and triangles are the results of brownian dynamics simulations obtained with the algorithm without variance reduction .
the dotted and dashed lines are plots of the following expression , @xmath281\ ] ] obtained from the first order perturbation expansion , and the continuous line is the well known result for a theta solvent @xcite , @xmath282 .
interestingly , the effect of excluded volume interactions on swelling increases with increasing shear rate . in the case of the first order perturbation expansion which appears to be accurate for @xmath243 , but not for @xmath261this can be understood by noting , in the expression above , that the term representing the correction to the theta solvent result due to the presence of excluded volume increases as the shear rate increases .
the use of a narrow gaussian potential , to describe the excluded volume interactions between the beads of a hookean dumbbell model , leads to the prediction of swelling and shear thinning for relatively small non - zero values of the extent of interaction @xmath0 .
this is essentially caused by an increase in the magnitude of the equilibrium moments relative to their @xmath49-solvent values .
a delta function description of the excluded volume potential , on the other hand , is found to predict neither swelling nor shear thinning for hookean dumbbells . for a given strength of the excluded volume interaction @xmath1 , the gaussian approximation is found to be reasonably accurate for values of @xmath0 larger than some threshold value .
the behavior of the gaussian approximation can be understood by comparing its predictions with those of a first order perturbation expansion in @xmath1 , since it is shown here to be exact to first order in @xmath1 .
the perturbation expansion reveals that the departure of the viscometric functions from their @xmath49-solvent values increases with increasing @xmath1 , but decreases with increasing @xmath0 , and increasing shear rate @xmath220 .
the use of a quadratic potential in fixman s theory leads to the prediction of viscometric functions which are considerably different from those of the narrow gaussian potential .
however , fixman s theory for dumbbells reproduces a number of universal features observed in good solvents . * acknowledgement * support for this work through a grant iii .
5(5)/98-et from the department of science and technology , india , to j. ravi prakash is acknowledged .
part of this work was carried out while jrp was a participant in the research programme _ jamming and rheology _ at the institute for theoretical physics , university of california , santa barbara , usa .
jrp would also like to thank the non - linear dynamics group at iit madras for providing the use of their computational facility .
99 ahn , k. h. ; lee , s. j. _ j. non - newtonian fluid mech . _ * 1992 * , _ 43 _ , 143 ; ahn , k. h. ; schrag , j. l. ; lee , s. j. _ j. non - newtonian fluid mech . _ * 1993 * , _ 50 _ , 349 .
andrews , n. c. ; doufas , a. k. ; mchugh , a. j. _ macromolecules _ * 1998 * , _ 31 _ , 3104 .
arfken , g. _ mathematical methods for physicists , _
3rd edn . ; academic press : boston , 1985
. bird , r. b. ; armstrong , r. c. ; hassager , o. _ dynamics of polymeric liquids .
fluid mechanics , _
; wiley - interscience : new york , 1987 ; vol . 1 . bird , r. b. ; curtiss , c. f. ; armstrong , r. c. ; hassager , o. _ dynamics of polymeric liquids . kinetic theory , _
; wiley - interscience : new york , 1987 ; vol . 2 .
buck , r. c. _ advanced calculus _ ; mcgraw - hill : new york , 1978 .
des cloizeaux , j. ; jannink , g. _ polymers in solution , their modelling and structure _ ; oxford university press : oxford , 1990 .
doi , m. ; edwards , s. f. _ the theory of polymer dynamics _ ; oxford university press : oxford , 1986 .
fixman , m. _ j. chem . phys . _ * 1966 * , _ 45 _ , 785 ; * 1966 * , _ 45 _ , 793 .
larson , r. g. _ constitutive equations for polymer melts and solutions _ ; butterworths : boston , 1988 .
press , w. h. ; teukolsky , s. a. ; vetterling , w. t. ; flannery , b. p. _ numerical recipes in fortran _ , 2nd edn . ; cambridge university press : cambridge , 1992 .
ttinger , h. c. _ j. non - newtonian fluid mech . _
* 1987 * , _ 26 _ , 207 .
ttinger , h. c. ; rabin , y. _ j. non - newtonian fluid mech . _
* 1989 * , _ 33 _ , 53 .
ttinger , h. c. _ phys .
rev . _ * 1989 * , _ a40 _ , 2664 .
ttinger , h. c. _ stochastic processes in polymeric fluids _ ; springer : berlin , 1996 .
prakash , j. r. ` the kinetic theory of dilute solutions of flexible polymers : hydrodynamic interaction ' . in _ advances in the flow and rheology of non - newtonian fluids _ ; siginer , d.a . ;
kee , d.d . ; chhabra , r.p . ,
eds . ; rheology series ; elsevier science ; in press .
schieber , j. d. ; _ j. rheol . _ * 1993 * , _ 37 _ , 1003 .
wagner , n. j. ; ttinger , h. c. _ j. rheol . _ * 1997 * , _ 41 _ , 757 .
wedgewood , l. e. _ rheol . acta _ * 1993 * , _ 32 _ , 405 .
yamakawa , h. _ modern theory of polymer solutions _ ; harper and row : new york , 1971 .
zylka , w. ; ttinger h. c. _ j. chem .
phys . _ * 1989 * , _ 90 _ , 474 .
zylka , w. _ j. chem . phys . _
* 1991 * , _ 94 _ , 4628 .
zylka , w. ; ttinger h. c. _ macromolecules _ * 1991 * , _ 24 _ , 484 .
the equilibrium average of any quantity @xmath151 is given by , @xmath283 where , @xmath284 . by multiplying and dividing the integrand with the gaussian distribution @xmath285 [ defined by equation ( [ uemgau ] ) ] , and using the normalisation conditions on @xmath286 and @xmath287 , equation ( [ xavg ] ) can be rewritten in the form , @xmath288 where , @xmath289 \ , q^2 + e^*$ ] .
for @xmath290 to be a good approximation to @xmath291 , @xmath292 must be small . on expanding @xmath293 in a taylor s series
, one can then write equation ( [ xavgmod ] ) , to first order in @xmath292 as , @xmath294 if @xmath156 is chosen such that @xmath295 it then follows that @xmath296 is well approximated by @xmath297 .
the choice of @xmath156 that best optimises the approximation clearly depends on the quantity @xmath151 that is averaged . in our case
we require averages of the quantities @xmath298 . from the well known result for the moments of a gaussian distribution , @xmath299^m \ , \gamma ( m + { 3 \over 2})\ ] ] where , @xmath300 represents the value of @xmath301 corresponding to @xmath302 .
it follows that the optimum value of @xmath303 is found by solving the following non - linear algebraic equation for @xmath304 , @xmath305^{3/2 } } \ , \bigl\lbrace \ , { \mu^{2 m } \over [ \alpha_{(m)}^2 + \mu^2 ] ^m } - 1 \ , \bigr\rbrace = 0\ ] ]
where , @xmath306 . in terms of @xmath304 , @xmath307 | a dilute polymer solution is modeled as a suspension of non - interacting hookean dumbbells and the effect of excluded volume is taken into account by incorporating a narrow gaussian repulsive potential between the beads of each dumbbell .
the narrow gaussian potential is a means of regularising a delta - function potential
it tends to the delta - function potential in the limit of the width parameter @xmath0 going to zero .
exact predictions of viscometric functions in simple shear flow are obtained with the help of a retarded motion expansion and by brownian dynamics simulations .
it is shown that for relatively small _ non - zero _ values of @xmath0 , the presence of excluded volume causes a _ swelling _ of the dumbbell at equilibrium , and _ shear thinning _ in simple shear flow . on the other hand ,
a delta function excluded volume potential does not lead to either swelling or to shear thinning .
approximate viscometric functions , obtained by assuming that the bead - connector vector is described by a gaussian non - equilibrium distribution function , are found to be accurate above a threshold value of @xmath0 , for a given value of the strength of excluded volume interaction , @xmath1 .
a first order perturbation expansion reveals that the gaussian approximation is exact to first order in @xmath1 .
the predictions of an alternative quadratic excluded volume potential suggested earlier by fixman ( j. chem .
phys . , 1966 , 45 , 785 ; 793 )
are also compared with those of the narrow gaussian potential . |
a rational difference equation is a nonlinear difference equation of the form @xmath1 where the initial conditions @xmath2 are such that the denominator never vanishes for any @xmath3 .
+ consider the equation @xmath4 where all the parameters and the initial conditions @xmath5 and @xmath6 are arbitrary complex number . this second order rational difference equation eq.([equation : total - equationa ] )
is studied when the parameters are real numbers and initial conditions are non - negative real numbers in @xcite . in this present article
it is an attempt to understand the same in the complex plane . + here , a very brief review of some well known results which will be useful in order to apprehend the behavior of solutions of the difference equation ( [ equation : total - equationa ] ) .
let @xmath7 where @xmath8 be a continuously differentiable function .
then for any pair of initial conditions @xmath9 , the difference equation @xmath10 with initial conditions @xmath11 + + then for any _ initial value _ , the difference equation ( 1 ) will have a unique solution @xmath12 .
+ + a point @xmath13 @xmath14 @xmath15 is called _ * equilibrium point * _ of eq.([equation : introduction ] ) if @xmath16 the _ linearized equation _ of eq.([equation : introduction ] ) about the equilibrium @xmath17 is the linear difference equation @xmath18 where for @xmath19 and @xmath20 .
@xmath21 the _ characteristic equation _ of eq.(2 ) is the equation @xmath22 the following are the briefings of the linearized stability criterions which are useful in determining the local stability character of the equilibrium @xmath13 of eq.([equation : introduction ] ) , @xcite .
let @xmath13 be an equilibrium of the difference equation @xmath23 . * the equilibrium @xmath17 of eq .
( 2 ) is called * locally stable * if for every @xmath24 , there exists a @xmath25 such that for every @xmath26 and @xmath5 @xmath14 @xmath27 with @xmath28 we have @xmath29 for all @xmath30 . * the equilibrium @xmath17 of eq .
( 2 ) is called * locally stable * if it is locally stable and if there exist a @xmath31 such that for every @xmath26 and @xmath5 @xmath14 @xmath27 with @xmath32 we have @xmath33 . *
the equilibrium @xmath17 of eq .
( 2 ) is called * global attractor * if for every @xmath26 and @xmath5 @xmath14 @xmath27 , we have @xmath33 . *
the equilibrium of equation eq .
( 2 ) is called * globally asymptotically stable / fit * is stable and is a global attractor . *
the equilibrium @xmath17 of eq .
( 2 ) is called * unstable * if it is not stable . *
the equilibrium @xmath17 of eq .
( 2 ) is called * source or repeller * if there exists @xmath34 such that for every @xmath26 and @xmath5 @xmath14 @xmath27 with @xmath35 we have @xmath36 .
clearly a source is an unstable equilibrium .
* result 1.1 : ( clark s theorem ) * the sufficient condition for the asymptotic stability of the difference equation ( 1 ) is @xmath37
the following difference equation is considered to be studied here . @xmath38 where all the parameters are complex number and the initial conditions @xmath5 and @xmath6 are arbitrary complex numbers . + we will consider three different cases of the eq.([equation : introduction ] ) which are as follows : + by the change of variables , @xmath40 , the difference equation @xmath41 reduced to the difference equation @xmath42 where @xmath43 and @xmath44 . by the change of variables , @xmath46 , the difference equation @xmath47 reduced to the difference equation @xmath48 where @xmath49 and @xmath50 . by the change of variables , @xmath52 , the difference equation @xmath53 reduced to the difference equation @xmath54 where @xmath55 and @xmath50 .
+ without any loss of generality , we shall now onward focus only on the three difference equations ( 6 ) , ( 7 ) and ( 8) .
in this section we establish the local stability character of the equilibria of eq.([equation : total - equationa ] ) in three difference cases as stated in the section 2 .
the equilibrium points of eq.(6 ) are the solutions of the quadratic equation @xmath57 eq.(6 ) has the two equilibria points @xmath58 and @xmath59 respectively
. the linearized equation of the rational difference equation(6 ) with respect to the equilibrium point @xmath60 is @xmath61 with associated characteristic equation
@xmath62 the following result gives the local asymptotic stability of the equilibrium @xmath63 of the eq .
the equilibriums @xmath64 of eq.(6 ) is + + locally asymptotically stable if @xmath65 the zeros of the characteristic equation ( 10 ) has two zeros which are @xmath66 and @xmath67 . therefore by _
clark s theorem _ , the equilibrium @xmath64 is _ locally asymptotically stable _ if the sum of the modulus of two coefficients is less than @xmath20 .
therefore the condition of the polynomial ( 10 ) reduces to @xmath68 .
the linearized equation of the rational difference equation ( 6 ) with respect to the equilibrium point @xmath69 is @xmath70 with associated characteristic equation @xmath71 the equilibriums @xmath72 of eq.(6 ) is + + locally asymptotically stable if @xmath73 proof the theorem follows from _
clark s theorem _ of local asymptotic stability of the equilibriums .
the condition for the local asymptotic stability reduces to @xmath74 .
here is an example case for the local asymptotic stability of the equilibriums .
+ for @xmath75 and @xmath76 the equilibriums are @xmath77 and @xmath78 . for the equilibrium @xmath79 , the coefficients of the characteristic polynomial ( 10 ) are @xmath80 and @xmath81 with same modulus @xmath82 .
therefore the condition as stated in the _ theorem 3.1 _ does not hold . therefore the equilibrium @xmath79 is _ unstable_. + for the equilibrium @xmath78 , the coefficients of the characteristic polynomial ( 12 ) are @xmath83 and @xmath84 with same modulus @xmath85 . therefore the condition as stated in the _ theorem 3.2 _ is hold good .
therefore the equilibrium @xmath78 is _ locally asymptotically stable_. + it is seen that in the case of real positive parameters and initials values , the positive equilibrium of the difference equation ( 6 ) is globally asymptotically stable @xcite . but the result is not holding well in the complex set up .
the equilibrium points of eq.(7 ) are the solutions of the quadratic equation @xmath87 the eq.(7 ) has only the zero equilibrium .
the linearized equation of the rational difference equation(7 ) with respect to the zero equilibrium is @xmath88 with associated characteristic equation
@xmath89 the following result gives the local asymptotic stability of the zero equilibrium of the eq .
the zero equilibriums of the eq .
( 7 ) is non - hyperbolic .
the zeros of the characteristic equation ( 14 ) has two zeros which are @xmath90 and @xmath20 .
therefore by definition , the zero equilibrium is non - hyperbolic as the modulus of one zero is @xmath20 .
it is nice to note that in the case of real positive parameters and initials values , the zero equilibrium of the difference equation ( 7 ) is globally asymptotically stable for the parameter @xmath91 @xcite .
but in the case of complex , the zero equilibrium is non - hyperbolic as we have seen the previous theorem .
the equilibrium points of eq.(8 ) are the solutions of the quadratic equation @xmath93 the eq.(8 ) has two equilibriums which are @xmath90 and @xmath94 .
the linearized equation of the rational difference equation(8 ) with respect to the zero equilibrium is @xmath95 with associated characteristic equation
@xmath96 the following result gives the local asymptotic stability of the zero equilibrium of the eq .
the zero equilibriums of the eq .
( 8) is locally asymptotically stable if @xmath97 and repeller if @xmath98 .
the zeros of the characteristic equation ( 15 ) has two zeros which are @xmath90 and @xmath99 .
therefore by definition , the zero equilibrium is locally asymptotically stable if @xmath100 and unstable ( repeller ) if @xmath101 .
hence the required is followed .
the linearized equation of the rational difference equation(8 ) with respect to the equilibrium @xmath94 is @xmath102 with associated characteristic equation
@xmath103 the following result gives the local asymptotic stability of the equilibrium @xmath94 of the eq .
the zero equilibriums of the eq .
( 8) is locally asymptotically stable if @xmath104 the equilibrium @xmath94 of the characteristic equation ( 18 ) would be _ locally asymptotically stable _ if the sum of the modulus of the coefficients of the characteristic equation ( 18 ) is less than 1 .
that is by _
clark s theorem _
, @xmath105 , that is @xmath104 here is an example case for the local asymptotic stability of the equilibriums .
+ for @xmath106 @xmath107 ) and @xmath108 ( @xmath109 ) the equilibriums are @xmath90 and @xmath110 . for the equilibrium @xmath110 , the coefficients of the characteristic polynomial ( 18 ) are @xmath111 and @xmath112 with modulus @xmath113 and @xmath114 respectively .
therefore the condition as stated in the _ theorem 3.5 _ hold good . therefore the equilibrium @xmath79 is _ locally asymptotically stable_. + in the case of real positive parameters and intimal values of the difference equation ( 8) , the positive equilibrium is locally asymptotically stable if @xmath115 and @xmath116 but in the complex set , it is encountered through the example above is that the equilibrium @xmath94 is locally asymptotically stable even though @xmath117 and @xmath118 @xcite .
in this section we would like to explore the boundedness of the solutions of the three difference equations ( 6 ) , ( 7 ) and ( 8) .
+ now we would like to try to find open ball @xmath119 such that if @xmath120 and @xmath121 then @xmath122 for all @xmath123 .
in other words , if the initial values @xmath124 and @xmath125 belong to @xmath126 then the solution generated by the difference equations would essentially be within the open ball @xmath126 . [
result : positive - local - stability2 ] for the difference equation ( 6 ) , for every @xmath24 , if @xmath127 and @xmath128 @xmath129 then @xmath122 provided @xmath130 let @xmath131 be a solution of the equation eq.(6 ) .
let @xmath132 be any arbitrary real number .
consider @xmath133 .
we need to find out an @xmath134 such that @xmath135 for all @xmath3 .
it is follows from the eq.(6 ) that for any @xmath24 , using triangle inequality for @xmath136 in order to ensure that @xmath137 , ( assuming @xmath138 ) it is needed to be @xmath139 that is @xmath140 . therefore the required is followed .
[ result : positive - local - stability2 ] for the difference equation ( 7 ) , for every @xmath24 , if @xmath127 and @xmath128 @xmath129 then @xmath122 provided @xmath141 let @xmath131 be a solution of the equation eq.(7 ) .
let @xmath132 be any arbitrary real number .
consider @xmath133 .
we need to find out an @xmath134 such that @xmath135 for all @xmath3 .
it is follows from the eq.(7 ) , that for any @xmath24 , using triangle inequality for @xmath142 therefore , @xmath143 .
we need @xmath144 to be less than @xmath134 .
therefore , @xmath141 is followed .
[ result : positive - local - stability2 ] for the difference equation ( 8) , for every @xmath24 , if @xmath127 and @xmath128 @xmath129 then @xmath122 and @xmath98 provided @xmath145 let @xmath131 be a solution of the equation eq.(8 ) .
let @xmath132 be any arbitrary real number and @xmath117 .
consider @xmath133 .
we need to find out an @xmath134 such that @xmath135 for all @xmath3 .
it is follows from the eq.(8 ) , that for any @xmath24 , using triangle inequality for @xmath146 therefore , @xmath147 in order to ensure that @xmath148 , it is needed to be @xmath149 that is @xmath145 therefore the required is followed .
a solution @xmath150 of a difference equation is said to be _ globally periodic _ of period @xmath151 if @xmath152 for any given initial conditions .
solution @xmath150 is said to be _ periodic with prime period _
@xmath153 if p is the smallest positive integer having this property . + we shall first look for the prime period two solutions of the three difference equations ( 6 ) , ( 7 ) and ( 8) and their corresponding local stability analysis .
let @xmath154 , @xmath155 be a prime period two solution of the difference equation @xmath56
. then @xmath156 and @xmath157 .
this two equations lead to the set of solutions ( prime period two ) except the equilibriums as @xmath158 and @xmath159 .
+ let @xmath154 , @xmath155 be a prime period two solution of the equation ( 6 ) .
we set @xmath160 @xmath161 then the equivalent form of the difference equation ( 6 ) is @xmath162 @xmath163 let t be the map on @xmath164 to itself defined by @xmath165 then @xmath166 is a fixed point of @xmath167 , the second iterate of @xmath168 .
+ @xmath169 @xmath170 where @xmath171 and @xmath172 .
clearly the two cycle is locally asymptotically stable when the eigenvalues of the jacobian matrix @xmath173 , evaluated at @xmath166 lie inside the unit disk .
+ we have , @xmath174 where @xmath175 and @xmath176 + @xmath177 and @xmath178 + + now , set @xmath179 @xmath180 in particular for the prime period @xmath181 solution , + + @xmath158 , we shall see the local asymptotic stability for some example cases of parameters @xmath153 and @xmath182 .
the general form of @xmath183 and @xmath184 would be very complected .
consider the prime period two solution of the difference equation ( 6 ) , @xmath185 corresponding two the parameters @xmath186 and @xmath187 .
+ + in this case , @xmath188 and @xmath189 .
therefore , by the linear stability theorem ( @xmath190 ) the prime period @xmath181 solution @xmath185 is _ locally asymptotically stable_. let @xmath154 , @xmath155 be a prime period two solution of the difference equation @xmath86
. then @xmath191 and @xmath192 .
this two equations lead to the set of solutions ( prime period two ) except the equilibriums as @xmath193 .
+ let @xmath154 , @xmath155 be a prime period two solution of the equation ( 7 ) .
we set @xmath160 @xmath161 then the equivalent form of the difference equation ( 7 ) is @xmath162 @xmath194 let t be the map on @xmath164 to itself defined by @xmath195 then @xmath166 is a fixed point of @xmath167 , the second iterate of @xmath168 .
+ @xmath196 @xmath170 where @xmath197 and @xmath198 .
clearly the two cycle is locally asymptotically stable when the eigenvalues of the jacobian matrix @xmath173 , evaluated at @xmath166 lie inside the unit disk .
+ we have , @xmath174 where @xmath199 and @xmath200 + @xmath201 and @xmath202 + + now , set @xmath203 @xmath204 in particular for the prime period @xmath181 solution , + + @xmath205 , we shall see the local asymptotic stability for some example cases of parameters @xmath153 and @xmath182 .
the general form of @xmath183 and @xmath184 would very complected .
consider the prime period two solution of the difference equation ( 7 ) , @xmath206 corresponding two the parameters @xmath207 and @xmath208 .
+ + in this case , @xmath209 and @xmath210 .
therefore , by the linear stability theorem ( @xmath190 ) the prime period @xmath181 solution @xmath206 is _ locally asymptotically stable_. let @xmath154 , @xmath155 be a prime period two solution of the difference equation @xmath92 . then @xmath211 and @xmath212 .
this two equations lead to the set of solutions ( prime period two ) except the equilibriums as @xmath213 .
+ let @xmath154 , @xmath155 be a prime period two solution of the equation ( 8) .
we set @xmath160 @xmath161 then the equivalent form of the difference equation ( 8) is @xmath162 @xmath214 let t be the map on @xmath164 to itself defined by @xmath215 then @xmath166 is a fixed point of @xmath167 , the second iterate of @xmath168 .
+ @xmath216 @xmath170 where @xmath217 and @xmath218 .
clearly the two cycle is locally asymptotically stable when the eigenvalues of the jacobian matrix @xmath173 , evaluated at @xmath166 lie inside the unit disk .
+ we have , @xmath174 where @xmath219 and @xmath220 + @xmath221 and @xmath222 + + now , set @xmath223 @xmath224 for the prime period @xmath181 solution , @xmath225 , @xmath226 and @xmath227 . therefore , by the linear stability theorem ( @xmath190 )
the prime period @xmath181 solution is _ locally asymptotically stable _ if and only if @xmath228 .
it turns out that @xmath229 .
in other words , the condition reduces to @xmath98 and @xmath118 which is same condition as it was for the real set up .
this is something which is absolutely new feature of the dynamics of the difference equation ( 1 ) which did not arise in the real set up of the same difference equation .
computationally we have encountered some chaotic solution of the difference equation ( 8) for some parameter values which are given in the following table .
+ the method of lyapunov characteristic exponents serves as a useful tool to quantify chaos .
specifically lyapunav exponents measure the rates of convergence or divergence of nearby trajectories .
negative lyapunov exponents indicate convergence , while positive lyapunov exponents demonstrate divergence and chaos .
the magnitude of the lyapunov exponent is an indicator of the time scale on which chaotic behavior can be predicted or transients decay for the positive and negative exponent cases respectively . in this present study
, the largest lyapunov exponent is calculated for a given solution of finite length numerically @xcite .
+ from computational evidence , it is arguable that for complex parameters @xmath153 and @xmath182 which are stated in the following table the solutions are chaotic for every initial values .
* interval of lyapunav exponent * + @xmath230 , @xmath231 & @xmath232 + @xmath233 , @xmath234 & @xmath235 + @xmath236 , @xmath237 & @xmath238 + @xmath239 , @xmath240 & @xmath241 + the chaotic trajectory plots including corresponding complex plots are given the following fig .
1 . in the fig .
1 , for each of the four cases ten different initial values are taken and plotted in the left and in the right corresponding complex plots are given . from the fig .
1 , it is evident that for the four different cases the basin of the chaotic attractor is neighbourhood of the centre @xmath242 of complex plane .
does the difference equation have higher order periodic cycle ?
if so , what is the highest periodic cycle ?
find out the set of all parameters @xmath153 and @xmath182 for which the difference equation ( 8) has chaotic solutions .
find out the subset of the @xmath15 of all possible initial values @xmath26 and @xmath243 for which the solutions of the difference equation are chaotic for any complex parameters @xmath153 and @xmath182 .
does the neighbourhood of @xmath242 is global chaotic attractor ? if not , are there any other chaotic attractors ?
in continuation of the present work the study of the difference equation @xmath244 where @xmath245 , @xmath246 , @xmath247 , @xmath248 , @xmath249 and @xmath250 are all convergent sequence of complex numbers and converges to @xmath251 , @xmath252 , @xmath253 , @xmath254 , @xmath255 and @xmath256 respectively is indeed would be very interesting and that we would like to pursue further .
also the most generalization of the present rational difference equation is @xmath257 where @xmath258 and @xmath259 are delay terms and it demands similar analysis which we plan to pursue in near future .
the author thanks _ dr . pallab basu _ for discussions and suggestions .
saber n elaydi , henrique oliveira , jos manuel ferreira and joo f alves , _ discrete dyanmics and difference equations , proceedings of the twelfth international conference on difference equations and applications _ , world scientific press , 2007 .
s. atawna , r. abu - saris , e. s. ismail , and i. hashim , stability of nonhyperbolic equilibrium solution of second order nonlinear rational difference equation , _ journal of difference equations _
volume 2015 ( 2015 ) , article i d 486985 .
e. camouzis , e. chatterjee , g. ladas and e. p. quinn , on third order rational difference equations open problems and conjectures , _ journal of difference equations and applications _ , 10(2004 ) , 1119 1127 .
grove , y. kostrov and s.w .
schultz , on riccati difference equations with complex coefficients , _ proceedings of the 14th international conference on difference equations and applications , istanbul , turkey _ , march 2009 . | the dynamics of the second order rational difference equation @xmath0 with complex parameters and arbitrary complex initial conditions is investigated . in the complex set up , the local asymptotic stability and boundedness are studied vividly for this difference equation .
several interesting characteristics of the solutions of this equation , using computations , which does not arise when we consider the same equation with positive real parameters and initial conditions are shown .
the chaotic solutions of the difference equation is absolutely new feature in the complex set up which is also shown in this article
. some of the interesting observations led us to pose some open interesting problems regarding chaotic and higher order periodic solutions and global asymptotic convergence of this equation . |
the problem of estimating parameters of selected populations has wide practical applications in estimation of experimental data in agriculture , industry and medicine .
some of the real world applications of this theory are the problem of estimating the average yield of a selected variety of plant with maximum yield ( kumar and kar , 2001 ) , estimating the average fuel efficiency of the vehicle with minimum fuel consumption ( kumar and gangopadhyay , 2005 ) and selecting the regimen with maximal efficacy or minimal toxicity from a set of regimens and estimating a treatment effect for the selected regimen ( sill and sampson , 2007 ) .
the problem of estimation after selection has received considerable attention by many researches in the past three decades .
interested readers are referred to , for example , gibbons et al .
( 1977 ) for more details .
some other contributions in this area include sarkadi ( 1967 ) , dahiya ( 1974 ) , kumar and kar ( 2001 ) , misra et al .
( 2006a , b ) , kumar et al .
( 2009 ) and nematollahi and motammed - shariati ( 2012 ) . for a summary of results , as well as a list of references until 2006 ,
see misra et al .
( 2006 a , b ) . in this paper
, we introduce and develop the problem of estimation of the parameters of a dynamically selected population from a sequence of infinite populations which is not studied in the literature , according to the best of our knowledge .
let @xmath0 be a sequence of random variables where @xmath1 is drawn from population @xmath2 with corresponding cumulative distribution function ( cdf ) @xmath3 and probability density function ( pdf ) @xmath4 .
the traffic volume trend , daily temperatures , sequences of stock quotes , or sequences of estimators of interior water volume in a dam reservoir are examples of such sequences . suppose we want to estimate the parameter of the population corresponding to the largest value of the sequence @xmath0 yet seen , that is @xmath5}^u=\theta_{t_n},\ ] ] where @xmath6 , with probability one , and for @xmath7 @xmath8 or similarly the parameter of the population corresponding to the smallest value of the sequence @xmath0 yet seen , that is @xmath5}^l=\theta_{t'_n},\ ] ] where @xmath9 , with probability one , and for @xmath7 @xmath10 we want to estimate @xmath11}^u$ ] , and similarly the lower ones @xmath11}^l$ ] .
this happens for example , when we want to estimate the largest value of traffic volume or stock quotes yet seen , the temperature of the coldest day or the largest volume of the coming water into the dam reservoir , up to now . for simplicity , we denote @xmath11}^u$ ] by @xmath11}$ ] hereafter .
we may write @xmath12}=\sum_{j = n}^{\infty}\theta_ji_j(x_1,x_2,\ldots),\end{aligned}\ ] ] where @xmath13 the statistics @xmath14 and @xmath15 are called upper and lower records , respectively . in the sequence @xmath16 ,
the sequences of partial maxima and upper record statistics are defined by @xmath17 and @xmath18 , respectively , where @xmath6 with probability 1 , and @xmath19 for @xmath20 .
the record statistics @xmath21 could be viewed as the dynamic maxima of the original random variables .
so , we call the problem of estimating @xmath11}$ ] as the estimation of the parameter of a dynamically selected population
. there is a vast literature on records for iid as well as non - stationary random variables . a thorough survey of available results , until 1998 ,
is given in the book of arnold et al .
more recent articles on record values include , among others , amini and balakrishnan ( 2013 , 2015 ) , doostparast and emadi ( 2013 ) , salehi et al .
( 2013 ) , ahmadi and balakrishnan ( 2013 , 2010 ) , psarrakos and navarro ( 2013 ) , raqab and ahmadi ( 2012 ) , zarezadeh and asadi ( 2010 ) , kundu et al .
( 2009 ) and baklizi ( 2008 ) .
this problem is related to the so - called _
general record model_. the geometrically increasing populations , the pfeifer , the linear drift and the @xmath22 record models are some of the generally used record models .
the basics of non - stationary schemes for the record values are due to nevzorov ( 1985 , 1986 ) and pfeifer ( 1989 , 1991 ) , who considered the so - called @xmath22-scheme , that is the sequences of independent random variables with distribution @xmath23 , where @xmath24 is a continuous cdf and @xmath25 s are positive parameters .
further generalization of the @xmath22-scheme was suggested by ballerini and resnick ( 1987 ) .
although non - stationary schemes could be employed in the most general setting , the special case of improving populations is usually of special interest .
alternative non - stationary schemes include geometrically increasing populations , linear trend and pfeifer models . in all the above models ,
strict assumptions are made on the sequence of parameters @xmath26 .
for instance , in @xmath22 record model , the sequence of the parameters is assumed to be known or depend on a fixed unknown parameter . in the linear drift model
, a linearly increasing population is assumed as the underlying population .
however , certain natural phenomena may behave otherwise .
for example , an earthquake is produced by a natural phenomenon which has a pivotal parameter that varies based on an unknown model . in order to predict extremely destructive earthquakes ,
a very important question is on the value of the parameters which cause a new record in the sequence of earthquakes ?
this motivates us to study the problem of dynamic after - selection estimation .
the rest of this paper is organized as follows .
the theoretical results of the dynamic after - selection problem , consisting unbiased estimation of the parameters of the model as well as unbiased estimation of the risk of the estimators are presented in sections 2 and 3 . in section 4 , we compare the proposed estimators with some natural estimators .
asymptotic distributional results for studying the limiting behavior of the risks of the estimators are studied in section 5 .
finally , a real data example is considered in section 6 to illustrate the applicability of the results .
let @xmath27 , @xmath28 and @xmath29 be a random parameter ( a function of @xmath30 and @xmath31 ) .
suppose that @xmath29 is estimated by @xmath32 .
following lehmann ( 1951 ) , the estimator @xmath32 is said to be risk unbiased for @xmath29 under the loss function @xmath33 , if it satisfies @xmath34 under the squared error loss ( sel ) function @xmath35 the condition reduces to @xmath36 in this section , we use the u - v method of robbins ( 1988 ) , to find uniformly minimum variance unbiased ( umvu ) estimator of @xmath11}$ ] under the two models 1 and 2 , presented below .
* model 1 : * let @xmath0 be a sequence of independent absolutely continuous random variables with pdf @xmath37 where @xmath38 is a complete sufficient statistic with the gamma@xmath39)-distribution .
some well - known members of the above family are : \1 .
exponential(@xmath40 ) , with @xmath41 , @xmath42 and @xmath43 ; \2 .
gamma(@xmath44 ) , with @xmath42 and @xmath45 ; \3 .
normal(0,@xmath46 ) , with @xmath47 , @xmath48 , @xmath49 and @xmath50 ; \4 .
inverse gaussian(@xmath51 ) , with @xmath52 , @xmath48 , @xmath53 and @xmath54 ; \5 .
weibull(@xmath55 ) , with known @xmath56 , @xmath57 , @xmath41 , @xmath58 and @xmath59 ; \6 .
rayleigh(@xmath60 ) , with @xmath61 , @xmath41 , @xmath49 and @xmath62 . to estimate @xmath11}$ ] in the family of distributions ,
we first consider the estimation of @xmath11}$ ] under the gamma(@xmath44)-distribution with pdf @xmath63 by using the u - v method of robbins ( 1988 ) , we have the following lemma ( see also vellaisamy and sharma , 1989 ) .
[ lem1 ] let @xmath0 be a sequence of independent random variables with densities defined in ( [ gamma ] ) .
let @xmath64 be a real - valued function such that for @xmath65 \(ii ) @xmath67 then the functions @xmath68 satisfy @xmath69 = e_{{\boldsymbol\theta}}\left [ \theta_j u_j(\mathbf{x})\right],~j=1,2,\cdots.\end{aligned}\ ] ] the next result obtains the unbiased estimator of @xmath11}$ ] , under the sel function , for the gamma(@xmath44 ) distribution with the pdf of @xmath1 as in . [ t1 ] for the gamma(@xmath44 ) distribution with the pdf of @xmath1 as in , an unbiased estimator of @xmath11}$ ] , under sel function , which satisfies with @xmath70}$ ] , is @xmath71 where @xmath21 is the @xmath72 upper record value of the sequence @xmath16 . from ( [ tet ] ) , ( [ ij ] ) and lemma 1 , an unbiased estimator of @xmath11}$ ] , under sel function , based on @xmath16
is given by @xmath73 where @xmath74 is defined in .
thus , @xmath75 @xmath76 to find an unbiased estimator of @xmath11}$ ] under the model 1 with the pdf of @xmath1 as in , let @xmath77 , @xmath78 and @xmath79 . then , by replacing @xmath1 with @xmath80 in theorem [ t1 ] , an unbiased estimator of @xmath11}$ ] , under the sel function , for the general family , can be obtained as @xmath81 where @xmath82 is the @xmath72 upper record value of the sequence @xmath83 . for a monotone function @xmath84 ( available in all of the above examples , except in the normal distribution ) ,
@xmath82 can be obtained simply as @xmath85 for an increasing @xmath86 and as @xmath87 for a decreasing @xmath86 .
for example , for the rayleigh(@xmath60)-distribution , an unbiased estimator for @xmath88}$ ] is @xmath89}=\frac{u_n^2/2}{1}\left(1-\left(\frac{u_{n-1}^2/2}{u_{n}^2/2}\right)^{1}\right)=\frac{u_n^2}{2}\left(1-\left(\frac{u_{n-1}}{u_{n}}\right)^2\right)=\frac{u_n^2-u_{n-1}^2}{2}.\end{aligned}\ ] ] * model 2 : * for @xmath90 , consider two families of distributions , the first with @xmath1 having the survival function @xmath91 and the second with @xmath1 having the cdf @xmath92 in which @xmath93 is a cdf , free of @xmath40 , and @xmath94 .
we assume @xmath95 to be known .
these are called _ proportional hazard rate _ and _ proportional reversed hazard rate _ families , or simply @xmath22 models in the context of record values .
some well - known members of the above families are : \1 .
exponential(@xmath40 ) , a member of with @xmath96 ; \2 .
rayleigh(@xmath40 ) , a member of with @xmath97 ; \3 .
beta(@xmath98 ) , a member of with @xmath99 ; \4 .
pareto(@xmath100 ) , a member of with @xmath101 , + and \5 .
burr(@xmath102 ) , a member of with @xmath103 . by making use of u - v method of robbins ( 1988 ) for the family
, we have the following lemma . [ lem2 ]
let @xmath0 be a sequence of independent random variables with survival function defined in ( [ phr ] ) .
let @xmath64 be a real - valued function such that for @xmath65 \(ii ) @xmath104 in which @xmath105 is the hazard function of @xmath95 and @xmath106 is the corresponding pdf of @xmath95 .
then the functions @xmath107 satisfy @xmath69 = e_{{\boldsymbol\theta}}\left [ \theta_j u_j(\mathbf{x})\right],~j=1,2,\cdots.\end{aligned}\ ] ] for one component problem ( i.e. , a single random variable @xmath108 , @xmath109 ) , let @xmath110 . then , we have @xmath111^{\theta_j^{-1}-1}g(x)\;{\rm d}x\\ & = & \int_{-\infty}^{+\infty } u(x )
\bar{f}_{\theta_j}(x ) h(x)\;{\rm d}x\\ & = & \int_{-\infty}^{+\infty }
u(x ) h(x)\left\ { \int_{x}^{+\infty } \;{\rm d}f_{\theta_j}(y ) \right\ } \;{\rm d}x \\ & = & \int_{-\infty}^{+\infty } \int_{-\infty}^{y } u(x ) h(x)\;dx\ ; { \rm d}f_{\theta_j}(y)=\int_{-\infty}^{+\infty } \nu(x){\rm d } f_{\theta_j}(x).\end{aligned}\ ] ] for the sequence @xmath16 , the result follows by a similar calculation .
@xmath76 the next result gives the unbiased estimator of @xmath11}$ ] , under sel function , for the general family .
assume @xmath95 to be known and let @xmath112 be the cumulative hazard function of @xmath95 .
for the general family , an unbiased estimator of @xmath11}$ ] , under the sel function , is @xmath113 from ( [ tet ] ) , ( [ ij ] ) and lemma 2 , an unbiased estimator of @xmath11}$ ] is given by @xmath114 @xmath76 similarly , for the family , an unbiased estimator for @xmath11}$ ] , under the sel function , is @xmath115 where @xmath116 is the cumulative reversed hazard function of the known cdf @xmath95 .
note that @xmath117 is a complete sufficient statistic for @xmath118 .
hence , the above unbiased estimators of @xmath11}$ ] are indeed umvu estimators of @xmath11}$ ] .
to compare the umvu estimator with other estimators , we need to compute the risk function of the proposed estimators . under the sel function , the risk of an estimator @xmath119 is @xmath120})={\rm e}(v^2)+{\rm e}(\theta_{[n]}^2)-2{\rm e}(v\theta_{[n]}).\ ] ] the umvu estimators obtained in section 3 are functions of @xmath121 .
suppose we want to estimate the risk of an estimator of @xmath11}$ ] which depend on @xmath30 only through @xmath21 and @xmath122 , i.e. @xmath123 .
then , we have the following results , under models 1 and 2 , respectively . [ risk ] under the model 1 and the sel function ,
an unbiased estimator of the risk of an estimator @xmath124 of @xmath11}$ ] is @xmath125 from lemma 1 with @xmath80 , we have @xmath126}^2)&=\sum_{j = n}^{\infty}\theta_j^2\e(i_j(\mathbf{y}))=\sum_{j = n}^{\infty}\theta_j\e\left[\nu_j(\mathbf{y})\right]\\ & = \sum_{j = n}^{\infty}\e\left[\nu_j^*(\mathbf{y})\right],\end{aligned}\ ] ] where @xmath127 therefore @xmath126}^2)&=\e\left[\sum_{j = n}^{\infty}\frac { i_j(\mathbf{y})}{y_j^{p-1}}\int_{u_{n-1}^s}^{y_j}\int_{u_{n-1}^s}^{s}t^{p-1}\;{\rm d}t\;{\rm d}s\right]\\ & = \e\left[\frac{1}{\left(u_{n}^s\right)^{p-1}}\int_{u_{n-1}^s}^{u_n^s}\int_{u_{n-1}^s}^{s}t^{p-1}\;{\rm d}t\;{\rm d}s\right]\\ & = \e\left[\frac{\left(u_{n}^s\right)^{p+1}-\left(u_{n-1}^s\right)^{p+1}-(p+1)\left(u_{n-1}^s\right)^p(u_n^s - u_{n-1}^s)}{p(p+1)\left(u_{n}^s\right)^{p-1}}\right].\end{aligned}\ ] ] furthermore @xmath126}v(u_{n}^s , u_{n-1}^s))&=\sum_{j = n}^{\infty}\theta_j\e(i_j(\mathbf{y})v(y_j , u^s_{n-1}))\\ & = \sum_{j = n}^{\infty}\e\left[\frac{1}{y_j^{p-1}}\int_{0}^{y_j}t^{p-1}v(t , u^s_{n-1})\right.\\ & \qquad\qquad\times\left.i_j(y_1,\ldots , y_{j-1},t , y_{j+1},\ldots)\;{\rm d}t\right]\\ & = \e\left[\frac{1}{(u^s_n)^{p-1}}\int_{u_{n-1}^s}^{u_n^s}t^{p-1}v(t , u_{n-1}^s)\;{\rm d}t\right].\end{aligned}\ ] ] which completes the proof .
@xmath76 an immediate corollary of theorem [ risk ] is as follows . under the model 1 and the sel function ,
an unbiased estimator of the risk of @xmath128 is @xmath129 [ risk2 ] for the general family , and under the sel function , an unbiased estimator of the risk of an estimator @xmath123 of @xmath11}$ ] is @xmath130 from lemma 2 and using similar argument as in the proof of theorem [ risk ] , we have @xmath126}^2)&=\sum_{j = n}^{\infty}\theta_j^2\e(i_j(\mathbf(x)))\\ & = \sum_{j = n}^{\infty}\theta_j\e\left[\int_{-\infty}^{x_j}h(t)i_j(x_1,\ldots , x_{j-1},t , x_{j+1},\ldots)\;{\rm d}t\right]\\ & = \sum_{j = n}^{\infty}\e\left[\int_{-\infty}^{x_j}h(s)\int_{-\infty}^{s}h(t)i_j(x_1,\ldots , x_{j-1},t , x_{j+1},\ldots)\;{\rm d}t\;{\rm d}s\right]\\ & = \e\left[\int_{u_{n-1}}^{u_n}h(s)\int_{u_{n-1}}^{s}h(t)\;{\rm d}t\;{\rm d}s\right]\\ & = \e\left[\frac{h^2(u_n)-h^2(u_{n-1})}{2}-h(u_{n-1})(h(u_n)-h(u_{n-1}))\right]\\ & = \e\left[\frac{(h(u_n)-h(u_{n-1}))^2}{2}\right].\end{aligned}\ ] ] furthermore @xmath126}v(u_{n},u_{n-1}))&=\sum_{j = n}^{\infty}\theta_j\e(i_j(\mathbf{x})v(x_j , u_{n-1}))\\ & = \sum_{j = n}^{\infty}\e\left(\int_{0}^{x_j}h(t)v(t , u_{n-1})\right.\\ & \qquad\times\left.i_j(x_1,\ldots , x_{j-1},t , x_{j+1},\ldots)\;{\rm d}t\right)\\ & = \e\left(\int_{u_{n-1}}^{u_n}h(t)v(t , u_{n-1})\;{\rm d}t\right).\end{aligned}\ ] ] this completes the proof . @xmath76
an immediate corollary of theorem [ risk2 ] is as follows .
[ c2]for the general family and under the sel function , + ( i ) an unbiased estimator of the risk of @xmath131 is @xmath132 ( ii ) the risk of @xmath133 is @xmath134})=\e(\theta_{[n]}^2).\ ] ] the results for the general family can be obtained by replacing @xmath135 with @xmath136 in theorem [ risk2 ] and corollary [ c2 ] . since @xmath117 is a complete sufficient statistic for @xmath118 , the above unbiased estimators of @xmath137})$ ] are indeed , umvu estimators of @xmath137})$ ] .
the following result presents the distribution of the unbiased estimator in the family .
[ expnons ] in the general family , the following identities hold : \(i ) for every @xmath20 and @xmath138 , @xmath139 \(ii ) for every @xmath140 , @xmath141 and @xmath142 , @xmath143 let @xmath144 and @xmath145 , @xmath146 .
we only prove part ( i ) . part ( ii )
is proved in a simillar way .
using the fact that @xmath147 and the lack of memory property of the exponential distribution , @xmath148 which is the required result.@xmath76
for the general family with pdf , we have @xmath149 thus , a natural estimator for @xmath11}$ ] , for this family of distributions is @xmath150 . for the general family with the survival function
, we have @xmath151 which candidates @xmath152 as a natural estimator of @xmath11}$ ] .
so a risk comparison of the natural estimators with umvues of @xmath11}$ ] , for both families of distributions is considered .
the following corollary of theorem [ risk2 ] states that , under model 2 , the umvue dominates the natural estimator . for the general family and under the sel function
, we have @xmath153})>\r(h(u_n)-h(u_{n-1}),\theta_{[n]}).\ ] ] first , we have @xmath154})&=\sum_{j = n}^{\infty}\theta_j\e(i_j(\mathbf{x})h(u_{n-1}))\\ & = \e\left ( h(u_{n-1})\sum_{j = n}^{\infty}\int_{0}^{x_j}h(t)i_j(x_1,\ldots , x_{j-1},t , x_{j+1},\ldots)\;{\rm d}t\right)\\ & = \e\left ( h(u_{n-1})\int_{u_{n-1}}^{u_n}h(t)\;{\rm d}t\right)\\ & = \e(h(u_{n-1})(h(u_n)-h(u_{n-1}))).\end{aligned}\ ] ] consequently , @xmath155})-\r(h(u_n)-h(u_{n-1}),\theta_{[n]})\\ & \hspace{-2cm}=2\e(h(u_n)h(u_{n-1}))-\e(h^2(u_{n-1}))\\ & \hspace{-2cm}\quad-2\e(h(u_{n-1})\theta_{[n]})\\ & \hspace{-2cm}=2\e(h(u_n)h(u_{n-1}))-\e(h^2(u_{n-1}))\\ & \hspace{-2cm}\quad-2\e(h(u_{n-1})(h(u_n)-h(u_{n-1})))\\ & \hspace{-2cm}=\e\left(h^2(u_{n-1})\right)>0.\end{aligned}\ ] ] this completes the proof.@xmath76 however , under model 1 , no explicit results can be obtained for domination of the umvue or the natural estimator with respect to the other , since we have similarly @xmath156})-\r(u_n^s / p,\theta_{[n]})\\ & \hspace{-5cm}=\e\left(\frac{\left(u_{n-1}^s\right)^{2p}-2\left(u_{n-1}^s\right)^p\left(u_{n}^s\right)^p+ 2p\left(u_{n}^s\right)^{p-1}\left(u_{n-1}^s\right)^{p}(u_n^s - u_{n-1}^s)}{p^2\left(u_n^s\right)^{2p-2}}\right).\end{aligned}\ ] ] to compare the umvue and the natural estimator under model 1 , we run a simulation study , which is described in the following section .
we assume @xmath157 . to compare the risks of the umvue @xmath158}^1=\frac{u_n}{p}\left(1-\left(\frac{u_{n-1}}{u_n}\right)^p\right)$ ] , with that of the natural estimator @xmath158}^2=\frac{u_n}{p}$ ] , for @xmath159 , @xmath160 , we consider three different models for the sequence of parameters as follows : model 1 ( an stochastic , positive error auto - regressive model ) : @xmath161 model 2 ( an stochastic geometrically increasing population ) : @xmath162 model 3 ( white noise model ) : @xmath163 the simulated bias and risks of the estimators are tabulated in table [ sim ] . as one can observe from table [ sim ] , the simulated risks of @xmath164}}^1 $ ] are less than those of @xmath164}}^2 $ ] .
also , biases and risks are increasing in @xmath165 , except the risks of @xmath164}}^1 $ ] , under the white noise model 3 .
.simulated bias and risk of the umvue and the natural estimator of @xmath11}$ ] under three different models from gamma distribution for different values of @xmath165 and @xmath166.[sim ] [ cols="^,^,^,^,^,^",options="header " , ] to generate random variables identically distributed as @xmath167 , one may generate an iid sample form standard exponential , namely , @xmath168 , and return @xmath169 .
table [ tct ] presents the simulated values of @xmath170-critical values of @xmath167 , @xmath171 , for @xmath172 , and @xmath173 , which are generated using r.14.1 package with @xmath174 iterations .
the hypothesis @xmath175 is rejected at level @xmath170 as @xmath176 for the rainfall data we obtain @xmath177 , which is less than @xmath178 .
therefore , @xmath175 is not rejected in favor of @xmath179 at level @xmath180 .
the problem of estimating parameters of the dynamically selected populations can be extended to the bayesian context .
moreover , the problem of unbiased estimation of the selected parameters under other loss functions is of interest .
the distributional models which are not members of studied families can be studied separately , specially the discrete distribution .
another problem is to find the two stage ( conditionally ) unbiased estimators of the parameters of the dynamically selected populations .
these problems are treated in an upcoming work , to appear in subsequent papers .
the authors thank the anonymous referee for his / her useful comments and suggestions on an earlier version of this manuscript which resulted in this improved version .
doostparast , m. and emadi m. ( 2013 ) .
evidential inference and optimal sample size determination on the basis of record values and record times under random sampling scheme , _ statistical methods & applications _ , doi:10.1007/s10260 - 012 - 0228-x .
misra n. , vander meulen e.c . and branden k.v .
( 2006a ) . on estimating the scale parameter of the selected gamma population under the scale invariant squared error loss function .
_ journal of computational and applied mathematics _ ,
* 186 * , 268 - 282 .
misra n. vander meulen e.c . and brandan k.v .
( 2006b ) . on some inadmissibility results for the scale parameters of selected gamma populations . _ journal of statistical planning and inference _ , * 136 * , 2340 - 2351 .
nematollahi , n. and motammed - shariati , f. ( 2012 ) .
estimation of the parameter of the selected uniform population under the entropy loss function , _ journal of statistical planning and inference _ , * 142 * , 2190 2202 .
salehi m. ahmadi j. and balakrishnan , n. ( 2013 ) .
prediction of order statistics and record values based on ordered ranked set sampling , _ journal of statistical computation and simulation _ , doi = 10.1080/00949655.2013.803194 .
sill , m. w. and sampson , a. r. ( 2007 ) .
extension of a two - stage conditionally unbiased estimator of the selected population to the bivariate normal case , _ communications in statistics - theory and methods _ , * 36 * , 801 813 . | we introduce the problem of estimation of the parameters of a dynamically selected population in an infinite sequence of random variables and provide its application in the statistical inference based on record values from a non - stationary scheme .
we develop unbiased estimation of the parameters of the dynamically selected population and evaluate the risk of the estimators .
we provide comparisons with natural estimators and obtain asymptotic results .
finally , we illustrate the applicability of the results using real data .
* keywords : * extreme value theory , general record models , partial maxima , pfeifer model , selected population , uniformly minimum variance unbiased estimator . |
in recent years , interest in studying graphene has increased steeply due to its exceptional electronic properties and potentially significant applications in nanotechnology .
graphene , a single layer of carbon atoms arranged in a hexagonal honeycomb lattices , is the thinnest known material in the universe and the strongest ever measured @xcite . in the low energy approximation ,
the tight - binding description of the electron transport in graphene is reduced to the massless dirac - weyl equation with fermi velocity @xmath3 .
the relatively low fermi velocity and the gapless linear energy dispersion near the dirac points @xmath4 and @xmath5 , enabled one to simulate relativistic effects like anomalous landau - hall effect @xcite , the zitterbewegung and the klein paradox @xcite , in condensed matter systems .
the optical - like behavior of electron waves in graphene - based waveguide structure has also been experimentally investigated @xcite .
these phenomena are expected to play a needful role in the future fabrication of desirable nanoelectronic devices .
since dirac electrons can not be confined by electrostatic potentials due to klein s paradox , it was suggested that magnetic confinement should be considered @xcite .
inhomogeneous magnetic fields on the nanometer scale can suppress the klein tunneling and produce graphene based structures to confine dirac fermions , but the parameters of the magnetic barriers can not be changed fast enough .
electron confinement in the single @xcite , double @xcite , and multiple @xcite square - well magnetic barriers and in delta function magnetic barriers @xcite have been investigated .
heaviside - step - function - like magnetic profile is used in @xcite to study the magnetic induced waveguide structure . and
it is shown that the guided modes are strongly confined by this gradually varied magnetic field .
recently a series of studies concerning the interaction of graphene electrons moving in magnetic fields perpendicular to the graphene surface @xcite and/or including electrostatic fields parallel to the surface @xcite have been carried out in order to find a way for confining the charges . in all these works ,
the dirac - weyl equation is considered , where a minimal coupling with the vector potential describes the interaction with the external field . in general ,
some kinds of numerical computation or semiclassical methods were needed to find the energy levels of confined states or transmission coefficients for scattering states under arbitrary magnetic fields .
however , attempts have been made in @xcite to investigate the same by analytical method like supersymmetric quantum mechanics ( susy - qm ) @xcite .
this was possible due to the fact that the hamiltonian of graphene in a background static magnetic field can be mathematically cast into one with quantum mechanical factorization properties used in susy - qm .
it is shown that except for the ground state , the hamiltonians describing the dirac fermions in the two adjacent sites in the unit cell of the graphene lattice , have the same spectrum . using the factorization technique ,
the exact expressions of discrete spectra and wave functions of graphene dirac electron in various magnetic fields having translational symmetry are obtained . in all these cases the effective partner hamiltonians are essentially shape invariant @xcite .
it is worth mentioning here that susy - qm has been applied as well to solvable models in the low energy approximation of carbon nanotubes @xcite and spherical molecules including fullerenes @xcite . in ref .
@xcite , the effect of homogeneous transverse magnetic fields on the energy spectrum of a carbon nanotube is explained .
whereas , the construction of new solvable systems within the susy framework is considered in @xcite where the results are relevant not only for carbon nanotubes but also for graphene with potential barriers that possess translational invariance in one direction . a class of exactly solvable metallic and maximally semiconducting nanotubes in the presence of an external inhomogeneous magnetic field is shown to possess nontrivial integrals of motion , which establishes the @xmath6 nonlinear supersymmetry @xcite .
the nonlinear hidden supersymmetry generated by local supercharges is also shown to appear naturally in the finite - gap and time - reversal invariant configurations of twisted nanotubes @xcite . in this paper
, we will obtain new one parameter families of inhomogeneous external magnetic fields @xmath1 for which the dirac electron possesses exact solutions .
the inhomogeneity appearing in @xmath1 can be controlled by the parameter @xmath2 . to this aim
we revisit supersymmetric factorization approach to dirac - weyl equation and precisely exploit the nonuniqueness property of the superpotential ( which is directly related to the vector potential , see section [ s2 ] ) .
the one parameter family of magnetic fields so obtained are shown to be the deformed variants of some well known physically interesting magnetic fields .
the corresponding effective hamiltonians are no longer shape invariant and have altogether different functional forms with interesting features .
the effect of these deformations on the energy spectrum and the probability current density associated with the lowest energy states of the dirac electron shall also be investigated . for illustration purposes
we will consider two examples : the dirac electron being placed first in a harmonic well associated to a uniform magnetic field , and then in a morse oscillator characterizing an exponentially decaying magnetic field .
it is shown that the deformation of a uniform magnetic field enables one to generate the gradually varying magnetic fields which were achieved in @xcite earlier with the help of heaviside - step function .
the paper is organized as follows . in section
[ s2 ] , a brief overview of the quantum mechanical treatment of the dirac - weyl equation for dirac fermion and its connection to first order susy - qm will be done .
next in section [ s3 ] , the most general form of the magnetic fields , bearing exactly solvable effective potentials , are going to be derived .
section [ s4 ] deals with two concrete examples : deformation of uniform and exponentially decaying magnetic fields , respectively . summary and future prospects concerning the possible experimental implication of the work are stated in section [ s5 ] .
we consider dirac electron moving in graphene under static orbital magnetic field acting perpendicular to the graphene @xmath7 plane .
moreover , we disregard the electronic spin degree of freedom and the presence of any disorder in the system . under these circumstances the low energy excitation of the electron around a dirac point in the brillouin zone is described by the following dirac - weyl equation @xcite @xmath8 \psi(x , y ) = e \psi(x , y ) , \label{e2}\ ] ] where the fermi velocity @xmath9 , @xmath10 are the pauli matrices , and @xmath11 is the two dimensional momentum operator . here
@xmath12 is the charge of the electron and the vector potential @xmath13 is related to a slowly varying magnetic field by @xmath14 there is a lot of freedom in choosing @xmath13 and a convenient choice is landau gauge @xmath15 .
now , taking into account the translational invariance along ( say ) @xmath16 direction we assume the two component spinor @xmath17 in the form @xmath18,\label{e11}\ ] ] @xmath19 being the wave number in the @xmath16 direction and @xmath20 describe the electron amplitude on two adjacent sites in the unit cell of graphene .
consequently , the dirac - weyl equation ( [ e2 ] ) yields two coupled first order linear differential equations @xmath21 which readily can be decoupled to get two non - relativistic schrdinger equations @xmath22 , where @xmath23 @xmath24 hence , the determination of a complete solution of the dirac - weyl equation is now reduced to the problem of solving the above two schrdinger equations .
indeed , the very form of the hamiltonians @xmath25 in equation ( [ e4 ] ) indicates that they can be embraced by the formalism of susy - qm if one defines the intertwining operators @xmath26 where @xmath27 is the superpotential .
now , for the two operators @xmath28 , the equations in ( [ e3 ] ) reduce to @xmath29 respectively and the two hamiltonians @xmath25 become factorized in the following way @xmath30 equations ( [ e15 ] ) and ( [ e16 ] ) imply that @xmath25 are supersymmetric partners connected by the following intertwining relationships @xmath31 .
this means that knowing all the eigenfunctions of @xmath32 one can determine the ones of @xmath33 using the operator @xmath34 , and vice versa , using @xmath35 one can construct all the eigenfunctions of @xmath32 from those of @xmath33 except possibly for the ground state .
specifically , in standard unbroken first order supersymmetric quantum mechanics the eigenvalues and eigenfunctions of the two partner hamiltonians are related by @xcite : @xmath36 at this point it should be mentioned that the superpotential @xmath37 is related directly to the vector potential [ see equation ( [ e40 ] ) ] and hence to the magnetic field by the relation @xmath38 .
thus , for a particular choice of magnetic field the problem of determining the energy spectrum and the corresponding spinor wave functions of the dirac electron reduces to solving the schrdinger eigenvalue equation for either @xmath32 or @xmath33 .
and once we have the complete solutions of @xmath32 or @xmath33 , we can obtain the solution of the dirac electron by using equations ( [ e11 ] ) , ( [ e10 ] ) and ( [ e7 ] ) .
so far , this technique has been implemented in order to solve the dirac - weyl equation under several magnetic fields which give rise to the so - called shape invariant effective potentials @xmath39 . in the next section , departing from these cases we are going to generate new inhomogeneous magnetic fields for which the corresponding effective potentials are no longer shape invariant yet the dirac - weyl equation is exactly solvable .
here we will obtain the most general form of the magnetic field departing from a given initial one @xmath40 , keeping intact the exact solvability of the dirac - weyl equation .
to this end , we introduce a generalized superpotential for the usual quantities and @xmath41 for those which correspond to the deformation . ]
@xmath42 which satisfies now a modified version of the riccati equation ( [ e16 ] ) @xmath43 and whose solutions are going to generate new deformed magnetic fields @xmath44
so that the corresponding dirac - weyl equation is once again exactly solvable . besides , as it will be shown below , the effective potential @xmath45 becomes now a one ( or two ) parameter family of potentials @xmath46 which can be considered as deformations of @xmath47 . in general , for arbitrary @xmath48 the riccati equation ( [ e12 ] ) is not always easy to solve to find the unknown function @xmath49 . in the following we first will consider @xmath50 ; the case of @xmath51 will be discussed subsequently .
+ @xmath52 * @xmath50 ( one parameter family of magnetic fields ) : * for @xmath50 the equation ( [ e12 ] ) reduces to a bernoulli differential equation whose solution is @xmath53^{-1}$ ] and hence @xmath54 where @xmath2 is an arbitrary real constant and @xmath55 is the ground state eigenfunction of @xmath45 .
the partner potentials corresponding to the generalized superpotential @xmath56 become now @xmath57 , \end{array}\label{e22}\ ] ] and the corresponding one parameter family of magnetic fields @xmath1 can be obtained using equation ( [ e14 ] ) @xmath58 . \label{e31}\ ] ] several remarks are in order at this stage : * the family of magnetic fields @xmath1 can be considered as a deformation of the usual magnetic field @xmath40 by an amount proportional to @xmath59 .
this infinite family of deformed magnetic fields generates an infinite family of effective partner potentials , @xmath60 and @xmath61 , which in general are not longer shape invariant yet exactly solvable .
the shape invariance property is retained only in the limit @xmath62 . in this limit both the effective potential and the magnetic field reduce to the undeformed ones , which shows that @xmath45 and @xmath40 themselves are members of @xmath1 and @xmath61 , respectively .
* @xmath1 is in general inhomogeneous in nature , where the amount of additional inhomogeneity can be controlled by the parameter @xmath2 .
most interestingly , if the initially given magnetic field @xmath40 is uniform , the deformed magnetic fields @xmath1 become inhomogeneous .
this case will be discussed in detail in section [ s4.1 ] . * although @xmath2 is an arbitrary real number , one has to avoid those values of @xmath2 for which both the magnetic field @xmath63 and the potential @xmath46 become unphysical ( singular ) . since the values of the integral @xmath64 lie between @xmath65 and @xmath66 , the only acceptable values are @xmath67 or @xmath68 . * all the potentials
@xmath61 have the same partner @xmath60 , so they are isospectral to @xmath45 , that is , @xmath69 .
the bound state wave functions @xmath70 are obtained from @xmath71 , and vice versa , using relations ( [ e7 ] ) with the modified intertwining operators @xmath72 . specifically , @xmath73}{\mathcal{e}_{n+1}^-[\lambda+\int_{-\infty}^{x } \psi_0 ^ -(t)^2 dt ] } , \end{array}\label{e33}\ ] ] where @xmath74 = f_1 f_2'- f_2 f_1'$ ] is the wronskian of the two functions @xmath75 . * for @xmath76 and @xmath77 ,
the ground state wave function @xmath78 is no longer normalizable .
therefore , @xmath79 should be removed from the spectrum of @xmath61 . as a result
, supersymmetry is broken in these two limit cases .
hence , for @xmath80 and @xmath81\}$ ] , the energy spectrum and spinor wave function of the dirac electron moving in graphene under the supersymmetry deformed magnetic field @xmath1 are given by @xmath82 , ~~ \widetilde{\psi}_{n+1 } = e^{i k y } \left[\begin{array}{c } { \psi}_{n}^+ \\
i ~ \widetilde{\psi}_{n+1}^- \end{array}\right ] .
\end{array}\label{e34}\ ] ] note that the energy spectrum is not altered by the deformation , and it contains both positive and negative eigenvalues which correspond to the electron and hole respectively , in a symmetric way . to probe the effect of the deformation on the nature of a massless dirac electron
, one should investigate the probability distribution @xmath83 and current density @xmath84 .
for a given state @xmath85 , they are determined by @xmath86 and @xmath87 respectively . since @xmath88
, the zero energy state does not carry any current irrespective of the magnetic field .
both the probability and current densities , which corresponds to stationary states , are independent of @xmath16 , due to the translational symmetry in this direction .
+ @xmath52 * @xmath51 ( two parameter family of magnetic fields ) : * the case with @xmath51 is more general and can be investigated in an analogous way as for @xmath89 .
but in that case one has to solve the general riccati equation ( [ e12 ] ) to evaluate @xmath49 .
this nonlinear equation can be linearized by assuming @xmath90 , where @xmath91 satisfies the following homogeneous second order linear differential equation @xmath92 the general solution of equation ( [ e17 ] ) is given by a linear combination of two linearly independent solutions @xmath93 as @xmath94 .
for all shape - invariant potentials , the equation ( [ e17 ] ) can be reduced to a hypergeometric or confluent hypergeometric differential equation .
that is , the two solutions @xmath93 can be expressed in terms of hypergeometric or confluent hypergeometric functions .
once the analytical form of @xmath91 is found , the partner potentials @xmath95 corresponding to the generalized superpotential , @xmath96 ( @xmath97 ) become now @xmath98 and the corresponding deformed magnetic field can be obtained using equation ( [ e14 ] ) .
both the effective potential @xmath46 and the magnetic field @xmath63 depend now on the parameter @xmath48 and the quotient @xmath99 or @xmath100 .
it is worth mentioning here that in order that the generalized superpotential , the deformed magnetic field and the effective potential @xmath101 are free from singularities , it is necessary that the function @xmath91 should be nodeless .
hence , we have to choose @xmath48 and the parameters appearing in @xmath91 such that @xmath91 becomes either strictly positive or strictly negative functions , which first of all implies that @xmath102 .
in addition , once a @xmath48 consistent with this restriction has been chosen , we will have to analyze further the asymptotic behavior of the function @xmath91 in order to select the nodeless one . the eigenfunctions and bound state energies of @xmath103
are easily obtained from those of @xmath60 , namely , @xmath104 on the other hand , the energies of @xmath46 are given by @xmath105 and the corresponding eigenfunctions @xmath106 are obtained from @xmath107 using relation ( [ e7 ] ) with the modified intertwining operator @xmath108 replacing the unmodified one . moreover ,
consistent with the fact that the state @xmath109 which is annihilated by @xmath110 could be either of square integrable or not , then the zero eigenvalue @xmath111 could be or not in the spectrum of @xmath112 respectively @xcite .
note that @xmath113 is a first - order differential equation with solution given by @xmath114 in contrast to the case with @xmath89 , here the energy spectrum of the dirac electron in the usual and deformed magnetic fields are not identical .
nonetheless , the square of the energy eigenvalues are related by @xmath115
it is worth mentioning that the mechanism discussed in the previous section is very general and can be applied to any magnetic field which initially yield exactly solvable shape - invariant effective potentials .
however , for illustration purposes here we will consider just two examples : the dirac electron being placed first in a harmonic well which feels a uniform magnetic field , and then in a morse oscillator experiencing an exponentially decaying magnetic field . since the analytical expressions of the deformed magnetic field , the corresponding effective potentials and the excited state wave functions are too involved for @xmath51 , we shall only consider the case with @xmath50 .
let us first choose a uniform magnetic field perpendicular to the graphene plane in the positive @xmath116 direction , @xmath117 , @xmath118 .
it can be generated by choosing the standard landau gauge @xmath119 . in this case the superpotential @xmath120 yields the shape - invariant effective potentials @xmath121 , respectively . here
@xmath122 whose dimension is ( length)@xmath123 .
the eigenvalues and eigenfunctions of these shifted harmonic oscillators turn out to be @xmath124 , \end{array}\label{e18}\ ] ] where @xmath125 and @xmath126 denotes the hermite polynomial of order @xmath127 .
hence , the complete solution of the corresponding dirac - weyl equation for the constant magnetic field can be found using equations ( [ e11 ] ) , ( [ e7 ] ) and ( [ e18 ] ) .
the above example is the well known case of equidistant landau levels @xcite .
now for @xmath89 , the deformation of the uniform magnetic field can be generated using the normalized ground state wave function @xmath128 and using equation ( [ e31 ] ) .
the explicit expression of the one parameter family of magnetic fields and effective potentials @xcite are obtained as @xmath129\right ) , \\
\widetilde{v}^-(x,\lambda ) = v^- - 2 \partial_{xx } \ln\left ( 2 \lambda + 1 + { \operatorname{erf}}\left[\sqrt{\frac{\omega}{2}}\left(x+\frac{2 k}{\omega}\right)\right]\right ) , \end{array}\label{e32}\ ] ] respectively . in order to avoid the singularity in both the magnetic field and the potential one has to choose either @xmath130 or @xmath68 . since the energy spectra of the potentials
@xmath46 remain the same as their undeformed version @xmath47 , therefore the landau levels still exist although the magnetic fields are no longer uniform . the bound state wave functions of the potential @xmath61 are obtained with the help of equation ( [ e33 ] ) @xmath131},\\ \widetilde{\psi}_{n+1}^- = n_{n+1 } e^{-\frac{\omega}{4}(x + \frac{2k}{\omega})^2 } \left [ h_{n+1}\left[\sqrt{\frac{\omega}{2}}\left(x + \frac{2k}{\omega}\right)\right ] + \frac{2 e^{-\frac{\omega}{2}(x + \frac{2k}{\omega})^2}}{\sqrt{\pi}\left(2 \lambda + 1 + { \operatorname{erf}}\left[\sqrt{\frac{\omega}{2}}\left(x+\frac{2 k}{\omega}\right)\right]\right ) } h_n\left[\sqrt{\frac{\omega}{2}}\left(x + \frac{2k}{\omega}\right)\right]\right ] .
\end{array } \label{e44}\ ] ] hence the complete solution , @xmath132 and @xmath133 , of the dirac electron moving under the deformed magnetic field ( [ e32 ] ) can be obtained using equation ( [ e34 ] ) .
interestingly , the energy spectrum @xmath134 together with @xmath135 , does not depend on the wave number @xmath19 .
in figures [ f1](a ) we have plotted the deformed magnetic fields for @xmath136 and @xmath137 .
the corresponding effective potentials @xmath60 and @xmath61 and ground state probability density are drawn in figures [ f1](b ) and [ f1](c ) respectively , while the current density @xmath138 is shown in figure [ f1](d ) . from these figures
it is clear that the larger the value of @xmath139 the smaller the effect of deformation . on the other hand , for a particular @xmath2
the deformation is strong for small @xmath140 , whereas the deformation is negligible for sufficiently large values of @xmath140 .
the position and value of the potential minima in @xmath46 are also different relative to those of @xmath47 .
this has a direct influence on the position and value of the maxima of the ground state probability distribution .
the changes of phase in the current density @xmath138 are also due to the deformed magnetic fields . from figure [ f1](a ) , it is also clear that the deformed magnetic fields are very similar to the gradually varying magnetic fields which were achieved earlier ( see figure 1(b ) of @xcite ) with the help of heaviside - step like function . in this case
we consider a magnetic field which is exponentially decaying in the positive @xmath140-direction @xcite , @xmath141 hence , the non - zero component of the vector potential is given by @xmath142 , which yields the superpotential @xmath143 and shape - invariant morse potential @xmath144 with @xmath145 .
the bound state solutions of these two potentials are given by @xcite @xmath146 where @xmath147 , @xmath148 , @xmath149 , and @xmath150 are the associated laguerre polynomials . here
the form of the one parameter family of deformed exponentially decaying magnetic fields and corresponding effective potentials is given by @xmath151,\\ \widetilde{v}^-(x,\lambda ) = v^-(x ) -2\partial_{xx } \ln \left [ \lambda ( 2k-1 ) !
+ \gamma(2k,2d e^{-x})\right ] , \end{array}\ ] ] where @xmath152 is the incomplete gamma function .
the normalized bound state wave functions of @xmath46 are given by : @xmath153}{(n+1)(2k - n-1)[\lambda ( 2k-1 ) !
+ \gamma(2k,2d e^{-x})]}\right ] .
\end{array}\ ] ] in figure [ f2](a ) we plot the deformed exponentially decaying magnetic fields for @xmath154 and @xmath137 . whereas the effect of these deformed magnetic fields on the nature of effective potentials , ground state probability density and on the first excited state current density are shown in figure [ f2](b),(c ) and ( d ) respectively .
to conclude , we have suggested a mechanism to find the most general form of a one parameter family of inhomogeneous magnetic field @xmath155 generated by applying the supersymmetric transformation to a given initial magnetic field @xmath40 .
it is shown that @xmath155 is a deformed variant of @xmath40 . the discrete spectrum and eigenfunctions of the corresponding dirac - weyl equation
are obtained analytically in closed form .
the mechanism is illustrated with the help of two examples : the dirac electron being placed first in a harmonic well associated to a uniform magnetic field , and then in a morse oscillator characterizing an exponentially decaying magnetic field .
the effects of the deformation on the probability and current densities are also investigated .
moreover , the construction of two parameter family of magnetic field is proposed .
since the aim of this work is to determine the most general form of the inhomogeneous magnetic fields for which the dirac - weyl equation can be solved exactly , a deeper investigation to uncover the new physics has been left for further studies .
however , as the inhomogeneity appearing in @xmath155 is controllable by the parameter @xmath2 , then the obtained magnetic fields can be used in graphene waveguide experiment of charge confinement .
note that the technology for doing this already exists as the profile of magnetic field can be generated experimentally by using ferromagnetic gates located beneath the substrate on which the graphene layer is deposited @xcite .
hence we expect that apart from their intrinsic mathematical interest , the results reported here will be useful for analyzing the dirac electron confinement by inhomogeneous magnetic fields .
bm thanks the physics department of cinvestav , mexico , for a visit and hospitality at the outset of this work
. he also acknowledges the postdoctoral grant from the belgian federal science policy office co - funded by the marie curie actions from the european commission .
df acknowledges the financial support of conacyt ( mxico ) , project 152574 .
a. kormanyos , p. rakyta , l. oroszlany , and j. cserti , phys .
b * 78 * , 045430 ( 2008 ) ; a. jellal and a.e .
mouhafid , j. phys .
a * 44 * , 015302 ( 2011 ) ; p. roy , t.k .
ghosh , k. bhattacharya , j. phys .
* 24 * , 055301 ( 2012 ) ; g. giavaras , p.a .
maksym and m. roy , j. phys .
* 21 * , 102201 ( 2009 ) ; s. hernandez - ortiz , g. murguia and a. raya , j. phys . :
cond . mat .
* 24 * , 015304 ( 2012 ) ; m.r .
masir , p. vasilopoulos and f.m .
peeters , j. phys .
* 23 * , 315301 ( 2011 ) ; r.r .
hartmann and m.e .
portnoi , phys .
rev . a * 89 * , 012101 ( 2014 ) .
m. ezawa , phys .
a * 372 * , 924 ( 2008 ) ; e. milpas , m. torres and g. murguia , j. phys .
: cond . mat .
* 23 * , 245304 ( 2011 ) ; m.r . setare and o. hatami , chin
. phys . lett . * 25 * , 3848 ( 2008 ) ; k.s .
park and k.s .
yi , j. kor .
. soc . * 50 * , 1678 ( 2007 ) . | we use supersymmetry transformations to obtain new one parameter family of inhomogeneous magnetic fields @xmath0 for which the massless dirac electron possesses exact solution .
the inhomogeneity appearing in @xmath1 can be controlled by the parameter @xmath2 .
the obtained magnetic fields are interpreted as deformed variants of some physically attainable well known magnetic fields .
a particular example , when a constant magnetic field is deformed , is considered to show that equidistant landau levels exist even in the presence of an infinite number of specially designed inhomogeneous magnetic fields . |
the dynamical masses of spiral galaxies are known to differ significantly from their visible masses .
the commonly accepted cause is the existence of a ellipsoidal halo of unseen matter in addition to the stars and gas .
the exact density distribution of these halos have become an increasingly important issue .
on one side , n - body simulations of the cosmological evolution of the cold dark matter ( cdm ) have now reached a sufficient resolution to predict the dark halo density profiles down to a scale corresponding to the inner parts of the spiral galaxies ( fukushige et al . 1997 ; navarro et al .
1997 ; moore et al . 1998 ; ghigna et al .
they almost always predict dense cuspy halos . on the other side , observations of dwarf
spiral galaxies show shallow inner rotation curves , compatible with a flat density core ( blais - ouellette et al .
2000 , hereafter paper ii ) .
the first step in showing the reality of this discrepancy is to eliminate the known possibilities of systematic observational biases .
two classes of errors could contribute to underestimate the velocities , hence the computed density , in the inner parts of spiral galaxies .
the prime culprit in radio observations is the `` beam smearing '' effect due to the relatively low angular resolution of 21 cm data with sufficient sensitivity to detect in the outer part of spiral galaxies .
combining the density gradient with the velocity gradient will lead to underestimate the velocity at a given radius .
h@xmath0observations always easily reach an angular resolution where any beam smearing effect can be neglected .
a less often commented source of uncertainties though is found in long slit observations , where most of h@xmath0data come from .
the lack of 2d coverage makes the alignment of the slit crucial to retrieve the real kinematics of a galaxy . missing the kinematical center , which is not always the photometric center ( paper ii ) , or just a few degrees between the slit and the galaxy position angle will also lead to an underestimation of the velocities .
inclination estimation , which has to be photometrically determined , is another major source of uncertainties .
in addition , the presence of a bar would hardly be noticed and its effect would most probably be confounded with the rotational kinematics . in order to hedge ourselves against these biases , we use fabry - perot high resolution h@xmath0spectroscopy combined with published radio synthesis observations to study the mass distribution of 8 dwarf and spiral galaxies . in paper ii , with a smaller sample , we focused on modeling the mass distribution using different shape of dark matter halos . here ,
in addition , we address the more precise question of what inner density slope ( @xmath5 ) halo can have for a given galaxy type or mass . in section 2 , we first briefly review the mass modeling used in the study . then , in section 3 , we look at detailed mass models of a few galaxies . in section 4 ,
the relation between @xmath5 and galaxy mass is discussed followed by concluding remarks .
to investigate in details the mass distribution of dark matter halos without a few assumptions on the matter content of spiral galaxies , one would have to adjust a good dozen of parameters .
first , the luminous matter distribution depends on the disk and bulge mass - to - light ratios ( and their radial gradient ) , and on the bulge - to - disk ratio .
the contribution have to be corrected for helium fraction .
the dark halos can be non - spherical , in addition to the five parameters usually used to describe the radial density distribution .
this general distribution function can be expressed as ( see paper ii for details ) : @xmath6 where @xmath7 and @xmath8 are characteristic density and radius , where c , included for ease of comparison with other works , can force the presence of a flat density core , where @xmath0 and @xmath5 are respectively the inverse outer and inner logarithmic slopes , and @xmath9 the transition parameter .
one could add the distance that is used to calculate the light distribution , and a central mass which is suspected to exist in most spirals .
for some of these constraints , a fixed value is well accepted .
the mild or absent color gradient in spiral galaxies lead to a radially constant mass - to - light ratio .
the helium fraction can be approximate to its primordial abundance .
distances are hopefully well constrained by independent means .
otherwise , essentially three data sets can be used to constraint these parameters : luminosity distributions ( visible an ) and velocity field or , assuming axisymmetry , light profiles and rotation curve . the visible light profile ,
is used to determine the bulge to disk ratio .
the other parameters are all to the charge of the rotation curve . most of the luminous contributions are heavily constrained by the most inner parts of the curve leading to a possible degeneracy between central mass , bulge , inner disk and halo contributions .
that is why , many studies including this one , tend to focus on dwarf galaxies where bulges are negligible and large central masses excluded by the rotation curve . only there ,
can one put strong limits on the inner slope of dark halo density distribution . for these galaxies ,
luminous matter is dependent only on the disk mass - to - light ratio , and dark matter from the central density , the core radius , and the three shape parameters . from the latter , only @xmath5 ,
the inner slope , has a significant impact on halo shape at the scale of dwarf galaxies , @xmath0 and @xmath9 being very poorly constrained . in our procedure , the visible light profile ( eventually decomposed in its bulge and disk components )
is inverted in a mass profile scaled by the mass - to - light ratios .
the contribution is estimated from its distribution and multiply by 1.33 to account for helium .
the halo shape is , in general , fixed to a known profile .
we choose four profiles , widely used in the literature , that span , with some redundancy , through most possibilities . the cuspy profile from navarro et al .
( 1997 ) is a canonical example of cdm simulation prediction , the pseudo - isothermal sphere , the profile from kravtsov et al .
( 1998 ) , and burkert s profile ( burkert et al .
1995 ) are all cusp - less profile with different shape parameters .
rotational velocities from all the contributions are then added in quadrature and compare to the rotation curve using a standard @xmath10 minimization .
beside the dwarf galaxies ngc 3109 and ic 2574 from paper ii , the whole sample includes 3 fairly bulge - less spirals ( ugc 2259 , ngc 2403 , and ngc 6946 ) , and 3 earlier type spirals ( ngc 5055 , ngc 2841 , and ngc 5985 ) from paper iii .
details are given in table [ tab : sample ] .
.parameters of the sample . from the main references unless otherwise specified .
[ cols="<,<,^,^,^,^,^,^,^,^,^ " , ] complete modeling , using the four models , of the most relevant cases are presented here .
ngc 3109 and ic 2574 ( from paper ii ) are re - analyze to correct for a numerical problem which underestimated the central density .
they are good examples of cdm incompatible dwarfs
. are also modeled , ngc 2403 , as an example of bulge - less small spiral compatible with any model , and ngc 5055 , as an earlier type spiral with a bulge .
analysis of the whole sample will be presented in a future paper ( paper iv ) .
error bars are the quadratic sum of : half the velocity difference between receding and approaching sides , half the correction for asymmetric drift , error cause by uncertainty on inclination , and statistical error(@xmath11 ) .
+ + l l l c c c c c model & galaxy & type & ( /l@xmath12)@xmath13bulge & ( /l@xmath12)@xmath13disk & r@xmath14 & @xmath7 & @xmath10 + & & & & & kpc & 10@xmath15@xmath16/pc@xmath15 & + ' '' '' iso & ngc 3109 & sbm & no bulge & 0.1 & 2.4 & 2.4 & 0.44 + & ic 2574 & sabm & no bulge & 0.3 & 5.0 & 0.75 & 2.7 + & ngc 2403 & sabcd & no bulge & 2.5 & 4.8 & 1.7 & 1.9 + & ngc 5055 & sabbc & 2.3 & 0.8 & 8.2 & 3.4 & 0.55 + ' '' '' burkert & ngc 3109 & & no bulge & 0.4 & 3.0 & 1.8 & 0.22 + & ic 2574 & & no bulge & 0.3 & 8.9 & 0.8 & 2.5 + & ngc 2403 & & no bulge & 2.5 & 8.0 & 1.9 & 1.8 + & ngc 5055 & & 3.2 & 1.6 & 8.2 & 2.8 & .56 + ' '' '' kkbp & ngc 3109 & & no bulge & 0.0 & 4.0 & 1.7 & 0.5 + & ic 2574 & & no bulge & 0.3 & 10.5 & 0.46 & 2.5 + & ngc 2403 & & no bulge & 2.5 & 8.5 & 1.2 & 1.8 + & ngc 5055 & & 3.1 & 2.0 & 11.2 & 1.0 & .56 + ' '' '' nfw & ngc 3109 & & no bulge & 0.0 & 605 & 0.0035 & 9.0 + & ic 2574 & & no bulge & 0.0 & 295 & 0.005 & 32 + & ngc 2403 & & no bulge & 1.4 & 11.8 & 0.9 & 1.4 + & ngc 5055 & & 2.6 & 1.1 & 10.8 & 1.8 & .52 + + + it is rather clear that nfw profile is incompatible with the two dwarf galaxies .
these results are in line with most similar studies using ( e.g. ct et al . 2000 ) or long slit observations ( e.g. swaters et al . 2000 ,
de blok et al . 2001 ) but without the related uncertainties ( van den bosch et al .
it has to be noticed that higher resolution n - body simulations tend to give even steeper inner density slopes .
the present discrepancy is therefore genuine and probably involve subtle phenomena or new physics .
mass - to - light ratios tend to be unrealistically low for the smallest galaxies .
halo contributions are therefore an upper limit and are in fact probably shallower .
the range of possible observational biases that could explain the core - cusp problem is now significantly reduced , rather close to nil . the question of the physical cause of this shallow inner density distribution
have been addressed many times leading to proposals ranging from multi - component dark matter ( burkert & silk 1997 ) to self - interacting dark matter ( spergel & steinhardt ) and baryonic feedback processes .
these scenarios are though to behave differently under different gravitational potentials .
for example , feedback processes can hardly have an impact in galaxies with very deep potential .
plotting @xmath5 , the density profile inner slope , against the asymptotic rotational velocity would precisely show the evolution of this behavior .
figure [ fig : gammax ] plots @xmath5 against the observe maximum velocity which , for ngc 3109 and ic 2574 , is lower than the true asymptotic velocity .
the steepness of ngc 5055 rotation curve leads to an important degeneracy between @xmath5 , r@xmath14 and @xmath17 .
it is therefore not clear from the plot if there is a canonic value of @xmath5 around 1.2 that breaks down below 100 km / s or if @xmath5 is intrinsically increasing in massive galaxies .
we have modeled 8 dwarf and spiral galaxies using cusp - less and cuspy halos . the latter , predicted by cdm n - body simulations are clearly incompatible with dwarf galaxy kinematics .
more precisely , galaxies with less than 100 km / s of maximum rotation velocity have a inner density logarithmic slope ( @xmath5 ) of less than .7 as opposed to @xmath18 predicted by the n - body simulations . presently , @xmath0 , the outer density slope is poorly constrained by rotation curves .
one need to reach a radius where the luminous disk contribution is negligible while the rotation curve is flat or decreasing ( carignan & purton 1998 ) . in our sample ,
ugc 2259 is a good candidate .
future studies should in part extend to massive bulge - less galaxies to see if @xmath5 is intrinsically rising with galaxy mass or if it is constant , with a break down at low mass . the ghasp survey ( garrido & amram 2002 ) should allow this kind of study .
aaronson , m. & mould , j. 1983 , , 265 , 1 begeman , k 1987 , phd thesis , rijksuniversiteit groningen blais - ouellette , s. , carignan , c. & amram , p. , & ct , s. 1999 , 118 , 2123 blais - ouellette , s. , amram , p. & carignan , c. 2001 , 121 , 1952 burkert , a. 1995 , , 447 , l25 carignan , c. & freeman , k.c .
1988 , , 332 , l33 carignan , c. , charbonneau , p. , boulanger , f. & viallefond , f. 1990 , , 234 , 43 carignan , c. & purton , c. 1998 , , 506 , 125 c^ ot ' e , s. , carignan , c. , & freeman , k. c. 2000 , , 120 , 3027 ct , s. , carignan , c. & sancisi , r. 1991 , , 102 , 904 corradi , r.l.m . , boulesteix , j. , bosma , a. , capaccioli , m. , amram , p. & marcelin , m. 1991 , , 244 , 27 de blok , w. j. g. , mcgaugh , s. s. , bosma , a. & rubin , v. c. 2001 , , 552 , l23 freedman , w. l. 1990 , , 335 , l35 fukushige & t. , makino , j. 1997 , , 477 , l9 garrido , o. , amram , p. , this conference ghigna , s. , moore , b. , governato , f. , lake , g. , quinn , t. & stadel , j. 2000 , 544 , 616 jobin , m. & carignan , c. 1990 , , 100 , 648 kent , s. m. 1984 , , 56 , 105 kravtsov , a. v. , klypin , a. a. , bullock , j. s. & primack , j. r. 2000 , , 502 , 48 martimbeau , n. , carignan , c. & roy , j .-
1994 , , 107 , 543 moore , b. , governato , f. , quinn , t. , stadel , j. & lake , g. 1998 , , 498 , l5 musella , i. , piotto , g. & capaccioli , m. 1997 , , 114 , 976 navarro , j.f .
, frenk c.s . &
white , s.d.m .
1997 , , 490 , 493 swaters , r. a. , madore , b. f. & trewhella , m. 2000 , , 531 , l107 thornley , m. d. & mundy , l. g. 1997 , , 484 , 202 van den bosch , f. c. , robertson , b. e. , dalcanton , j. j. & de blok , w. j. g. 2000 , , 119 , 1579 | we use fabry - perot h@xmath0spectroscopy , complemented with published radio synthesis observations to derive high resolution rotation curves of spiral galaxies .
we investigate precisely their inner mass distribution and compare it to cdm simulations predictions .
having verified the existence of the so - called core - cusp problem , we find that the dark halo density inner slope is related to the galaxy masses .
dwarf galaxies with @xmath1 km / s have halo density inner slope @xmath2 while galaxies with @xmath3 km / s are best fitted by @xmath4 . |
true random numbers , or more precisely nondeterministic random number generators , seem to be of an ever increasing importance .
random numbers are essential in cryptography ( both classical and quantum ) , monte carlo numerical simulations and calculations , statistical research , randomized algorithms , lottery etc . + historically , there are two approaches to random number generation : algorithmic ( pseudorandom ) and by hardware ( nondeterministic ) .
pseudorandom number generators are well known in the art @xcite .
a pseudorandom generator is nothing more than a mathematical formula which produces deterministic , periodic sequence of numbers which is completely determined by the initial state called _ seed_. by definition such generators are not provably random . in contrast to that , hardware generators extract randomness form physical processes believed to behave in nondeterministic way which makes them better candidates for true random number generation .
+ in applications where provability is essential , randomness sources ( if involved ) must also be provably random . for example the famous bb84 quantum key distribution protocol described in @xcite
would be completely insecure if only an eavesdropper could calculate ( or predict ) either alice s random numbers or bob s random numbers or both . from analysis of the secret key rate
presented therein it is obvious that any guessability of random numbers by the eavesdropper would leak relevant information to him , thus diminishing the effective key rate .
it is intriguing ( and obvious ) that in the case that the eavesdropper could calculate the numbers exactly , the cryptographic potential of the bb84 protocol would be zero .
this example shows that the local random number generators assumed in bb84 are essential for its security and should not be taken for granted .
+ it has also been noticed that a random bit generator combined with an ordinary classical digital channel ( this combination being referred to as `` symmetric binary channel with noise '' ) gives raise to cryptographic potential of such a channel @xcite .
+ at present state of the art probably the best way to realize scientifically provable random number generator is to rely on intrinsic randomness of certain simple quantum systems .
one such system , frequently exploited for random number generation , is based upon a photon passing through a beam splitter , as schematically shown in fig .
[ splitter ] . + a light source emits photons which pass through a beam splitter
whenever a photon is emitted it takes one of the two possible paths ( with a small probability to get reflected or absorbed ) .
if it ends up in the detector d1 then we generate binary value `` 1 '' , whereas if it ends up in d2 we generate binary value `` 0 '' . because non - polarized light can be understood as an equal mixture of two orthogonal polarizations , there is a theoretical reason why zeros and ones would appear with the same probability in an idealized apparatus ,
that is why the bias ( defined as @xmath0 where @xmath1 is probability of ones ) would be zero .
other realizations of optical random quantum systems include : photons incident on a semi - transparent mirror @xcite , polarized photons incident on a rotateable polarizing beamsplitter which offers a possibility to tune the bias @xcite and an apparatus using polarization entangled photons to suppress false photon detections @xcite .
practical realizations of quantum random number generators , of course , necessarily suffer from imperfections .
primary imperfection of such systems is that the probability of ones is not exactly equal to the probability of zeros . in the systems mentioned above this
is firstly because photons which produce zeros and ones traverse different physical paths ( which might have different transmission probabilities ) , and secondly because photons are detected with two ( or more ) detectors .
detection efficiency of single a photon detector can vary significantly form one physical device to another and is typically very sensitive to temperature drift , supply voltage fluctuations , component tolerances and aging . in order to achieve a low bias one needs to cope with two problems : first how to achieve that the detectors have exactly the same efficiencies ( i.e. precisely such efficiencies as to compensate for different photon paths ) and second how to keep the detectors efficiencies extremely constant . to minimize the bias ,
quantum random number generators described so far must be fine tuned for a low bias prior to use .
this is not convenient because this fine tuning is a very time consuming procedure due to the statistical nature of measurement of bias .
furthermore , due to stability problems one can not expect that the bias once adjusted would stay constant for a long time .
+ the source of problems with previous art is that it uses spatial information of photons randomly ariving at different places as a source of randomness .
this necessarily requires two ( or more ) photon detectors . in our approach
we use temporal information of photons emmited at random times by a light emitting diode or a laser .
all photons traverse the same optical path and arrive always at the same single detector .
because now both zeros and ones are detected by the same detector , our random number generating method ( described below ) does not require any fine tuning of the bias .
furthermore the method is highly immune to stability problems because any slow change in the detector s performance influences by the same amount both zeros and ones and thus precisely cancels out .
the price for improved randomness quality is paid in a somewhat more complex bit extraction method , but still easily manageable by logic electronics circuits .
+ the random number generator presented here consists of two principal parts : a physical random pulse generator ( rpg ) and a method of extracting random bits from such a generator .
it will be shown that the method produces perfect random numbers ( bits ) if fed by perfectly random events .
one practical realization will be presented and statitically tested .
for the purpose of this paper we define a random bit generator as a device which produces bits independently of each other and with equal probability of the two outcomes , i.e. @xmath2 . we assume existence of a physical setup which produces poissonian random events , typically in form of digital pulses . by definition
, poisonian events are occuring independently of each other and in such a manner that time intervals between subsequent events follow the exponential probability distribution function ( pdf ) .
+ basic idea of the method for extracting random bits is to consider a pair of non overlapping random time intervals ( @xmath3 , @xmath4 ) which are defined with subsequent random events , as shown in fig .
[ restart]a , and generate either binary value `` 0 '' if @xmath5 , or `` 1 '' if @xmath6 .
next two intervals will be considered to generate the next random bit .
since the events which determine time intervals are by definition independent of each other it is not possible that @xmath5 would appear with any different probability than @xmath6 , consequently the probability to generate `` 0 '' is exactly equal to the probability to generate `` 1 '' . in other words ,
the distribution of @xmath7 is symmetric .
furthermore , the bits are mutually independent ( i.e. uncorrelated ) since independent pairs of events are used to generate different bits .
+ this `` basic algorithm '' readily ( and obviously ) yields perfect random bits if fed by perfectly random events .
however , it needs to be modified in order to cope with common limitations of physical realizations , notably detectors and electronics , and still give good random numbers .
+ first modification deals with the fact that measurement of time intervals usually results with an integer number which is produced , for example , by counting of a high frequency periodic signal ( eg . quartz clock ) or digizing a time to amplitude converter ( tac ) output voltage .
the time intervals @xmath8 are then represented by integer numbers @xmath9 . with this effect of `` time discretization ''
it is still true that @xmath10 but now it can happen that @xmath11 . in that case , in order to preserve equalness of probabilities of zeros and ones , the two intervals ( @xmath3 and @xmath4 ) are discarded and no bit is generated .
+ next modification is crucial and deals with the fact that the time discretization leads to appearance of correlations among bits .
the serial autocorrelation coefficient with lag @xmath12 of the sequence of bits @xmath13 is defined according to @xcite as : @xmath14 since different bits correspond to different random events we do not expect long - range correlations among bits .
therefore , we are mostly concerned with the lowest lag coefficient , @xmath15 , which for simplicity we henceforth denote with @xmath16 . + a technologically appealing way to measure ( approximately ) the time interval between subsequent random pulses is by counting periodic pulses from a continuous ( i.e. non - restartable ) quartz - controlled clock , as shown in fig . [ restart]b .
the time intervals @xmath8 are then represented by integer numbers @xmath9 . due to
the symmetry between generating zeros and ones the method ensures zero bias .
however since time intervals are only measured approximately it is not anymore obvious that bits obtained from such measurements will be random , as it was for the basic algorithm .
indeed , simulations show that a sequence of bits will have a non - negative autocorrelation ( fig . [
fig : anord ] ) which vanishes in the limits of a very fast clock or a very slow clock .
this behavior can be understood as follows . for non - restartable clock method ( dead time = 0 ) .
@xmath17 is the clock period , @xmath18 is the mean time interval between random pulses.,width=88 ] in the fast clock limit the time interval measurement becomes `` exact '' and this brings us back to the basic algorithm which provides unbiased and uncorrelated bits . + in the slow clock limit the probability of having more than one count during a random period becomes negligible .
therefore bit `` 0 '' becomes mainly represented by @xmath19 whereas `` 1 '' becomes mainly represented by @xmath20 .
in other words , in the limit , one bit is generated with every clock count and if the clock count occurs during an even random interval then `` 0 '' is produced , whereas if the clock count occurs during an odd random interval then `` 1 '' is produced . during a clock period of length @xmath17 , there are on average @xmath21 random intervals , where @xmath18 is the mean time interval between random pulses . according to the central limit theorem ,
the number of intervals becomes normally distributed as @xmath21 goes to infinity and since the normal distribution is symmetric even and odd periods ( and thus 0 s and 1 s ) appear with equal probability and at random .
+ apart from the correlation which characterizes the quality of randomness , another technologically important issue is the `` bit efficiency '' which we define as the number of bits generated per random event .
it reaches its maximum of 1/2 in the fast clock limit and continuously goes to zero at the slow clock limit ( fig .
[ fig : enord ] ) .
since good quality random events from the rpg are `` technologically expensive '' , fast clock limit is preferred for practical realizations of random number generators .
where @xmath17 is the clock period and @xmath18 is the mean time interval between random pulses.,width=88 ] since the fast clock limit could never be reached in practice , it is interesting to understand what happens in its vicinity .
if we look at a pair of subsequently generated bits ( for some @xmath22 ) , a process which involves measurements of 4 subsequent random time intervals , it turns out that this process slightly prefers pair values `` 11 '' and `` 00 '' ( symmetrically ) over the other two possibilities .
this has a direct impact on autocorrelation since eq .
( [ eq : auto ] ) for @xmath23 is equivalent to : @xmath24 where @xmath25 is probability of `` 11 '' etc . by simulation and symmetry arguments
we have found that following relations hold for probabilities of pairs : @xmath25=@xmath26 ; @xmath27=@xmath28 and @xmath29 for finite @xmath21 . acording to eq .
( [ altcorel ] ) this gives rise to a positive correlation .
understanding of this autocorrelation effect from combinatorics and basic laws of probability is involved , but for our purpose it suffices to understand that the correlation appears as a consequence of the way in which random time intervals @xmath30 are measured .
figure [ restart]b illustrates this problem .
if time interval @xmath3 ends up close to the beginning of a clock period it is more likely that @xmath31 will be smaller then in the case when @xmath3 ends up close to the end of a clock period .
it means that measurement of the time interval @xmath4 depends on the length of @xmath3 , that is @xmath32 and @xmath31 are somehow correlated .
+ by help of extensive simulations we have derived an empirical asymptotic expansion for autocorrelation in the fast clock limit : @xmath33 where @xmath34 .
this formula describes autocorrelation very well for @xmath35 .
for example , an autocorrelation of @xmath36 is achieved if the ratio of the clock frequency and mean frequency of random events is about 90 ( @xmath37 ) . for a very modest bit production speed of 2mhz
this already requires clock frequency of 360mhz ( bit efficiency @xmath38 ) which corresponds to fastest available logic circuits family .
the required minimum clock frequency scales linearly with the bit production speed and inverse of the square root of the desired autocorrelation which leaves no room for substantial improvement , either in the bit production rate or in correlation , with the present technology .
however autocorrelation can be completely removed by a different approach that we call `` restartable clock '' .
+ namely , it is our finding that in order to avoid the correlations completely it suffices to synchronize clock pulses with beginning of each random interval , as shown in fig .
[ restart]c . in this case ,
result of measurement of each time interval is only a function of its length and obviously it is not dependent on any previous measurements , which completely eliminates correlations .
because all time intervals are measured in the same way , it is again true that @xmath10 , meaning there is no bias either .
the quality of radomness is now completely independent of ratio @xmath21 .
this `` restartable clock method '' theoretically yields perfectly random bits if fed by perfectly random events .
+ the efficiency is again approximately described with the graph fig .
[ lecroy]b .
but now the departure from the maximum efficiency does not indicate any problem with the quality of randomness .
it simply means that there are too few clocks within an average interval and that because of that we often have @xmath11 in which case we must drop the whole pair of events .
good news is that with restartable clock method we can use clock of any speed : the only penalty for using a slow clock is efficiency lower than the maximum . + it is important to understand here that the dead time caused by the photon counting system does not introduce correlations ( nor bias ) provided that the dead time following each detected photon is constant in length i.e. independent of any previously detected photons
this can be understood as follows . in the basic algorithm
, the bit generating process makes use only of the difference of two time interval lengths . since each random time interval contains exactly one dead time ( namely the one following the first event ) the two dead times cancel out completely . as a result ,
the distribution of time interval differences @xmath39 is independent of the dead time .
the same cancellation argument works also for restartable clock because the dead time is always measured in exactly the same way and produces always the same number of counts in the counter .
however the cancelation does not work for the non - restartable clock and indeed in that case ( as we have checked by simulation ) a non - zero dead time introduces additional correlation .
the random pulse generator is made of a photon source followed by a single photon detector . as a source of photons
we have used a standard low - efficiency red light emitting diode ( led ) .
le diodes are direct bandgap devices which produce incoherent light by spontaneous emission @xcite which is essentially a random process . if operated at sufficiently low power , led emits photons which are virtually independent of each other , that is the photon emission is then a poissonian proces @xcite .
the important parameter here is the coherence time @xmath40 , a time scale at which photons are becoming to be correlated .
the coherence time can be estimated by help of the heisenberg uncertainty relation , assuming gaussian emission spectrum : @xmath41 the type of led we used has an exceptionally broad spectrum , approximately gaussian centered at @xmath42 688 nm and width @xmath43 35 nm . if the mean frequency of photon emissions is kept well below the coherence frequency of @xmath44hz , photons are practically mutually independent and consequently time intervals between subsequent photon emissions are exponentially distributed .
namely , the exponential distribution of time intervals between photon emissions and mutual independence of photons are two equivalent characterizations
. + alternatively , instead of a led diode one could as well use appropriately attenuated laser ( diode ) .
namely , for a continuous , constant - power source of coherent light , a random distribution of photon emissions in time yields a poissonian photon - counting distribution @xcite .
+ photons from the led are directed toward photocathode of a photomultiplier tube . each time a photon hits the photocathode
, there is a probability that an electron is emitted via the photoelectric effect .
creation of this so called `` photoelectron '' is a process governed by the binomial statistics .
therefore , the statistics of time intervals between emission of subsequent photoelectrons is governed by convolution of exponential and differential binomial distributions
. however , the convolution of any distribution and exponential distribution gives again the exponential distribution .
this is easy to understand qualitatively : if we have a source of random events and choose randomly only some of these events , then the chosen events are again random
. even more favorably , the convolution of an approximately exponential distribution and any other distribution yields an improved exponential distribution , meaning that randomness of the photoelectric effect can only enhance overall randomness of produced photoelectrons .
therefore , creation of a photoelectron is virtually truly random event .
creation of useable electric signal comprises : electron collection , electron multiplication , signal amplification , level detection and pulse shaping .
the effect of this signal processing is that following a photon detection the detector becomes insensitive to further incoming photons during a certain period of time called the `` dead time '' . for good quality detectors
the dead time has a well defined , fixed length ( which is important for our method ) .
but even if this is not the case , that is if for a certain detector dead time varies from event to event , it is easy to modify such detector to achieve a constant dead time
. this can be done , for example , by pasing its output through a non - retriggerable one - shot pulse shaping circuit which produces pulses of a constant length and which itself has the dead time slightly above the longest dead time observed in the detector .
such pulse shaping circuits , whose dead time can be set at a desired length , are well known in the art , see for example @xcite .
+ in our experiments we have used hamamatsu photomultiplier with bialkali photocathode r5611 , high voltage source c3830 , and the photon counting unit c3866 ( amplifier + discriminator + pulse shaper ) .
dead time of this photon counting system is about 25 ns , very stable and independent of brightness of the led .
the low quantum efficiency of only about 0.05 percent of the bialkali photocathode at the wavelength of the led presents no problem in our application .
we have intentionally used low efficiency red led because its widest wavelength spectrum in comparison to other leds indicates highest randomness of emited photons .
+ the probability density function ( pdf ) of measured time intervals between subsequent pulses from the rpg operated at 2 mhz as measured by the fast digital osciloscope lecroy waverunner 6100a , is shown in fig .
[ lecroy]a .
an almost perfect fit of measured data to the exponential distribution extends over nearly five decades of the random variable .
the effect of dead time is visible at the left end of the distribution as a sharp fall of the distribution at 25 ns .
a zoom of this region is shown in fig .
[ lecroy]b .
small wrinkles wisible at left end are caused by the binning effect due to the finite time resoution of the oscilloscope .
good agreement of the measured pdf with exponential distribution confirms that the photon emission of the led is indeed random and that afterpulsing in the photon detector is negligible at the mean output frequency of 2 mhz .
if , in some case , randomness of a single led source would be suspected , one could use several independently driven led sources to enhance the randomness of the light source .
however , having in mind very small coherence time @xmath40 this is unlikely to be an issue in practice . on the other hand , the afterpulsing in photomultipliers could be a potential problem at high bit extraction rates .
afterpulsing in pmt s creates correlations in the distribution of random intervals which dies off tipically with a time constant of the order of @xmath45 ns @xcite . at large photon incidence frequencies
this effect becomes negligible in comparison with abundant signal caused by real photons .
nevertheles , in principle , afterpulsing could cause serial correlations among random bits when bit extraction rate exceeds @xmath46 no matter what the extraction method .
this effect should be studied in more detail preferably by simulations which greatly over emphasize the level of afterpulsing . in any case
, bits produced by our prototype at 1 mbit / sec do not show any measurable correlations , as will be presented later .
+ the non - zero dead time , as will be shown , does not create bias nor correlations among random bits in our method .
using the above rpg and logic circuitry embedding our bit extraction method we have built and tested a laboratory prototype of a quantum random number generator which operates upon the above described method .
big advantage of this method is that it requires only one photon detector and is insensitive to slow variations in the detector s gain and quantum efficiency , in contrast to some other methods which crucially depend on extreme stability of these parameters over a full lifetime of the detector .
simplified schematic diagram of the setup is shown in fig .
it contains a quantum random pulse generator described in the previous section ( fig .
[ gsp ] ) . for the construction of the prototype
we have used a controlable bi - directional ( up / down ) counter which operates in the following way . at the beginning of any first time interval , in a pair of subsequent random time intervals
as explained earlier , the counter state is set to 0 .
during each first random time interval its control input is held at high logic state so it counts forward , and during the second random interval the control input is set low , so it continues counting backward until the end of the second random time interval .
the required difference @xmath47 is thus effectively obtained without a need to actually perform any subtraction .
if the counter state represents negative number we generate `` 0 '' , if positive we generate `` 1 '' and if zero we generate nothing .
+ implementing the extraction algorithm in the discrete 74hc logic family and operating the random event generator at a mean frequency @xmath48mhz , we were able to extract random bits at almost 1 million bits per second . with a 48mhz restartable clock we have measured bit efficiency of @xmath49 , which is quite close to the theoretical limit of 0.5 , and in a good agreement with theoretical prediction of 0.490 .
namely it can be shown that the bit efficiency of our method @xmath50 , defined as the number of bits produced on average per 1 random event asymptotically behaves like : @xmath51 where @xmath34 .
the error of this asymptotic expansion is less than 0.25 percent for @xmath52 . since in our case
@xmath53=2/48 this above formula gives @xmath54 .
+ although theoretically eliminated , the main problem with the practical realization of this prototype was a non - vanishing bias .
this seems to be a general problem with physical random number generators . in our case
the bias was probably mainly due to the fact that switchings of the bi - directional counter to `` up '' direction and to `` down '' direction are not equally fast .
even the slightest systematic difference in the length of `` up '' and `` down '' counting time intervals has a measurable effect on bias .
it can be shown that in the case of zero dead time the bias is given by : @xmath55 where @xmath56 is the difference between transition times for `` up '' and `` down '' counting directions .
however this biasing effect has no bearing on correlations which were indeed consistent with zero .
the bias of this setup was measured to be very small , of the order of @xmath36 .
this is readily by far the smallest bias reported for the raw bits from a quantum random bit generator @xcite .
however , bias quickly raised with random events freqency arriving to @xmath57 at @xmath58mhz , the scaling with @xmath59 being quite in accordance with eq .
( [ finetune ] ) . + in order to lower the bias even further or make possible operation at higher random events frequencies one could add a small tuneable delay to the clock input of the up / down counter and adjust it for a minimum bias .
alternative approaches include use of faster electronics which has smaller internal circuit delays and software debiasing procedures .
quality of randomness of long sequences of bits produced by our prototype was evaluated by three batteries of statistical tests : ent @xcite , diehard @xcite and sts version 1.5 for linux @xcite .
ent is a collection of `` standard '' basic statistical tests , including autocorrelation test .
we have modified this program in order to be able to check correlation coefficients with lags higher than 1 .
notably , we have checked serial correlation coefficients ( eq . [ eq : auto ] ) with lags from 1 up to 32 to be consistent with zero within statistical error bars which were of the order of @xmath36 .
below is a result for a typical sequence of 1e9 bits . for sake of brevity only the first 8 serial correlation coefficients
are shown in table [ entresults ] .
+ .results of ent tests suite for a typical sequence of 1e9 bits .
[ cols= " < " , ] nist has recently put forward `` a statistical test suite for random and pseudorandom number generators for cryptographic applications '' ( sts ) which aims to become first industry standard for testing of random number generators .
sts tests consist of total of 38 individual statistical tests , such as entropy , chisquare , long runs , dna tests , birthday spacing tests , lempel - ziv complexity , maurer s test etc . the written output from sts is lengthy and therefore not suitable for presenting here .
it suffices to say that we have repeatedly tested sequences as long as @xmath60 bits , which passed all of the tests .
we have presented the concept and the development of a new type of a fast nondeterministic random number generator whose randomness relies on intrinsic randomness of the quantum physical process of photonic emission in semiconductors and subsequent detection by the photoelectric effect .
timing information of detected photons , or more precisely the time interval between detected random events , is used to generate binary random digits - bits .
+ presented prototype consist of a light source , one single - photon detector and fast electronics for timing analysis of the detected photons providing random output numbers ( bits ) at 1mbit / sec . by using only one photodetector
there is no need to perform any fine tuning of the generator , moreover method is immune to detector instability problems , which are improved characteristics compared to most of the systems based on the spatial photon detection used today .
another originality is that for the purpose of eliminating correlations time interval measurement principle has been modified to a restartable clock method . the collection of statistical tests applied to random numbers produced with our quantum random number generator present results which demonstrate the high quality of randomness resulting in bias less than @xmath36 and autocorrelation consistent with zero , with efficiency of @xmath49 bits per detected random event .
+ with appearance of fast quantum random generators in the market [ ] and obvious declining of speed of computer processors , the speed gap between software and hardware generators has become much smaller especially if one takes into account that hardware generators do not put such a load on the processor as software generators do .
we predict that quantum random number generators will find their applications in basic scientific research , sensitive data protection and encryption , as well as in simulations in science and technology .
+ we are grateful to prof
. b. basrak form the mathematical faculty of the university of zagreb for enlightenig discussions .
parts of this work are subject of the patent application wo2005106645 .
building of the prototype was supported by world bank technical assistance program 2004 for croatia ( tal-2 ) .
100 t. ritter , cryptologia .
* 15 * , 81 ( 1991 ) .
bennett et .
al . , j. cryptology , * 5 * 3 , ( 1992 ) .
a. d. wyner , bell syst .
j. * 54 * , 1355 ( 1975 ) .
i. csiszar , j. koerner , ieee t. inform .
theory * 24 * , 339 ( 1978 ) .
u. maurer , ieee t. inform .
theory * 39 * , 733 ( 1993 ) .
a. stefanov , n. gisin , o. guinnard , l. guinnard and h. zbinden quant - ph/9907006 , 1999 t. jennewein et .
al . , rev .
instrum . * 71 * , 1675 ( 2000 ) .
ma hai - qiang et al , chinese phys .
* 21 * , 1961 ( 2004 ) . c. e. shannon , bell system technical journal , * 28 * , 656 ( 1949 ) .
s. nakamura , et al .
, japan j. appl .
phys . , * 34 * l1332 ( 1995 ) .
74hc221 dual non - retriggerable monostable multivibrtor with reset - product specification , philips semiconductor , december 1990 .
e. gulari , b. chu , rev .
instrum.*48 * , 1560 ( 1977 ) .
p. b. coates , the origins of afterpulses in photomultipliers , j. phys .
phys . , * 6 * 1159 ( 1995 ) .
d. e. knuth , the art of computer programming , vol .
2 , third edition , ( addison - wesley , reading , 1997 ) web url ` http://www.idquantique.com/products/quantis.htm ` web url ` http://qrbg.irb.hr ` m. stipevi ' c , rev .
instrum . * 75 * , 4442 ( 2004 ) .
g. marsaglia , diehard battery of tests of randomness , internet url : ` http://stat.fsu.edu/pub/diehard/ ` g. marsaglia , statistics & probability letters , * 9 * , 345 ( 1990 ) .
a. rukhin et .
al . , a statistical test suite for random and pseudorandom number generators for cryptographic applications , nist special publication , ( nist , gaithersburg , 2001 ) , internet url : ` http://csrc.nist.gov/rng/ ` j. walker , hotbits , http://www.fourmilab.ch/hotbits/how.html
1 . beam splitter is a frequently used component for random number generators .
the two photon detectors d1 and d2 are used to detect two possible outcomes corresponding to one of the two possible paths that photon can take .
thus each photon entering the beam splitter generates one random binary digit - bit .
time interval measurement methods ( a ) with continuous ( b ) and restartable ( c ) clock .
the time counter advances with every positive going edge of the clock pulse .
3 . serial autocorrelation coeficient @xmath16 for non - restartable clock method ( dead time = 0 ) .
@xmath17 is the clock period , @xmath18 is the mean time interval between random pulses .
bit efficiency , defined as the number of bits generated per random event , for non - restartable clock method as a function of @xmath21 where @xmath17 is the clock period and @xmath18 is the mean time interval between random pulses .
schematic diagram of the random pulse generator ( rpg ) composed of : light emitting diode ( led ) and single photon detector setup .
mean frequency of output pulses can be set by potentiometer p1 .
( a ) measured probability density function ( pdf ) of time intervals between subsequent events from our setup ; ( b ) zoom of the dead - time region which is characterized by a sharp transition from complete insensitivity of the detector following a photon detection to fully restored sensitivity some 25 ns later . + fig .
diagram of the quantum random number generator consistsing of random pulse generator and curcitry for implementation of the restartable clock method . | ' '' '' * we report upon a novel realization of a fast nondeterministic random number generator whose randomness relies on intrinsic randomness of the quantum physical processes of photonic emission in semiconductors and subsequent detection by the photoelectric effect .
timing information of detected photons is used to generate binary random digits - bits . the bit extraction method based on the restartable clock method
theoretically eliminates both bias and autocorrelation while reaching efficiency of almost 0.5 bits per random event .
a prototype has been built and statistically tested . * |
in this paper , we study the temporal behavior of the distribution of stock prices for 24 stocks in the dow jones industrial average ( djia ) .
this is done using a new method of measuring changes in the volatility and drifts of stocks with time . when this method is applied to time - series constructed from the daily close of stocks , changes as fast as one day
can be seen in both .
given that it is not possible to accurately _ measure _ ( as oppose to _ predict _ ) intraday changes in the volatility using only daily - close data , for two of the 24 stocks we have been able to reach the maximum resolution ( known as the nyquist criteria ) of one day in the rate that the volatility can change , while for the great majority of the remaining stocks , we have come within one day of this maximum .
we believe that this method can measure changes in the volatility and drift that occur during the trading day as well if intraday price data is used . but
even with only daily - close data , we have been extraordinarily successful at determining the temporal behavior of stocks in general , and of the volatility in particular , and in the process , we have furthered our understanding of the behavior of stock prices as a whole .
we find that the stock prices of these 24 stocks can be well described by a stochastic process for which the volatility changes _ deterministically _ with time . on the one hand
, this is a process where the yield at any one time is not correlated with the yield at any other time ; the process thus describes an efficiently priced stock . on the other hand , this is a process where the predicted kurtosis agrees with the sample kurtosis of the stock ; the process thus also provides a solution to the long standing problem of explaining how an efficiently priced stock can have a kurtosis that is so different from what is expected for a gaussian distribution .
indeed , we find that abnormally large kurtoses are due solely to changes in the volatility of the stock with time .
when this temporal behavior is accounted for in the daily yield , the kurtosis reduces dramatically in value , and now agrees well with model predictions .
this finding is in agreement with rosenberg s ( 1972 ) observation that the kurtosis for nonstationary random variables is larger than than the kurtosis of individual random variables .
we have also determined changes in the volatility of these stocks , and for three of the 24 stocks , variations of as fast as one day can be seen . for another 16 stocks ,
this temporal resolution was two days or less , and for only five of the 24 stocks is this resolution longer than 2.5 days .
the behavior of the drifts for all 24 stocks can also be determined using this method , and with the same resolution as their volatility .
we find that the drift for the majority of the stocks is positive ; these drifts thus tend to augment the increase of the stock price caused by the random - walk nature of the stochastic process .
this finding is not surprising , nor is it surprising that we find that the drift is much smaller than the volatility for all 24 stocks .
what is surprising is that for three of the 24 stocks the drift is uniformly _
negative_. for these stocks , the drift tends not to increase the stock price , but to depress it . that the stock price for these three stocks increase at all is because this drift is much smaller in the magnitude than the volatility . over the short term , growth in the prices of these stocksas they are for all 24 stocksis due to a random walk , and thus driven more by the volatility than the drift .
indeed , this is the only reason that the prices of these stocks increase with time .
finally , the distribution of the stock prices for the 24 djia stocks has been determined .
when the temporal variation in the volatility is corrected for in the daily yield , we find that the resultant distribution for all but four of the stocks is described by a rademacher distribution with the probability that the yield increases on any one day being 1/2 .
for the four other stocks , the distribution is described by a generalized rademacher distribution with the probability that the yield increases on any one day being slightly greater than the probability that it decreases .
in 2005 , 403.8 billion shares were traded on the new york stock exchange ( nyse ) with a total value of $ 14.1 trillion dollars ( see nyse ) . during the same period ,
468 million contracts were written on the chicago board options exchange ( cboe ) with a total notional value of $ 12 trillion dollars . at the nyse , traders , investors , and speculatorsbig and smallplace bets on the movement of stock prices , whether up or down .
profits are made , or losses are reconciled , based on the changing price of the stock . as such , great effort is made to predict the movements of stock prices in the future , and thus much attentionwith attending analysisis focused on the price of stocks . in the cboe ,
traders , investors , and speculators write or enter into contacts to purchase or sell a predetermined amount of stocks at a set time in the future .
profits here are made , or losses reconciled , based on the degree of risk that the movement of the stock will be down when expected to be up , or up when expected to be down . here , it is not so much the price of the stock that matters .
it is the amount of volatility in the stock , and predicting how stock prices may move in the future is much less important .
indeed , the pricing of optionsthrough the black - scholes equation and its variantsis based on the argument that it is _ not _ possible to predict how the price of stocks will change in the future . in this pricing , it is taken for granted that the markets are efficient , and that earning returns which are in excess of the risk - free interest rate is not possible .
all is random , and the increase in stock prices seen is due to a simple random walk with a ( small ) drift .
great interest is thus paid in modeling the _ distribution _ of stock prices , and the application of these models to the pricing of options and derivatives .
given the $ 26.1 trillion dollars in trades and contracts in the nyse and cboe in 2005 , it is not surprising that much effort has been expended in determining the properties of the stock market . given the precipitous drop in stock market prices in october of 2008__which occurred over period of days__accurate determination of how these properties change with time has become even more important . since
the work by bachelier ( 1900 ) at the turn of the 20th century , a great deal of these efforts have been focused on determining the distribution of the daily yields of stock prices ( osborne 1959a and osborne 1959b ) .
inherent in this determination is determining the volatility of the distribution .
use of this volatility is now pervasive in modern finance , and is a critical ingredient in such endeavors as the pricing of options , the general assessment of risk and the determination the value of assets at risk , and the construction of optimal portfolios . that this effort continues today is indicative of the difficulty in determining this distribution , its importance in modern finance , and the financial impact that its determination can have . while bachelier ( 1990 ) characterized the distribution as a random walk with the prices of the stock having a given drift and a constant volatility , it has been known since the detail analysis of the behavior of stock prices by fama ( 1965 ) that the distribution of daily yields is only approximately gaussian ; the distribution calculated by famawhich does not take into account variations in the volatility with timehas a fatter tail than expected for a gaussian distribution .
indeed , it is typically found that the kurtosis can be as high as 100 , while by comparison the kurtosis of a gaussian distribution is only three .
this discrepancy between the distribution of daily yields as they are traditionally calculated and the gaussian distribution , while seemingly an inconsequential detail , nonetheless has wide - ranging consequences .
mathematics tells us that if the distribution of daily yields of a stock is a gaussian distribution , then the daily yield on any one day can not depend on the daily yield on any other .
this is the central limit theorem ( clt ) , and it is embodied in a number of waysthe various forms of the efficient market hypothesis ( emh ) ( see fama 1970 and fama and french 1988 ) , and the no - arbitrage conditionin modern finance .
this lack of predictability is one of the underlying assumptions used in the pricing of derivatives .
mathematics says we can also turn the statement of the clt around , however .
namely , if the daily yield on any one day does not depend on the daily yield on any other , then the distribution of daily yields must _ necessarily _ be a gaussian distribution as long as the number of days used in its determination is large enough , and as long as the distribution is well behaved . in the face of this mathematical result , there are two possibilities .
the first possibility is that the distribution of daily yields for stocks is not gaussian .
the daily yield on one day does depend on the daily yield on some other day , and it is possible , in principle , to predict future stock prices by looking at historical prices .
the second possibility is that the emh nevertheless holds , and there are good , albeit unknown , reasons for the unexpectedly large kurtosis .
the situation is further muddied when the autocorrelations of the daily yield of stocks are calculated .
it is well known from these calculations that the value of the daily yield on different days are uncorrelated with each other , and we have seen this behavior for the stocks studied here as well .
this independence extends also to other asset classes , as shown by kendall ( 1953 ) .
there have been numerous attempts at using other distributionsthe levy and its generalization , the pareto , proposed by mandelbrot ( 1963 ) , the student t - distribution proposed by blattberg and gonedes ( 1974 ) , and the discrete mixture of gaussian distributions model proposed by kon ( 1984)to describe the distribution of stock prices ( see tyli , sysi - aho , and kaski 2004 for an overview and assessment ) .
these attempts are based on the belief that the second of the two possibilities holds , and that the reason for the overly large kurtosis is because the distribution used to describe the stock was not correct . as such , for these distributions the daily yield on any one day also does not depend on the daily yield on any other day , and the consequences of the clt is instead evaded in various ways .
these approaches have had various degrees of success .
for example , while the pareto distribution does have a fatter tail than the gaussian distribution and has a kurtosis that can agree with observations , all moments with order greater than an integer @xmath0which determines the power - law behavior of the distributionis ill defined ; in this way , the distribution is not well behaved , and thus does not fall within the class of distributions for which the clt is applicable . for the levy distribution , the volatility itself ( as well as all higher moments ) is ill - defined , requiring the truncation of the distribution to price options using this model , as described in kleinert ( 2002 ) .
the student - t distribution differs significantly from the gaussian distribution only when the number of data points are small ( thereby evading the clt ) , which begs the question of what happens when this distribution is applied to time - series with more than , say , 200 terms in it . kon s model is a of mixture of gaussian distributions , and thus the moments of his distribution are all finite .
however , while the model is effective at describing the large kurtosis of stocks , it is nonetheless an empirical model ; the origin of the discrete mixture is not known , and the parameters used in its construction are determined only after the model is fitted to the stock data . our approach is also based on the belief that the second of the two possibilities hold .
but unlike the previous attempts at describing the distribution of stock prices mentioned above , we find that the underlying reason for the overly large kurtosis is because time variations in the distribution of stocks have not been properly taken account of .
as observed by rosenberg ( 1972 ) , it is often assumed that the distribution of stock prices being analyzed does not change during the period of interest .
this assumption was certainly made for all the models described above .
in contrast to these approaches , we will take time variations in the distribution explicitly into account .
doing so results in a distribution that can both explain the abnormally kurtosis , and still have the property that the yield on any one day is not correlated with the yield on any other . in the process
, we will also be able to determine , for the first time , how the volatility and the drift changes instantaneously with time . that the volatility of stocks changes with time is not a new observation .
this behavior has been known since at least the work by osborne ( 1962 ) ( see also lo 1988 ) , and analyzed explicitly by rosenberg ( 1972 ) .
indications of this have been reported by many others since then ( ball and torous 1985 , french and roll 1986 , conrad and kaul 1988 , andersen and bollerslev 1997 , kullmann , tyli , kertesz , kanto , and kaski 1999 , nawroth and peinke 2006 ) .
much effort has since been made to determine how this volatilityand thus necessarily how the distributionchanges with time , with the main focus of this effort on extending the usual random walk description of stock market prices .
this has lead to the introduction of the jump - diffusion model proposed by merton ( 1976 ) , where discrete , random jumps in the prices of a stock in time are incorporated in continuous stochastic processes , and to stochastic volatility models developed over a number of years by praetz ( 1972 ) , christie ( 1982 ) , hull and white ( 1987 ) , scott ( 1987 ) and heston ( 1993 ) where the volatility itself is modeled as a stochastic process with its own drift and volatility ( see also muzy , delour , and bacry 2000 for a multifractal - inspired , stochastic volatility model ) .
however , as it was pointed out by hull and white ( 1987 ) , methods for directly _ measuring _ time - varying volatilities were not , at the time , known . this inability to directly measure variations in the volatility
has greatly constrained efforts in studying how the volatility of real market data varies with time . to a great extent
, it has also driven the development of stochastic volatility models . by characterizing the volatility as a stochastic process
, a time varying volatility can be modeled using a comparatively simple choice of a constant drift and a constant volatility for the process .
even then , however , parameters in stochastic volatility models are determined not by a direct analysis of the daily yields of stock prices , but are instead determined indirectly .
namely , the price of an option for a stock is calculated for the process in terms of a set of model parameters , and these parameters are then set by adjusting their values until the calculated price agrees with the market price of the option . the inherent difficulty in determining from market data how a distribution changes with time is described in boyle and anathanarayanan ( 1977 ) , and is straightforward to understand . to determine the distribution of a stock
, a collection of stock prices is required ; the larger the collection , the better .
since stock prices change sequentially in time , this collection has to be done over a period of time , and because a relatively large collection is needed , this period must be correspondingly long .
for example , most distributions are calculated using the daily close of a stock , if for no other reason then because these prices are readily available in the public domain .
if the collection of prices used is as large as 500 daily closes , then the stock prices in this collection must span a period of nearly two years ; fama ( 1965 ) , for example , used stock prices that span a period of up to six years in his analysis . using a collection of 500 stock prices to determine the distribution of the stock through standard methods means that one is tacitly assuming that the price of the stock two years ago belongs to the same distribution__with the same volatility__as the price of the stock today .
this strains credibility , especially given the rapid movements in the markets during the last quarter of 2008 .
while it is possible to calculate the distribution with a fewer number of stock prices and thus shorten the period of time over which they are collected , statistical errors inherent in determining the distribution are proportional to @xmath1 , where @xmath2 are the number of data points in the sample , and will thus be correspondingly larger . at one point ,
the period of time would be so short that we would not be able to say whether distribution is gaussian or not .
indeed , we only have to look at the extreme case where the period is so short that there are only three stock prices collected over three days , resulting in only two daily yields to determine the whole distribution ; this clearly can not be done with any certainty whatsoever ! ( this inherent difficulty has lead to the development of other approaches to calculating volatility such as those found in ball and torous 1984 , parkinson 1980 , and longin 2005 ) where the number of daily close needed is reduced . )
mathematics does not require that the distribution remains constant .
the general theory of stochastic processes allows for volatilities that change with time .
in fact , we will show below that even though the volatility of a gaussian distribution may change with time , the daily yield on any one day need not depend on the daily yield on any previous day ; the emh still holds for this case .
instead , what has been lacking up to now is a method for _ calculating _ the statistical properties of a stock when the volatility changes with time .
this we have been able to do .
our approach is based on the observation that when the volatility depends solely on time , we can remove the time dependence of the distribution by dividing the daily yield by the volatility .
this standardizes the daily yield , and a gaussian distribution with a time - varying volatility is mapped into a gaussian distribution with unit volatility .
the volatility of this distribution is clearly constant , and thus the standardized daily yields all belong to the _ same _ distribution .
the inherent difficulty in determining a distribution that changes with time mentioned above is thus circumvented .
indeed , large collections of stock prices are now a benefitthey result in smaller standard errorsand not a detriment . that the volatility of the mapped distribution is _ known _ then allows us to determine how the volatility of the original daily yield changes with time .
in addition , it is readily apparent from our analysis that the distribution of standardized yields is equivalent to a special case of the binomial distribution , and this observation allows us to extract easily the temporal behavior of the drift of the yield as well .
this approach is straightforward , and at its heart resembles the process one goes through in using a table of values for the cumulative standardized gaussian distribution : the random variable at hand is scaled with its volatility to get the standardized gaussian distribution with unit volatility .
the difference is that in our case the volatility is not know a prior ; it must be determined .
this is done using a combination of statistical methods , fourier analysis , and signal processing techniques . while prevalent in other fields , many of these techniques are not commonly found in the finance or business literature , and it would be easy to become too involved with the mathematics while neglecting the finance when presenting our results . to avoid doing so , we will focus on finance in the main body of the paper , and when our model of stock prices is constructed , it will be motivated by , and justified with , an analysis of the time - series of the stocks at hand .
importantly , a validation of each step taken will made .
only enough of the underlying mathematical analysis needed to explain the essential ideas behind our approach will be presented in the main body of the paper ; we will refer the reader to the appendices for many of the details .
our analysis will be applied explicitly to coca cola stock in this paper to demonstrate the underlying ideas behind the approach .
this stock is chosen out of the 24 because for our purposes its underlying behavior is representative of all the others .
analysis of the other 23 stocks studied here follow in much the same way , and we will only present a summary of the results for them , along with graphs of the volatilities for all 24 stocks as a function of trading day .
it would not be an exaggeration to say that the only characterizations of a stock that is not model dependent is the price per share that it was sold at , the day and time it was sold , and the total number of shares of the stock that was sold over a given period of time .
these are the only characterizations that are objective and verifiable , and for whom all can agree on how they are obtained .
the distribution of the daily yields of a stock certainly is not , and herein lies the problem : how should any model of stock prices be validated ? to see how difficult the problem of validation is to resolve ( this issue was explicitly studied by magdon - ismail and abu - mostafa 1998 for volatility models ) , consider the volatility of the 24 stocks considered here . as we will calculate the volatility of these stocks
, it would seem that a comparison of the volatilities we obtain here with the volatilities calculated using any one of the many other approaches in the literature would be an effective way of assessing the validity of our model .
however , irrespective of the approach taken to make this calculation , assumptions about the behavior of the stock will have already been made .
the historical volatility , for example , uses a moving average to calculate the volatility on any given trading day .
it implicitly assumes that the volatility does not change significantly over the window of time used when calculating the average , and thus can not effectively measure changes in the volatility that occur within this window .
the implicit volatility , developed over a series of papers by latan and rendleman ( 1976 ) , schmalensee and trippi ( 1978 ) , and beckers ( 1981 ) , can measure instantaneous changes in the volatility , but it is calculated by inverting the black - scholes ( or any other ) equation for pricing options , and thus implicitly assumes that the particular pricing equation used accurately prices the option at any given time .
autoregression approaches to calculating the volatilitysuch as the exponentially weighted moving average ( ewma ) , the autoregregressive conditional heteroskedasticity ( arch ) proposed by engle ( 1982 ) , the generalized autoregregressive conditional heteroskedasticity ( garch ) proposed by bollerslev ( 1986 ) , and a new approach that combines autoregressive and fourier ( spectral ) analysis techniques proposed by bollerslev and wright ( 2001)are designed more to _ manage _ volatilities that change with time than to _ characterize _ them .
they depend on one or more parameters that must subsequently be set using some property of a stock , and are not designed to _ determine _ how the volatility changes .
stochastic volatility models explicitly consider volatilities that change ( randomly ) with time , but to determine how this volatility changes , the approach adjusts the parameters that determine the volatility until the predicted option prices agree with market prices ( see lamoureux and lastrapes 1993 for a test of this approach ) . using a comparison of volatilities to validate models is therefore more a comparison of the underlying models of the market or methods of calculation than it is of the volatilities themselves . indeed , the question of which approach to calculating
the volatility is the better one is one that has been address many times over the years ( see day and lewis 1992 , canina and figlewski 1993 , jorion 1995 , figlewski 1997 , andersen and bollerslev 1998 , chong , ahmad , and abdullah 1999 , szakmary , ors , kim , and davidson ii 2003 , and mcmillan and speight 2004 ) , apparently without consensus .
this difficulty in validating models is particularly inopportune here .
while many of the techniques we have used in this paper has been long used in other fields , our approach in this paper is novel , and have not been used in the analysis of stock market prices before .
we therefore take a particularly stringent approach to validating our model .
first , the model must be able to explain the observed properties of the stocks .
this we accomplish by construction .
properties of the stock price are presented _ first _ , and the model is then constructed explicitly to describe them .
second , the model must be self - consistent , and must be able to predict some property the stock price , which can subsequently be verified .
all models of stock market prices make a certain set of underlying assumptions about properties of the price ; these assumptions have consequences .
these consequences can in turn be used to predict properties of the stock price that can then be used to validate it . for our model ,
the distribution of standardized daily yields is described by a rademacher distribution or its generalization .
this distribution gives specific values for the population skewness and kurtosis , and they provide a simple and statistically meaningful approach to validating our model .
specifically , we calculate the sample skewness and kurtosis from each stock s time - series .
we then compare this sample skewness and kurtosis to the population skewness and kurtosis predicted by our model .
if the sample skewness and kurtosis agree with the population skewness and kurtosis of our model at the 95% confidence level ( cl ) , we assert that our model is valid .
in fact , we find that this agreement holds for all 24 stocks considered here , and it does so over the whole of the time period spanned by their time series . indeed , for a number of the stocks , _ this period spans over 80 years_. it is because of this operational approach to validating our model that we chose to analyzed stocks from the djia
. first , all the stocks in the djia are large caps , and have a large daily trading volume ; they are precisely the type of stocks for which we expect the market to be efficient .
they are in this way similar , and we would expect they can be described by the same type of model .
second , each of these companies has been publicly traded for a number of years .
we therefore have access to a large collection of daily close prices for these stocks with which to construct their time - series . these time series , for example , range in time from as short as 5,090 trading days for citigroup , to as long as 21,527 trading days for exxon - mobil .
the availability of a large sample of daily close is particularly important as we will be numerically assessing the validity of each step in the construction of our model . with such large collections of stock prices , standard errors in our calculation
can be as small as 0.7% , and as such , we are able to say with a great deal of certainty whether or not our approach is self - consistent .
third , the 24 chosen were the simplest , in terms of their ownership , of the 30 stocks listed in the djia .
the six djia not chosen were recently involved in mergers or acquisitions , which introduces unwanted complications ; an assessment of the temporal of these stock prices may not be as clear cut as the 24 stocks considered here .
a detailed description of how the time - series are constructed is given in appendix @xmath3 , where any particularities in the analysis of stocks are listed as well .
a list of these stocks given in terms of their stock symbol is presented in table i along with the starting date of the time - series and the total number of daily yields in each .
the ending date for all 24 time - series is december 29 , 2006 .
we begin our study of the temporal behavior of stock market prices with an application to finance .
specifically , for the 24 stocks considered here we study whether the daily yield on december 29 , 2006 depends on the daily yield on any day previous to it .
this property of the market , which has direct implications in finance , will be used as the starting point for the construction of our model of stock prices .
indications of an efficiently priced stock shown in fig .
@xmath4 is a graph of the autocorrelation function of the daily yield for coca cola using eq .
@xmath5 from the appendix @xmath6 .
this autocorrelation is calculated between the daily yield of the stock on december 29 , 2006 , and the daily yield @xmath7 days _ before _ the 29th .
the graph thus shows the dependency of the yield on the 29th on the yield on any previous day .
if the yield on the 29th depends on the yield on day @xmath7 , then the autocorrelation function will not vanish on that day at the 95% cl .
if , on the other hand , the yield on the 29th does not depend on the yield on day @xmath7 , then the autocorrelation function will be within statistical error of zero .
also shown on the graph in fig .
@xmath4 is the errorbar for each of the calculated values of the autocorrelation function .
these errorbars are set at the 95% cl , which is @xmath8 times the standard error calculated using eq .
@xmath9 for the autocorrelation function on that day .
they thus set the 95% confidence interval ( ci ) about the calculated value for the autocorrelation function .
if the value of the autocorrelation function falls within its errorbar of zero , there is a 95% probability that the autocorrelation on this day equals zero . with 21,522 total trading days in the time - series for coca cola ,
the standard error for the values of the autocorrelation function shown in the graph is roughly 0.7% , and is thus quite small ; the errorbars shown are correspondingly small . the standard error for the majority of the stocks studied here are equally small .
all but one of the errorbars for the autocorrelation shown in fig .
@xmath4 straddles the @xmath10-axis . as such
, we can say that the value of the autocorrelation function for @xmath11 is within a 95% cl of zero for all but one day . indeed ,
when we continue this calculation all the way back to december 31 , 1925 , the starting date for the time - series , we find that the autocorrelation function for the daily yield on 20,279 out of a total of 21,522 trading days fall within the 95% ci of zero ; the autocorrelation function on 1,243 trading days , or 6% of the trading days , fall outside of the 95% ci ( see table i ) .
this does not necessarily mean that there is a correlation between the 29th and these 1,243 trading days , however .
statistically , we would expect values of the autocorrelation function to exceed the 95% ci on 5% , or 1,076 , of the trading days .
we can only conclude that on at least 1% , or 215 , of the trading days the autocorrelation function does not vanish for @xmath11 .
if instead a 99% ci in chosen , we find that the value of the autocorrelation function falls within the 99% ci of zero for 21,172 out of 21,522 trading days ; they fall outside of the 99% ci on only 2% , or 350 , of the trading days .
we can therefore still conclude that for at least 1% , or 215 , of the trading days the autocorrelation function may not vanish for @xmath11 .
the autocorrelation function of the daily yield for all 24 stocks have been calculated for the length their time - series , and we have found that the autocorrelation function for these stocks behave similarly to coca cola s .
namely , the autocorrelation function is maximum at @xmath12 , and it does not vanish for at least 1% to 3% of the trading days for each stock ; for citigroup and verizon , this percentage is even lower .
we may conclude from this analysis that for the vast majority of the time the daily yield of these stocks on any one day is not correlated with the daily yield on any subsequent day ; the market is thus extremely efficient for these 24 stocks . in addition , we will show below that for the 1% to 3% of the trading days when the autocorrelation function does not vanish , this is due to changes in the volatility of the stock with time , and not to correlations between daily yields .
based on the above analysis , it would seem that the usual stochastic process with a constant volatility would be a good model for these stocks .
the lack of dependence of the daily yield on any one day from any other is precisely the property inherent in such a model .
there are , however , other properties of the distribution of daily yields for stocks that any model would have to explain as well , and it is here that the constant - volatility model of stocks is lacking . shown in the insert of fig .
@xmath4 is the sample skewness of the daily yields for coca cola calculated using a 251-day moving average .
if indeed the stock price of the stock is well described by a stochastic process with a drift and a constant volatility , then we would expect the skewness of the daily yield to be zero . for coca cola ,
we find that the skewness ranges from @xmath13 to @xmath14 .
although the skewness is large in magnitude , its standard error is correspondingly large , and we find that the skewness exceeds the 95% ci of zero on only 943 out of 21,022 days , or 4% , of the time . thus , the sample skewness calculated using a 251-day moving average agrees with what is expected from modeling the yield of the stock using a stochastic process with constant volatility .
the situation is quite different for the kurtosis , however .
shown also in the insert of fig .
@xmath4 is the sample kurtosis of the daily yields for coca cola calculated with the same 251-day moving average .
although the kurtosis for a daily yield described by a stochastic process with a constant volatility is expected to be three , what we find instead is that the sample kurtosis calculated for the coca cola time - series ranges in value from @xmath15 to @xmath16 . like the skewness ,
the standard error for the kurtosis is large when the kurtosis is large , but unlike the skewness , the error is not overwhelmingly large .
we find that the kurtosis exceeded the 95% ci of three on 15,393 out of 21,022 days , or 72% , of the time .
for the great majority of the trading days in the time - series , the kurtosis is different from that expected for a stochastic process with constant volatility .
we have done this calculation for all 24 stocks , and these results are not unique to coca cola .
this , then , is the contradiction inherent in using a stochastic process with constant volatility to model stock market prices .
on the one hand , calculations of the autocorrelation function indicate that the market is extremely efficient for these stocks , which is consistent with a stochastic process with constant volatility .
on the other hand , calculation of the kurtosis for these stocks are much larger than expected for such a process .
we will use this contradiction to guide the construction of our model in the analysis below .
lrrrrrr c & 10/29/86 & 5088 & 227 ( 4% ) & 53 ( 1% ) & 256 ( 5% ) & 45 ( 1% ) + msft & 03/13/86 & 5248 & 221 ( 4% ) & 63 ( 1% ) & 292 ( 6% ) & 60 ( 1% ) + vz & 02/16/84 & 5770 & 236 ( 4% ) & 58 ( 1% ) & 306 ( 5% ) & 57 ( 1% ) + intc & 12/14/72 & 8592 & 516 ( 6% ) & 159 ( 2% ) & 466 ( 5% ) & 93 ( 1% ) + axp & 12/14/72 & 8592 & 412 ( 5% ) & 101 ( 1% ) & 391 ( 5% ) & 78 ( 1% ) + aig & 12/14/72 & 8592 & 470 ( 5% ) & 137 ( 2% ) & 470 ( 5% ) & 96 ( 1% ) + & & & & & & + wmt & 11/20/72 & 8608 & 458 ( 5% ) & 118 ( 1% ) & 415 ( 5% ) & 85 ( 1% ) + hpq & 03/03/61 & 11524 & 610 ( 5% ) & 169 ( 1% ) & 575 ( 5% ) & 112 ( 1% ) + dis & 11/12/57 & 12386 & 741 ( 6% ) & 182 ( 1% ) & 624 ( 5% ) & 135 ( 1% ) + aa & 06/11/55 & 14014 & 688 ( 5% ) & 153 ( 1% ) & 710 ( 5% ) & 158 ( 1% ) + mrk & 05/15/46 & 15444 & 946 ( 6% ) & 233 ( 2% ) & 835 ( 5% ) & 190 ( 1% ) + mmm & 01/15/46 & 15544 & 733 ( 5% ) & 253 ( 2% ) & 724 ( 5% ) & 145 ( 1% ) + & & & & & & + jnj & 09/25/44 & 15920 & 759 ( 5% ) & 163 ( 1% ) & 734 ( 5% ) & 145 ( 1% ) + pfe & 01/17/44 & 16126 & 812 ( 5% ) & 181 ( 1% ) & 812 ( 5% ) & 181 ( 1% ) + ba & 09/04/34 & 18946 & 1281 ( 7% ) & 360 ( 2% ) & 980 ( 5% ) & 200 ( 1% ) + cat & 12/02/29 & 20358 & 1691 ( 8% ) & 702 ( 3% ) & 1027 ( 5% ) & 239 ( 1% ) + pg & 08/12/29 & 20442 & 1644 ( 8% ) & 678 ( 3% ) & 1026 ( 5% ) & 223 ( 1% ) + ge & 12/31/25 & 21518 & 1474 ( 7% ) & 545 ( 3% ) & 1104 ( 5% ) & 242 ( 1% ) + & & & & & & + gm & 12/31/25 & 21518 & 1370 ( 6% ) & 488 ( 2% ) & 1120 ( 5% ) & 251 ( 1% ) + dd & 12/31/25 & 21520 & 1520 ( 7% ) & 532 ( 2% ) & 1071 ( 5% ) & 215 ( 1% ) + mo & 12/31/25 & 21520 & 1804 ( 8% ) & 762 ( 4% ) & 1047 ( 5% ) & 242 ( 1% ) + ibm & 12/31/25 & 21522 & 1387 ( 6% ) & 461 ( 2% ) & 1095 ( 5% ) & 193 ( 1% ) + ko & 12/31/25 & 21522 & 1243 ( 6% ) & 350
( 2% ) & 1124 ( 5% ) & 230 ( 1% ) + xom & 12/31/25 & 21526 & 1498 ( 7% ) & 477 ( 2% ) & 1096
( 5% ) & 220 ( 1% ) + [ greensummary ] in this section , we show that for continuous stochastic models of stock prices with a _
volatility that changes with time , the yield at time , @xmath17 , does not depend on the yield at any other time , @xmath18 .
such a stock price is thus able to model the properties of the autocorrelation function found for the 24 stocks above . in a later section
, we will show that the time - variation in the volatility can also explain the abnormally large sample kurtosis .
take as the price of the stock at any time , @xmath17 , the continuous function @xmath19 .
this is an approximation , of course .
stocks are bought and sold in discrete time periods , and the prices of these transactions are always recorded in discrete units .
it is , however , easier to develop an understanding of the model , and to show a number of properties of it , using this continuous approximation instead of using a discrete time - series of stock prices . in the next section ,
when we develop a recursion relation for the volatility , we will consider real - world data , and will discretize the continuous model presented here .
our model for @xmath19 is a stochastic process with a drift , @xmath20 , and a volatility , @xmath21 , that change only with time : @xmath22 where @xmath23 is a gaussian random variable such that @xmath24 = 0 , \qquad \hbox{and}\qquad e[\xi(t)\xi(t ' ) ] = \delta(t - t ' ) . \label{delta}\ ] ] here , @xmath25 $ ] is the expectation value of @xmath26 over a gaussian distribution , and @xmath27 is the dirac delta function .
we emphasize that while eq .
@xmath28 may have a form that is similar to various stochastic volatility models of the stock market , for us @xmath21 is a _ deterministic _ function of time ; it does not have the random component that is inherent in stochastic volatility models . as usual , it is more convenient to work with @xmath29 $ ] ; for continuous compounding , @xmath30 is then the instantaneous yield of the stock . in terms of @xmath31 , eq .
@xmath28 reduces to @xmath32 where @xmath33 .
it is straightforward to show that for this stochastic process the instantaneous yield at time , @xmath17 , does not depend on the yield at any other time , @xmath18 .
to do so , consider the expectation value @xmath34 = e[\sigma(t)\sigma(t')\xi(t)\xi(t')].\ ] ] because @xmath21 is a deterministic function , it can be moved outside the expectation value so that @xmath35 = \sigma(t)\sigma(t')e[\xi(t)\xi(t')]$ ] . using eq .
@xmath36 , we then conclude that @xmath34 = \sigma(t)^2\delta(t - t ' ) , \label{cont - auto}\ ] ] so that the autocorrelation function of the instantaneous yield vanishes unless @xmath37 ; the yield of the stock at any one time does not depend on the yield at any other time .
our model thus describes a market for the stock that is efficient .
this is to be expected . at each instant , @xmath17 ,
@xmath38 describes a gaussian distribution with drift , @xmath39 and volatility , @xmath21 , and it is well known that for a gaussian distribution the daily yield on any one day is not correlated with the daily yield on any other . note that if the volatility was a function of @xmath40 as well as @xmath17 , or if it was itself a stochastic process , as it is taken to be in stochastic volatility models , we could not have moved the volatilities outside the expectation value to obtain eq .
@xmath41 . in these cases , it is not clear whether the yield of the stock at any one time depends on the yield at any other time . formally , the solution to eq .
@xmath38 is straightforward .
if @xmath42 for all @xmath17 , divide through by @xmath21 , and then reparametize time by taking @xmath43 equation @xmath38 then simplifies to @xmath44 where @xmath45 and @xmath26 is still a gaussian random variable , but now in @xmath46 .
equation @xmath47 is simply a stochastic process with drift @xmath48 and unit volatility ; its solution in terms of @xmath46 is well known .
the solution to the original equation , eq .
@xmath38 , can then be obtained , at least in principle , by integrating eq .
@xmath49 , and then replacing @xmath46 with resulting function of @xmath17 . in practice , our task is much more difficult .
we are _ not _ given a @xmath21 , and then asked to find the price , @xmath19 , of the stock at subsequent times .
we are instead given a collection of stock prices collected over some length of time , and then asked to find the volatility .
this is a much more difficult problem , but surprisingly , it is a solvable one , as we will see in the next subsection .
the drift for the standardized yield in this subsection , we derive a recursions relation that is used to solve for the volatility as a function of time .
this derivation is most conveniently done using a discretized version of the continuous stochastic process eq .
@xmath38 considered above , and we consider @xmath19 as a continuous approximation to the discrete time series , @xmath50 , for @xmath51 , of stock prices collected at equal time intervals , @xmath52 ; this @xmath52 is usually taken as one trading day .
the subscript @xmath53 enumerates the time step when the price of the stock was collected , and is an integer that runs from 0 to the total number of data points , @xmath54 . as such , @xmath55 , @xmath56 , @xmath57 , @xmath58 , and @xmath59
is the volatility at @xmath60 .
the instantaneous daily yield is then @xmath61 where @xmath62 .
it is clear that @xmath63 is the yield of @xmath19 over the time period @xmath52 ; when @xmath52 is one trading day , @xmath64 is the daily yield .
our task in this section is to determine @xmath65 _ given _ the time - series @xmath50 , and we do so by making use of the analysis in the previous section .
we call @xmath66 the _ standardized _ yield of the stock price over a time period @xmath52 , and if @xmath52 is one trading day , we call it the standardized daily yield . since @xmath67 then from the discretized versions of eqs
. @xmath47 and @xmath68 , we see that the distribution of standardized yields has a volatility of @xmath69 , or one , if @xmath52 is set to one trading day .
the collection of standardized yields has a _ known _ volatility .
consider now a subset of the time - series with @xmath70 elements , and the corresponding collection of standardized yields , @xmath71 , where @xmath53 runs now runs from @xmath72 to @xmath2 . because this subset was arbitrarily selected from a collection of standardized yields that has a volatility of @xmath69
, this subset must also have a volatility of @xmath69 .
as such @xmath73 where we have included for completeness in eq .
@xmath74 the standard error for the volatility given @xmath2 data points ( see stuart and ord 1994 ) to emphasize that the accuracy of eq .
@xmath74 depends on @xmath2 .
equation @xmath74 must be true for each @xmath75 .
in particular , it must hold for @xmath76 , and thus we can write @xmath77 which is similar in form to eq . @xmath74 .
this self - similar property of the distribution is used to determine @xmath65 , as we show below .
we first expand eq .
@xmath74 , and single out the @xmath78 terms , @xmath79 where we have dropped the error terms in eq .
@xmath74 for clarity .
using eq .
@xmath80 in the second term of eq .
@xmath81 and completing a square , we arrive at a surprisingly simple equation for @xmath82 , @xmath83 this is easily solved to give , @xmath84 where the sign of the root must be chosen so that @xmath85 for all @xmath2 .
the standardized yield , @xmath86 , can then be calculated using eq .
@xmath87 for each time step .
equation @xmath88 gives a recursion relation for @xmath82 .
a recursive approach to calculating the volatility similar in spirit to the one above is described in stuart and ord ( 1994 ) .
that calculation is for volatilities that do not change with time , however , while in ours the volatility can do so explicitly .
as we will see below , this introduces a number of complications .
we note also that eq .
@xmath88 differs markedly from autoregression approaches such as the ewma , arch , and garch in that @xmath82 depends nonlinearly on @xmath89 .
equation @xmath88 gives a first - order recursion relation for @xmath65 , and thus given an initial @xmath90 , the values for @xmath65 for @xmath91 is determined . to determine this initial @xmath90
, we note that in the continuous process eq .
@xmath68 holds .
a similar relation must hold for the discretized yields @xmath64 . to determine this relation , we follow the same approach that led to eq .
@xmath74 , and consider the following function @xmath92 the first term in eq .
@xmath93 is the average of the standardized yield over the first @xmath2 terms in the time - series , and it corresponds to the discretization of the first term in continuous constraint eq .
the second term is the quotient of the average daily yield calculated over the same period with the volatility evaluated at the end of this period , and it corresponds to the discretization of the second term in continuous constraint eq .
@xmath68 . if @xmath90 can be chosen so that the mean of @xmath94 , @xmath95 can be minimized to zero at the 95% cl , then eq .
@xmath68 will hold on average for the discretized yield .
as usual , the 95% cl for this mean is calculated through the standard error , @xmath96^{1/2}$ ] , where @xmath97 is the standard deviation of @xmath94 . we have successfully applied the recursion relation , eq .
@xmath88 , to the 24 djia considered here , and have obtained for each stock time - series for @xmath65 and @xmath71 .
this was done by determining the quotient @xmath98 through an iterative search algorithm that was implemented with a simple c++ program .
this algorithm searches for a @xmath90 that drives the mean of @xmath94 to zero while in the process minimizing @xmath99 .
in addition , since the volatility must be non - negative , this search is done under the constraint that all calculated values for @xmath65 must be greater or equal to zero , and it was stopped once the mean of @xmath94 has been calculated to sufficient accuracy . a @xmath100 that minimizes @xmath94 while at the same time giving a non - negative value for the volatility can be found for all 24 stocks .
indeed , we found that the mean of @xmath94 can be driven as close to zero as needed .
the results of this calculation is given in table ii , which lists for each stock the value of @xmath100 , the mean of @xmath94 for this @xmath98 , and the standard error of the mean .
while values for @xmath100 is only given to an accuracy of @xmath101which is sufficient given the accuracy of the @xmath50 for the stocks as noted in appendix @xmath3we have been able to drive the mean value of @xmath94 to as far down as @xmath102 by increasing the accuracy of @xmath100 to @xmath103 .
it is clear from the standard errors given in table ii that the mean of @xmath94 vanishes within standard error at the 95% cl .
this validates the recursion relation for @xmath65 for all 24 stocks .
implicit in the derivation of eq .
@xmath74 is that @xmath2 is large , and yet since @xmath82 starts at some initial point @xmath90 , @xmath82 are necessarily generated at small @xmath2 .
we would thus expect that there is a transient interval marked by some @xmath104 for which the solution to eq .
@xmath88 for @xmath105 is markedly different from the solution when @xmath106 .
this is seen .
for all 24 stocks , the behavior of @xmath71 for @xmath53 near one is different than its behavior for large @xmath53 .
this difference is similar for all of the stocks , indicating that it is due to the recursion process itself , and not to any underlying behavior of the markets .
we would thus hesitate to use the calculated values for @xmath65 when @xmath107 to draw conclusions about the behavior of the stock .
the length of this interval , @xmath108 , varies from stock to stock , but typically ranges between 100 to 400 trading days .
given that the shortest time - series considered here contains 5,088 trading days , this interval is extremely short for all 24 stocks , and is not relevant in practice .
this is yet another reason why we have chosen stocks that have a long track record to analyze .
lrrr ge & -1.29986151 & 3.70@xmath109 & 2.51@xmath110 + axp & -1.11059569 & -4.56@xmath111 & 7.60@xmath110 + pfe & -0.69103252 & 6.37@xmath112 & 6.02@xmath110 + dis & -0.62361094 & -1.66@xmath113 & 1.08@xmath114 + msft & -0.23500530 & -4.61@xmath113 & 3.36@xmath114 + ko & -0.00633264 & 8.28@xmath109 & 3.39@xmath110 + & & & + pg & 0.29288976 & -2.68@xmath114 & 2.47@xmath115 + gm & 0.29925834 & 1.19@xmath113 & 5.08@xmath110 + aig & 0.30999620 & -2.73@xmath113 & 5.90@xmath110 + mmm & 0.36915581 & -4.67@xmath113 & 5.32@xmath110 + aa & 0.37469224 & 1.30@xmath109 & 4.77@xmath110 + hpq & 0.64209911 & 1.55@xmath113 & 6.24@xmath110 + & & & + jnj & 0.96565296 & -1.96@xmath113 & 5.54@xmath110 + cat & 0.98788045 & -2.20@xmath113 & 5.27@xmath110 + intc & 0.99040135 & -6.99@xmath109 & 1.58@xmath114 + wmt & 0.99093623 & 5.31@xmath116 & 1.53@xmath114 + c & 1.05691370 & 1.41@xmath117 & 5.76@xmath110 + ibm & 1.12634844 & -1.08@xmath113 & 5.73@xmath110 + & & & + mo & 1.18101555 & 1.15@xmath113 & 1.35@xmath110 + vz & 1.30751627 & 2.03@xmath111 & 1.40@xmath114 + dd & 1.34345528 & -4.88@xmath109 & 1.17@xmath114 + ba & 1.41323601 & -2.74@xmath113 & 1.36@xmath114 + xom & 1.41414287 & -1.88@xmath111 & 5.08@xmath118 + mrk & 1.41417225 & -9.05@xmath111 & 7.72@xmath118 + [ recursion ]
although the recursion relation , eq .
@xmath88 , has been successfully solved for all 24 stocks , we will delay until sec . @xmath119 to present the solutions to this equation .
instead , we will first validate our model by showing that the stochastic process introduced in the previous sections solves the overly large kurtosis problem raised in sec .
@xmath120 . in the process
, we will find that the distribution of standardized yield is a generalized rademacher distribution , and will show that the simple skewness and kurtosis agree with the values for population skewness and kurtosis for this distribution . by doing so , we will also have validated our model of stock market prices using the criteria outlined in sec . @xmath121 . as part of this process
, we will be able to determine the drift of the yield as a function of time as well . to complement the autoregression calculation for the daily yield shown in fig .
@xmath4 , we have calculated autocorrelation function for the standardized daily yield , @xmath122 , for all 24 djia stocks .
a plot of @xmath122 as a function of @xmath123 has the same shape as that shown in fig .
@xmath4 , but with @xmath124 for all the stocks instead of a range of values . like the daily yield
, the standardized daily yield on any one day is not correlated with the standardized yield on any other day ; this is to be expected if the volatility is a function of time only .
we have also determined the number of trading days for which the value of the autocorrelation function falls within the 95% and 99% ci of zero .
the results are shown in table i , and for for all but one stock , the results as as expected .
on 5% of the trading days , the value of the autocorrelation function exceeds the 95% ci , and on 1% of the trading days , the value exceeds the 99% ci .
the only exception is microsoft at the 95% cl when on 6% of the trading days the autocorrelation function exceeds the 95% ci of zero . in sec .
@xmath120 , we noted that on at least 1% of the trading days the value of the autocorrelation function for the daily yield exceeds either the 95% or 99% ci of zero , and we can say with a degree of statistical certainty that for these days , the autocorrelation function does not vanish . with the exception of microsoft at the 95% cl , such days are not found in the autocorrelation function of the standardized yields .
since the standardized yield is obtained from the yield by removing the time - dependent volatility , we conclude from the results in table i for the standardized yield that , with the possible exception of microsoft , this 1% is not due to correlations in the daily yield , but rather to temporal variations in the volatility .
next , shown in fig .
@xmath125a is a plot of the daily yield with respect to trading day for coca cola . in comparison ,
@xmath125b is the is the plot of the standardized daily yield for the stock over the same period .
it is readily apparent that instead of taking a range of values between @xmath126 as the daily yield does , the standardized yield jumps between two values , one near @xmath127 and one near @xmath128 .
notice also that while the standardized yield is not precisely @xmath127 or @xmath128 , any changes in the standardized yield near @xmath127 are accompanied by the _
same _ variations of the yield near @xmath128 ; the variations in the standardized yield near @xmath127 and near @xmath128 would seem to move up or down in unison .
indeed , using a 251-day moving average , we find that the average of the difference in the value of the standardized daily yield , @xmath129 , near @xmath127 and its value , @xmath130 , near @xmath128 ranges from a minimum of @xmath131 to a maximum of @xmath132 ; both are within the 95% ci of one .
this binomial behavior for the yield is not surprising for the same reasons that binomial trees are effective at pricing options .
as noted by cox and ross ( 1976 ) , a continuous stochastic process with constant volatility can be approximated as a discrete random walk where at each time step , @xmath133 , there is a probability , @xmath134 , that the stock price will increase at the next time step , and a probability @xmath135 that it will decrease .
the discrete stochastic process can thus be approximated by a binomial distribution , and as the binomial distribution is known to approach the gaussian distribution in the large @xmath53 limit , the discrete random walk approaches a continuous stochastic process for the stock price .
indeed , this limit is the reason why binomial trees are effective in the first place . in this subsection
, we will determine the drift of the standardized daily yield as a function of time .
we do so by noting that it is apparent from fig .
@xmath125 that the distribution of standardized daily yields is a binomial distribution .
thus , at each time step , @xmath53 , there is a probability , @xmath134 , that the standardized yield will increase by an amount @xmath136 on that day , and probability , @xmath135 , that it will decrease by an amount @xmath137 . while in principle @xmath134 may be different at different time steps , the fact that @xmath136 and @xmath137 change in unison while keeping the average distance between position and negative standardized yields constant suggests that any variation in time is due to an overall shift in the distribution .
variations in @xmath136 and @xmath137 are not due to a time - dependent @xmath134 , but rather to a drift for the standardized yield that changes with time . with this realization , the drift can easily be determined for all 24 stocks . from eqs.@xmath47 and @xmath138
, we can express the standardized daily yield as @xmath139 where @xmath140 is the discretized drift of the standardized yield , and @xmath141 is a random variable with zero mean and unit volatility such that @xmath142 = \delta_{nm}$ ] . while for the continuous process @xmath143 would be a gaussian random variable , for the discrete process we will show that @xmath141 is a random variable for the generalized rademacher distribution described below .
the standard way of calculating @xmath144 is to use a moving average over a window of @xmath145 days .
however , just like for the volatility , calculating @xmath144 with a moving average will mean that variations in the drift faster than @xmath145 can not be clearly seen .
we will instead calculate @xmath144 directly from @xmath71 , which is possible to do because the distribution of standardized yield is so simple .
we first note that since changes to @xmath136 and @xmath137 is due to shifts in the distribution of standardized yield with time , these shifts must be due to the drift , @xmath144 , of the standardized yield .
shifts in random variables are trivial changes to the distribution , however , and a drift that changes with time will not materially change the distribution of standardized yields .
we next note that @xmath146 = 0 $ ] . as the values of @xmath71 lie close to @xmath147
, we conclude that @xmath141 can only take the values @xmath148 .
any deviation by @xmath149 from @xmath150 must be due to the drift .
this drift can be determined by solving the equation @xmath151 for @xmath144 by taking the sign of @xmath141 to be the same as the sign of @xmath71 .
this solution is straightforwardly implemented , and the results for coca cola is shown in fig .
@xmath125c . for clarity
, we have only shown the values of the drift between @xmath152 .
while there are values that lie outside of this range , they occur in the first 10 time steps in the series , and are part of the transient behavior mentioned above .
the drift for the standardized yield of all 24 stocks have be found using this approach . not surprisingly , we find that 21 out of the 24 stocks have a drift that positive for the great majority of the time - series .
what is surprising is that for three of the 24 stocks ( exxon - mobil , merck , and proctor and gamble ) the drift of the yield of the stock is _ negative _ outside the transient region .
for these stocks , the only reason why their price increases is due to the random walk , and because the volatility is so much greater than the drift .
while in the last subsection we determined the drift of the daily yield , in this subsection we will show that the distribution of the standardized daily yield is a generalized rademacher distribution shifted by the drift , @xmath144 .
this will be done by comparing the skewness and kurtosis of the rademacher distribution with the sample skewness and kurtosis for @xmath71 after the drift has been removed .
we will show that the two agree at the 95% cl , and doing so will both determine the distribution and validate our model of stock market prices as a stochastic process with a time - dependent volatility .
we begin by describing the properties of the generalized rademacher distribution .
a generalize rademacher distribution consists of a random variable , @xmath153 , that takes the value @xmath127 with probability @xmath134 , and the value @xmath128 with probability @xmath135 .
we denote the expectation value for this distribution as @xmath154 $ ] , and find that the population mean , @xmath155 $ ] , is simply @xmath156 this vanishes for @xmath157 . the @xmath0th population moment ,
@xmath158)^k]$ ] , is easily calculated to be @xmath159 .
\label{mk}\ ] ] the population variance is thus @xmath160 , while the population skewness is @xmath161 and the population kurtosis for the distribution is @xmath162 clearly , if @xmath157 , then @xmath163 , @xmath164 , and @xmath165 ; this is the rademacher distribution , which is a special case of the binomial distribution .
when @xmath166 , we call this the generalized rademacher distribution . given the plot in fig .
@xmath125b , we would expect that the distribution of standardized yields to be a rademacher distribution with @xmath167 at all time steps . to show that that this is the case , we have calculated the sample skewness and the kurtosis of the standardized yield after the drift , @xmath144 , has been removed from @xmath71 .
this has been done for all 24 stocks using the entire time - series for each .
we then compared these sample skewness and kurtosis with the population skewness and kurtosis for the rademacher distribution using the t - test . for completeness
, we have also calculated the probability , @xmath134 , for each stock by counting the total number of @xmath168 , and compared it to the rademacher value of @xmath167 using the chi - squared test .
the results of these calculations and tests are given in table iii .
we see that for all but four of the stocks the fit is exceedingly good ; the skewness , the kurtosis , and the probability all agree at the 95% cl .
lrrrrrr gm @xmath169 & @xmath170 & @xmath1710.07@xmath171 & @xmath172 & @xmath1710.92@xmath171 & 0.500 & @xmath1710.01 + dd @xmath169 & @xmath173 & @xmath1710.15@xmath171 & @xmath174 & @xmath1710.92@xmath171 & 0.499 & @xmath1710.02 + vz @xmath169 & @xmath175 & @xmath1710.18@xmath171 & @xmath176 & @xmath1710.92@xmath171 & 0.501 & @xmath1710.03 + dis @xmath169 & @xmath177 & @xmath1710.54@xmath171 & @xmath178 & @xmath1710.92@xmath171 & 0.498 & @xmath1710.29 + axp @xmath169 & @xmath179 & @xmath1710.58@xmath171 & @xmath180 & @xmath1710.92@xmath171 & 0.497 & @xmath1710.34 + mmm @xmath169 & @xmath181 & @xmath1710.66@xmath171 & @xmath182 & @xmath1710.93@xmath171 & 0.503 & @xmath1710.43 + & & & & & & + pg @xmath169 & @xmath183 & @xmath1710.67@xmath171 & @xmath184 & @xmath1710.93@xmath171 & 0.502 & @xmath1710.45 + jnj @xmath169 & @xmath185 & @xmath1710.78@xmath171 & @xmath186 & @xmath1710.94@xmath171 & 0.503 & @xmath1710.60 + hpq @xmath169 & @xmath187 & @xmath1710.78@xmath171 & @xmath188 & @xmath1710.94@xmath171 & 0.504 & @xmath1710.61 + ko @xmath169 & @xmath189 & @xmath1710.94@xmath171 & @xmath190 & @xmath1710.96@xmath171 & 0.503 & @xmath1710.89 + c @xmath169 & @xmath191 & @xmath1710.98@xmath171 & @xmath192 & @xmath1710.97@xmath171 & 0.507 & @xmath1710.96 + ba @xmath169 & @xmath193 & @xmath1711.28@xmath171 & @xmath194 & @xmath1711.04@xmath171 & 0.495 & @xmath1711.64 + & & & & & & + aig @xmath169 & @xmath195 & @xmath1711.32@xmath171 & @xmath196 & @xmath1711.04@xmath171 & 0.507 & @xmath1711.73 + intc @xmath169 & @xmath197 & @xmath1711.34@xmath171 & @xmath198 & @xmath1711.05@xmath171 & 0.507 & @xmath1711.79 + pfe @xmath169 & @xmath199 & @xmath1711.42@xmath171 & @xmath200 & @xmath1711.07@xmath171 & 0.506 & @xmath1712.01 + cat @xmath169 & @xmath201 & @xmath1711.42@xmath171 & @xmath202 & @xmath1711.07@xmath171 & 0.505 & @xmath1712.01 + ge @xmath169 & @xmath201 & @xmath1711.49@xmath171 & @xmath203 & @xmath1711.10@xmath171 & 0.505 & @xmath1712.21 + aa @xmath169 & @xmath204 & @xmath1711.50@xmath171 & @xmath205 & @xmath1711.70@xmath171 & 0.494 & @xmath1712.26 + & & & & & & + msft @xmath169 & @xmath206 & @xmath1711.76@xmath171 & @xmath207 & @xmath1711.19@xmath171 & 0.512 & @xmath1713.12 + wmt @xmath169 & @xmath208 & @xmath1711.77@xmath171 & @xmath209 & @xmath1711.19@xmath171 & 0.510 & @xmath1713.13 + mrk @xmath169 & @xmath210 & @xmath1712.32@xmath171 & @xmath211 & @xmath1711.40@xmath171 & 0.509 & @xmath1715.37 + ibm @xmath169 & @xmath212 & @xmath1712.68@xmath171 & @xmath213 & @xmath1711.55@xmath171 & 0.509 & @xmath1717.21 + xom @xmath169 & @xmath214 & @xmath1712.86@xmath171 & @xmath215 & @xmath1711.63@xmath171 & 0.510 & @xmath1718.20 + mo @xmath169 & @xmath216 & @xmath1712.94@xmath171 & @xmath217 & @xmath1711.66@xmath171 & 0.510 & @xmath1718.67 + [ statistics ] although this agreement is not as good for altria , exxon , ibm , and merck , this is only because we were comparing it with the rademacher distribution with @xmath167 .
we find that the distribution of standardized yields for these stocks is instead the generalized rademacher distribution with a @xmath134 slightly greater than @xmath218 .
using eq .
@xmath219 and the sample mean for these stocks , we have solved for a probability , @xmath134 , for the stocks .
we find that this probability is in close agreement with those listed in table iii for these stocks , and when this @xmath134 is then used in eqs .
@xmath220 and @xmath221 to predict values for the skewness and kurtosis , the predicted values are now in agreement with the sample skewness and kurtosis at the 95% cl .
we also note that variance of the distribution calculated using the values for @xmath134 given in table 4 in eq .
@xmath222 ranges from 0.9994 to 1.000 .
this is in excellent agreement with the requirement that the variance of the standardized daily yield is one when @xmath52 is one trading day .
we thus conclude that the distribution of standardized daily yields is a generalized rademacher distribution shifted by the drift , @xmath144 . for 20 of these stocks
, we find that @xmath223 at the 95% cl .
the probability that the daily yield increases is the same as the probability that it decreases . for the other four stocks ,
@xmath134 is slightly greater than @xmath218 , and the probability that the daily yield increases is slightly larger than the probability that it decreases .
having determined the distribution for the standardized daily yield , we now turn our attention to determining the volatility of the stock .
we find that while the recursion relation , eq .
@xmath88 , is straightforwardly solved using the @xmath90 given in table ii , there is a great deal of noise associated with the resultant values for @xmath65 .
this can be seen in fig .
@xmath224a where we have plotted as a function of trading day the volatility obtained from eq . @xmath88 .
although we can discern that there is an inherent structure in graph , this structure is buried within random fluctuations of @xmath65 .
these fluctuations are due to random noise generated when eq
. @xmath88 is solved , and they mask the functional dependence of @xmath225 on @xmath17 . in this section , we will extract this dependence from the noise .
the presence of the noise in @xmath65 is inherent , but not because @xmath21 itself obeys a stochastic process , as is assumed in stochastic volatility models .
if it were , then there will necessarily be a second stochastic differential equation for @xmath21 to augment eq .
@xmath47 , and the two coupled equations would have to be solved simultaneously .
certainly , eq .
@xmath47 and the recursion relation eq .
@xmath88 would not , in general , be solutions of the coupled stochastic differential equations , and it is this recursion relation that was used to obtain fig .
rather , this noise is inherent in determining the volatility itself .
note from eq .
@xmath88 that @xmath226 . for a stochastic process of the form eq .
@xmath38 where the volatility changes with time , at each time step , @xmath227 , @xmath64 is a random variable from a distribution with volatility @xmath65 . as @xmath65 need not equal @xmath228 for any two @xmath53 and @xmath229
, each @xmath64 can come from a _ different _ distribution . in the worst case
, we will have only _ one _ @xmath64 out of any distribution with which to determine @xmath65 , and this @xmath64 can take any value from @xmath230 to @xmath231 with a probability @xmath232 determining @xmath65 would thus seem to be an impossible task . that it can nevertheless be done
is due to three observations .
first , because @xmath233 is gaussian , there is a 68% probability that any value of @xmath234 will be within @xmath235 .
it is for this reason that it is still possible to discern an overall functional dependence of @xmath65 on the trading day through the noise in fig .
second , @xmath21 is a _ deterministic _ function of @xmath17 , and thus the value of the volatility at time step @xmath227 is related to its value at time step @xmath236 . given a sufficient number of @xmath64and thus a sufficient number of @xmath65it must be possible to construct a functional form for @xmath21 .
third , using fourier analysis ( also called spectral analysis ) and signal processing techniques , it is possible to remove from fig .
@xmath224 the noise that is obscuring the details of how @xmath65 depends on the trading day , and obtain a functional form for the volatility .
noise in the discrete volatility , @xmath65 that fourier analysis provides an efficient way of removing the noise from fig .
@xmath224 is based on the following theorem : * theorem : * if @xmath237 is a time series where @xmath238 is a gaussian random variable with zero mean , and volatility , @xmath225 , then the fourier sine , @xmath239 , and fourier cosine , @xmath240 , coefficients of the fourier transform of @xmath238 are gaussian random variables with zero mean and volatility @xmath241 .
this theorem is well - known in signal analysis , and is an immediate consequence of parseval s theorem . a proof of this theorem , as well as a review of the discrete fourier transform , is given in appendix @xmath242 .
it is because the volatility of the fourier sine and cosine coefficients for gaussian random variables are reduced by a factor of @xmath1 that it is possible to remove from @xmath65 the random noise . in general
, this reduction in the coefficients does not occur if the @xmath238 are _ not _ random variables , and thus the structure in fig .
@xmath224 can be resolved once the fourier transform of @xmath65 is taken . after this removal
is accomplished , we can then take the inverse fourier transform to obtain @xmath21 , which we call the instantaneous volatility to differentiate it from the @xmath65 that comes directly from eq .
@xmath88 .
figure @xmath224b and c are plots of the fourier sine and fourier cosine coefficients of the discrete fourier transform of @xmath65 defined as @xmath243 they depend on an integer @xmath0 , which runs from @xmath244 to @xmath245 .
as @xmath246 the fourier transform decomposes the time - series , @xmath65 , into components that oscillate with frequency @xmath247 day@xmath248 ( or , equivalently , with period @xmath249 days ) for @xmath250 ; the coefficients @xmath251 and @xmath252 are the amplitudes of these oscillations . in the graphs shown in figs .
@xmath224b and@xmath224c , we can readily see that there is a component of the fourier coefficients for coca cola that varies randomly between @xmath253 .
this is the noise floor .
coefficients in this floor are the result of the fourier transform of the noise that mask the functional behavior of @xmath65 on the trading .
this noise floor is similar for both the fourier sine and cosine coefficients , as can be see in detail in inset plot in fig .
@xmath224c where the features of the plot of @xmath254 are magnified for @xmath0 between @xmath255 .
it is also apparent from the graph that there are fourier coefficients that rise above the noise .
while the most prominent of these is @xmath256 ( which is the average of @xmath257 over all trading days ) , such points exist for other coefficients as well .
this is due to the structure in @xmath65 shown in fig .
@xmath224a ; if there were no structure at all in the plot , then there would be no fourier coefficients that rise above the noise floor . the instantaneous volatility , @xmath21 , after noise removal by combining this observation with the near uniformity of the noise floor , we are able to filter out the noise component of @xmath65 , and construct a ( approximately ) noise - free instantaneous volatility , @xmath21 .
a description of the process that we used , along with the statistical criteria used to determine the noise floor for the fourier sine and cosine coefficients , is given in detail in appendix @xmath258 .
the effectiveness of the noise removal process can be seen in fig .
@xmath259 where a plot of the instantaneous volatility for coca cola is shown . when this plot is compared to fig .
@xmath224a , the amount of noise removed , and the success of the noise removal procedure , is readily apparent .
indeed , out of a total of 21,523 fourier sine and cosine coefficients for @xmath65 , 10,734 fourier sine and 10,728 of fourier cosine coefficients were removed as noise ; only 59 points were kept to construct @xmath21 . while the graph of @xmath21 may appear to be noisy , this is because eight decades of trading days are plotted in the figure .
much of this apparent noise disappears when the range of trading days plotted is narrowed , as can be seen in the inset figure . here ,
the instantaneous volatility over a one - year period from december 29 , 2005 to december 29 , 2006 has been plotted . to compare the instantaneous volatility with the historical volatility , we have included in fig .
@xmath259 a graph of historical volatility calculated from the daily yield using a 251-day moving average .
it is immediately apparent that the historical volatility is generally larger than the instantaneous volatility ; at times it is dramatically so .
it is also readily apparent that the historical volatility does not show nearly as much detail as the instantaneous volatility , as can be seen in the inset figure .
a functional form for @xmath21 can be found for all 24 stocks .
for coca cola , this expression has 59 terms ; we give only four of them here , @xmath260 t ) \bigg ] , \label{instantaneoussigma}\end{aligned}\ ] ] where @xmath261 rad / day is the fundamental angular frequency .
the amplitude of the first term in the expression is the largest ; it is the average of @xmath65 over all the trading days in the time - series .
the second largest amplitude is the sine term in eq .
@xmath262 , and it is 25% the size of the first .
all other amplitudes are smaller then this term , for most by a factor of 5 , and yet notice from fig .
@xmath259 that these amplitudes are nonetheless sufficient to generate a instantaneous volatility that is far from a constant function . from the last term in eq .
@xmath262 , we see the that shortest frequency of oscillations that make up @xmath21 is @xmath263 day@xmath248 .
this is very close to the nyquist criteria of @xmath264 day@xmath248 for @xmath21 , which is the upper limit on the frequencies of the fourier components of @xmath21 .
the underlying reason for such a limit is because the the original time - series , @xmath50 , was acquired once each trading day .
we therefore can not _ measure _ oscillations with a period shorter than two trading days ; there simply is not enough information about the stocks to determine what happens within the trading day .
( this is in contrast to _ predicting _ how the volatility may behave during the trading day , which certainly can be done . ) for each of the 24 stocks , the shortest period of the fourier components that make up the instantaneous volatility are listed in table iii , and we see that for all but 3 of the stocks our expression for @xmath21 comes very close to nyquist criteria . in the case of alcoa , caterpillar , and johnson & johnson , the shortest period has even reached it .
lrrrrrr cat & 2.0 & 3.33 & 3.07 & 3.36 & -0.04 & 3.04 + jnj & 2.0 & 3.20 & 2.99 & 3.23 & -0.03 & 2.99 + aa & 2.0 & 3.40 & 3.00 & 3.47 & -0.02 & 3.00 + ge & 2.1 & 3.47 & 3.01 & 3.44 & -0.04 & 3.00 + hpq & 2.1 & 3.45 & 3.00 & 3.31 & -0.08 & 3.01 + dis & 2.1 & 3.42 & 2.99 & 3.64 & -0.03 & 3.03 + & & & & & & + xom & 2.2 & 3.79 & 3.11 & 3.61 & -0.05 & 3.03 + msft & 2.3 & 3.51 & 2.99 & 3.27 & -0.08 & 3.00 + pfe & 2.3 & 3.38 & 3.00 & 3.40 & -0.05 & 3.00 + axp & 2.4 & 3.56 & 3.17 & 3.33 & -0.04 & 3.14 + mrk & 2.4 & 3.33 & 3.03 & 3.95 & -0.03 & 3.00 + intc & 2.4 & 3.26 & 3.00 & 3.36 & -0.05 & 2.99 + & & & & & & + wmt & 2.5 & 3.47 & 2.96 & 3.29 & 0.11 & 2.99 + ko & 2.5 & 3.89 & 3.30 & 3.53 & 0.02 & 3.23 + ba & 2.7 & 3.37 & 3.00 & 3.30 & -0.01 & 2.99 + pg & 3.2 & 3.31 & 2.95 & 3.40 & -0.04 & 3.00 + aig & 3.2 & 3.44 & 3.00 & 3.32 & 0.01 & 3.00 + mmm & 3.5 & 3.03 & 2.54 & 3.06 & -0.04 & 2.58 + & & & & & & + ibm & 4.0 & 3.41 & 3.00 & 3.64 & 0.03 & 3.00 + vz & 5.2 & 3.28 & 2.99 & 3.23 & -0.09 & 3.00 + mo & 5.5 & 3.56 & 3.14 & 3.64 & -0.07 & 3.21 + c & 6.2 & 3.36 & 3.02 & 3.33 & 0.05 & 3.12 + dd & 9.6 & 3.81 & 3.31 & 3.64 & 0.00 & 3.12 + gm & 14.2 & 3.59 & 3.00 & 3.64 & -0.06 & 3.00 + [ fouriersummary ] in figs .
@xmath265 and @xmath266 , we have graphed the instantaneous volatility as a function of trading day for all 24 stocks .
they have been ordered into graphs where the degree of volatility are similar , with the stocks with roughly the highest volatility graphed last .
analytical expressions for the other 23 stocks are not given as they are too lengthy .
with @xmath21 now known and the drift for the standardized daily yield obtained previously , the drift for the daily yield , @xmath267 , can be found for all 24 stocks using the discretized version of eq .
@xmath68 , @xmath268 , where the @xmath21 is the instantaneous volatility obtained above .
since @xmath269 for all 24 stocks , @xmath270 .
thus for all of the 24 stocks , the drift of the stock is smaller than the volatility of it .
this is to be expected .
if the drift of a stock is larger than the volatility , then future trends in the stock can be predicted with a certain degree of certainty ; the drift is , after all , a _ deterministic _ function of time . such trends could be seen by investors , and nearly riskless profits could be made .
this clearly does not happen .
it is instead very difficult to discern future trends in the price of stocks , and this is precisely because the volatility of the stock is so large . the instantaneous volatility for the djia stocks , i the instantaneous volatility for the djia stocks , ii
as a continuous process , we have found that the 24 djia stocks can be described as a stochastic process with a volatility that changes deterministically with time .
it is a process for which the autocorrelation function of the yield vanishes at different times , and thus one that describes a stock whose price is efficiently priced . from the results of our calculation of the autocorrelation function of the daily yield for the 24 stocks ,
this property of our stochastic process is in very good agreement with how these stocks are priced by the market .
it is also a process for which the solution of the stochastic differential can be , at least formally , solved .
this solution is valid only because the volatility is a deterministic function of time , however .
if the volatility is depends on the stock price , or if the volatility itself is a stochastic process , the solution of the stochastic differential equation will not be so simple , and the autocorrelation function need not vanish at different times .
it is , however , only after using the discretized stochastic process that we are able to validate our model . after correcting for the variability of the volatility by using the standardized daily yield
, we have shown that for all 24 stocks the distribution of standardized daily yields is well described by the general rademacher distribution . indeed , we found that the abnormally large kurtosis is due to a volatility that changes with time . for 20 of the 24 stocks , the sample skewness , kurtosis , and probability distribution agrees with a rademacher distribution where @xmath167 at the 95% cl ; the probability that these stocks will increase on any one day is thus equal to the probability that it will decrease .
the other four stocks agree with a generalized rademacher distribution and have a @xmath134 slightly greater than @xmath218 .
for these stocks , the probability that the yield will increase on any one day is slightly higher than the probability that it will decrease .
we conclude that our model is a very good description of the behavior of these stocks . that the kurtosis for the standardized daily yield is smaller than the kurtosis for the daily yield is in agreement with the results found by rosenberg ( 1972 ) .
the daily yield is time dependent , and is thus a nonstationary random variable , while for the standardized daily yield , the time dependence due to the volatility has been taken account of .
indeed , in many ways rosenberg ( 1972 ) presages the results of this work . by combining the properties of our continuous stochastic process for the stocks with noise removal techniques , we have been able to determine the time dependence of both the volatility and the drift of all 24 stocks . unlike the implied volatility
, the volatility obtained here was obtained from the daily close directly without the need to fit parameters to the market price of options .
the theory is thus self - contained . for alcoa , caterpillar , and johnson&johnson
, the time dependence of the volatility can be determined down to a resolution of a single trading day , while for another 13 stocks , they can be determined to a resolution of less than 1 1/2 trading days .
while other , more sophisticated signal analysis techniques can be used , given that the time - series is based on the daily close and thus the resolution is ultimately limited to a period of two trading days , we do not expect that it will be possible to dramatically improve on these results .
only when intraday price data is used will we expect significant improvement to this resolution . indeed , with intraday data we expect that changes to the volatility that occur during the trading day can be seen .
we have deliberately used large cap stocks in our analysis , and we take care to note that this approach to the analysis of the temporal behavior of stocks have only been shown to be valid for the 24 stocks we analyzed here . while we would expect it to be applicable to other large - cap stocks , whether our approach will also be valid when applied to mid- or small - cap stocks is still an open question .
indeed , it will be interesting to see the range of stocks for which the volatility depends solely on time . with both the drift and the volatility determined down nearly to the single trading day level for most of the stocks , it is now be possible to calculate the autocorrelation function for both , as will as the correlation function between the drift and the volatility .
in particular , the degree of influence that the volatility or drift on any one day has on the volatility or drift on any future day can be determined .
this analysis is currently being done .
the time - series for the 24 djia stocks analyzed here were obtained from the center for research in stock prices ( crsp ) .
while the ending date for each series is december 29 , 2006 , the choice of the starting date is often different for different stocks .
this choice of starting dates was not governed by a desire for uniformity , but rather by the desire to include as many trading days in the time series as possible , and thereby minimize standard errors .
in addition , by maximizing the number of trading days included , we also demonstrate that our model is valid over the entire period for which the prices of the stock are available .
although the daily close of stocks between the starting and ending dates are used as the basis of the time - series ( with dividends included in the price ) , a series of adjustments to the crsp data were made when the series were constructed .
if the closing price of stock is listed by crsp as a negative numberan indication that the closing price was not available on that day , and the average of the last bid and ask prices was used insteadwe took the positive value of this number as the daily close on that day . if no record of the daily close was given for a particular trading day at allan indication that the bidding and asking prices were also not available for that datewe used the average of the closing price of the stock on the day preceding and the day following as the daily close for that day .
next , the daily close of the stock prices were scaled to adjust for splits in the stock . for example , although the daily close for coca cola on december 12 , 1925 is listed by crsp as $ 153.625 , this price was scaled by a factor of 6745.134 to account for the accumulated splits that the stock has gone through since 1925 .
the price recorded in the time - series is instead 0.022776 .
because of this scale factor , the prices of stocks are listed in all time - series to an accuracy of at least @xmath271 to ensure that the daily close on any day can be reconstructed from the time - series .
this level of accuracy or higher was then used in all the calculations in this paper .
while we could have avoided this subtlety by scaling the daily close , $ 48.25 , of the stock on december , 29 , 2006 by @xmath272 , doing so would result in stock prices that are $ 300k , which is deceivingly large . finally , from eq .
@xmath88 we see that @xmath65 vanishes if @xmath64 vanishes , and yet from eq .
@xmath273 , it is implicit that the quotient @xmath274 must be well defined .
indeed , the reduction of eq .
@xmath38 to eq .
@xmath47 is only valid if @xmath21 is nowhere zero . in practice , there are trading days on which the daily yield vanishes ; for coca cola , this occurred 2070 out of a total of 21,523 trading days . to ensure that @xmath274 is well defined on these days ,
we have added to the daily close a random number less than @xmath275 if the close on successive days are equal .
this is done before the daily close is scaled to adjust for stock splits . since the stock price changes by at least $ 0.01 increments , doing so does not materially change the stock price , while still insuring that @xmath276 .
we have not adjusted for inflation in our time - series , nor have we accounted for weekends , holidays , or any other days on which trading did not take place .
we have instead concatenated the daily close on each trading day , one after another , when constructing the time - series .
the time - series are thus a sequence of _ trading _ days , and not calendar days . while this concatenation is natural , issues of bias such as those studied by fleming , kirby , and ostdiek ( 2006 ) have not been taken into account .
whether these issues are relevant for the stocks considered here we leave for further study .
our focus is instead on the gross features of the stock price .
finally , we list here the following particularities that occurred in our analysis of the 24 djia stocks . _
boeing : _ when solving for @xmath65 , the @xmath277 term was greater than 27 , while all other terms was less than @xmath264 .
this data point was an outlier , and since it is in the transient region for the stock , we have set this term equal to @xmath278 , which is the typical size of @xmath65 for @xmath279 . _
merck : _ when solving for @xmath65 , the @xmath277 term was greater than 168 , while all other terms were three orders of magnitude smaller .
this data point was replaced by @xmath280 , which is the typical size of @xmath65 for @xmath279 . _ exxon - mobil : _ when solving for @xmath65 , the @xmath277 term was greater than 158 , while the @xmath281 term was greater than 1500 .
the @xmath277 data point was replaced by @xmath278 , and the @xmath281 data point was replaced by @xmath282 .
in this section , we collect the expressions used here in calculating the mean , variance , skewness , kurtosis , and autocorrelation of the time - series , along with their respective standard errors . with the exception of the autocorrelation function ,
these expressions are taken from stuart and ord ( 1994 ) .
given a collection of @xmath2 data points , @xmath257 , the sample moments , @xmath283 , of order , @xmath0 , that are used in our analysis are defined as follows @xmath284 as usual , the sample skewness and kurtosis are defined as @xmath285 while the standard error of the mean and the variance is well known , @xmath286 the standard error in the sample skewness and kurtosis are not .
for the skewness , this error is @xmath287 while for the sample kurtosis , the standard error is @xmath288 although standard errors are defined in terms of the population moments , these moments are not known a prior . following stuart and ord ( 1994 ) ,
we have used instead the sample moments listed in eq .
@xmath289 when calculating standard errors . for the time - series , @xmath257 , where @xmath290 , we define the autocorrelation of @xmath291 to be @xmath292 equation
@xmath5 measures the correlation of the time - series at time step @xmath2 with the time - series at time step @xmath293 .
this definition differs somewhat from the one given in kendall ( 1953 ) and in kendall , stuart and ord ( 1983 ) in that they divide @xmath294 by the product of the volatility of the time series , @xmath295 with the volatility of of the time - series , @xmath296
. it also differs substantially from the expression used in alexander ( 2001 ) , where a simplified expression for the autocorrelation function in kendall , stuart , and ord ( 1983 ) is used .
we use eq .
@xmath5 instead of the expressions given in kendall , stuart , and ord ( 1983 ) and alexander ( 2001 ) for two reasons .
first , @xmath297 is simply the variance of the time - series , so that the volatility for a stock can be read off easily from its graph , as can be seen in fig .
second , we will see below that the variance for @xmath298 is easily calculated when @xmath257 are gaussian random variables , and the standard error for @xmath299 can be readily determined .
derivations of the standard error for the autocorrelation functions given in kendall , stuart , and ord ( 1983 ) and alexander ( 2001 ) , on the other hand , are more involved . to determine the standard error for @xmath298 , consider a time - series where the @xmath257 are gaussian random variables with mean zero and standard deviation , @xmath300
. then @xmath301=0 $ ] while @xmath302 = \sigma^2\delta_{ik}$ ] .
( here , @xmath303 is the kronecker delta with @xmath304 if @xmath305 while @xmath306 otherwise . ) consequently , @xmath307 = 0 $ ] when @xmath308 , as can be seen from eq .
we thus only have to calculate @xmath309= \frac{1}{(n - m)^2}\sum_{i = m+1}^n \sum_{k = m+1}^n e[x_ix_{i - m}x_kx_{k - m}].\ ] ] since the @xmath257 are gaussian random variables , @xmath310 = \sigma^4\left(\delta_{ij}\delta_{kl}+\delta_{ik}\delta_{jl}+\delta_{il}\delta_{jk}\right).\ ] ] thus , @xmath311 & = & \frac{\sigma^4}{(n - m)^2}\sum_{i , k = m+1}^n \big(\delta_{i , i - m}\delta_{k , k - m}+\delta_{ik}\delta_{i - m , k - m}+ \delta_{i , k - m}\delta_{i - m , k}\big )
. \nonumber \\ \label{var}\end{aligned}\ ] ] the first term vanishes since @xmath308 , while the third term vanishes because it requires that @xmath312 and @xmath313 ; this can only happen when @xmath314 .
we are thus left with only the second term , so that @xmath309= \frac{\sigma^4}{n - m}.\ ] ] the standard error , @xmath315 , for @xmath294 when @xmath308 is then simply @xmath316 where we have used the fact that @xmath297 is the variance of the time - series . the standard error for @xmath298
when @xmath314 can then be found from eq .
@xmath317 after remembering that @xmath318 for a gaussian distribution .
note that differences between eq .
@xmath9 and the standard error found in kendall , stuart , and ord ( 1983 ) are due mainly to our defining @xmath298 with the factor @xmath319 instead of the factor @xmath320 used by them .
although the standard error for @xmath298 when @xmath257 is not a gaussian random variable can be found for special cases ( see kendall , stuart , and ord 1983 ) , eq .
@xmath9 is sufficient for our purposes .
if the market is efficient , we expect the autocorrelation function for the standard yield to vanish for @xmath308 . as this expectation
is borne out by fig .
@xmath4 , we hypothesize that the reason why the autocorrelation function in fig .
@xmath4 is not identically zero when @xmath11 is due to sample errors , which in turn is due to gaussian random variables with zero mean .
we would therefore expect eq .
@xmath9 to be a good description of the standard error of this autocorrelation function , and indeed , this expectation is consistent with the results listed in table i.
in this appendix , we review the properties of the fourier transform needed in the analysis we present in this paper .
while much of this is well - known , our purpose here is to establish the notation used in the paper , and to review the properties of fourier series needed . at the end of this section
, we will also show that the fourier transform of a gaussian random variable is once again a gaussian random variable , and describe the method used to remove the noise from @xmath65 .
consider a times - series , @xmath257 , such that @xmath321 ; the total number of data points in the time - series is then @xmath322 . in the analysis below
, we will assume that @xmath2 is an odd number .
as there are at least 5,000 trading days in our time - series , we can always change the starting point of a time - series by one trading day to insure that there are an odd number of terms in the series ; this assumption is thus not restrictive .
the expansion of the time - series in a fourier series is defined as @xmath323 where @xmath324 .
the quantity , @xmath325 , is called the fourier transform of @xmath257 . as @xmath326 in taking the fourier series of @xmath257 we have decomposed @xmath257 into terms that oscillate with definite period , @xmath327 , for @xmath250 , and have a definite amplitude , @xmath328 .
this transform is thus a natural method of characterizing how a time - series changes with time .
the amplitude of these oscillations , @xmath325 , is a complex number in general .
the original time - series , @xmath257 , is real , however , and this fact must also be reflected in @xmath325 .
how it is reflected can be seen by taking the complex conjugate of eq .
@xmath329 , @xmath330 where the complex conjugate is denoted by a bar .
since @xmath331 , by comparing eq .
@xmath329 with eq .
@xmath332 we find after taking @xmath333 in eq .
@xmath329 the reality condition @xmath334 that the fourier transform must satisfy .
the transform eq .
@xmath329 is invertible .
namely , we can express @xmath325 in terms of @xmath257 by taking the following sum @xmath335^n , \label{inverse}\ ] ] of eq .
@xmath329 . to evaluate this sum
, we consider first the case where @xmath336 for any integer @xmath337 .
the series on the right can then be summed to give @xmath338}\left(\frac{1-e^{2\pi i(\hat{k}-k)}}{1-e^{2\pi i(\hat{k}-k)/n}}\right ) , \label{sum2}\ ] ] after using the following identity for the geometric series , @xmath339 as @xmath340 , while @xmath341 , we conclude that eq .
@xmath342 vanishes in this case .
we next consider the case when @xmath343 .
each term in the sum is then one , and eq .
@xmath342 is easily summed to give @xmath2 .
combining these two results , we find that @xmath344 we then conclude from eq . @xmath345 that @xmath346 this is the inverse fourier transform of @xmath257 . in particular ,
notice that when @xmath347 , @xmath348 is simply the average of @xmath257 over the whole time - series .
the fourier series eq .
@xmath329 can also be expressed as an explicitly real expansion , @xmath349 where the amplitudes @xmath350 are called the fourier cosine and fourier sine coefficients , respectively . while analytical calculations are more easily done with eq .
@xmath329 , numerical calculations are necessarily done with eq .
@xmath351 , and it is on the fourier sine and cosine coefficients that we will focus most of our analysis in this paper . because each @xmath238 in the time - series is a gaussian random variable with zero mean and variance , @xmath300 , the probability distribution for the time - series is just @xmath352 where @xmath52 is the time interval between successive points in the time series , and in the last equality we have used @xmath353 = 0 $ ] for @xmath354 . expanding @xmath238 in a fourier series using eq .
@xmath329 , we find that @xmath355 = \sum_{k , k'=-(n-1)/2}^{(n-1)/2}\xi^\omega_k \xi^\omega_{k'}n \delta_{k , -k ' } , \label{krondelta}\ ] ] where the last equality holds from eq . @xmath356 .
thus , @xmath357 which is parseval s theorem for a discrete fourier series .
following eq .
@xmath358 , we express @xmath359 in eq .
then eq . @xmath361 can be written as @xmath362\right\ } , \label{fprob}\end{aligned}\ ] ] and the theorem is proved .
it is straightforward to see that the converse is also true .
namely , if @xmath363 and @xmath364 are gaussian random variables with zero mean and volatility @xmath241 , then @xmath257 is a gaussian random variable with zero mean and volatility , @xmath225 .
notice from eq .
@xmath365 that @xmath366 = 0 $ ] , and thus the two random variables are independent .
notice also that while we began with @xmath2 degrees of freedom with the random variables , @xmath257 , we seem to have ended up with @xmath367 degrees of freedom for the random variables , @xmath240 and @xmath368 . from eq .
@xmath358 we see , however , that @xmath369 and @xmath370 ; not all the variables in eq
. @xmath365 are independent . when this redundancy is taken account of , we arrive back to @xmath2 degrees of freedom . given that the noise floor associated with the fourier sine and cosine coefficients is constant over all frequencies , @xmath371 , noise removal is straight forward .
we need only remove from @xmath372 , the set of all fourier sine coefficients for @xmath65 , and @xmath373 , the set of all fourier cosine coefficients for @xmath65 , those coefficients whose amplitudes is less than the amplitudes , @xmath374 and @xmath375 , of the noise floor for the fourier sine and fourier cosine coefficients , respectively .
the coefficients left over@xmath376 for the fourier sine coefficients , and @xmath377 for the fourier cosine coefficientscan then be used to construct the instantaneous volatility , @xmath21 , by summing the fourier series eq .
@xmath378 .
the noise floor amplitudes , @xmath374 and @xmath375 , are determined statistically .
consider the set of coefficients that are removed : @xmath379 for the fourier sine coefficients and @xmath380 for the fourier cosine coefficients . because the distribution of the noise floor is gaussian , @xmath374 and @xmath375
must be chosen so that the distributions of coefficients in @xmath381 and @xmath382 are gaussian as well .
if either amplitude is chosen too large , then coefficients from @xmath372 or @xmath373 that make up the signal , @xmath21 , would be included in the noise distributions as noise .
as these coefficients are supposed to be above the noise , they will skew and flatten the distribution ; the skewness and the kurtosis for the distribution of @xmath381 and of @xmath382 will then differ from their gaussian values if these coefficients are included . on the other hand , if either amplitude for the noise floor is chosen too _
, then coefficients from @xmath372 or @xmath373 that make up the noise would be _ excluded _ from the noise distributions .
as these coefficients would have populated the tails of the gaussian distribution , their removal will tend to _ narrow _ the distribution , and the kurtosis of the noise distributions will differ once again from its gaussian value .
( because the coefficients are remove symmetrically about the horizontal zero line , a choice of the amplitude for the noise floor that is too small will not tend to change the skewness significantly . ) thus , @xmath374 and @xmath375 must be chosen so that the skewness and kurtosis of the distribution of coefficients in @xmath381 and @xmath382 is as close to their gaussian distribution values as possible . while the above procedure is straightforward , there is an additional constraint .
the volatility can not be negative , and thus the resultant instantaneous volatility , @xmath21 , obtained after the noise floor is removed must the positive as well .
this constraint is not trivial . for a number of stocks
, a choice of @xmath374 and @xmath375 that results in noise distributions that are closest to a gaussian distribution also results in a @xmath21 that is negative on certain days . to obtain a @xmath21 that is non - negative , slightly larger amplitudes for the noise floors
were chosen , which resulted in a slightly larger skewness and kurtosis .
this approach to removing the noise from the volatility , @xmath65 , has been successfully applied to all 24 stocks using a simple c++ program that implements an iterative search algorithm to determine @xmath374 and @xmath375 .
the results of our numerical analysis are shown in table iv . there , we have listed the noise floor amplitudes , @xmath374 and @xmath375 , used for each of the 24 stocks .
their values are given as multiples of the standard deviation of the distribution of the fourier coefficients in @xmath381 and @xmath382 .
as these values range from 3.030 times the standard deviation to 3.948 times the standard deviation , 99.756% to 99.992% of the data points that make up a gaussian distribution can be included in these distributions if they are present in either @xmath381 or @xmath382 .
listed also in table iv are the kurtosis for @xmath381 and @xmath382 . we have found that they range in value from 2.95 to 3.31 , and are thus very close to the gaussian distribution value of three for the kurtosis .
the skewness of the noise of the distribution of @xmath382 was calculated as well , and was found to vary in value from -0.03 to 0.11 ; this also is very close to the gaussian distribution value of zero for the skewness .
the skewness for the distribution of @xmath381 was also calculated , but we find that their values are @xmath383 to @xmath101 times smaller than the skewness for the distribution of @xmath382 , and there was no need to listed these values in the table .
this extremely close agreement with the skewness of the gaussian distribution is because the fourier sine coefficients are antisymmetric about @xmath347 : @xmath384 .
the average of any odd power of @xmath385 over @xmath0and in particular , the skewness of the distribution of @xmath372thus automatically vanishes . for this reason ,
the skewness for @xmath381 is exceedingly small .
the author is also an adjunct professor at the department of physics , diablo valley college , pleasant hill , ca 94523 , and a visiting professor at the department of physics , university of california , berkeley , ca 94720 .
he would like to thank peter sendler for his support and helpful criticisms while this research was being done .
it is doubtful that this work would have been completed without his encouragement and , yes , prodding . | while the use of volatilities is pervasive throughout finance , our ability to determine the instantaneous volatility of stocks is nascent . here , we present a method for measuring the temporal behavior of stocks , and show that stock prices for 24 djia stocks follow a stochastic process that describes an efficiently priced stock while using a volatility that changes deterministically with time .
we find that the often observed , abnormally large kurtoses are due to temporal variations in the volatility .
our method can resolve changes in volatility and drift of the stocks as fast as a single day using daily close prices .
keywords : spectral analysis , noise reduction , rademacher distribution |
fluctuations of the order parameter near the critical temperature @xmath4 are much larger in high-@xmath4 superconductors than in classical low temperature superconductors .
one of the reasons lies in the higher thermal energy @xmath5 which provides the excitations , and the other in a very short coherence lengths which occur in high-@xmath4 cuprate superconductors . with these properties ,
the region of critical fluctuations was estimated from the ginzburg criterion to be of the order of 1k , or more , around @xmath4 , which renders the critical region accessible to experimental investigations.@xcite farther above @xmath4 , one expects to observe the transition from critical to noninteracting gaussian fluctuations which are the lowest order fluctuation corrections to the mean field theory.@xcite the layered structure of high-@xmath4 superconductors requires some theoretical sophistication .
one could treat these superconductors with various models from three - dimensional ( 3d ) anisotropic to coupled layers lawrence - doniach , or purely two - dimensional ( _ 2d _ ) ones . due to the temperature variation of the coherence lengths
one could even expect a dimensional crossover in some systems .
the fluctuation conductivity is altered by dimensionality in various models so that a detailed comparison of model calculations and experimental data could address the dimensionality problem . for the reasons stated above
, the fluctuation conductivity in high-@xmath4 superconductors was studied experimentally by many authors .
@xcite most of them used @xmath6 resistivity measurements .
@xcite the reports were controversial in the conclusions about the dimensionality of the system , and the critical exponents .
it has been shown that , in a wide temperature range above @xmath4 , the fluctuation conductivity did not follow any of the single exponent power laws predicted by scaling and mean - field theories.@xcite the data in the gaussian regime could be fitted by an expression derived within the ginzburg - landau ( gl ) theory with a short wavelength cutoff in the fluctuation spectrum .
recently , silva et al.@xcite have proven that the gl approach with an appropriate choice of the cutoff parameter yields result which is identical to that of the microscopic aslamazov - larkin ( al ) approach with reduced excitations of the short wavelength fluctuations.@xcite it has been further shown that the detailed temperature dependence of the fluctuation conductivity was not universal , but sample dependent . in this respect ,
the gl approach has practical advantage since the cutoff parameter can be readily adjusted in fitting the experimental data .
silva et al.@xcite could fit very well the data on a number of thin films in the gaussian region from @xmath4 + 1 k to @xmath4 + 25 k. when critical fluctuations are studied , it becomes essential to know accurately the value of @xmath4 .
however , the determination of @xmath4 from @xmath6 resistivity measurements brings about some uncertainties .
one should avoid the use of unjustifiable definitions of @xmath4 such as e. g. : ( i ) zero resistance temperature , ( ii ) midpoint of the transition , ( iii ) maximum of the derivative @xmath7 , ( iv ) intersection of the tangent to the transition curve with the temperature axis , etc .
the correct value of the critical temperature can be determined as an additional fitting parameter in the analysis of the fluctuation conductivity .
usually one assumes that a well defined power law holds in a given narrow temperature range and then determine both , @xmath4 and the critical exponent from the selected segment.@xcite however , the experimental data usually show an almost continuous change of the slope so that the uncertainty in the determination of @xmath4 is an unsolved problem . besides
, the effects of the cutoff have been neglected in the analysis of the data close to @xmath4 .
even though the values of the fluctuation conductivity near @xmath4 are not much affected by the introduction of the cutoff , the slopes can be considerably changed,@xcite and the analysis may become uncertain . a number of microwave studies have been reported showing clear signs of fluctuations in both , the real and imaginary parts of the @xmath0 conductivity.@xcite the real part @xmath8 of the complex conductivity @xmath9 has a sharp peak at @xmath4 , which is not observed in e.g. nb as a representative of low temperature classical superconductors .
@xcite the salient feature of the @xmath0 case is that the fluctuation conductivity does not diverge at @xmath4 because a finite frequency provides a limit to the observation of the critical slowing down near @xmath4 .
the real part @xmath8 has a maximum at @xmath4 .
it is also important to note that @xmath8 and @xmath10 have individually different temperature and frequency dependences , even though they result from the same underlying physics . testing
a given theoretical model becomes more stringent when two curves have to be fitted with the same set of parameters .
the expressions for the @xmath0 fluctuation conductivity in the gaussian regime have been deduced within the time dependent ginzburg - landau ( tdgl ) theory by schmidt.@xcite using general physical arguments , fisher , fisher , and huse @xcite provided a formulation for the scaling of the complex @xmath0 conductivity as @xmath11 where @xmath12 is the correlation length , @xmath13 is the dynamical critical exponent , @xmath14 is the dimensionality of the system , and @xmath15 are some complex scaling functions above and below @xmath4 .
this form of the fluctuation conductivity was claimed to hold in both , the gaussian and critical regimes .
dorsey @xcite has deduced the scaling functions in the gaussian regime above @xmath4 , and verified the previous results of schmidt.@xcite more recently , wickham and dorsey @xcite have shown that even in the critical regime , where the quartic term in the gl free energy plays a role , the scaling functions preserve the same form as in the gaussian regime .
the above mentioned theoretical expressions of the @xmath0 fluctuation conductivity did not take into account the slow variation approximation which is required for the validity of the ginzburg - landau theory.@xcite it was noted long time ago that the summation over the fluctuation modes had to be truncated at a wavevector which corresponded roughly to the inverse of the intrinsic coherence length @xmath16.@xcite the improved treatment with a short wavelegth cutoff was applied in fluctuation diamagnetism,@xcite and @xmath6 paraconductivity far above @xmath4.@xcite this approach was also applied in @xmath6 fluctuation conductivity of high-@xmath4 superconductors where one encounters a large anisotropy.@xcite the introduction of the short wavelength cutoff was found to be essential in fitting the theoretical expressions to the experimental data . in view of the great potential of the microwave method described above
, we find motivation to elaborate in this paper the improved theory of @xmath0 fluctuation conductivity including the short wavelength cutoff .
we find that the resulting expressions can be written in the form of eq .
( [ 1 - 1 ] ) .
however , the cutoff introduces a breakdown of the scaling property in the variable @xmath17 . also , we find that the phase @xmath1 of the complex conductivity ( @xmath18 ) evaluated at @xmath4 departs from the value @xmath3 when cutoff is introduced .
values of @xmath1 larger than @xmath3 were observed experimentally,@xcite but were attributed to an unusually large dynamic critical exponent . also , deviation of the scaling in the variable @xmath17 was observed already at 2 k above @xmath4,@xcite but no analysis was made considering the short wavelength cutoff in the fluctuation spectrum .
the present theory is developed for different dimensionalities which facilitates comparison with experimental data .
frequency dependent conductivity can be calculated within the kubo formalism from the current correlation function . for the fluctuation conductivity one
has to consider the current due to the fluctuations of the order parameter .
the resulting expression for the real part of the conductivity is @xcite @xmath19 where the current is assumed to be in the @xmath20-direction .
@xmath21 is the fourier component of the order parameter , and @xmath22 is the relaxation time of the @xmath23-th component .
the relaxation time for the @xmath24 mode is given by @xmath25 where @xmath13 is the dynamic critical exponent .
an alternative approach is to calculate the response of the system to an external field through the expectation value of the current operator averaged with respect to the noise.@xcite however , the introduction of the short wavelength cutoff in this approach leads only to selfconsistent implicit expressions.@xcite eq . ( [ 3 - 1 ] ) is obtained from the time dependent ginzburg - landau theory and represents the equivalent of the aslamazov - larkin fluctuation conductivity obtained from microscopic calculations . in the following
, we present the results which take account of the short wavelength cutoff in this contribution to the @xmath26 fluctuation conductivity .
the other contributions such as maki - thomson ( mt ) and one - electron density of states ( dos ) renormalization@xcite can not be treated within the time dependent ginzburg - landau theory but require microscopic calculations .
it has been shown@xcite that mt anomalous contribution in high-@xmath4 superconductors is almost temperature independent while dos contribution is strongly temperature dependent , and contains a number of parameters which have to be determined through a complex fitting procedure in an experimental data analysis.@xcite since the three terms in the fluctuation conductivity are additive it is important to have the aslamazov - larkin term corrected for short wavelength cutoff which then allows to fit the mt and dos contributions properly from the rest of the total experimental fluctuation conductivity .
the sum in eq .
( [ 3 - 1 ] ) can be evaluated by integration considering the appropriate dimensionality . in this section
we discuss the simplest case of an isotropic _
3d _ superconductor .
the integration in @xmath27-space needs a cutoff since the order parameter can not vary appreciably over distances which are shorter than some minimum wavelength .
the cutoff in @xmath28 can be expressed as @xmath29 , where @xmath30 is a dimensionless cutoff parameter .
obviously , @xmath31 would imply no cutoff in the integration , whereas for @xmath32 one obtains the usually assumed cutoff at @xmath33 . in the _ 3d _
isotropic case , the same cutoff applies also to @xmath34 and @xmath35 so that for the _
3d _ integration in the @xmath27-space one has to set the cutoff limit for the modulus @xmath36 . with the change of variable @xmath37 one obtains @xmath38}\,dq \,\,\ , , \label{3 - 2}\ ] ] where @xmath39 is the temperature dependent cutoff limit in the @xmath40-space , and
@xmath41 is a dimensionless variable which depends on frequency and temperature as independent experimental variables . for the @xmath6 case ( @xmath42 ) , and no cutoff ( @xmath31 ) , one finds from eq .
( [ 3 - 2 ] ) @xmath43 which reduces to the well known aslamazov - larkin result @xcite provided that relaxational dynamics is assumed ( @xmath44 ) , and @xmath45 is taken only in the gaussian limit as @xmath46
. however , with a finite cutoff parameter @xmath30 one obtains @xmath47 \,\,\ , .
\label{3 - 5}\ ] ] this result has been obtained by hopfengrtner et .
@xcite except that they used only the gaussian limit @xmath46 for the reduced correlation length @xmath45 .
their analysis has shown that the cutoff plays no role exactly at @xmath4 since @xmath48 regardless of @xmath30 . however , at any temperature above @xmath4 one gets a finite @xmath49 and the value of the conductivity is lowered with respect to the result given by eq .
( [ 3 - 4 ] ) .
their conclusion was that the gaussian fluctuations with no cutoff yield an overestimated fluctuation conductivity .
in this paper we are primarily interested in the ac case . before integrating eq .
( [ 3 - 2 ] ) with @xmath50 , we find the corresponding expression for the imaginary part @xmath10 .
we can apply kramers - kronig relations to each of the fourier components in eq .
( [ 3 - 1 ] ) , and carry out the summation .
this is equivalent to a calculation of the kernel @xmath51 for @xmath10 from the kernel @xmath52 used in eq .
( [ 3 - 2 ] ) , namely @xmath53 with the kernel @xmath54 , the imaginary part of the fluctuation conductivity can be calculated for any cutoff parameter @xmath30 @xmath55}\,dq \,\,\ , .
\label{3 - 7}\ ] ] finally , the complex fluctuation conductivity can be written in the form @xmath56 \,\,\ , .
\label{3 - 8}\ ] ] the prefactor is equal to the @xmath6 result with no cutoff effect as in eq .
( [ 3 - 4 ] ) .
the functions @xmath57 are given by the following expressions @xmath58 \,\,\ , , \label{3 - 9}\ ] ] @xmath59 \,\,\ , , \label{3 - 10}\ ] ] where we used the following shorthand notations @xmath60 @xmath61 @xmath62 it can be easily verified that the @xmath57-functions given by eq .
( [ 3 - 9 ] ) and eq . ( [ 3 - 10 ] ) have proper limits . in the @xmath6 limit ( @xmath63 )
, one finds that @xmath64 , and @xmath65 leads to the @xmath6 result of eq .
( [ 3 - 5 ] ) .
one can also verify that the @xmath0 results obtained previously by schmidt @xcite and dorsey @xcite can be recovered from our eq .
( [ 3 - 9 ] ) and eq .
( [ 3 - 10 ] ) in the limit @xmath31 , i. e. when no cutoff is made
. the effects of the cutoff are not trivial in the @xmath0 case .
it is essential to examine those effects in detail as they have strong bearing on the analysis of the experimental data .
the prefactor in eq .
( [ 3 - 8 ] ) depends only on temperature while the cutoff parameter @xmath30 is found only in the @xmath66-functions .
therefore , the effects of the cutoff can be studied through the @xmath67-functions alone .
we can look at the temperature and frequency dependences of these functions with and without the cutoff .
[ fig1](a ) shows a set of @xmath68 curves as functions of @xmath45 for three different frequencies . far above @xmath4
the relaxation time @xmath69 is so short that @xmath70 for any of the chosen frequencies .
therefore the response of the system is like in the @xmath6 case . with no cutoff
, @xmath68 saturates to unity ( dashed lines in fig .
[ fig1](a ) ) .
this limit is required in order that @xmath8 from eq .
( [ 3 - 8 ] ) becomes equal to @xmath71 in eq .
( [ 3 - 4 ] ) .
if a cutoff with a finite @xmath30 is included , @xmath68 decays at higher temperatures ( solid lines in fig .
[ fig1](a ) ) .
the reduction of @xmath68 is more pronounced at smaller values of @xmath72 since the integration in the @xmath40-space is terminated at a lower value @xmath73 . at higher temperatures
the conductivity @xmath8 at any frequency behaves asymptotically as @xmath74 given by eq .
( [ 3 - 5 ] ) . at temperatures closer to @xmath4 the relaxation time @xmath69 increases as @xmath75 with increasing correlation length according to eq .
( [ 3 - 1a ] ) which is usually termed critical slowing down . when @xmath76 for a given frequency , @xmath68 is sharply reduced and vanishes in the limit of @xmath4 . with the diverging prefactor in eq .
( [ 3 - 8 ] ) it can still yield a finite @xmath8 at @xmath4 .
it may appear from fig .
[ fig1](a ) that cutoff makes no effect when @xmath4 is approached , but we show later that an important feature still persists in @xmath8 . obviously , at lower operating frequencies one needs to approach @xmath4 closer so that the critical slowing down could reach the condition @xmath76 .
one can see from fig .
[ fig1](a ) that for frequencies below 1 ghz one would have to approach @xmath4 closer than 1 mk in order to probe the critical slowing down in fluctuations .
the higher the frequency , the farther above @xmath4 is the temperature where the crossover @xmath76 occurs .
this feature expresses the scaling property of the conductivity in frequency and temperature variables .
however , the scaling property holds strictly only in the absence of the cutoff .
namely , if one sets @xmath77 , the function @xmath68 depends only on the scaling variable @xmath78 .
[ fig1](b ) shows the same set of curves as in fig .
[ fig1](a ) , but plotted versus @xmath78 .
the three dashed curves from fig .
[ fig1](a ) coalesce into one dashed curve in fig .
[ fig1](b ) , thus showing the scaling property in the absence of cutoff . however , the full lines representing the functions @xmath68 with a finite cutoff parameter @xmath30 do not scale with the variable @xmath78 .
the reason is that the function @xmath68 then depends also on @xmath49 , which itself is not a function of @xmath78 .
namely , the cutoff in the @xmath40-space depends on the properties of the sample , and on the temperature , but not on the frequency used in the experiment .
hence , the cutoff brings about a breakdown of the scaling property in frequency and temperature .
the effect is more pronounced at temperatures farther above @xmath4 where the cutoff is stronger .
the properties of the function @xmath79 are shown in fig .
[ fig2 ] for the same set of three measurement frequencies as in fig .
[ fig1 ] . when plotted versus @xmath45 , the function @xmath79 exhibits a maximum at the point where the corresponding function @xmath68 shows the characteristic crossover due to @xmath80 as discussed above .
when @xmath4 is approached , @xmath79 tends to zero .
when @xmath79 is multiplied with the diverging prefactor in eq .
( [ 3 - 8 ] ) , one finds a finite @xmath10 at @xmath4 . far above @xmath4 ,
the function @xmath79 vanishes , regardless of the cutoff .
this is consistent with the behavior of @xmath68 .
namely , at high enough temperatures , @xmath68 acquires asymptotically the @xmath6 value , as seen in fig .
obviously , the imaginary part of the conductivity must vanish when @xmath6 like limit is approached .
the decrease of the function @xmath79 at higher temperatures is very rapid so that the effects of the cutoff are unnoticeable on the linear scale . only with the logarithmic scale used in the inset to fig .
[ fig2](a ) , one observes that the cutoff effects are present also in @xmath79 , though by a very small amount .
[ fig2](b ) shows the scaling property of @xmath79 with no cutoff and its breakdown when cutoff is included .
we have noted above that both , @xmath68 and @xmath79 tend to zero when @xmath4 is approached .
also , the effect of cutoff is seen to be small in that limit . yet , these functions are multiplied by the diverging prefactor in eq .
( [ 3 - 8 ] ) , and then may yield finite @xmath8 and @xmath10 . a careful analysis is needed in order to find the phase @xmath1 of the complex conductivity ( @xmath81 ) at @xmath4 . for a _
3d _ isotropic superconductor , dorsey @xcite has predicted @xmath82 , i. e. @xmath83 at @xmath4 .
his result was obtained with no cutoff and it remains to be seen if this property is preserved even when a finite cutoff is made .
[ fig3](a ) shows @xmath68 and @xmath79 as functions of @xmath84 for 100 ghz frequency .
the effects of cutoff on @xmath79 are noticeable only far above @xmath4 .
closer to @xmath4 , the curves for @xmath79 calculated with , and without cutoff , are indistinguishable . in contrast
, a finite cutoff reduces the values of @xmath68 even in the limit of @xmath4 . as a result ,
the final cutoff parameter @xmath30 yields a crossing of the curves for @xmath68 and @xmath79 at some temperature slightly above @xmath4 .
it is better seen on an enlarged scale in fig .
[ fig3](b ) .
this is a surprising result which has bearing on the experimental observations . due to the cutoff ,
the condition @xmath85 ( @xmath86 ) is reached at a temperature slightly above @xmath4 . since both , @xmath68 and @xmath79
are multiplied with the same prefactor in eq .
( [ 3 - 8 ] ) , one finds that the crossing of @xmath87 and @xmath88 does not occur at the peak of @xmath89 , but at a slightly higher temperature .
exactly at @xmath4 , @xmath10 is higher than @xmath8 since a finite cutoff parameter reduces @xmath8 , but makes no effect on @xmath10 .
the observation that the cutoff brings about a reduction of @xmath8 at @xmath4 is worth further investigation since it can be measured experimentally .
[ fig4](a ) shows the ratio @xmath90 ( equal to @xmath91 ) at temperatures approaching @xmath4 . with no cutoff ( dashed lines in fig .
[ fig4](a ) ) , this ratio reaches unity regardless of the frequency used . a finite cutoff parameter ( @xmath92 in fig .
[ fig4](a ) ) makes the ratio equal to unity at a temperature slightly above @xmath4 , and in the limit of @xmath4 the ratio saturates at some higher value .
the saturation level is seen to be higher when a higher frequency is used .
one can find analytical expansions of the @xmath67-functions in the limit of @xmath4 ( @xmath93 ) .
the leading terms are @xmath94 where we used the notation @xmath95 @xmath96 the parameter @xmath97 depends on the frequency @xmath98 and the cutoff parameter @xmath30 @xmath99 both functions tend to zero in the limit of @xmath4 ( @xmath100 ) , but their ratio is finite and depends on the parameter @xmath97 . fig .
[ fig4](b ) shows the plot of @xmath101 at @xmath4 as the function of @xmath97 .
one can observe that for a given cutoff parameter @xmath30 , the ratio @xmath101 at @xmath4 increases at higher frequencies ( lower @xmath97 ) .
the limits at @xmath4 in fig .
[ fig4](a ) represent only three selected points on the curve for @xmath101 in fig .
[ fig4](b ) . in a given experiment
, the ratio @xmath102 at @xmath4 can be directly determined from the experimental data so that the corresponding value of the parameter @xmath97 can be found uniquely from the curve of @xmath103 in fig .
[ fig4](b ) , and the cutoff parameter @xmath30 is obtained using eq .
( [ 3 - 18 ] ) .
we should note that @xmath30 is a temperature independent parameter .
it can be determined by the above procedure from the experimental data at @xmath4 , but it controls the cutoff at all temperatures
. one may observe from eq .
( [ 3 - 14 ] ) that in the limit of @xmath4 the leading terms in the expansions of the @xmath67-functions behave as @xmath104 .
taking into account the prefactor in eq .
( [ 3 - 8 ] ) one finds that @xmath8 and @xmath10 can have finite nonzero values at @xmath4 only if @xmath105 , i. e. for the purely relaxational dynamics .
we have assumed this case in all the figures of this section . from the experimental data at @xmath4
one can determine also the parameter @xmath16 .
using eq .
( [ 3 - 8 ] ) and the @xmath106-functions in eq .
( [ 3 - 14 ] ) , one obtains finite conductivities at @xmath4 @xmath107 where @xmath108 as explained above , from the ratio of the experimental values @xmath109 at @xmath4 one can determine the parameter @xmath97 , and the values of @xmath110 can then be calculated from eq .
( [ 3 - 20 ] ) .
the remaining unknown parameter @xmath16 can be obtained using eq .
( [ 3 - 19 ] ) and either of the experimental values of @xmath111 or @xmath112 .
it is also interesting to look at the plots of @xmath113 in fig .
[ fig4](b ) .
one can observe that @xmath114 saturates to unity already at small values of @xmath115 . on the other hand
, @xmath116 is smaller than unity at any finite value of @xmath97 , in conformity with the ratio @xmath117 at @xmath4 . at this point
it is useful to find the expected range of the values of @xmath97 encountered in the experiments . for the microwave frequencies in the range 1 - 100 ghz , with @xmath92 , and @xmath118 , one finds that @xmath97 is in the range 9 - 90 .
according to fig .
[ fig4](b ) , @xmath119 in this range .
this means that the cutoff makes no effect on @xmath79 , and only @xmath68 is reduced , in conformity with the calculated curves shown in fig .
most high-@xmath4 superconductors are anisotropic , some of them even having a high value of the anisotropy parameter @xmath120 . therefore ,
for practical purposes one needs adequate expressions for the @xmath0 fluctuation conductivity .
the real part of the fluctuation conductivity in the ab - plane is obtained using the kubo formalism as in the isotropic case .
one obtains @xmath121 } \,\,\ , .
\label{4 - 1}\ ] ] taking @xmath122 , and substituting the variables @xmath123 and @xmath124 , one can evaluate the sum in eq .
( [ 4 - 1 ] ) by integration in the @xmath125-plane , and along the @xmath126-axis @xmath127 } \,d q_{ab}\,d q_c \,\,\ , , \label{4 - 2}\ ] ] where we allowed a cutoff @xmath128 in the @xmath125-plane , and a possibly different cutoff @xmath129 along the @xmath126-axis .
the dimensionless parameter @xmath78 is the same as given by eq .
( [ 3 - 3 ] ) .
we use the notation @xmath84 for both @xmath130 and @xmath131 .
we may briefly examine the @xmath6 case ( @xmath132 ) . with no cutoff one obtains @xmath133 which reduces to the aslamazov - larkin result for @xmath134 ( relaxational dynamics ) and @xmath84 taken in the gaussian limit .
note that the fluctuation conductivity in the @xmath135-plane depends on @xmath136 .
finite cutoff parameters reduce the fluctuation conductivity when the temperature is increased above @xmath4 @xmath137 \,\,\ , .\end{aligned}\ ] ] this expression has not been reported in the previous literature .
the analysis of a @xmath6 fluctuation conductivity is difficult because of the number of unknown parameters .
the @xmath0 fluctuation conductivity can be obtained from the integral in eq .
( [ 4 - 2 ] ) for the real part while the imaginary part is obtained by the procedure analogous to that of the isotropic case described in the preceding section @xmath138}\ , d q_{ab}\,d q_c \,\,\ , .
\label{4 - 4a}\ ] ] the full expression can again be written in the form @xmath139 \,\,\ , .
\label{4 - 5}\ ] ] the @xmath67-functions for the _ 3d _ anisotropic case are found to be @xmath140 \,\,\
, , \end{aligned}\ ] ] @xmath141 \,\,\ , , \end{aligned}\ ] ] where we used the shorthand notations for @xmath142 as in eq .
( [ 3 - 11 ] ) , and the following @xmath143 @xmath144 } { ( 1+q_{ab}^2+q_c^2)^2\,[\omega^2+(1+q_c^2)^2]}\right ) } \,\,\ , , \label{4 - 9}\ ] ] @xmath145 @xmath146 @xmath147 @xmath148 @xmath149 @xmath150 @xmath151 the effects of cutoff are similar as those described at length in the preceding section for the simpler case of _
3d _ isotropic superconductors . in this section
we discuss only the modifications in the limit @xmath152 where the relevant parameters can be determined .
the @xmath106-functions can be expanded in the limit of @xmath4 ( @xmath153 ) , and the leading terms are @xmath154 \displaystyle \frac{1}{\sqrt{\omega } } \,\,\ , , \end{aligned}\ ] ] @xmath155 \displaystyle \frac{1}{\sqrt{\omega } } \,\,\ , , \end{aligned}\ ] ] where we used the following shorthand notations @xmath156 @xmath157 @xmath158 @xmath159 @xmath160 @xmath161 @xmath162 @xmath163 } { ( w_{ab}^2+w_c^2)^2(1+w_c^4)}\right ) } \,\,\ , .
\label{4 - 25}\ ] ] the cutoff parameters appear in @xmath164 @xmath165 we note that in the anisotropic case , the @xmath67-functions behave also as @xmath166 when @xmath167 .
as already discussed in the previous section , this implies that finite nonzero @xmath168 and @xmath169 can be obtained only for @xmath134 ( relaxational model ) .
since the available experimental data in anisotropic high-@xmath4 superconductors @xcite show finite nonzero @xmath168 and @xmath169 , we can adopt @xmath44 in the remainder of this section . in analogy to the _ 3d _
isotropic case described in the preceding section , one may define the functions @xmath170 so that the conductivities at @xmath4 are given by @xmath171 the ratio of experimental values @xmath172 at @xmath4 does not define uniquely the cutoff parameters @xmath173 and @xmath174 .
it puts , however , a constraint on their choice .
[ fig5](a ) shows the plot of @xmath175 given by eq .
( [ 4 - 28 ] ) as a function of two variables , @xmath176 and @xmath177 .
it is evident that a fixed value of @xmath178 defines a simple curve of the possible choices of ( @xmath176,@xmath177 ) .
[ fig5](b ) shows a selection of such curves for @xmath1791.05 , 1.1 , 1.15 , and 1.2 .
the dashed line marks the condition @xmath180 ( @xmath181 ) .
experimentally , one has to probe the possible choices for ( @xmath176 , @xmath177 ) and look at the fits of the theoretical curves to the experimental data at @xmath182 .
the parameter @xmath136 in eq .
( [ 4 - 5 ] ) can be obtained once the choice for ( @xmath176 , @xmath177 ) is made .
note that in practical applications of the above theory one needs measurements where the microwave current flows only in the ab - plane . particularly suitable for this purpose
are the measurements in which the superconducting sample is placed in the antinode of the microwave electric field @xmath183 in the cavity . @xcite
superconducting transition does not occur in a strictly _ 2d _ system .
however , if the sample is a very thin film so that its thickness is much smaller than the correlation length , the fluctuations will be restricted within the film thickness @xmath184 in one direction , and develop freely only in the plane of the film . using the formalism described in the preceding sections , we find that the fluctuation conductivity is given by @xmath185}\,dq \,\,\ , , \label{5 - 1}\ ] ] @xmath186}\,dq \,\,\ , , \label{5 - 2}\ ] ] @xmath187 \,\,\ , , \label{5 - 3}\ ] ] where @xmath188 and @xmath189 .
the prefactor is the aslamazov - larkin result for the _
2d _ case with no cutoff , and the gaussian form @xmath190 replaced by the more general expression @xmath191 .
the @xmath106-functions are given by @xmath192 } { ( 1+q^2+q_{n}^2)^2[\omega^2+(1+q_{n}^2)^2]}\right)}\right ] \,\,\
, , \end{aligned}\ ] ] @xmath193\right .
\label{5 - 6}\\ \nonumber & \left .
\displaystyle -\ln{\left(\frac{(1+q_{n}^2)^2[\omega^2+(1+q^2+q_{n}^2)^2 ] } { ( 1+q^2+q_{n}^2)^2[\omega^2+(1+q_{n}^2)^2]}\right ) } -2\,\frac{q^2}{1+q^2+q_{n}^2}\right ] \,\,\ , .\end{aligned}\ ] ] the summation over @xmath194 in eq .
( [ 5 - 3 ] ) has to carried out until the factor @xmath195 reaches some cutoff value @xmath30 , which is of the order of unity .
if the film thickness is large ( @xmath196 ) , one has to sum up to a high @xmath197-value . in such cases ,
the summation is well approximated by an integration , and one retrieves the _
3d _ case of the preceding section .
the _ 2d _ character is better displayed when the film thickness is comparable to @xmath16 .
then , only a few terms have to be taken into account . in the extreme case of @xmath198 , only the @xmath199 term is found below the cutoff limit . the zero frequency limit ( @xmath200 ) yields @xmath201 the @xmath199 term yields the previous result of hopfengrtner et al .
@xcite and gauzzi et al .
@xcite in the limit of @xmath4 ( @xmath93 ) one obtains @xmath202 } { ( 1+w_{n}^4)(w^2+w_{n}^2)^2}}\right ] \,\,\ , , \label{5 - 8}\ ] ] @xmath203 } { ( 1+w_{n}^4)(w^2+w_{n}^2)^2}}\right ] \,\,\ , , \label{5 - 9}\ ] ] where we used the notation @xmath204 @xmath205 one may observe that for @xmath199 the real part of the conductivity is finite , but the imaginary part diverges .
it is due to the logarithmic term in eq .
( [ 5 - 9 ] ) .
this is an unphysical result .
it may indicate that the @xmath199 term is not physically acceptable , or that the _ 2d _ model should not be applied exactly at @xmath4 .
the relevance of the theoretical expressions derived in the preceding sections can be demonstrated by comparison of the calculated and experimentally measured @xmath206 fluctuation conductivity . as an example we present here an analysis of the data in @xmath207 thin film .
the experimental results of the complex conductivity measured at 9.5 ghz are shown in fig .
[ fig6](a ) .
the main features are the same as reported previously in single crystals of high-@xmath4 superconductors.@xcite we have to note that in our measurement the thin film was positioned in the center of an elliptical microwave cavity resonating in @xmath208te@xmath209 mode , and oriented in such a way that the electric field @xmath183 was in the _ ab_-plane .
thus the in - plane conductivity was measured and the application of the theoretical expressions of the preceding sections is appropriate .
other experimental details have been reported previously.@xcite in this section we are interested in the fluctuation conductivity near @xmath4 which is shown on an enlarged scale in fig . [ fig6](b ) .
the real part of the conductivity has a maximum when the coherence length diverges . since the critical temperature of a phase transition is characterized by the divergence of the coherence length , we use the maximum of @xmath8 in fig .
[ fig6](b ) to determine @xmath210 k. one can also observe in fig .
[ fig6](b ) that the imaginary part of the conductivity crosses the real part at a temperature slightly above @xmath4 .
this is a direct experimental evidence of the short wavelength cutoff as discussed in section [ sect2 ] .
the experimental values of @xmath8 and @xmath10 at @xmath4 can be used in the evaluation of the parameters which enter the theoretical expressions of the preceding sections .
[ fig7](a ) shows the experimental data above @xmath4 plotted against the reduced temperature @xmath211 .
we can analyze this data first by the theoretical expressions which have no cutoff on the fluctuation wavevector . if one sets @xmath212 for the _ 3d _ anisotropic case , it is straightforward to evaluate @xmath136 using eq .
( [ 4 - 29 ] ) and the experimental value of @xmath10 at @xmath4 .
we have obtained @xmath213 nm in @xmath207 thin film .
once this parameter is determined , the fluctuation conductivity at all temperatures above @xmath4 follows from eq .
( [ 4 - 5 ] ) .
the real and imaginary parts of the @xmath214 fluctuation conductivity have to be mutually consistent
. this can be exploited in the data analysis .
we insert the experimental values of @xmath10 into the imaginary part of eq .
( [ 4 - 5 ] ) , and solve numerically for @xmath215
. then we exploit these same values of the reduced correlation length in the real part of eq .
( [ 4 - 5 ] ) .
the calculated @xmath8 is shown by the dotted line in fig .
[ fig7](a ) .
the calculated line lies far from the experimental points .
note in particular that the calculated @xmath8 meets @xmath10 at @xmath4 when no cutoff is included ( cf . section [ sect2 ] ) .
besides , the shape of the calculated @xmath8 differs from that of the experimental one .
we may conclude that with no cutoff on the fluctuation wavevector the theoretical expression does not describe properly the experimental fluctuation conductivity .
a finite cutoff on the fluctuation wavevector improves greatly the agreement of the theory and experiment . from fig .
[ fig6](b ) we can evaluate the ratio @xmath216 at @xmath4 .
this value yields a constraint on the choices of @xmath173 and @xmath174 as described in section [ sect3 ] and fig .
the actual choices are presented in the inset of fig .
[ fig7](b ) . for a given choice ( @xmath173 , @xmath174 ) from this constraining line
, one has to determine first @xmath136 using eq .
( [ 4 - 29 ] ) and the experimental value of @xmath10 at @xmath4 .
then , the temperature dependence of the reduced correlation length @xmath215 is evaluated numerically from the imaginary part of eq .
( [ 4 - 5 ] ) with the selected pair ( @xmath173 , @xmath174 ) and the experimental values of @xmath10 .
the obtained values of @xmath215 are finally used to calculate @xmath8 from the real part of eq .
( [ 4 - 5 ] ) .
the results vary with the possible choices of the pairs ( @xmath173 , @xmath174 ) from inset of fig .
[ fig7](b ) .
the best fit of the calculated @xmath8 to the experimental points is shown by the solid line in fig .
[ fig7](a ) .
it is obtained with the choice ( @xmath217 , @xmath218 ) .
it is physically reasonable . with @xmath173 being of the order of unity ,
the minimum wavelength of the superconducting fluctuations in the _ ab_-plane is given by @xmath219 , which is much larger than the atomic size and could be accepted as a mesoscopic quantity .
in contrast , the value of @xmath220 is below the atomic size and , hence , could not be physically accepted for the lower limit of the fluctuation wavelength along the c - axis .
therefore one needs a value of @xmath221 so that the minimum wavelength of the superconducting fluctuations @xmath222 along the c - axis becomes also an acceptable mesoscopic quantity .
we have tested also a number of other choices of the cutoff parameters . by shifting the choice of the parameters along the constraining curve in the inset of fig .
[ fig7]b , one degrades the fit of the calculated @xmath8 to the experimental points . the dashed line in fig . [ fig7](a ) is the calculated @xmath8 using the choice of equal cutoff parameters ( @xmath223 ) permitted by the constraint in the inset of fig .
[ fig7](b ) .
the fit to the experimental points is seen to be much worse than that of the full line . for the sake of completeness , we show also the result for a choice of the cutoff parameters on the other branch of the constraining curve in the inset of fig .
[ fig7](b ) .
if one takes ( @xmath224 , @xmath225 ) the calculated @xmath8 is as shown by the dashed - dotted line in fig .
[ fig7](a ) .
the fit is unsatisfactory .
moreover , this choice has to be refuted on the grounds of physically unacceptable minimum wavelength of the fluctuations along the c - axis . also shown in fig .
[ fig7](a ) by the short - dotted line is the result obtained by the isotropic _
3d _ expression in eq .
( [ 3 - 8 ] ) . in this case
the parameters @xmath226 nm and @xmath227 are obtained straightforwardly from eq .
( [ 3 - 19 ] ) and the experimental values of @xmath8 and @xmath10 at @xmath4 .
the fit in fig .
[ fig7](a ) is obviously not good .
it is also seen that the expressions for the anisotropic _
3d _ case always yield curves which are different from that of the isotropic _
3d _ case .
the complexity of the anisotropic _
3d _ expressions elaborated in section [ sect3 ] is not futile .
indeed , we find that these expressions must be used when analyzing an anisotropic superconductor . fig .
[ fig7](b ) shows an enlarged view of the high temperature part where the same curves as in fig .
[ fig7](a ) are better distinguished .
we have analyzed also the _
2d _ expressions of section [ sect4 ] . fig .
[ fig8](a ) shows again the same experimental data as in fig .
[ fig7](a ) , but fitted with @xmath228 term of the _
2d _ expansion in eq .
( [ 5 - 3 ] ) . the parameter @xmath184 has been chosen so as to optimize the fit to the experimental values of @xmath8 . the resulting curve in fig .
[ fig8](a ) was obtained with @xmath229 nm .
the _ 2d _ results are not so sensitive to the fluctuation wavevector cutoff as those of the _
3d _ case .
the curves obtained with no cutoff ( @xmath230 ) and with @xmath231 are practically indistinguishable in fig .
[ fig8](a ) .
one may conclude that closer to @xmath4 the bscco superconductor clearly does not behave as a _
2d _ system .
however , at higher temperatures both , _
2d _ and _ 3d _ expressions yield almost equally good fits to the experimental values , as seen in fig .
[ fig8](b ) .
thus the dimensionality of the fluctuations at higher temperatures can not be resolved from the @xmath26 fluctuation conductivity .
finally , we remark that the above analysis could explain very well the experimental @xmath26 fluctuation conductivity above @xmath4 in the bscco thin film using the aslamazov - larkin type expressions with wavevector cutoff as deduced in the preceding sections of this paper .
the other contributions such as maki - thomson and one electron density of states , mentioned in section [ sect2 ] , are not necessary over most of the temperature range covered in the experiment .
this is in accord with the recent microscopic calculation@xcite proving that these contributions may cancel in the ultraclean case of nonlocal electrodynamics .
however , they may play a role at high enough temperatures where the above calculated curves depart from the experimental data .
their analysis is beyond the scope of the present paper .
we have presented full analytical expressions for the @xmath0 fluctuation conductivity in _
isotropic , _ 3d _ anisotropic , and _ 2d _ superconductors .
the effects of the short wavelength cutoff in the fluctuation spectrum on the @xmath6 and @xmath0 conductivities were discussed in detail .
the short wavelength cutoff brings about a breakdown of the scaling property in frequency and temperature .
it also has a small , but experimentally very important effect on @xmath8 . due to a finite cutoff parameter @xmath30 , the value of @xmath8 at @xmath4 is lower than that of @xmath10 . in the simpler _
isotropic case , this observation can be used to determine @xmath30 directly from the experimental data at @xmath4 . in the _ 3d _
anisotropic case , one obtains a constraint on the choices of ( @xmath173 , @xmath174 ) .
the useful feature of @xmath0 fluctuation conductivity measurements is that @xmath4 can be determined directly from the experimental data .
moreover , the expressions derived in this paper enable the determination of @xmath16 ( _ 3d _ isotropic ) , or @xmath232 ( _ 3d _ anisotropic ) from the experimental values @xmath233 .
thus , we establish that the analysis of @xmath0 fluctuation conductivity requires no free fit parameters in the _ 3d _ isotropic case , and only one ( @xmath173 , @xmath174 ) in the _ 3d _ anisotropic case .
we have shown that the anisotropic _
3d _ expressions with an appropriate cutoff of the fluctuation wavevector can account for the experimental fluctuation conductivity in a bscco thin film within a large temperature range above @xmath4 .
the _ 2d _ expression is less sensitive to the cutoff and was found to match the experimental data only at higher temperatures .
n . peligrad and m. mehring acknowledge support by the deutsche forschungsgemeinschaft ( dfg ) project nr .
me362/14 - 2 .
a. duli acknowledges support from the croatian ministry of science , and dlr stiftung ( project nr .
kro00498 ) .
-curves for the _ 3d _
isotropic case calculated from eq .
( [ 3 - 9 ] ) for a finite cutoff parameter @xmath234 ( full lines ) , and for @xmath235 ( dashed lines ) .
the variable is @xmath45 in ( a ) , and @xmath78 in ( b ) .
the curves are labelled by the frequency @xmath236 used in the calculations.,scaledwidth=80.0% ] -curves for the _ 3d _
isotropic case calculated from eq .
( [ 3 - 10 ] ) for the same parameters as in fig . [ fig1 ] .
the effects of a finite cutoff are small and can be seen only on logarithmic scales used in the insets.,scaledwidth=80.0% ] and @xmath79 of the _
isotropic case calculated with the choice @xmath237 ghz in the limit @xmath238 .
the dashed lines are the @xmath239-curves calculated with no cutoff ( @xmath235 ) , and the full lines include a finite cutoff ( @xmath234 ) .
( b ) enlarged section which shows the crossing of @xmath68 and @xmath79 ( @xmath240 , see text ) at a temperature slightly above @xmath4 .
the two dashed lines are indistinguishable in this temperature range.,scaledwidth=80.0% ] ( equal to @xmath241 ) for the _ 3d
_ isotropic case at temperatures approaching @xmath4 . with no cutoff ( dashed lines )
the ratio tends to unity for all frequencies . with a finite cutoff ( @xmath234 ) the ratio equals unity at a temperature slightly above @xmath4 , dependent on the frequency . in the limit
@xmath238 the ratio saturates to a frequency dependent value larger than unity .
( b ) the ratio @xmath242 and @xmath243 given by eq .
( [ 3 - 20 ] ) .
the variable @xmath97 is defined by eq .
( [ 3 - 18]).,scaledwidth=80.0% ] for _ 3d _ anisotropic case as a function of @xmath176 and @xmath177 ( cf .
( [ 4 - 28 ] ) ) .
( b ) selection of curves for fixed ratios of @xmath244 indicated by numbers .
the dashed line marks the condition @xmath245 ( @xmath246).,scaledwidth=80.0% ] above @xmath4 plotted versus the reduced temperature @xmath247 .
various lines are the conductivities calculated in the _ 3d _ cases as described in the text .
( b ) enlarged view of the high temperature part of the same curves as in ( a ) .
the constraining curve for the choices of the cutoff parameters @xmath173 and @xmath174 resulting from the experimental ratio @xmath248 at @xmath4 is shown in the inset.,scaledwidth=80.0% ] | the short wavelength cutoff has been introduced in the calculation of @xmath0 fluctuation conductivity of superconductors .
it is shown that a finite cutoff leads to a breakdown of the scaling property in frequency and temperature .
also , it increases the phase @xmath1 of the complex conductivity ( @xmath2 ) beyond @xmath3 at @xmath4 .
detailed expressions containing all essential parameters are derived for 3d isotropic and anisotropic fluctuation conductivity . in the _
2d _ case we obtain individual expressions for the fluctuation conductivity for each term in the sum over discrete wavevectors perpendicular to the film plane . a comparison of the theory to the experimental microwave fluctuation conductivity is provided . |
in astrophysics , the observation of a splitting of spectral lines in the visible and uv ranges for a few white dwarfs @xcite confirmed the existence of intense magnetic fields ( 0.1 - 10@xmath2 mg ) as predicted by blackett @xcite .
the influence of a magnetic field on an atom modifies its emission or absorption lines .
thanks to this property , known as zeeman effect , the detection of magnetic fields is possible at large distances , through the measured radiation .
the linear and quadratic zeeman effects @xcite explain the separation of spectral lines and enable one to determine a value of the magnetic field .
in the same way , pulsars and neutron stars having an even more intense magnetic field ( 10@xmath3 - 10@xmath4 mg ) have been discovered through their spectrum in the range of radio - frequencies and x - rays .
there are numerous astrophysical applications , either direct or indirect , and requiring sometimes a sophisticated theoretical modeling .
the methods differ according to the nature of the objects studied ( see table [ tab1 ] ) , the magnitude and the geometry of the magnetic fields , and to the quality of the observation in terms of sensitivity and spectral resolution .
moreover , the variations of the magnetic field of stars during their rotation bring some information about their global geometry .
the `` spectro - polarimetric '' methods exploit the additional recording of the circular polarization with respect to the wavelength .
this enables one to obtain a detailed map of the field @xcite through a separation of its components parallel or perpendicular to the line of sight .
strong magnetic fields are also encountered , for instance , in magneto - inertial fusion @xcite .
inserting a magnetic field into inertial - confinement - fusion capsules before compressing them @xcite presents the advantages to suppress the electron thermal - conduction losses and to better control the @xmath5-particle energy deposition .
the magnetic fields generated inside a hohlraum can reach a few mg .
.orders of magnitude of magnetic fields encountered in astrophysics ( 1 mg=10@xmath6 g=100 t ) . [ cols="^,^",options="header " , ] if needed , the evaluation of @xmath7 can be refined .
for instance , it is possible to calculate an average value of @xmath8 depending only on @xmath9 .
this can be achieved using the sum rule @xcite @xmath10 which states that the sum of the land factors for any given @xmath9 is independent of the coupling conditions .
such a property stems from the fact that the trace of a matrix is invariant under an orthogonal transformation .
one can thus define an average land factor associated to a given value of @xmath9 : @xmath11 where @xmath12 is the number of levels having angular momentum @xmath9 @xcite , which can be evaluated recursively @xcite , in a similar manner to @xmath13 ( see eq .
( [ recsl ] ) ) .
the same methodology can be applied in order to determine analytically the moments of the hyperfine components of a line .
the hyperfine operator in the subspace corresponding to the relevant nucleus and atomic level reads : @xmath14 where @xmath15 is the magnetic hyperfine - structure constant of the level @xmath16 . the @xmath17-order moment of the hyperfine components
is provided by the expression @xmath18^n\nonumber\\ & & \times\langle\gamma jifm|\mathcal{z}_q^{(1)}|\gamma jifm\rangle^2,\end{aligned}\ ] ] where @xmath19 is the @xmath20-component of the dipole operator @xmath21 .
the @xmath9-file sum rule @xcite enables one to simplify the expression of the strength : @xmath22=\frac{1}{3}[i , j],\end{aligned}\ ] ] and therefore @xmath23}\sum_{f , f'}\left(a_jx_{fij}-a_{j'}x_{f'ij'}\right)^n\nonumber\\ & & \times\langle f||\mathcal{z}^{(1)}||f'\rangle^2,\end{aligned}\ ] ] where @xmath24 .
equation ( [ mn0 ] ) can be written @xmath25}\sum_{f , f'}[f , f']\left(a_jx_{fij}-a_{j'}x_{f'ij'}\right)^n\nonumber\\ & & \times\langle ( ij)f||\mathcal{z}^{(1)}||(ij)f'\rangle^2,\end{aligned}\ ] ] or @xmath26}\sum_{f , f'}[f , f']\left(a_jx_{fij}-a_{j'}x_{f'ij'}\right)^n\nonumber\\ & & \times{\left\{\begin{array}{ccc}f & 1 & f ' \\ j ' & i & j \end{array}\right\}}^2.\end{aligned}\ ] ] in the case where @xmath27 or @xmath28 is equal to 0 , the calculation is very simple @xcite . in the general case , using @xmath29}{\left\{\begin{array}{ccc}f & j & i \\ 1 & i & j \end{array}\right\}},\ ] ] one has to calculate : @xmath30{\left\{\begin{array}{ccc}f & j & i \\ 1 & i & j \end{array}\right\}}^{k_1}\nonumber\\ & & \times{\left\{\begin{array}{ccc}f ' & j ' & i \\ 1 & i & j ' \end{array}\right\}}^{k_2}{\left\{\begin{array}{ccc}f & 1 & f ' \\ j ' & i & j \end{array}\right\}}^2,\end{aligned}\ ] ] which can be done using graphical methods @xcite .
another approach consists in adopting another point of view , leading to the evaluation of quantities of the type : @xmath31(\bar{f}-a)^n\ ] ] where @xmath32 is a constant ( depending on other quantum numbers ) .
such a quantity can be expressed , as for the zeeman effect , in terms of bernoulli numbers ( see appendix b ) : @xmath33 the splitting of @xmath27 components in a weak magnetic field @xcite is in every way similar to the splitting of @xmath9 levels .
the scale of the splitting is determined by the factor @xmath34 , which is defined by @xmath35 and connected with the land factor by @xmath36
in this work , a statistical modeling of electric dipolar lines in the presence of an intense magnetic field was proposed .
the formalism requires the moments of the zeeman components of a line @xmath37 , which can be obtained analytically in terms of the quantum numbers and land factors .
it was found that the fourth - order a - type gram - charlier expansion series provides better results than the usual development in powers of the magnetic field often used in radiative - transfer models . using
our recently published recursive method for the numbering of ls - terms of an arbitrary configuration , a simple approach to estimate the contribution of a magnetic field to the width ( and higher - order moments ) of a transition array of e1 lines was presented .
we hope that such results will be useful for the interpretation of z - pinch absorption or emission spectra , for the study of laser - induced magnetic fields in inertial - fusion studies , for the modeling of magnetized stars as well as for any application involving magnetic fields in spectroscopic studies of atomic and molecular systems .
the authors would like to thank c. bauche - arnoult , j. bauche and r. karazija for helpful discussions .
@xmath40=(2a+1)(2b+1)(2c+1)\cdots\ ] ] @xmath41 @xmath42 @xmath43{\left(\begin{array}{ccc}1 & j '' & 1 \\ 0 & 0 & 0 \end{array}\right)}{\left(\begin{array}{ccc}1 & j '' & 1 \\ -q & 0 & q \end{array}\right ) } & & \nonumber\\ \times{\left\{\begin{array}{ccc}j & j & 1 \\ j '' & 1 & j \end{array}\right\}}{\left\{\begin{array}{ccc}j ' & j & 1 \\ j '' & 1 & j \end{array}\right\}}.\nonumber\\ & & \end{aligned}\ ] ] @xmath44 @xmath45 @xmath46}}.\ ] ] @xmath47\bar{j}}}.\ ] ] @xmath48(2j-1)(2j+3)\bar{j}}}.\ ] ] @xmath49\bar{j}}}.\ ] ]
the bernoulli polynomials can be obtained by successive derivation of a generating function : @xmath50 one can write @xmath51 where @xmath52 is the @xmath53-order bernoulli number , which is non - zero only if @xmath53 is even and which can be obtained from the relation : @xmath54 the first bernoulli polynomials are @xmath55 @xmath56 @xmath57 @xmath58 and @xmath59 the bernoulli polynomials obey the following identity : @xmath60 and the bernoulli numbers have the explicit laplace s determinantal formula @xcite : @xmath61
using second - order ts expansion ( [ iq2 ] ) and assuming the knowledge of the variance @xmath62 of the other broadening mechanisms , it becomes possible to estimate the magnitude of the magnetic field from the measurement of the full width at half maximum ( fwhm ) of the line @xmath63 . this simple formula ( [ est ] ) can provide an estimation of the magnetic field , even if the other broadening mechanisms ( stark , electron collisions , doppler , autoionization ) are dominant .
however , it is not as efficient as the method proposed by stambulchik _
@xcite , which is applicable in situations where the magnetic field has various directions and amplitudes ( or if they vary in time ) .
n. zeldes , arch .
exact sci . * 63 * , 289 ( 2009 ) .
g. racah developed a method for fitting by least squares to the theoretical formulas the measured g - factors of atomic levels together with the energy values ( see [ n. zeldes , arch .
exact sci . * 63 * , 289 ( 2009 ) ] and references therein ) .
unfortunately , the work was never presented ( g. racah died in august 1965 , before the conference ) . | the modeling of complex atomic spectra is a difficult task , due to the huge number of levels and lines involved . in the presence of a magnetic field ,
the computation becomes even more difficult .
the anomalous zeeman pattern is a superposition of many absorption or emission profiles with different zeeman relative strengths , shifts , widths , asymmetries and sharpnesses .
we propose a statistical approach to study the effect of a magnetic field on the broadening of spectral lines and transition arrays in atomic spectra . in this model ,
the @xmath0 and @xmath1 profiles are described using the moments of the zeeman components , which depend on quantum numbers and land factors .
a graphical calculation of these moments , together with a statistical modeling of zeeman profiles as expansions in terms of hermite polynomials are presented .
it is shown that the procedure is more efficient , in terms of convergence and validity range , than the taylor - series expansion in powers of the magnetic field which was suggested in the past .
finally , a simple approximate method to estimate the contribution of a magnetic field to the width of transition arrays is proposed .
it relies on our recently published recursive technique for the numbering of ls - terms of an arbitrary configuration .
characterization of anomalous zeeman patterns in complex atomic spectra jean - christophe pain and franck gilleron cea , dam , dif , f-91297 arpajon , france |
the field of nanotribology evolved around attempts to understand the relationship between frictional forces and microscopic properties of systems.@xcite recent experimental and theoretical studies@xcite have suggested that the observed frictional phenomena might originate from the formation and rupture of microscopic bonds ( junctions ) that form between surfaces in close vicinity .
furthermore , these findings indicate that stick - slip motion is connected to a collective behavior of the bonds @xcite .
the formation and rupture of bonds are thermally activated processes and , thus , temperature may play an important role in the dynamics of friction at the nanoscale.@xcite friction is not simply the sum of single - bond responses , but is influenced by temporal and spatial dynamics across the entire ensemble of bonds that form the frictional interface .
the way how individual bonds can be averaged to yield friction response has been the focus of intense research in the past decade,@xcite however many key aspects of the friction dynamics and its relation to the kinetics of bond formation and rupture are still not well understood .
one of the main difficulties in understanding and predicting frictional response is a lack of information on the nature of mediating bonds and their kinetic characteristic rates of formation and rupture .
the phenomenological models @xcite that successfully described velocity and temperature dependencies of friction through thermally activated rupture , formation , and strengthening of molecular bonds involved a large number of empirical parameters , which limits their insight and predictive power . in most frictional force microscopy ( ffm )
experiments the tip apex termination remains unknown in term of its structure and chemical nature , a fact that may severely restrict the interpretation of the data .
a significant progress in understanding microscopic mechanism of friction in terms of dynamical rupture and formation of molecular bonds can be achieved through investigations of model systems , where the tip and substrate are functionalized by assemblies of host molecules and adhesion between the contacting surfaces is caused by supramolecular host - guest interactions.@xcite in this case , in contrast to previous studies of frictions , the usage of single - molecule techniques allows more detailed insight in binding forces and rates of bond formation and rupture.@xcite first experimental study of friction and adhesion caused by cooperative rupture of supramolecular bonds@xcite discovered a remarkable difference in the dynamics of these processes : the pull - off force increased dramatically with unloading rate , while the friction force was found to be constant , over a sliding velocity range of more than three orders of magnitude .
moreover , it was suggested that different connector molecules can be bound to the same surface functionalization , allowing to control friction and adhesion using switchable connector molecules @xcite or to adapt to surface roughness by varying the lengths of the connectors . in this article
we present results of simulations of dynamics of adhesion and friction caused by cooperative rupture of supramolecular bonds .
particular attention is given to the effect of a non - equal load sharing between the bonds , which results from the curved shape of the afm tip , on the dynamics of adhesion and friction .
the rate dependence of the pull - off force and velocity dependence of friction have been calculated in a broad range of rates of formation and rupture of molecular bonds , and different dynamical regimes have been discovered .
we employ a model in which the host complexes , represented by ideal rubber bands with spring - constant @xmath0 and rest length @xmath1 , are attached with homogeneous density to the surface of a hemispherical 3d tip , and to the counter - surface substrate ( see fig .
[ figure2 ] ) . the tip , with mass @xmath2 , diameter @xmath3 , and coordinates @xmath4 , is connected to the cantilever holder at @xmath5 through springs of lateral and normal stiffnesses @xmath6 and @xmath7 , respectively .
the tip is then subject to a viscous damping @xmath8 , accounting for water environment .
tip and holder are initially placed at @xmath9=@xmath10=0 , @xmath11=@xmath12=@xmath13 .
the resulting equations of motion are @xmath14 [ eq.motion ] where @xmath15 connected complexes ( up to @xmath16 ) , forming angles @xmath17 with the substrate , have elongations @xmath18 modulated by the heaviside step - function @xmath19 in order to suppress repulsive contributions , that are negligible within the rubber - band assumption .
the term @xmath20 describes a short - range repulsion between the tip and substrate.@xcite the bonds form vertically , and follow the motion of the tip along @xmath10 direction ; thus no forces along the transverse @xmath21 direction is generated .
the complexes bind or unbind with thermally - activated rates @xmath22 and @xmath23 expressed by @xmath24 , \\
r_{off } & = r_{off}^0\exp[\theta(l - l_0)k(l - l_0)\delta_l / k_bt ] , \end{aligned}\label{eq.ron_roff}\ ] ] in which we have introduced the characteristic attempt rates @xmath25 and @xmath26 , the rupture barrier length @xmath27 , and the thermal energy @xmath28 .
it should be noted that , because of the curved shape of the afm tip , the rates @xmath22 and @xmath23 depend on the position of the attached bond with respect to the tip apex .
this configuration leads to a non - equal load sharing between the bonds .
we also note that due to the nature of the problem and the symmetry of such a geometrical configuration , the addition of a lateral direction @xmath21 in the model would make the computational efforts just more demanding without adding any valuable physical insight .
the equations of motion are integrated using a velocity - verlet algorithm with a time step @xmath29=1@xmath30s , and formation / rupture of the complexes is stochastically governed according to the probabilities @xmath31=@xmath22@xmath29 and @xmath32=@xmath23@xmath29 at each time step . if not differently stated , the employed parameters are : @xmath16=100 , @xmath0=0.16n / m , @xmath1=10 nm , @xmath27=0.1 nm , @xmath13=0.1 nm , @xmath28=25.85mev , @xmath6=50n / m , @xmath7=0.23n / m , @xmath8=13.3ms@xmath33 , @xmath2=1.15@xmath3410@xmath35 kg , @xmath3=20 nm .
reference values have been taken from typical afm setup and from recent experimental data we intend to reproduce within our modeling approach.@xcite
= @xmath26=0.1khz ; ( b ) instantaneous number of connected complexes .
( c ) sketch of the two different stretching scenarios during the vertical pulling : fast rate ( left ) , slow rate(right),title="fig : " ] + = @xmath26=0.1khz ; ( b ) instantaneous number of connected complexes .
( c ) sketch of the two different stretching scenarios during the vertical pulling : fast rate ( left ) , slow rate(right),title="fig : " ] the friction measurements alone do not allow to establish a microscopic mechanism of energy dissipation . in order to gain better understanding of the kinetics of cooperative rupture and formation of supramolecular bonds ,
the mechanical response of the system upon retraction from the surface has been measured.@xcite to understand the experimentally observed behavior , we have performed numerical simulations of the normal force as a function of the distance between the cantilever holder and the surfaces , @xmath11 .
the normal force is determined by fixing the vertical velocity of the cantilever holder , @xmath36 , corresponding to a unloading rate @xmath37 , and by measuring the force exerted on the tip , i.e. @xmath38 . the normal force trend as a function of the tip height @xmath11 for different unloading rates is reported in figure [ figure3](a ) . before the complete detachment of the complexes , and the consequent release of the tip , we can identify two different regimes : a linear ( negative ) increase of the force with increasing @xmath11 , and a plateau region with normal force almost independent of @xmath11 .
the linear growth of the force becomes particularly noticeable for sufficiently high unloading rates , for which the number of connected complexes is very slowly vaying in time until complete detachment occurs ; here , the detailed unbinding dynamics is unable to respond fast enough during the cantilever retraction [ see figure [ figure3](b ) ] .
this almost constant number of bonds is that at the initial equilibrium , when @xmath11=@xmath12=@xmath13 , and is solely determined by the ratio @xmath39 . as depicted in figure [ figure3](c ) , when the cantilever holder gets displaced the system must equilibrate the tension among a parallel of many complexes connected in series with the cantilever spring , so that the latter must burden most of the overall stretching .
therefore , the slope of the linear increment of the normal force is mostly determined by the cantilever normal stiffness @xmath7 .
the plateau region in the retraction force traces is a characteristic feature of the discussed experiments with supramolecular bonds , and as far as we know this effect was not observed in previous force measurements . as shown in fig .
[ figure3](b ) in the plateau region the number of connected complexes decreases with higher @xmath11 , thus diminishing the effective stiffness of the complexes in parallel , and compensating for the stress of the cantilever spring .
the rupture starts at the periphery of the tip and propagates towards its apex .
the progressive unbinding of the complexes gives rise to a saw - tooth - shaped fine structure of the normal force in plateaus , as recently observed in experiments and exploited to determine the effective stiffness of a single complex.@xcite as anticipated above , the maximum normal force the pull - off force ( or rupture force )
is therefore strongly dependent on the unloading rate , and on its relation with the binding / unbinding rates in eq .
[ eq.ron_roff ] . to investigate this aspect ,
we have reported in figure [ figure4 ] the pull - off forces calculated for a hemispherical tip , as a function of the unloading rate for a set of @xmath26,@xmath25 values . in agreement with experimental observations,@xcite
our results exhibit two distinct regimes : ( i ) at low unloading rates the rupture force is approximately constant , ( ii ) with increasing unloading rates @xmath40 , the pull - off force grows as @xmath41 . a comparison of our simulated rate - dependence of pull - off force with the experimental observations of ref .
is provided in fig .
[ figure4 ] inset . for increasing unbinding rates @xmath42 , the transition separating these two regimes shifts at higher unloading rates
. regime ( i ) is usually attributed to a quasi - equilibrium state in which the number of bonds is maintained approximately constant by a fast rebinding of broken bonds.@xcite in our system the origin of this regime is different , it results from the progressive unbinding of bonds . besides
, since 1/@xmath23 represents a characteristic time for bond rupture , more bonds are connected at lower @xmath26 values , resulting in larger pull - off forces .
instead , @xmath25 produces negligible effects in its considered range , due to the fact that the formation of bonds with large elongation demands energies much larger than @xmath28 , thus requiring huge @xmath25 prefactors in eq .
[ eq.ron_roff ] to compensate over the exponential term . here
, we note that thermal energy is overcome already at elongations of @[email protected] nm and @[email protected] nm for bond formation and bond rupture , respectively . beyond these elongations , very large ( small ) @xmath25 ( @xmath26 ) prefactors are required to compensate the exponential terms in eq .
[ eq.ron_roff ] .
thus , within the considered range for @xmath25 , the formation of a bond becomes impossible already at modest elongations , so that the bonds will never reform after rupture at the typical unbinding time - scales .
however , when a bond is intact , the elongation before rupture is ( roughly speaking ) determined by the rupture time 1/@xmath23 times the cantilever holder velocity . clearly , larger unloading rates correspond to larger rupture lengths .
this is why , as shown in fig .
[ figure4 ] , in the investigated range of parameters the pull - off forces much increase at lower @xmath26 , while @xmath25 acts as a mild pull - off shifter . ) as a function of the unloading rate for different @xmath46 pairs as reported in legend ( in khz ) .
control curve is with free tip , i.e. subject only to viscous damping .
( inset ) best fit of experimental data from ref . , obtained with @xmath25@xmath47100khz , @xmath26@xmath4725hz , @[email protected] nm , @[email protected] / m . ] under the condition of identical bonds , having the same rest lengths and the same binding / unbinding probabilities , which is characteristic for a blind tip with an extended flat surface
, one might expect an almost simultaneous rupture of all the bonds when the applied force exceeds a critical value , rather than a one - by - one rupture as occurring during the normal force plateaus .
however , in a common experimental setup , because of the curved shape of the afm tip , some bonds are more stretched than others , thus giving rise to a non - uniform distribution of unbinding probabilities in the ensemble of bonds . to explore this aspect
we have performed simulations with different tip shapes . in particular , in figure [ figure5 ] we report normal force curves obtained for flat , hemispherical , and conical tips .
clearly , both spherically- and conically - shaped tips produce similar curves , with well pronounced plateaus of the normal force extending for more than 5 nm .
conversely , in the case of a flat tip no plateu forms , supporting the idea that the one - by - one rupture occurring at the plateaus is intimately connected to a non - uniform distribution of bond elongations induced by the tip geometry .
the bonds connected at the periphery of the tip become more stretched , and they broke first .
= @xmath26=0.1khz ) ; ( b ) instantaneous number of connected complexes . ]
+ in sliding experiments , the friction force is determined by fixing the height , @xmath11 , and the driving velocity , @xmath48 , of the cantilever holder , and measuring its lateral torsion , @xmath49 .
figure [ figure6](a ) reports the lateral force traces as a function of the cantilever position for different sliding velocities . we notice two opposite limiting behaviors : a smooth almost - constant force at small sliding velocity , and a chaotic stick - slip regime , with strongly increased average value ( friction ) , at fast sliding .
as reported in figure [ figure6](b ) , the average number of connected bonds during sliding is roughly constant at low velocity , revealing a steady - state regime in which the complexes bind and unbind at rates faster than the characteristic sliding frequency @xmath50 . in this quasi - equilibrium regime
the time - averaged friction force grows with the driving velocity [ see fig .
[ figure6](c ) ] . on the contrary , at large velocities , we find pronounced sudden drops in the force traces of fig .
[ figure6](a ) , revealing the occurrence of avalanches of breaking contacts . similar to the pull - off case ,
the latter regime onsets when the characteristic sliding frequency is higher than the rate of the natural unbinding timescale , thus increasing the lateral force until some bonds induce the collective rupture that leads to an abrupt slip of the tip .
this mechanism is sketched in figure [ figure6](d ) . by plotting the average lateral ( friction ) force as a function of the driving velocity for different binding / unbinding rates [ fig .
[ figure6](c ) ] , we observe a multiplicity of possible distinctive regimes , with increasing or decreasing friction as a function of velocity , and , in some case , the occurrence of a blunt maximum . to understand such scenario we compare the characteristic sliding time with the two microscopic characteristic times of our model , namely @xmath51 , the minimum bond formation time , and @xmath52 , the maximum bond rupture time .
during sliding , the bonds are continuously broken and reformed in an unstressed state and the friction force can be estimated as @xmath53,@xcite where @xmath54 is the number of intact bonds in equilibrium .
a characteristic feature of low sliding velocity regime is the essential absence of correlation between individual rupture events , which is a manifestation of thermal bond dissociation . with increase of the speed , processes of spontaneous ( thermal ) and shear - induced bond dissociation start to compete , and we observe an erratic stick - slip dynamics .
the stick - slip motion becomes more pronounced at larger velocities , @xmath55@xmath56@xmath52 , where the rupture is completely determined by the effect of shear stress on the height of the unbinding potential barriers . in contrast to the low - velocity sliding regime where rupture events are uncorrelated , the stick - slip motion is characterized by a cooperative behavior of bond subsystem .
it should be noted that this high friction regime occurs only under the condition that bonds can quickly reattach , that happens when @xmath55@xmath57@xmath58 .
instead , when the sliding velocity is too high ( @xmath59 ) , only a small number of bonds survives , and the friction force is reduced and becomes mainly determined by the contribution coming from the viscous damping of the aqueous solution .
it is worth to note that , differently from the previous simulations of friction in terms of rupture and rebinding of interfacial junctions,@xcite figs .
[ figure6](a),(b ) do not show a clear correlation between the stick - slip oscillations of friction force and the variation of number of attached bonds with the cantilever displacement .
the important difference of the present system from those considered before is that the equilibrium length of the supramolecular bonds is significantly larger than the characteristic distance between the tip apex and the substrate . as a result , during the frictional motion most of attached bonds are in the unstretched state , and only a small percentage of them gets ruptured at the stick - to - slip transition .
our simulations demonstrate that , by changing the ratio @xmath60 we can tune the position of the maximum of the friction force as a function of velocity [ see figure [ figure6](c ) ] , and even push it to range of low velocities , or of very large velocities being left with a monotonically increasing trend only . besides , there exist values of the ratio @xmath60 which lead to regimes of quasi - constant friction over a broad velocity range , reproducing pretty similar trends of recent experimental data.@xcite the above - described variety of frictional trends has been experimentally observed through the adoption of different molecular sets.@xcite as shown by previous studies,@xcite slide - hold - slide ( shs ) experiments are able to access the characteristic bonding time @xmath58 of a certain molecular complex . in this kind of experiments the sliding motion of the tip
is suddenly stopped and restarted after a certain holding time @xmath61 . these measures
are usually carried out in order to estimate the ageing of the tip - substrate junctions , i.e. to which extent the static friction is increased by the formation of bonds or capillary bridges during the rest time . the friction force measured in our shs simulations
is reported in figure [ figure7](a ) for different holding times .
as expected , the number of intact bonds increases with @xmath61 , as well as the resulting force maximum occurring upon resuming sliding [ fig .
[ figure7](b ) ] . however ,
when @xmath62 the number of connected complexes can reach equilibrium , and the maximum force saturates [ fig .
[ figure7](c ) ] .
the above argument can provide a way to directly estimate @xmath22 by shs experiments : in our case the maximum lateral force starts to increase at @xmath61@xmath6310ms , in agreement with the used @xmath22 value of 0.1khz . in fig .
[ figure7](c ) we report the maximum lateral force ( static friction ) also calculated for different distances between the holder and the surface , @xmath11 , from -1.0 nm to 5.0 nm . here
negative values of @xmath11 correspond to the regime of repulsive tip - surface interaction , while for @xmath11=5.0 nm there is a strong attraction between the tip and the surface . for a resting tip , while at zero or negative @xmath11 values the bonds can easily form across the whole tip surface , at larger @xmath11 the bonds form preferentially at the tip apex .
however , during lateral motion , the tip is subject to an attractive force , acting in the normal direction , that arises due to bond stretching as described by the last term of eq .
[ eq.motionz ] . this normal component of attractive force grows with the tip velocity .
therefore , even for large @xmath11 , during sliding the tip is attracted toward the surface , and its distance from the surface becomes very small , entering the repulsive regime .
supramolecular bonds newly - formed during the hold time do not produce an attractive force acting in normal direction , since typical tip - substrate distances are smaller than the equilibrium bond lengths ( this is a characteristic feature of long supramolecular bonds ) .
thus , the absence of the bond - induced attraction between the tip and substrate may lead to an increase of the tip - substrate distance during the hold time , and to the corresponding reduction of the number of attached bonds .
this counterintuitive `` anti - aging '' effect becomes evident for large values of @xmath11 , as shown in fig .
[ figure7](c ) for @xmath11=5.0 nm . such effect has not been observed so far experimentally , since repulsive regime is usually adopted in experiments as reference to assure a tip - substrate contact .
to summarize , a model for the description of dynamics of friction and adhesion caused by cooperative rupture and formation of supramolecular bonds has been proposed and investigated .
we found that a non - equal load sharing between the intact bonds , which results from the curved shape of the afm tip , strongly influences the dynamics of friction and adhesion .
in particular , this effect leads to the plateaus in the normal force as a function of tip - surface distance , which have been observed in the pull - off experiments with supramolecular bonds.@xcite in agreement with such experimental observations , the results of simulations exhibit two distinct regimes of forced rupture , where the maximum normal force is independent of the unloading rate at low driving , and it grows for higher unloading rates . the transitional unloading rate separating these regimes increases with the rate of bond rupture .
in lateral sliding , the model predicts a multiplicity of possible distinctive regimes of motion each corresponding to a specific molecular complex with increasing , decreasing , or constant friction force trends as a function of velocity .
our simulations demonstrate that the position of the maximum of the friction force as a function of velocity is determined by the ratio of the rates for bond formation and rupture .
we have also shown that slide - hold - slide experiments may provide a way to directly estimate the rate of bond formation .
finally , our model predicts an enhancement of adhesion during sliding that should be a characteristic feature of long supramolecular bonds .
the latter may lead to a novel `` anti - ageing '' effect , in which the number of connected bonds , and correspondingly the static friction force , decreases at larger hold times .
we acknowledge enlightening discussions with dr . j. blass and dr .
r. bennewitz from leibnitz institute for new materials , saarbrcken ( germany ) .
r. g. and a. v. acknowledge financial support from the erc advanced grant 320796 - modphysfrict and from the cost action mp1303 _ understanding and controlling nano and mesoscale friction_. m. u. acknowledges the financial support of the israel science foundation , grant no .
m. m. acknowledges financial support from a fellowship program for outstanding postdoctoral researchers from china and india in israeli universities . | * the adhesive and frictional response of an afm tip connected to a substrate through supramolecular host - guest complexes is investigated by dynamic monte carlo simulations .
the variation of the pull - off force with the unloading rate recently observed in experiments is here unraveled by evidencing a simultaneous ( progressive ) break of the bonds at fast ( slow ) rates .
the model reveals the origin of the observed plateaus in the retraction force as a function of tip - surface distance , showing that they ensue from the tip geometrical features . in lateral sliding ,
the model exhibits a wide range of dynamic behaviors ranging from smooth sliding to stick - slip at different velocities , with the average friction force determined by the characteristic formation / rupture rates of the complexes .
in particular , it is shown that for some molecular complexes friction can become almost constant over a wide range of velocities .
also , we show the possibility to exploit ageing effect through slide - hold - slide experiments , in order to infer the characteristic formation rate . finally , our model predicts a novel `` anti - ageing '' effect which is characterized by a decrease of static friction force with the hold time .
such effect is explained in terms of enhancement of adhesion during sliding , especially observed at high driving velocities . * |
hadronic parity - violation has been traditionally analyzed in terms of potential models ; specifically the ddh model@xcite , which is a parity violating single meson exchange picture containing seven phenomenological constants . however , there exist well known discrepancies between experimental measurements and the ddh model@xcite .
some of this discrepancy is no doubt due to nuclear physics uncertainties , but another source may be the use of the model - dependent ddh potential . a possible solution to these problems has recently been proposed by zhu et al.@xcite , restriction of experiments to nuclei with @xmath1 so that nuclear uncertainties are minimal , and analysis using a model - independent picture via effective field theory . at low energies , less than @xmath2
, such an approach is provided by pionless eft ( @xmath3 ) , which has been extremely successful at low energies in the two - body and three - body sector for parity - conserving interactions , including interactions with external currents@xcite . at low energies inclusion of parity - violation
requires only five additional low energy constants ( lec s ) in the nucleon - nucleon interaction .
these lec s involve all possible isospin structures that mix s and p waves with one derivative and are equivalent to the parameters originally posited by danilov@xcite .
the fact that five lec s are needed has also been specifically shown by girlanda , who does this by performing a non - relativistic reduction of all possible one derivative relativistic parity - violating structures that conserve cp@xcite .
calculations for parity - violation using eft methods have heretofore been primarily focused in the two - body sector .
such calculations include parity violation in nucleon - nucleon scattering and in the process @xmath4@xcite .
parity - violating eft calculations have been done in the three - body sector using a hybrid approach , wherein the parity - violating potential is given by @xmath3 , but is used with wavefunctions determined by either the variational method or by a fadeev integral equation technique@xcite .
such calculations include neutron spin rotation and beam asymmetry in @xmath0 interactions .
recently a paper by griesshammer , schindler , and springer predicted the spin rotation of a neutron on a deuterium target up to and including nlo effects in @xmath3@xcite .
however , they only included order of magnitude estimates for the parity violating coefficients and left open the calculation of other possible parity violating observables in @xmath0 interactions . in this paper we set out to obtain estimates for the parity violating coefficients by matching them to the ddh
best " value estimates . as well as calculating the neutron spin rotation on a deuteron target at lo in @xmath3
, we also calculate the beam and target asymmetry at lo in @xmath0 scattering .
estimation of these observables will then allow one to assess the feasibility of @xmath0 interactions as a realistic experimental probe for the five pv lec s .
below we calculate the lo amplitudes for s - p mixing in @xmath0 scattering due to the two - body parity - violating lagrangian .
( since griesshammer and schindler showed that no three - body parity - violating force occurs up to and including nlo , only five lec s exist at lo@xcite . )
predictions are made for pv observables and numerical estimates are given based on ddh best " value estimates . in a future publication we shall present higher order corrections .
the paper is organized as follows . in section
ii we give the form of the two - body parity - violating interaction .
then in section iii we show what diagrams are needed at lo and how to calculate them .
section iv shows how estimates for parity - violating lec s can be obtained , and in section v we show how to relate our amplitudes to observables . finally in section
vi we summarize the results .
the leading order two - body parity - conserving lagrangian in the auxiliary field formalism is given by @xmath5\\\nonumber & -s_{a}^{\dagger}\left(i\partial_{0}+\frac{{\vec{\mathbf{\nabla}}}^{2}}{4m_{n}}-\delta^{({}^1\!s_{0})}_{(-1)}-\delta^{({}^1\!s_{0})}_{(0)}\right)s_{a}+y_{t}\left[s_{a}^{\dagger}n^{t}\bar{p}_{a}n+h.c.\right]\end{aligned}\ ] ] where @xmath6 ( @xmath7 ) is the deuteron ( singlet ) auxiliary field@xcite . here
@xmath8 projects out the @xmath9 channel and @xmath10 projects out the @xmath11 channel .
the auxiliary field formalism is equivalent to the partial wave formalism in which only nucleon fields are used as can be seen by integrating over the auxiliary fields and using a field redefinition@xcite . in this formulation
the two - body parity - violating lagrangian amplitude is given by a form including five low energy constants@xcite .
@xmath12+h.c.\end{aligned}\ ] ] where @xmath13 , and @xmath14 $ ] projects out the isotensor contribution.the deuteron kinetic energy and the term @xmath15 are sub - leading with respect to @xmath16 .
thus at lo the bare deuteron propagator is given by @xmath17 which is then dressed by an infinite number of nucleon bubbles as seen inf fig . [
fig : lodeuteronpropagator ] .
the resulting propagator depends on @xmath18 and @xmath19 , with values adjusted such that the deuteron propagator has its pole at the correct value .
a similar calculation can be carried out for the propagator of the singlet auxiliary field .
this procedure has been carried out in many papers , the end results for the lo deuteron and singlet propagator are listed below@xcite . also we include the constraints imposed on the coefficients @xmath18 , @xmath19 , and their @xmath11 counterparts at lo .
note the presence of the parameter @xmath20 , which is a cutoff imposed by using dimensional regularization with the pds subtraction scheme@xcite .
( here @xmath21 mev is the deuteron binding momentum and @xmath22 , where @xmath23 fm is the scattering length in the @xmath11 channel . )
( 325,56)(90,-10 ) ( 100,-1)(130,-1 ) ( 100,1)(130,1 ) ( 135,1.5)(144,1.5 ) ( 135,-1.5)(144,-1.5 ) ( 150,0)(180,0 ) ( 185,0)(195,0 ) ( 190,-5)(190,5 ) ( 200,0)(221,0 ) ( 244,0)(265,0 ) ( 270,0)(280,0 ) ( 275,-5)(275,5 ) ( 285,0)(306,0 ) ( 329,0)(341,0 ) ( 364,0)(385,0 ) ( 390,0)(400,0 ) ( 395,-5)(395,5 ) ( 405,0).5 ( 410,0).5 ( 415,0).5 @xmath24 @xmath25 @xmath26
as in the parity - conserving case for three - body interactions one needs to solve an infinite sum of diagrams for the parity - violating amplitude at lo@xcite , leading to a coupled set of integral equations .
numerical solution is necessary , as such integral equations can not be solved analytically . in general
we must solve a set of four coupled integral equations .
however , since parity - violation is so small , @xmath27 we can ignore second order pv terms .
then the integral equations for the parity - conserving amplitudes decouple@xcite , and are exactly the same as in previous papers@xcite . the remaining coupled parity - violating integral equations at lo are shown in fig .
[ fig : lo_integral_equations ] , where the boxes represent parity - violating vertices , the double line the dressed deuteron propagator , the double dashed lines the dressed singlet propagator , and the line with arrow the nucleon propagator .
the thick lines represent a sum over both deuteron and singlet propagators .
thus the thick line allows one to represent two feynman diagrams with a single diagram .
there are also diagrams where two dibaryon lines and two nucleon lines meet at a single vertex , due to the three - body force term in the lagrangian .
( 470,60)(90,-20 ) ( 100,25.5)(160,25.5 ) ( 100,23.5)(160,23.5 ) ( 100,-5)(130,-5 ) ( 130,-5)(160,-5 ) ( 130,10)(16.5,10)(0)blackyellow ( 130,10)(0)[c]pv ( 170,10)(179,10 ) ( 170,12.5)(179,12.5 ) ( 190,25.5)(210,25.5 ) ( 190,23.5)(210,23.5 ) ( 210,24.5)(240,24.5 ) ( 190,-5)(220,-5 ) ( 220,-6)(240,-6 ) ( 220,-4)(240,-4 ) ( 210,24.5)(220,-5 ) ( 210,24.5)(5,5)greengreen ( 250,10)(260,10 ) ( 255,5)(255,15 ) ( 270,25.5)(300,25.5 ) ( 270,23.5)(300,23.5 ) ( 300,24.5)(320,24.5 ) ( 320,24.5)(350,24.5 ) ( 270,-5)(300,-5 ) ( 300,-5)(330,-5 ) ( 330,-4)(350,-4 ) ( 330,-6)(350,-6 ) ( 320,24.5)(330,-4.5 ) ( 300,10)(16.5,10)(0)blackred ( 320,24.5)(5,5)greengreen ( 300,10)(0)[c]pc ( 360,10)(370,10 ) ( 365,5)(365,15 ) ( 380,25.5)(410,25.5 ) ( 380,23.5)(410,23.5 ) ( 410,24.5)(430,24.5 ) ( 430,24.5)(460,24.5 ) ( 380,-5)(410,-5 ) ( 410,-5)(440,-5 ) ( 440,-4)(460,-4 ) ( 440,-6)(460,-6 ) ( 430,24.5)(440,-4.5 ) ( 410,10)(16.5,10)(0)blackyellow ( 410,10)(0)[c]pv ( 250,-40)(260,-40 ) ( 255,-45)(255,-35 ) ( 270,-24.5)(300,-24.5 ) ( 270,-26.5)(300,-26.5 ) ( 305,-25.5)(320,-40.5 ) ( 320.5,-40)(335.5,-55 ) ( 318.5,-41)(332.5,-55 ) ( 305,-55)(320,-40 ) ( 320,-40)(335,-25 ) ( 270,-55)(300,-55 ) ( 300,-40)(16.5,10)(0)blackyellow ( 300,-40)(0)[c]pv ( 470,100)(90,-50 ) ( 100,25.5)(130,25.5 ) ( 100,23.5)(130,23.5 ) ( 130,25.5)(160,25.5)3 ( 130,23.5)(160,23.5)3 ( 100,-5)(130,-5 ) ( 130,-5)(160,-5 ) ( 130,10)(16.5,10)(0)blackyellow ( 130,10)(0)[c]pv ( 170,10)(179,10 ) ( 170,12.5)(179,12.5 ) ( 190,25.5)(210,25.5 ) ( 190,23.5)(210,23.5 ) ( 210,24.5)(240,24.5 ) ( 190,-5)(220,-5 ) ( 220,-6)(240,-6)3 ( 220,-4)(240,-4)3 ( 210,24.5)(220,-5 ) ( 210,24.5)(5,5)greengreen ( 250,10)(260,10 ) ( 255,5)(255,15 ) ( 270,25.5)(320,25.5 ) ( 270,23.5)(320,23.5 ) ( 300,24.5)(320,24.5 ) ( 320,24.5)(350,24.5 ) ( 270,-5)(300,-5 ) ( 300,-5)(330,-5 ) ( 330,-4)(350,-4)3 ( 330,-6)(350,-6)3 ( 320,24.5)(330,-4.5 ) ( 300,10)(16.5,10)(0)blackred ( 320,24.5)(5,5)greengreen ( 300,10)(0)[c]pc ( 360,10)(370,10 ) ( 365,5)(365,15 ) ( 380,25.5)(410,25.5 ) ( 380,23.5)(410,23.5 ) ( 410,24.5)(430,24.5 ) ( 430,24.5)(460,24.5 ) ( 380,-5)(410,-5 ) ( 410,-5)(440,-5 ) ( 440,-4)(460,-4)3 ( 440,-6)(460,-6)3 ( 430,24.5)(440,-4.5 ) ( 410,10)(16.5,10)(0)blackyellow ( 410,10)(0)[c]pv ( 250,-40)(260,-40 ) ( 255,-45)(255,-35 ) ( 270,-24.5)(300,-24.5 ) ( 270,-26.5)(300,-26.5 ) ( 305,-25.5)(320,-40.5 ) ( 320.5,-40)(335.5,-55)3 ( 318.5,-41)(332.5,-55)3 ( 305,-55)(320,-40 ) ( 320,-40)(335,-25 ) ( 270,-55)(300,-55 ) ( 300,-40)(16.5,10)(0)blackyellow ( 300,-40)(0)[c]pv this three - body force term enters at lo , only in the doublet s - wave channel .
( note we have not yet projected out any specific channel . )
the momentum integrals are regulated ( 350,200)(160,-160 ) ( 100,25.5)(130,25.5 ) ( 100,23.5)(130,23.5 ) ( 130,25.5)(160,25.5 ) ( 130,23.5)(160,23.5 ) ( 100,-5)(130,-5 ) ( 130,-5)(160,-5 ) ( 130,10)(16.5,10)(0)blackyellow ( 130,10)(0)[c]pv ( 170,10)(179,10 ) ( 170,12.5)(179,12.5 ) ( 175,10)[c]@xmath28 ( 190,25.5)(210,25.5 ) ( 190,23.5)(210,23.5 ) ( 210,24.5)(240,24.5 ) ( 190,-5)(220,-5 ) ( 220,-6)(240,-6 ) ( 220,-4)(240,-4 ) ( 210,24.5)(220,-5 ) ( 210,24.5)(5,5)greengreen ( 250,10)(260,10 ) ( 255,5)(255,15 ) ( 270,25.5)(320,25.5 ) ( 270,23.5)(320,23.5 ) ( 320,24.5)(350,24.5 ) ( 270,-5)(300,-5 ) ( 300,-5)(330,-5 ) ( 330,-4)(350,-4 ) ( 330,-6)(350,-6 ) ( 320,24.5)(330,-4.5 ) ( 300,10)(16.5,10)(0)blackred ( 320,24.5)(5,5)greengreen ( 300,10)(0)[c]pc ( 360,10)(370,10 ) ( 365,5)(365,15 ) ( 380,25.5)(410,25.5 ) ( 380,23.5)(410,23.5 ) ( 410,25.5)(430,25.5)3 ( 410,23.5)(430,23.5)3 ( 430,24.5)(460,24.5 ) ( 380,-5)(410,-5 ) ( 410,-5)(440,-5 ) ( 440,-4)(460,-4 ) ( 440,-6)(460,-6 ) ( 430,24.5)(440,-4.5 ) ( 410,10)(16.5,10)(0)blackred ( 430,24.5)(5,5)greengreen ( 410,10)(0)[c]pc ( 250,-40)(260,-40 ) ( 255,-45)(255,-35 ) ( 270,-24.5)(290,-24.5 ) ( 270,-26.5)(290,-26.5 ) ( 270,-55)(300,-55 ) ( 290,-25.5)(300,-55 ) ( 290,-24.5)(350,-24.5 ) ( 290,-26.5)(350,-26.5 ) ( 300,-55)(320,-55 ) ( 320,-55)(350,-55 ) ( 320,-40)(16.5,10)(0)blackred ( 290,-25.5)(5,5)greengreen ( 320,-40)(0)[c]pc ( 360,-40)(370,-40 ) ( 365,-45)(365,-35 ) ( 380,-24.5)(400,-24.5 ) ( 380,-26.5)(400,-26.5 ) ( 380,-55)(410,-55 ) ( 400,-25.5)(410,-55 ) ( 400,-24.5)(430,-24.5)3 ( 400,-26.5)(430,-26.5)3 ( 430,-24.5)(460,-24.5 ) ( 430,-26.5)(460,-26.5 ) ( 410,-55)(430,-55 ) ( 430,-55)(460,-55 ) ( 430,-40)(16.5,10)(0)blackred ( 400,-25.5)(5,5)greengreen ( 430,-40)(0)[c]pc ( 190,-90)(200,-90 ) ( 195,-95)(195,-85 ) ( 210,-74.5)(260,-74.5 ) ( 210,-76.5)(260,-76.5 ) ( 260,-75.5)(290,-75.5 ) ( 210,-105)(240,-105 ) ( 240,-105)(270,-105 ) ( 270,-104)(320,-104 ) ( 270,-106)(320,-106 ) ( 260,-75.5)(270,-104.5 ) ( 290,-75.5)(320,-75.5 ) ( 240,-90)(16.5,10)(0)blackred ( 290,-90)(16.5,10)(0)blackred ( 260,-75.5)(5,5)greengreen ( 240,-90)(0)[c]pc ( 290,-90)(0)[c]pc ( 330,-90)(340,-90 ) ( 335,-95)(335,-85 ) ( 350,-74.5)(400,-74.5 ) ( 350,-76.5)(400,-76.5 ) ( 400,-75.5)(430,-75.5 ) ( 350,-105)(380,-105 ) ( 380,-105)(410,-105 ) ( 410,-104)(430,-104)3 ( 410,-106)(430,-106)3 ( 430,-104)(460,-104 ) ( 430,-106)(460,-106 ) ( 400,-75.5)(410,-104.5 ) ( 430,-75.5)(460,-75.5 ) ( 380,-90)(16.5,10)(0)blackred ( 430,-90)(16.5,10)(0)blackred ( 400,-75.5)(5,5)greengreen ( 380,-90)(0)[c]pc ( 430,-90)(0)[c]pc ( 190,-140)(200,-140 ) ( 195,-145)(195,-135 ) ( 210,-124.5)(240,-124.5 ) ( 210,-126.5)(240,-126.5 ) ( 240,-124.5)(260,-124.5)3 ( 240,-126.5)(260,-126.5)3 ( 260,-125.5)(290,-125.5 ) ( 210,-155)(240,-155 ) ( 240,-155)(270,-155 ) ( 270,-154)(320,-154 ) ( 270,-156)(320,-156 ) ( 260,-125.5)(270,-154.5 ) ( 290,-125.5)(320,-125.5 ) ( 240,-140)(16.5,10)(0)blackred ( 290,-140)(16.5,10)(0)blackred ( 260,-125.5)(5,5)greengreen ( 240,-140)(0)[c]pc ( 290,-140)(0)[c]pc ( 330,-140)(340,-140 ) ( 335,-145)(335,-135 ) ( 380,-124.5)(400,-124.5)3 ( 380,-126.5)(400,-126.5)3 ( 350,-124.5)(380,-124.5 ) ( 350,-126.5)(380,-126.5 ) ( 400,-125.5)(430,-125.5 ) ( 350,-155)(380,-155 ) ( 380,-155)(410,-155 ) ( 410,-154)(430,-154)3 ( 410,-156)(430,-156)3 ( 430,-154)(460,-154 ) ( 430,-156)(460,-156 ) ( 400,-125.5)(410,-154.5 ) ( 430,-125.5)(460,-125.5 ) ( 380,-140)(16.5,10)(0)blackred ( 430,-140)(16.5,10)(0)blackred ( 400,-125.5)(5,5)greengreen ( 380,-140)(0)[c]pc ( 430,-140)(0)[c]pc using a sharp cutoff @xmath29 .
the three - body force term is cutoff dependent .
this cutoff is convenient because it can be implemented straightforwardly numerically . the three - body force term in the doublet s wave channel
is given by@xcite @xmath30 the cutoff dependence of three - body force term @xmath31 is chosen so that the doublet s wave amplitude produces the correct scattering length@xcite .
( as noted earlier there is no need to include a parity - violating three - body force , as it has been explicitly shown that no such term exists up to and including nlo@xcite . ) to first order in parity - violation the integral equation for the parity - violating amplitude is given by the sum of diagrams shown in fig .
[ fig : lo_diagrams ] .
the first diagram corresponds to a pure parity - violating transition with no scattering in the initial or final channel .
the next set of diagrams has a parity - violating transition with scattering either in the initial or final channel but not both .
( also note that for these diagrams the singlet field acts as an intermediate state , which can only exist in the doublet channel . )
finally we have the set of diagrams with a parity - violating transition and scattering in both the final and initial channels.summing all of these figures one finds the parity - violating amplitude given in eq .
( [ eq : lo_amp ] ) , where @xmath32 is the incoming nucleon momentum and @xmath33 is the outgoing nucleon momentum in the c.m . frame .
since our diagrams are on shell we have @xmath34 , and the total energy in the c.m .
frame is given by @xmath35 .
the vector index letter @xmath36 ( @xmath37 ) , represents the initial ( final ) deuteron auxiliary field polarization .
finally the greek index @xmath38 ( @xmath39 ) is the initial ( final ) spinor index and @xmath40 ( @xmath41 ) is the initial ( final ) isospinor index .
@xmath42 the vector @xmath43 projects out the nucleon - deuteron amplitude in cluster - configuration space and is defined as@xcite @xmath44 and the parity - conserving amplitudes @xmath45 are a vector defined as follows @xmath46 where @xmath47 is the amplitude for @xmath0 scattering and @xmath48 is the amplitude for @xmath0 going to a nucleon and a singlet combination of the remaining nucleons .
( note that we have not yet projected out quartet or doublet channels . )
the expressions @xmath49 and @xmath50 are both matrices defined via .
@xmath51 @xmath52 where the functions @xmath53 , which contain all of the parity - violating dependence are defined as [ seq : kequations ] @xmath54 @xmath55 @xmath56 @xmath57 ( note that the capital letters a , b , and c are used for the singlet auxiliary field polarization and the lowercase letters i , j , and k are used for the deuteron auxiliary field polarization . )
integrating over the energy and picking up the poles from the nucleon propagators in our diagrams eq .
( [ eq : lo_amp ] ) becomes .
@xmath58 where @xmath59 now that we have derived the parity - violating amplitude , we note that it contains the related scattering amplitudes from the parity - conserving sector .
such parity - conserving scattering amplitudes are calculated in@xcite , by numerically solving fadeev s equation in an angular momentum basis . however , as part of this solution one runs into singularities along the real axis . to overcome
this difficulty the method of hetherington and schick is employed , in which the axis of integration is rotated into the complex plane , therefore avoiding the singularities@xcite .
one can then use the solutions along the deformed contour to solve for the amplitudes along the real axis .
details of the procedure to calculate these amplitudes can be found in@xcite . in order to use the solutions to fadeev s equations we need to project out our parity - violating amplitude into an angular momentum basis .
however , unlike the parity - conserving sector , the pv amplitudes mix different angular momentum states . also , since at leading order , spin and orbital angular momentum mix , the appropriate angular momentum basis to use is the total angular momentum @xmath60 .
thus we express our parity - violating amplitude as @xmath61 where the spin angle functions are given by @xmath62 @xmath63 being the spinor part of the spin - angle functions , @xmath64 the appropriate clebsch - gordan coefficient , and the appropriate @xmath65 spherical harmonic . since the spin - angle functions are orthogonal , we can project out the amplitudes in our angular momentum basis via @xmath66 at sufficiently low energies s - p mixing will dominate .
thus we will calculate only the amplitudes @xmath67 ( note j = s here since l=0 ) for all possible values of @xmath68 and @xmath69 .
all spin and angle dependence is contained within the matrix @xmath70 , and the appropriate projections in @xmath71 , @xmath72,and @xmath73 can be found in the appendix .
going to a partial wave basis we finally obtain an expression for the parity - violating partial wave amplitudes .
@xmath74 this expression contains the parity - conserving amplitudes in the partial wave basis of total angular momentum @xmath60 .
( these are equivalent to the parity - conserving amplitudes in the partial wave basis of orbital angular momentum . )
it can be shown straightforwardly that the parity - conserving amplitudes are independent of total angular momentum @xmath71 .
thus we can use the parity - conserving amplitudes as calculated numerically by@xcite and perform the integration numerically in order to obtain the associated parity - violating amplitudes . before integrating eq .
( [ eq : lo_amp_projected ] ) we multiply by the lo deuteron renormalization @xmath75@xcite , and use the renormalized parity - conserving amplitudes .
we find that all the parity - violating lec s occur in the combinations .
@xmath76 ( note @xmath77 does not appear as a @xmath78 transition is not allowed for a first order parity - violating transition in @xmath0 scattering . ) for the sake of convenience we find it useful to make the following definitions .
@xmath79 since these coefficients are unknown , we will write the parity - violating partial wave amplitudes as follows , where @xmath80 is calculated by setting @xmath81 and @xmath82 in the parity - violating partial wave amplitude .
thus the parity - violating amplitude can be written as .
it is clear from eq .
( [ eq : pvamp ] ) that in order to obtain numerical values for the parity - violating amplitude one needs to know the size of the coefficients @xmath84 , which at this time are not determined from either theory or experiment .
nevertheless , we can obtain estimates by matching the @xmath84 onto the familiar ddh coefficients .
we will carry out this procedure in three steps .
first we match the ddh coefficients onto the coefficients of the zhu potential@xcite .
then we match the zhu potential on to the girlanda potential@xcite .
finally we project the coefficients of the girlanda potential onto the coefficients of the auxiliary field formalism .
we also show how all these formalisms can be matched to the familiar danilov parametersthe ddh model@xcite is a single - meson - exchange picture , limited to exchange of the lightest mesons @xmath85 , @xmath86 , and @xmath87 .
, @xmath88 , or @xmath89 by barton s theorem@xcite ] the strong hamiltonian is given by @xmath90 with the strong couplings given approximately by @xmath91 and @xmath92 , while the magnetic moment terms are approximately @xmath93 and @xmath94 .
the phenomenological weak interaction hamiltonian posited by ddh consists of seven weak coupling terms @xmath95 ddh attempted to obtain theoretical predictions for the seven constants using su(6 ) symmetry and quark model techniques .
however , due to the difficulty of this calculation they were only able to come up with reasonable ranges and best " values as shown in table [ tab : ddh ] .
( also shown are estimates by other groups . )
.[tab : ddh]weak nnm couplings .
all numbers are quoted in units of the `` sum rule '' value@xmath96 [ cols="^,^,^,^,^ " , ] finally we should note that the calculation for the beam and target asymmetries were done using both the standard cross section methods and the optical theorem .
plotting the results from both , we found they were indistinguishable .
this agreement confirms that our amplitudes are unitary and acts as a check on the validity of our results . for the values quoted in the table ,
it was found that for the beam and target asymmetries , as well as the values from either the cross sections or optical theorem agreed to less than one percent .
above we calculated the low energy parity - violating @xmath0 transition amplitudes using pionless eft . matching the auxiliary field formalism onto the ddh potential , we made predictions for the coefficients of the auxiliary field formalism . using these amplitudes , we were able to make predictions for the spin rotation , beam asymmetry , and target asymmetry in low energy @xmath0 interactions .
the values obtained for the neutron spin rotation and beam asymmetry are in agreement with values found by other authors@xcite . unfortunately due to the smallness of these values
they will require very precise experiments to measure . however , the five lec s used in parity - violation at lo are not very well determined .
thus upon further experiments it may be found that the values of the lec s are such that the observables for @xmath0 interactions are larger than predicted . the largest contribution to parity violation
was shown to come from the coefficient @xmath97 , which contains the pion contribution , and such experiments should then allow one to determine its value .
it is noted that to first order in parity violation the @xmath78 ( @xmath98 ) term does not contribute .
thus @xmath0 scattering is sensitive to four out of the five pv coefficients . in principle
we should be able to calculate to nlo in @xmath3without the need for parity - violating three - body forces@xcite .
griesshammer , schindler , and springer calculated the nlo parity violating amplitudes using the partially resummed approach which introduces higher order contributions at nlo@xcite .
however , to calculate the nlo contributions without higher order contributions , one must calculate the full off shell lo amplitude .
since a calculation of the full off shell lo amplitude is numerically expensive it will be left to a future publication .
in projecting out the amplitudes one has to perform angular integrations which are given by ( note the polarization and spin indices are summed over corresponding indices that are not explicitly shown in the spin angle functions ) .
each matrix element of @xmath104 has a different projection .
each of these four different projections has in turn four pieces given by eqs .
( [ seq : kequations ] ) .
fortunately many of these terms can be related by time reversal simplifying the projection considerably . for simplicity
we only show how to project out the first @xmath105 piece of the matrix element @xmath106_{11}$ ] as given in ( [ eq : krelation ] ) and ( [ eq : k11 ] ) , and simply quote the other results . in order to project this term out we use properties of spherical tensors and the wigner - eckart theorem to reduce the following expression to a sum over clebsch - gordan coefficients .
integration over the angular variable can be carried out trivially by using ( [ eq : angularprojection2 ] ) leaving a sum of products of clebsch - gordan coefficients . then using symmetry properties of the clebsch - gordan coefficients we find .
i would like to thank harald w. griesshammer and l. girlanda for useful discussions , barry r. holstein for guidance during this project , and u. van kolck for useful input .
this work is supported in part by the national science foundation under grant no .
nsf - phy 0855119 | we calculate the parity - violating amplitudes in the @xmath0 interaction with pionless effective field theory to lo . matching the parity violating low energy constants to the ddh coefficients we make numerical predictions for parity - violating observables .
in particular we give predictions for the spin rotation of a neutron on a deuteron target , and target and beam asymmetries in @xmath0 scattering . |
on the 25-th of november in 1915 , einstein presented his field equations , the basic equations of general relativity , to the prussian academy of sciences in berlin .
this equation had a tremendous impact on physics , in particular on cosmology .
the essence of the theory was expressed by wheeler by the words : _ spacetime tells matter how to move ; matter tells spacetime how to curve_. einsteins theory remained unchanged for about 40 years .
then one started to investigate theories fulfilling mach s principle leading to a variable gravitational constant .
brans - dicke theory was the first realization of an extended einstein theory with variable gravitational constant ( jordans proposal was not widely known ) .
all experiments are , however , in good agreement with einstein s theory and currently there is no demand to change it .
general relativity ( gr ) has changed our understanding of space - time . in parallel ,
the appearance of quantum field theory ( qft ) has modified our view of particles , fields and the measurement process .
the usual approach for the unification of qft and gr to a quantum gravity , starts with a proposal to quantize gr and its underlying structure , space - time .
there is a unique opinion in the community about the relation between geometry and quantum theory : the geometry as used in gr is classical and should emerge from a quantum gravity in the limit ( planck s constant tends to zero ) .
most theories went a step further and try to get a space - time from quantum theory .
then , the model of a smooth manifold is not suitable to describe quantum gravity , but there is no sign for a discrete space - time structure or higher dimensions in current experiments @xcite .
therefore , we conjecture that the model of spacetime as a smooth 4-manifold can be used also in a quantum gravity regime , but then one has the problem to represent qft by geometric methods ( submanifolds for particles or fields etc . ) as well to quantize gr .
in particular , one must give meaning to the quantum state by geometric methods .
then one is able to construct the quantum theory without quantization .
here we implicitly assumed that the quantum state is real , i.e. the quantum state or the wave function has a real counterpart and is not a collection of future possibilities representing some observables .
experiments @xcite supported this view .
then the wave function is not merely representing our limited knowledge of a system but it is in direct correspondence to reality ! then one has to go the reverse way : one has to show that the quantum state is produced by the quantization of a classical state .
it is , however , not enough to have a geometric approach to quantum gravity ( or the quantum field theory in general ) .
what are the quantum fluctuations ?
what is the measurement process ?
what is decoherence and entanglement ? in principle , all these questions have to be addressed too . here
, the exotic smoothness structure of 4-manifolds can help finding a way .
a lot of work was done in the last decades to fulfill this goal .
it starts with the work of brans and randall @xcite and of brans alone @xcite where the special situation in exotic 4-manifolds ( in particular the exotic @xmath0 ) was explained .
one main result of this time was the _ brans conjecture _ : exotic smoothness can serve as an additional source of gravity .
i will not present the whole history where i refer to carl s article . here
i will list only some key results which will be used in the following * exotic smoothness is an extra source of gravity ( brans conjecture is true ) , see asselmeyer @xcite for compact manifolds and sadkowski @xcite for the exotic @xmath0 .
therefore an exotic @xmath0 is always curved and can not be flat ! * the exotic @xmath0 can not be a globally hyperbolic space ( see @xcite for instance ) , i.e. represented by @xmath2 for some 3-manifold .
instead it admits complicated foliations @xcite .
using non - commutative geometry , we are able to study these foliations ( the leaf space ) and get relations to qft . for instance , the von neumann algebra of a codimension - one foliation of an exotic @xmath0 must contain a factor of type @xmath3 used in local algebraic qft to describe the vacuum @xcite . *
the end of @xmath0 ( the part extending to infinity ) is @xmath4 .
if @xmath0 is exotic then @xmath4 admits also an exotic smoothness structure .
clearly , there is always a topologically embedded 3-sphere but there is no smoothly embedded one .
let us assume the well - known hyperbolic metric of the spacetime @xmath4 using the trivial foliation into leafs @xmath5 for all @xmath6 .
now we demand that @xmath4 carries an exotic smoothness structure at the same time .
then we will get only topologically embedded 3-spheres , the leafs @xmath5 .
these topologically embedded 3-spheres are also known as wild 3-spheres . in @xcite
, we presented a relation to quantum d - branes .
finally we proved in @xcite that the deformation quantization of a tame embedding ( the usual embedding ) is a wild embedding so that the image @xmath7 is an infinite polyhedron or the triangulation needs always infinitely many simplices . ] .
furthermore we obtained a geometric interpretation of quantum states : wild embedded submanifolds are quantum states .
importantly , this construction depends essentially on the continuum , because wild embedded submanifolds admit always infinite triangulations .
* for a special class of compact 4-manifolds we showed in @xcite that exotic smoothness can generate fermions and gauge fields using the so - called knot surgery of fintushel and stern @xcite . in the paper @xcite we presented an approach using the exotic @xmath0 where the matter can be generated ( like in qft ) .
* the path integral in quantum gravity is dominated by the exotic smoothness contribution ( see @xcite or by using string theory @xcite ) . the paper is organized as follows . in the following three sections we will explain exotic 4-manifolds and motivate the whole approach by using the path integral for the einstein - hilbert action .
here we will also present how to couple the matter and gauge fields to this theory . for a 4-manifold , there are two main invariants the euler and pontrjagin class which determine the main topological invariant of a 4-manifold , the intersection form . in section
[ sec : action - induced - by - top ] , we will obtain the einstein - hilbert and holst action by using these two classes . at the first view , this section is a little bit isolated from the previous and subsequent sections but we will use this result later during the study of the scaling . in the main section [ sec : wild - embeddings ] , we will construct the foliation of an exotic @xmath0 of codimension ( equivalent to a lorentz structure ) . following connes , @xcite the leaf space is an operator algebra constructed from the geometrical information of the foliation ( holonomy groupoid ) .
this operator algebra is a factor @xmath8 von neumann algebra and we will use the tomita - takesaki modular theory to uncover the structure of the foliation .
it is not the first time that this factor was used for quantum gravity and we refer to the paper @xcite for a nice application .
states in this operator algebra are represented by equivalence classes of knotted curves ( element of the kauffman bracket skein module ) .
the reconstruction of the spatial space from the states gives a wild embedded 3-sphere as geometrical representation of the state .
surprisingly , it fits with the properties of the exotic @xmath0 .
if one introduces a global foliation of the exotic @xmath0 by a global time then one obtains a foliation into wild embedded 3-spheres .
in contrast , if one uses a local but complicate foliation then this wild object can be omitted and one obtains a state given by a finite collection of knotted curves .
interestingly , the operator algebra can be understood as observable algebra given by a deformation quantization ( turaev - drinfeld quantization @xcite ) of the classical observable algebra ( poisson algebra of holonomies a la goldman @xcite ) . in section [ sec :
action - at - the - quantum - level ] , we will use the splitting of the operator algebra ( [ eq : crossed - product - factor - iii ] ) given by tomita - takesaki modular theory to introduce the dynamics ( see connes and rovelli @xcite with similar ideas ) .
finally we will obtain a quantum action ( [ eq : quantum - action ] ) in the quantized calculus of connes @xcite .
then the scaling behavior is studied in the next section . for large scales ,
the action can be interpreted as a non - linear sigma model .
the renormalization group ( rg ) flow analysis @xcite gives the einstein equations for large scales .
the short - scale analysis is much more involved , yielding for small fluctuations the einstein - hilbert action and a non - minimally coupled scalar field .
in particular , we will obtain a @xmath9dimensional fractal structure . in section
[ sec : some - properties ] we will present some direct consequences of this approach : the nonlinear graviton @xcite , a relation to lattice gauge field theory with a discussion of discreteness and the appearance of dimensional reduction from 4d to 2d . in section [ sec : where - does - fluctuation ] we will discuss the answer to a fundamental question : where does the quantum fluctuations come from ?
the main result of this section can be written as : _ the set of canonical pairs ( as measurable variables in the theory ) forms a fractal subset of the space of all holonomies .
then we can only determine the initial condition up to discrete value ( given by the canonical pair ) and the chaotic behavior of the foliation ( i.e. the anosov flow ) makes the limit not predictable .
_ this interesting result is followed by a section where we will discuss the collapse of the wavefunction by the gravitational interaction by calculating the minimal decoherence time
. furthermore we will discuss entanglement and the measurement process . in section [ sec
: some - implications - for.cosmology ] we will list our work in cosmology which uses partly the results of this paper . in the last section [ sec : conclusion - and - open - questions ] , we will discuss some consequences and open questions .
some mathematical prerequisites are presented in three appendices .
this article is dedicated to my only teacher , carl h. brans for 20 years of collaboration and friendship .
he is the founder of this research area .
we had and will have many interesting discussions .
carl always asked the right question and put the finger on many open points . during the 7 years of writing our book
, we had a very fruitful collaboration and i learned so much to complete even this work .
carl , i hope for many discussions with you in the future .
i m very glad to count on your advice .
happy birthday !
why am i going to concentrate on a concept like exotic smoothness ?
einstein used the equivalence principle as a key principle in the development of general relativity .
every gravitational field can be locally eliminated by acceleration .
then , the spacetime is locally modeled as subsets of the flat @xmath0 or the equivalence principle enforces us to use the concept of a manifold for spacetime .
together with the smoothness of the dynamics ( usage of differential equations ) , we obtain a smooth 4-manifold as model for the spacetime in agreement with the current experimental situation .
a manifold consists of charts and transition functions forming an atlas which covers the manifold completely . _
the smooth atlas is called the smoothness structure _ of the manifold .
it was an open problem for a long time whether every topological manifold admits a unique smooth atlas . in 1957 , milnor found the first counterexample : the construction of a 7-sphere with at least 8 different smoothness structures .
later it was shown that all manifolds of dimension larger than 4 admit only a finite number of distinct smoothness structures .
the real breakthrough for 4-manifolds came in the 80 s where one constructed infinitely many different smoothness structures for many compact 4-manifolds ( countably infinite ) and for many non - compact 4-manifolds ( uncountably infinite ) including the @xmath0 . in all dimensions smaller than four
, there is only one smoothness structure ( up to diffeomorphisms ) , the standard structure .
the standard @xmath0 is simply characterized by the unique property to split smoothly like @xmath10 .
all other distinct smoothness structures are called _ exotic smoothness _ structures .
these structures are different , nonequivalent , smooth descriptions of the same topological manifold , a different atlas of charts . in case of the exotic @xmath0 ,
the difference is tremendous : the standard @xmath0 needs one chart ( and every other description can be reduced to it ) whereas every known exotic @xmath0 admits infinitely many charts ( which can not be reduced to a simpler description ) .
so , the spacetime exhibits a much larger complexity by using an exotic smoothness structure , but why is dimension 4 so special ?
there is a good description in @xcite and i will give a short account now .
at first we have to discuss the question : how do i build an atlas for a smooth manifold ?
the answer is given by considering the construction of diffeomorphisms .
every diffeomorphism is locally given by the solution of @xmath11 for a real function @xmath12 over the manifold .
the fixed points of this equation are the critical points of @xmath12 . in case of isolated critical points
, one can reproduce the structure of the manifold ( this is called morse theory ) .
every critical point leads to the attachment of a handle , a submanifold like @xmath13 , i.e. the @xmath14handle ( where @xmath15 is the @xmath14disk ) .
in many cases , the corresponding structure of the manifold , the handle body , can be very complicated but there are rules ( handle sliding ) to simplify them . in all dimensions except dimension 4 .
therefore , two handle bodies can be described by the same 4-manifold topologically but differ in the smooth description .
one of the most surprising aspects of exotic smoothness is the existence of exotic @xmath0 s . in all other dimensions @xcite ,
the euclidean space @xmath16 with @xmath17 admits a unique smoothness structure , up to diffeomorphisms .
beginning with the first examples @xcite , taubes @xcite and freedman / demichelis @xcite constructed countably many large and small exotic @xmath0 s , respectively .
a small exotic @xmath0 embeds smoothly in the 4-sphere whereas a large exotic @xmath0 can not be embedded in that way . for the following we need some simple definitions : the connected sum @xmath18 and the boundary connected sum @xmath19 of manifolds .
let @xmath20 be two @xmath21-manifolds with boundaries @xmath22 .
the _ connected sum _
@xmath23 is the procedure of cutting out a disk @xmath24 from the interior @xmath25 and @xmath26 with the boundaries @xmath27 and @xmath28 , respectively , and gluing them together along the common boundary component @xmath29 .
the boundary @xmath30 is the disjoint sum of the boundaries @xmath22 .
the _ boundary connected sum _
@xmath31 is the procedure of cutting out a disk @xmath32 from the boundary @xmath33 and @xmath34 and gluing them together along @xmath35 of the boundary .
then the boundary of this sum @xmath31 is the connected sum @xmath36 of the boundaries @xmath22 .
large exotic @xmath0 can be constructed using the failure to arbitrarily split a compact , simply - connected 4-manifold . for every topological 4-manifold one
knows how to split this manifold _ topologically _ into simpler pieces using the work of freedman @xcite .
donaldson @xcite , however , that some of these 4-manifolds do not exist as smooth 4-manifolds .
this contradiction between the continuous and the smooth case produces the first examples of exotic @xmath0 .
below we discuss one of these examples .
one starts with a compact , simply - connected 4-manifold @xmath37 classified by the intersection form @xcite @xmath38 a quadratic form over the second integer homology group . in the first construction of a large exotic @xmath0 ,
one starts with the k3 surface as 4-manifold having the intersection form @xmath39 with the the matrix @xmath40 : @xmath41.\ ] ] the work of donaldson @xcite shows that a closed , * smooth * , simply - connected , compact 4-manifold @xmath42 with intersection form @xmath43 does not exist .
freedman @xcite showed , however , that there is a topological splitting @xmath44 with the @xmath45times connected sum @xmath46 ( see above ) which fails to be smooth .
this splitting means that we glue together the two manifolds @xmath47 and @xmath48 along the common boundary @xmath49 ( @xmath50 is the 4-disk or 4-ball ) .
now we define the interior @xmath51 .
the splitting ( [ eq : splitting - k3 ] ) gives a way to represent the @xmath52 part of the intersection form ( [ eq : intersection - k3 ] ) by using @xmath37 but that fails smoothly .
so , choosing a topological splitting @xmath53\right)\cup\left(\#_{3}(s^{2}\times s^{2})\setminus d^{4}\right)\end{aligned}\ ] ] gives a @xmath54 $ ] inside the @xmath55 .
the interior of @xmath54 $ ] defines a manifold @xmath56 glued to a ( topological ) 4-disk @xmath57 along the common boundary , i.e. @xmath58 topologically .
@xmath59 is homeomorphic to @xmath0 but the non - existence of the smooth splitting implies that it is an exotic @xmath0 and there is no smooth embedded @xmath60 ( otherwise the topological splitting is smooth ) .
this failure for a smooth embedding implies also that such exotic @xmath0 s do not embed in the 4-sphere , i.e. it is a large exotic @xmath0 .
the details of the construction can be found in our book @xcite ( section 8.4 ) .
gompf @xcite introduced an important tool for finding new exotic @xmath61 from others , the end - sum @xmath62 .
let @xmath63 be two topological @xmath61 s .
the end - sum @xmath64 is defined as follows : let @xmath65 and @xmath66 be smooth properly embedded rays with tubular neighborhoods @xmath67 and @xmath68 , respectively . for convenience , identify the two semi - infinite intervals with @xmath69 and @xmath70 $ ] leading to diffeomorphisms , @xmath71 and @xmath72\times\mathbb{r}^{3}$ ] .
then define @xmath73 as the end sum of @xmath74 and @xmath75 . with a little checking ,
it is easy to see that this construction leads to @xmath64 as another topological @xmath76 however , if @xmath63 are themselves exotic , then so will @xmath64 and in fact , it will be a `` new '' exotic manifold , since it will not be diffeomorphic to either @xmath74 or @xmath75 .
gompf used this technique to construct a class of exotic @xmath61 s none of which can be embedded smoothly in the standard @xmath61 . by an extension of donaldson theory for a special class of open 4-manifolds , so - called end - periodic 4-manifolds ,
taubes @xcite gives a continuous family of exotic @xmath0 which was extended by gompf to a continuous 2-parameter family @xmath77 .
small exotic @xmath0 s are again the result of anomalous smoothness in 4-dimensional topology but of a different kind than for large exotic @xmath0 s . in 4-manifold topology @xcite , a homotopy - equivalence between two compact , closed , simply - connected 4-manifolds implies a homeomorphism between them ( a so - called h cobordism ) , but donaldson @xcite provided the first smooth counterexample , i.e. both manifolds are generally not diffeomorphic to each other .
the failure can be localized in some contractible submanifold ( akbulut cork ) so that an open neighborhood of this submanifold is a small exotic @xmath0 .
the whole procedure implies that this exotic @xmath0 can be embedded in the 4-sphere @xmath78 .
below we discuss the details for one of these examples . in 1975
casson ( lecture 3 in @xcite ) described a smooth 5-dimensional h - cobordism between compact 4-manifolds and showed that they `` differ '' by two proper homotopy @xmath61 s ( see below ) .
freedman knew , as an application of his proper h - cobordism theorem , that the proper homotopy @xmath61 s were @xmath61 .
after hearing about donaldson s work in march 1983 , freedman realized that there should be exotic @xmath61 s and , to find one , he produced the second part of the construction below involving the smooth embedding of the proper homotopy @xmath61 s in @xmath78 .
unfortunately , it was necessary to have a compact counterexample to the smooth h - cobordism conjecture , and donaldson did not provide this until 1985 @xcite .
the idea of the construction is simply given by the fact that every such smooth h - cobordism between non - diffeomorphic 4-manifolds can be written as a product cobordism except for a compact contractible sub - h - cobordism @xmath79 , the akbulut cork .
an open subset @xmath80 homeomorphic to @xmath81\times{{\mathbb{r}}^{4}}$ ] is the corresponding sub - h - cobordism between two exotic @xmath61 s .
these exotic @xmath61 s are called ribbon @xmath61 s .
they have the important property of being diffeomorphic to open subsets of the standard @xmath61 . that stands in contrast to the previous defined examples of kirby , gompf and taubes . to be more precise ,
consider a pair @xmath82 of homeomorphic , smooth , closed , simply - connected 4-manifolds .
the transformation from @xmath83 to @xmath84 visualized by a h - cobordism can be described by the following construction .
+ _ let @xmath59 be a smooth h - cobordism between closed , simply connected 4-manifolds @xmath83 and @xmath84 .
then there is an open subset @xmath85 homeomorphic to @xmath81\times{{\mathbb{r}}^{4}}$ ] with a compact subset @xmath86 such that the pair @xmath87 is diffeomorphic to a product @xmath81\times(x_{-}\setminus k ,
u\cap x_{-}\setminus k)$ ] .
the subsets @xmath88 ( homeomorphic to @xmath61 ) are diffeomorphic to open subsets of @xmath61 .
if @xmath83 and @xmath84 are not diffeomorphic , then there is no smooth 4-ball in @xmath89 containing the compact set @xmath90 , so both @xmath89 are exotic @xmath61 s . _
+ thus , remove a certain contractible , smooth , compact 4-manifold @xmath91 ( called an akbulut cork ) from @xmath83 , and re - glue it by an involution of @xmath92 , i.e. a diffeomorphism @xmath93 with @xmath94 and @xmath95 for all @xmath96 .
this argument was modified above so that it works for a contractible _ open _ subset @xmath97 with similar properties , such that @xmath98 will be an exotic @xmath61 if @xmath84 is not diffeomorphic to @xmath83 .
furthermore @xmath98 lies in a compact set , i.e. a 4-sphere or @xmath98 is a small exotic @xmath0 . in the next subsection
we will see how this results in the construction of handle bodies of exotic @xmath61 . in @xcite freedman and demichelis
constructed also a continuous family of small exotic @xmath0 .
one of the characterizing properties of an exotic @xmath0 ( all known examples ) is the existence of a compact subset @xmath99 which can not be surrounded by any smoothly embedded 3-sphere ( and homology 3-sphere bounding a contractible , smooth 4-manifold ) .
let @xmath100 be the standard @xmath0 ( i.e. @xmath101 smoothly ) and let @xmath1 be a small exotic @xmath0 with compact subset @xmath99 which can not be surrounded by a smoothly embedded 3-sphere .
then every completion @xmath102 of an open neighborhood @xmath103 is not bounded by a 3-sphere @xmath104 . however , @xmath1 is a small exotic @xmath0 and there is a smooth embedding @xmath105 in the standard @xmath0 .
then the completion of the image @xmath106 has the boundary @xmath107 as subset of @xmath100 .
so , we have the strange situation that an open subset of the standard @xmath100 represents a small exotic @xmath1 . in case of the large exotic @xmath0 ,
the situation is much more complicated .
a large exotic @xmath0 does not embed in any smooth 4-manifold which is simpler than the manifold used for the construction of this exotic @xmath0 .
above we considered the example of a large exotic @xmath0 constructed from a k3 surface .
therefore this large exotic @xmath0 embeds in the k3 surface but not in simpler 4-manifolds like @xmath108 . as of now , we only know of exotic @xmath0 s represented by an infinite number of
coordinate patches .
this naturally makes it difficult to provide an explicit description of a metric .
however , in @xcite , a suggestion to overcome this limitation is provided by the consideration of periodic explicitly described coordinate patches making use of more complex pieces , so - called handles , and even more complex gluing maps .
then one also gets infinite structures of handles but with a clear picture : the coordinate patches have a periodic structure .
* handles * every 4-manifold can be decomposed using standard pieces such as @xmath109 , the so - called @xmath110-handle attached along @xmath111 to the @xmath112handle @xmath113 . in the following we need two possible cases : the 1-handle @xmath114 and the 2-handle @xmath115 .
these handles are attached along their boundary components @xmath116 or @xmath117 to the boundary @xmath60 of the @xmath112handle @xmath50 ( see @xcite for the details ) .
the attachment of a 2-handle is defined by a map @xmath118 , the embedding of a circle @xmath119 into the 3-sphere @xmath60 , i.e. a knot . this knot into @xmath60
can be thickened ( to get a knotted solid torus ) .
the important fact for our purposes is the freedom to twist this knotted solid torus ( so - called dehn twist ) .
the ( integer ) number of these twists ( with respect to the orientation ) is called the framing number or the framing .
thus the gluing of the 2-handle on @xmath50 can be represented by a knot or link together with an integer framing .
the simplest example is the unknot with framing @xmath120 representing the complex projective space @xmath108 or with reversed orientation @xmath121 , respectively .
the 1-handle will be glued by the map of @xmath122 represented by two disjoint solid 2-spheres @xmath123 .
akbulut @xcite introduced another description .
he observed that a 1-handle is something like a cut - out 2-handle with a fixed framing .
we remark that all details can be found in @xcite .
now we are ready to discuss the handle body decomposition of an exotic @xmath0 by bizaca and gompf @xcite .
* handle decomposition of small exotic @xmath0 * first it is very important to notice that the exotic @xmath0 is the * interior * of the handle body described below ( since the handle body has a non - null boundary and is compact ) .
the construction of the handle body can be divided into two parts .
the first part is a submanifold consisting of a pair of a 1- and a 2-handle .
this pair can be canceled topologically by using a casson handle and we obtain the topological 4-disk @xmath50 with @xmath0 as interior . this submanifold is a smooth 4-manifold with a boundary that can be covered by a finite number of charts .
the smoothness structure of the exotic @xmath0 , however , depends mainly on the infinite casson handle .
* casson handle * now consider the casson handle and its construction in more detail .
briefly , a casson handle @xmath124 is the result of attempts to embed a disk @xmath125 into a 4-manifold . in most cases this attempt fails and
casson @xcite looked for a substitute , which is now called a casson handle .
freedman @xcite showed that every casson handle @xmath124 is homeomorphic to the open 2-handle @xmath126 but in nearly all cases it is not diffeomorphic to the standard handle @xcite .
the casson handle is built by iteration , starting from an immersed disk in some 4-manifold @xmath127 , i.e. an injective smooth map @xmath128 every immersion @xmath129 is an embedding except on a countable set of points , the double points .
one can kill one double point by immersing another disk into that point .
these disks form the first stage of the casson handle . by iteration
one can produce the other stages . finally consider not the immersed disk but rather a tubular neighborhood @xmath115 of the immersed disk including each stage .
the union of all neighborhoods of all stages is the casson handle @xmath124 .
so , there are two input data involved with the construction of a @xmath124 : the number of double points in each stage and their orientation @xmath130
. thus we can visualize the casson handle @xmath124 by a tree : the root is the immersion @xmath129 with @xmath110 double points , the first stage forms the next level of the tree with @xmath110 vertices connected with the root by edges etc .
the edges are evaluated using the orientation @xmath130 .
every casson handle can be represented by such an infinite tree .
the casson handle @xmath131 having an immersed disk with one ( positively oriented ) self - intersection ( or double point ) is the simplest casson handle represented by the simplest tree @xmath132 having one vertex in each level connected by one edge with evaluation @xmath133 .
one of the characterizing properties of an exotic @xmath0 ( all known examples ) is the existence of a compact subset @xmath99 which can not be surrounded by any smoothly embedded 3-sphere ( and homology
3-sphere bounding a contractible , smooth 4-manifold ) .
let @xmath100 be the standard @xmath0 ( i.e. @xmath101 smoothly ) and let @xmath1 be a small exotic @xmath0 with compact subset @xmath99 which can not be surrounded by a smoothly embedded 3-sphere .
then every completion @xmath102 of an open neighborhood @xmath103 is not bounded by a 3-sphere @xmath104 , but @xmath1 is a small exotic @xmath0 and there is a smooth embedding @xmath105 in the standard @xmath0 .
then the completion of the image @xmath106 has the boundary @xmath107 as subset of @xmath100 .
so , we have the strange situation that an open subset of the standard @xmath100 represents a small exotic @xmath1 .
now we will describe @xmath1 .
historically it was constructed by using a counterexample of the smooth h - cobordism theorem @xcite .
then the compact subset @xmath134 is given by a non - canceling 1-/2-handle pair .
the attachment of a casson handle @xmath124 cancels this pair topologically .
then one obtains the 4-disk @xmath50 with interior @xmath100 , but this cancellation of the 1/2-handle pair can not be done smoothly and one obtains a small exotic @xmath1 which is schematically given by @xmath135 .
remember @xmath1 is a small exotic @xmath0 , i.e. @xmath1 is embedded into the standard @xmath100 by definition .
the completion @xmath136 of @xmath137 has a boundary given by the 3-manifold @xmath138 .
there is also the possibility to construct @xmath138 directly as the limit @xmath139 of a sequence @xmath140 of 3-manifolds . to construct this sequence of 3-manifolds @xcite
, one can use the kirby calculus , i.e. one represents the compact subset @xmath134 by 1- and 2-handles pictured by a link say @xmath141 where the 1-handles are represented by a dot ( so that surgery along this link gives @xmath134 ) @xcite .
then one attaches a casson handle to this link @xcite . as an example see figure [ fig : link - picture - for - k ] .
the casson handle is given by a sequence of whitehead links ( where the unknotted component has a dot ) which are linked according to the tree ( see the right figure of figure [ fig : building - block - simplest - ch ] for the building block and the left figure for the simplest casson handle given by the unbranched tree ) .
for the construction of a 3-manifold which surrounds the compact @xmath134 , one considers @xmath142stages of the casson handle and transforms the diagram to a real link ( the dotted components are changed to usual components with framing @xmath143 ) . by handle manipulations one obtains a knot so that the @xmath21-th ( untwisted ) whitehead double of this knot represents the desired 3-manifold ( by using surgery ) .
then our example in figure [ fig : link - picture - for - k ] will result in the @xmath21-th untwisted whitehead double of the pretzel knot @xmath144 , figure [ fig : pretzel - knot ] ( see @xcite for the handle manipulations )
. then this sequence of 3-manifolds @xmath145 characterizes the exotic smoothness structure of @xmath1 .
the limit of this sequence @xmath139 gives a wild embedded 3-manifold @xmath138 whose physical relevance will be explained later .
here , we will motivate the appearance of exotic smoothness by discussing the path integral for the einstein - hilbert action . for simplicity ,
we consider general relativity without matter ( using the notation of topological qft ) .
space - time is a smooth oriented 4-manifold @xmath127 which is non - compact and without boundary . from the formal point of view ( no divergences of the metric ) one is able to define a boundary @xmath146 at infinity .
the classical theory is the study of the existence and uniqueness of ( smooth ) metric tensors @xmath147 on @xmath127 satisfying the einstein equations subject to suitable boundary conditions . in the first order hilbertpalatini formulation ,
one specifies an @xmath148-connection @xmath149 together with a cotetrad field @xmath150 rather than a metric tensor .
fixing @xmath151 at the boundary , one can derive first order field equations in the interior ( now called _ bulk _ ) which are equivalent to the einstein equations provided that the cotetrad is non - degenerate .
the theory is invariant under space - time diffeomorphisms @xmath152 . in the particular case of the space - time @xmath153 ( topologically )
, we have to consider smooth 4-manifolds @xmath154 as parts of @xmath127 whose boundary @xmath155 is the disjoint union of two smooth 3-manifolds @xmath156 and @xmath157 to which we associate hilbert spaces @xmath158 of 3-geometries , @xmath159 .
these contain suitable wave functionals of connections @xmath160 .
we denote the connection eigenstates by @xmath161 .
the path integral , @xmath162\right)\label{eq : path - integral}\ ] ] is the sum over all connections @xmath149 matching @xmath163 , and over all @xmath150 .
it yields the matrix elements of a linear map @xmath164 between states of 3-geometry .
our basic gravitational variables will be cotetrad @xmath165 and connection @xmath166 on space - time @xmath127 with the index @xmath167 to present it as 1-forms and the indices @xmath168 for an internal vector space @xmath79 ( used for the representation of the symmetry group ) .
cotetrads @xmath150 are square - roots of metrics and the transition from metrics to tetrads is motivated by the fact that tetrads are essential if one is to introduce spinorial matter .
@xmath165 is an isomorphism between the tangent space @xmath169 at any point @xmath170 and a fixed internal vector space @xmath79 equipped with a metric @xmath171 so that @xmath172 .
here we used the action @xmath173=\intop_{m_{i , f}}\epsilon_{ijkl}(e^{i}\wedge e^{j}\wedge\left(da+a\wedge a\right)^{kl})+\intop_{\partial m_{i , f}}\epsilon_{ijkl}(e^{i}\wedge e^{j}\wedge a^{kl})\label{eq : action - with - boundary}\ ] ] in the notation of @xcite .
the boundary term @xmath174 is equal to twice the trace over the extrinsic curvature ( or the mean curvature ) .
for fixed boundary data , ( [ eq : path - integral ] ) is a diffeomorphism invariant in the bulk .
if @xmath175 are diffeomorphic , we can identify @xmath176 and @xmath177 i.e. we close the manifold @xmath154 by identifying the two boundaries to get the closed 4-manifold @xmath178 .
provided that the trace over @xmath179 can be defined , the partition function , @xmath180\right)\label{eq : path - integral-1}\ ] ] where the integral is now unrestricted , is a dimensionless number which depends only on the diffeomorphism class of the smooth manifold @xmath178 . in case of the manifold @xmath154 , the path integral ( as transition amplitude ) @xmath181 is the diffeomorphism class of the smooth manifold relative to the boundary . the diffeomorphism class of the boundary , however , is unique and the value of the path integral depends on the topology of the boundary as well on the diffeomorphism class of the interior of @xmath154 . therefore we will shortly write @xmath182 and consider the sum of manifolds like @xmath183 with the amplitudes @xmath184 where we sum ( or integrate ) over the connections and frames on @xmath185 ( see @xcite )
. then the boundary term @xmath186=\intop_{\sigma_{f}}\epsilon_{ijkl}(e^{i}\wedge
e^{j}\wedge a^{kl})=\intop_{\sigma_{f}}h\sqrt{h}d^{3}x\label{eq : boundary - term - eh}\ ] ] is needed where @xmath187 is the mean curvature of @xmath157 corresponding to the metric @xmath188 at @xmath157 ( as restriction of the 4-metric ) . in the path integral ( [ eq : path - integral ] )
, one integrates over the frames and connections .
the possibility of singular frames was discussed at some places ( see @xcite ) . the cotetrad field @xmath189 changes w.r.t .
the smooth map @xmath190 by @xmath191
. the transformation matrix @xmath192 has maximal rank @xmath193 for every regular value of the smooth map , but at the critical points @xmath194 of @xmath12 , some derivatives vanish and one has a smaller rank at the point @xmath194 , called a singular point .
then there is no inverse frame ( or tetrad field ) at this point .
usually singular frames are of this nature and one can decompose every singular frame into a product of a regular frame and a ( singular ) transformation induced by a smooth map .
how can one interpret these singularities ?
at this point one needs some differential topology .
a homeomorphism can be arbitrarily and accurately approximated by smooth mappings ( see @xcite , theorem 2.6 ) , i.e. in a neighborhood of a homeomorphism one always finds a smooth map .
secondly , there is a special class of smooth maps , the stable maps . here , two smooth maps are stable equivalent if both maps agree after a diffeomorphism of the corresponding manifolds @xcite . here
we are interested into smooth mappings from 4-manifolds into 4-manifolds . by a deep result of mather @xcite , stable mappings for this dimension are dense in all smooth mappings of 4-manifolds . in @xcite
, we analyzed this situation : the approximation of a homeomorphism by a stable map .
if this smooth map has no singularities then we can perturb them to a diffeomorphism . for a singular map
, however , we showed that it induces a change of the smoothness structure .
then , a singular frame corresponds to a regular frame in a different smoothness structure .
the path integral changed the domain of integration : @xmath195 we remark that this change is unique for dimension four .
no other dimension has this plethora of smoothness structures which can be used to express the singular frames .
the inclusion of exotic smoothness changed the description of trivial spaces like @xmath0 completely . instead of a single chart , we have now an infinite sequence of charts or an infinite sequence of 4-dimensional submanifolds .
we will describe it more completely later .
each submanifold is bounded by a 3-manifold ( different from a 3-sphere ) and we obtain a sequence of 3-manifolds @xmath196 characterizing the smoothness structure .
the sequence of 3-manifolds divides the path integral into a product @xmath197 and we have to think about the boundary term ( [ eq : boundary - term - eh ] ) .
in @xcite we analyzed this term : the boundary @xmath198 seen as embedding into the spacetime @xmath127 can be described locally as spinor @xmath199 and one obtains for the boundary term @xmath200 the dirac action with the dirac operator @xmath201 and @xmath202(see @xcite for the construction of @xmath199 ) .
in particular we obtained the eigenvalue equation @xmath203 , i.e. the mean curvature is the eigenvalue of the dirac operator which has compact spectrum ( from the compactness of @xmath198 ) or we obtained discrete levels of geometry .
this result enforced us to identify the 3-manifolds ( or the parts ) with the matter content .
furthermore the path integral of the boundary can be carried out by an integration along @xmath199 ( see @xcite).with some effort @xcite , one can extend this boundary term to a tubular neighborhood @xmath204 $ ] of the boundary @xmath198 . however , the relation ( [ eq : boundary - term - dirac ] ) is only true for simple ( i.e. irreducible ) 3-manifolds , i.e. for complements of a knot admitting hyperbolic structure . for more complex 3-manifolds , we have the following simple scheme : the knot complements
are connected by torus bundles ( locally written as @xmath205 $ ] ) .
therefore we also have to describe these bundles by using the boundary term . in @xcite
we described this situation by using the geometrical properties of these bundles and we will give a short account of these ideas in subsection [ sub : the - graviton - propagator ] .
simply expressed , in this bundle one has a flow of constant curvature along the tube .
the constant curvature connections are given by varying the chern - simons functional .
now following floer @xcite , the 4-dimensional version of this flow equation is the instanton equation ( or the self - dual equation ) leading to the correct yang - mills functional ( chern - simons gives the pontrjagin class and the instanton equation makes it to the yang - mills functional ) .
more importantly , the three possible types of torus bundles fit very good into the current scheme of three gauge field interactions ( see @xcite ( section 8) ) .
now we have the following picture : fermions as hyperbolic knot complements and gauge bosons as torus bundles .
both components together are forming an irreducible 3-manifold which is connecting to the remaining space by a @xmath206boundary ( see the prime decomposition in appendix b ) .
this connection via @xmath207 $ ] ( @xmath206bundle ) is the only connection between matter and space . here , there is only one interpretation :
this @xmath206bundle must be interpreted as gravity . in this section
we will support this conjecture and construct the corresponding action .
at first we will fix the model , i.e. let @xmath208 and @xmath209 be the 3-manifolds for matter and space , respectively .
the connected sum @xmath18 of both components represents the whole spatial component @xmath210\right)\cup_{s^{2}}\sigma_{s}=\sigma_{m}\#s^{3}\#\sigma_{s}\ ] ] of the spacetime .
the decomposition above showed the geometry of the@xmath206bundle ( in the sense of thurston , see appendix b ) to be the spherical geometry with isometry group @xmath211 .
the idea of the following construction can be simply expressed : the 2-sphere @xmath212 explores locally the curvature of the space where the curvature is given by the inverse volume @xmath213 of the 2-sphere @xmath212 .
the 2-sphere can be written as a homogenous space @xmath214 also known as hopf bundle .
as mentioned above , the geometry of the bundle @xmath207 $ ] ( interpreted as an equator region of @xmath60 ) is the spherical geometry with isometry group @xmath211 .
so , as a local model we have an embedding of a 3-manifold ( as the spatial component for a fixed time ) into the spacetime with local lorentz symmetry ( represented by @xmath215 ) . from the mathematical point of view
, it is a reductive cartan geometry@xcite over the homogenous space @xmath216 , the 3-dimensional hyperbolic space . for the moment ,
let us extend this symmetry to the spacetime @xmath127 itself .
a cartan connection @xmath149 decomposes as a @xmath217valued connection @xmath218 ( @xmath219 denotes the lie algebra of @xmath211 ) and a coframe field @xmath150 ( with values in @xmath220 ) as @xmath221 by using the scale @xmath222(in agreement with the physical units ) and with curvature @xmath223 then for the spacetime ( as 4-manifold ) , we interpret the cartan connection @xmath149 as the connection of the frame bundle ( with respect to the lorentz structure ) .
now we have to think about what characterizes the @xmath206bundle in a 4-manifold , i.e. a surface bundle over a surface ( at least locally ) .
it is known that a surface bundle over a surface is topologically described by the euler class as well as the pontrjagin class ( via the hirzebruch signature theorem ) .
therefore we choose the sum of the euler and pontrjagin class for the frame bundle as action @xmath224 where the pontrjagin class is weighted by a parameter @xmath225 . using the rules above
, we obtain @xmath226 the einstein - hilbert action with cosmological constant and the holst action with immirizo parameter as well the euler and pontrjagin class for the reduced bundles . in this model ,
the curvature is changed locally by adding a @xmath206bundle .
then the scale @xmath227 has to agree with the volume of the @xmath212 . in the action we have the coupling constant @xmath228 which has to agree with @xmath229 ( @xmath230 planck length ) to get in contact with einsteins theory , i.e. we must set @xmath231 the agreement with the einstein - hilbert action showed that this approach can describe gravity but it does not describe the global geometry .
later we can show , however , that it must be the de sitter space @xmath232 globally .
in this section we will support our main hypothesis that an exotic @xmath0 has automatically a quantum geometry , but as noted in the introduction , we must implicitly assume that the quantum - geometrical state is realized in the exotic @xmath0 .
interestingly , it follows from the physically motivated existence of a lorentz metric which is induced by a codimension - one foliation .
therefore we will construct the foliation and the corresponding leaf space as the space of observables ( using ideas of connes ) .
this leaf space is a non - commutative @xmath233algebra with observable algebra a factor @xmath3 von neumann algebra .
a state in this algebra can be interpreted as a wild embedding which is also motivated by the exotic smoothness structure .
the classical state is the tame embedding .
then , the deformation quantization of this tame embedding is the wild embedding ( see @xcite ) . in principle
, the wild embedding determines the @xmath233algebra completely .
this algebra is generated by holonomies along connections of constant curvature .
it is known from mathematics that this algebra ( forming a so - called character variety @xcite ) determines the geometrical structure of the 3-manifold ( along the way of thurston @xcite ) .
the main structure in this approach is the fundamental group , i.e. the group of closed , non - contractible curves in a manifold .
the quantization of this group ( as an expression of the classical geometry ) gives the so - called skein algebra of knots in this manifold .
we will relate this skein algebra to the leaf space above . on the way to show this relation ,
we will obtain the generator of the translation from one 3-manifold into another 3-manifold , i.e. the time together with the hamiltonian . in section [ sec : pathint ] , we described the sequence of 3-manifolds @xmath145 characterizing the exotic smoothness structure of @xmath1 .
then @xmath112framed surgery along this pretzel knot produces @xmath234 whereas the @xmath21-th untwisted whitehead double will give @xmath198 . for large @xmath21 ,
the structure of the casson handle is contained in the topology of @xmath198 and in the limit @xmath139 we obtain @xmath138 ( which is now a wild embedding @xmath235 in the standard @xmath100 given by the embedding of the small exotic @xmath1 , see above ) .
what do we know about the structure of @xmath198 or @xmath138 in general ?
the compact subset @xmath134 is a 4-manifold constructed by a pair of one 1-handle and one 2-handle which topologically cancel .
the boundary of @xmath134 is a compact 3-manifold having the first betti number @xmath236 this information is also contained in @xmath138 . by the work of freedman @xcite ,
every casson handle is topologically @xmath126 ( relative to the attaching region ) and therefore @xmath138 must be the boundary of @xmath50 ( the casson handle trivializes @xmath134 to be @xmath50 ) , i.e. _ @xmath138 is a wild embedded 3-sphere @xmath60_. then we obtain two different descriptions of @xmath1 : 1 . as a sequence of 3-manifolds @xmath198 ( all having the first betti number @xmath237 ) as boundaries of the neighborhood of @xmath134 with increasing size and 2 . as a global hyperbolic space of @xmath238 written as @xmath239 where @xmath240 is a wild embedded 3-sphere ( which looks differently for different @xmath6 )
the first description gives a non - trivial but smooth foliation but there is no global spatial space .
in contrast to this highly non - trivial foliation , the second description gives a global foliated spacetime containing a global spatial component , the wild embedded 3-sphere .
in the first description we have a complex , relational description with no global time - like slices . here
, there only is a local coordinate system ( with its own eigenzeit ) .
this relational view has the big advantage that the simplest parts are also simple submanifolds ( only finite surfaces with boundary ) .
in contrast , the second description introduces a global foliation into equal time slices .
then the complexity is contained into the spatial component which is now a wild embedding ( i.e. a space with an infinite number of polygons ) .
this second approach will be described in the next subsection .
so , lets concentrate on the first approach . every 3-manifold @xmath198 admits a codimension - one ( @xmath241invariant ) foliation ( see @xcite for the details ) . by the description of the exotic @xmath1 using the sequence of 3-manifolds @xmath242 we also get a foliation of the exotic @xmath1 .
the foliation on @xmath198 is defined by a @xmath241invariant one - form @xmath218 which is integrable @xmath243 and defines another one - form @xmath244 by @xmath245 .
then the integral @xmath246 is known as godbillon - vey number @xmath247 with the class @xmath248 . from the physics point of view
, it is the abelian chern - simons functional .
the godbillon - vey class characterizes the codimension - one foliation for the 3-manifold @xmath198 ( see the appendix b for more details ) .
the foliation is very complicated . in @xcite
the local structure was analyzed .
let @xmath249 be the curvature and torsion of a normal curve , respectively .
furthermore , let @xmath250 be the frame formed by this vector field dual to the one - forms @xmath251 and let @xmath252 be the second fundamental form of leaf .
then the godbillon - vey class is locally given by @xmath253 where @xmath254 for @xmath255 invariant foliations i.e. @xmath256=z$ ] , @xmath257=t$ ] and @xmath258=n$ ] .
recall that a foliation @xmath259 of a manifold @xmath127 is an integrable subbundle @xmath260 of the tangent bundle @xmath261 .
the leaves @xmath262 of the foliation @xmath259 are the maximal connected submanifolds @xmath263 with @xmath264 .
we denote with @xmath265 the set of leaves or the leaf space .
now one can associate to the leaf space @xmath265 a @xmath266-algebra @xmath267 by using the smooth holonomy groupoid @xmath268 of the foliation ( see connes @xcite ) . according to connes @xcite
, one assigns to each leaf @xmath269 the canonical hilbert space of square - integrable half - densities @xmath270 .
this assignment , i.e. a measurable map , is called a random operator forming a von neumann @xmath271 .
a deep theorem of hurder and katok @xcite for foliations with non - zero godbillon - vey invariant states that this foliation has to contain a factor @xmath8 von neumann algebra .
as shown in @xcite , the von neumann algebra for the foliation of @xmath198 and for the exotic @xmath1 is a factor @xmath272algebra . for the construction of this algebra
, one needs the concept of a holonomy groupoid .
foliations are determined by the holonomies of closed curves in a leaf and the transport of this closed curve together with the holonomy from the given leaf to another leaf .
now one may ask why one considers only closed curves .
let @xmath273 the space of all paths in a manifold then this space admits a fibration over the space of closed paths @xmath274 ( also called loop space ) with fiber the constant paths ( therefore homeomorphic to @xmath127 ) , see @xcite .
then , a curve is determined up to deformation ( i.e. homotopy ) by a closed path .
consider now a closed curve @xmath225 in a leaf @xmath222 and let act a diffeomorphism on @xmath222 .
then the curve @xmath225 is modified as well to @xmath275 but @xmath225 and @xmath275 are related by a ( smooth ) homotopy .
therefore to guarantee diffeomorphism invariance in this approach , one has to consider all closed curves up to homotopy .
this structure can be made into a group ( using concatenation of paths as group operation ) called the fundamental group @xmath276 of the leaf .
above we spoke about holonomy but a holonomy needs a connection of some bundle which we did not introduce until now .
but connes @xcite described a way to circumvent this difficulty : given a leaf @xmath222 of @xmath259 and two points @xmath277 of this leaf , any simple path @xmath225 from @xmath278 to @xmath279 on the leaf @xmath222 uniquely determines a germ @xmath280 of a diffeomorphism from a transverse neighborhood of @xmath278 to a transverse neighborhood of @xmath279 . the germ of diffeomorphism @xmath280 only depends upon the homotopy class of @xmath225 in the fundamental group of the leaf @xmath222 , and is called the holonomy of the path @xmath225 .
all fundamental groups of all leafs form the fundamental groupoid .
the holonomy groupoid of a leaf @xmath222 is the quotient of its fundamental groupoid by the equivalence relation which identifies two paths @xmath225 and @xmath275 from @xmath278 to @xmath279 ( both in @xmath222 ) iff @xmath281 .
then the von neumann algebra of the foliation is the convolution algebra of the holonomy groupoid which will be constructed later for the wild embedding .
remember a von neumann algebra is an involutive subalgebra @xmath127 of the algebra of operators on a hilbert space @xmath187 that has the property of being the commutant of its commutant : @xmath282 .
this property is equivalent to saying that @xmath127 is an involutive algebra of operators that is closed under weak limits .
a von neumann algebra @xmath127 is said to be hyperfinite if it is generated by an increasing sequence of finite - dimensional subalgebras .
furthermore we call @xmath127 a factor if its center is equal to @xmath283 .
it is a deep result of murray and von neumann that every factor @xmath127 can be decomposed into 3 types of factors @xmath284 .
the factor @xmath285 case divides into the two classes @xmath286 and @xmath287 with the hyperfinite factors @xmath288 the complex square matrices and @xmath289 the algebra of all operators on an infinite - dimensional hilbert space @xmath187 .
the hyperfinite @xmath290 factors are given by @xmath291 , the clifford algebra of an infinite - dimensional euclidean space @xmath292 , and @xmath293 .
the case @xmath8 remained mysterious for a long time .
now we know that there are three cases parametrized by a real number @xmath294 $ ] : @xmath295 the krieger factor induced by an ergodic flow @xmath59 , @xmath296 the powers factor for @xmath297 and @xmath298 the araki - woods factor for all @xmath299 with @xmath300 .
we remark that all factor @xmath8 cases are induced by infinite tensor products of the other factors .
one example of such an infinite tensor space is the fock space in quantum field theory .
the modular theory of von neumann algebras ( see also @xcite ) has been discovered by m. tomita @xcite in 1967 and put on solid grounds by m. takesaki @xcite around 1970 .
it is a very deep theory that , to every von neumann algebra @xmath301 acting on a hilbert space @xmath179 , and to every vector @xmath302 that is cyclic , i.e. @xmath303 , and separating , i.e. for @xmath304 , @xmath305 , associates : * a one - parameter unitary group @xmath306 * and a conjugate - linear isometry @xmath307 such that : @xmath308 where the commutant @xmath309 of @xmath310 is defined by @xmath311_{-}\!=0,\forall\ , a\!\in\!\mathcal{b}(\mathcal{h})\}$ ] . more generally , given a von neumann algebra @xmath310 and
a faithful normal state is faithful if @xmath312 ; it is normal if for every increasing bounded net of positive elements @xmath313 , we have @xmath314 . ]
( more generally for a faithful normal semi - finite weight ) @xmath218 on the algebra @xmath310 , the modular theory allows to create a one - parameter group of @xmath315-automorphisms of the algebra @xmath310 , @xmath316 such that : * in the gelfand namark segal representation @xmath317 induced by the weight @xmath218 , on the hilbert space @xmath318 , the modular automorphism group @xmath319 is implemented by a unitary one - parameter group @xmath320 i.e. we have @xmath321 , for all @xmath322 and for all @xmath6 ; * there is a conjugate - linear isometry @xmath323 , whose adjoint action implements a modular anti - isomorphism @xmath324 , between @xmath325 and its commutant @xmath326 , i.e. for all @xmath322 , we have @xmath327 . the operators @xmath328 and @xmath329 are called respectively the modular conjugation operator and the modular operator induced by the state ( weight ) @xmath218 .
we will call `` modular generator '' the self - adjoint generator of the unitary one - parameter group @xmath330 as defined by stone s theorem i.e. the operator @xmath331 the modular automorphism group @xmath319 associated to @xmath218 is the unique one - parameter automorphism group that satisfies the kubo martin
schwinger ( kms - condition ) with respect to the state ( or more generally a normal semi - finite faithful weight ) @xmath218 , at inverse temperature @xmath332 , i.e. @xmath333 and for all @xmath334 . using tomita - takesaki - theory
, one has a continuous decomposition ( as crossed product ) of any factor @xmath8 algebra @xmath127 into a factor @xmath335 algebra @xmath336 together with a one - parameter group is the group of positive real numbers with multiplication as group operation also known as pontrjagin dual . ] @xmath337 of automorphisms @xmath338 of @xmath336 , i.e. one obtains @xmath339 that means , there is a foliation induced from the foliation producing this @xmath335 factor .
connes @xcite ( in section i.4 page 57ff ) constructed the foliation @xmath340 canonically associated to the foliation @xmath341 of factor @xmath3 above having the factor @xmath335 as von neumann algebra . in our case it is the horocycle flow : let @xmath342 the polygon on the hyperbolic space @xmath343 determining the foliation above .
@xmath342 is equipped with the hyperbolic metric @xmath344 together with the collection @xmath345 of unit tangent vectors to @xmath342 .
a horocycle in @xmath342 is a circle contained in @xmath342 which touches @xmath346 at one point , but from the classification of factors , we know that @xmath335 is also splitted into @xmath347 so that every factor @xmath8 is determined by the factor @xmath348 .
the factor @xmath287 are the compact operators in the hilbert space . with an important observation we will close this intermezzo .
the factor @xmath335 admits an action of the group @xmath349 by automorphisms so that the crossed product ( [ eq : crossed - product - factor - iii ] ) is the factor @xmath3 .
the corresponding invariant , the flow of weights @xmath350 , was determined by connes @xcite to be the godbillon - vey invariant .
therefore _ the modular generator above is given by the godbillon - vey invariant , i.e. this invariant is the hamiltonian of the theory . _
then the @xmath266-algebra @xmath351 of the foliation @xmath259 is the @xmath266-algebra @xmath352 of the smooth holonomy groupoid @xmath268 . for completeness
we will present the explicit construction ( see @xcite sec .
the basic elements of @xmath351 ) are smooth half - densities with compact supports on @xmath268 , @xmath353 , where @xmath354 for @xmath355 is the one - dimensional complex vector space @xmath356 , where @xmath357 , and @xmath358 is the one - dimensional complex vector space of maps from the exterior power @xmath359 , @xmath360 , to @xmath283 such that @xmath361 for @xmath362 , the convolution product @xmath363 is given by the equality @xmath364
then we define via @xmath365 a @xmath315-operation making @xmath366 into a @xmath315-algebra . for each leaf @xmath262 of @xmath259 one
has a natural representation of @xmath366 on the @xmath367 space of the holonomy covering @xmath368 of @xmath262 . fixing a base point @xmath369
, one identifies @xmath368 with @xmath370and defines the representation @xmath371 the completion of @xmath366 with respect to the norm @xmath372 makes it into a @xmath266-algebra @xmath351 . among all elements of the @xmath266-algebra
, there are distinguished elements , idempotent operators or projectors having a geometric interpretation in the foliation . for later use
, we will construct them explicitly ( we follow @xcite sec .
ii.8.@xmath373 closely ) .
let @xmath374 be a compact submanifold which is everywhere transverse to the foliation ( thus @xmath375 ) .
a small tubular neighborhood @xmath376 of @xmath336 in @xmath127 defines an induced foliation @xmath340 of @xmath376 over @xmath336 with fibers @xmath377 .
the corresponding @xmath266-algebra @xmath378 is isomorphic to @xmath379 with @xmath380 the @xmath266-algebra of compact operators . in particular
it contains an idempotent @xmath381 , @xmath382 , where @xmath12 is a minimal projection in @xmath380 .
the inclusion @xmath383 induces an idempotent in @xmath351 which is given by a closed curve in @xmath127 transversal to the foliation . in case of the foliation above ( of the 3-manifolds @xmath198 )
, one has the foliation of the polygon @xmath342 in @xmath343 and a circle @xmath119 attached to every leaf of this foliation .
therefore we have the leafs @xmath384 $ ] and the @xmath119is the closed curve transversal to the foliation
. then every leaf defines ( using the isomorphism @xmath385)=\pi_{1}(s^{1})=\mathbb{z}$ ] ) an idempotent represented by the fiber @xmath119 forming a base for the gns representation of the @xmath233algebra .
now we are able to construct a state in this algebra .
a state is a linear functional @xmath386 so that @xmath387 and @xmath388 .
elements of @xmath351 are half - densities with a support along some closed curve ( as part of the holonomy groupoid ) . in a first step
, one can use the gns - representation of the @xmath389algebra @xmath351 by a map @xmath390 in to the bounded operators of a hilbert space . by the theorem of frchet - riesz ,
every linear functional can be represented by the scalar product of the hilbert space for some vector . to determine the linear functionals
, we have to investigate the geometry of the foliation .
the foliation was constructed to be @xmath241invariant , i.e. fixing the upper half space @xmath343 .
then we considered the unit tangent vectors of the tangent bundle over @xmath343 defining the @xmath391geometry .
but more is true .
every part of the 3-manifold @xmath198 is a knot / link complement with hyperbolic structure with isometry group @xmath392 where the other geometric structures like @xmath393 and @xmath255 embed .
here we remark the known fact that every @xmath392-geometry lifts uniquely to @xmath394 ( the double cover ) .
therefore , to model the holonomy , we have to choose a flat @xmath395connection and write it as the well - known integral of the connection 1-form along a closed curve .
the linear functional is the trace of this integral ( seen as matrix using a representation of @xmath394 ) known as wilson loop .
one can use the well - known identity @xmath396 for @xmath394 which goes over to the wilson loops .
let @xmath397 $ ] be the wilson loop of a connection @xmath149 along the closed curve @xmath225 .
then the relation of the wilson loops @xmath398\cdot w_{\eta}[a]=w_{\gamma\circ\eta}[a]+w_{\gamma\circ\eta^{-1}}[a]\ ] ] for two intersecting curves @xmath225 and @xmath244 is known as the mandelstam identity for intersecting loops , see fig .
[ fig : mandelstam - identity - as - skein ] for a visualization . and subsection [ sub : drinfeld - turaev - quantization]).[fig : mandelstam - identity - as - skein ] ] this relation is also known from another area : knot theory . there , it is the kauffman bracket skein relation used to define the kauffman knot polynomial
. therefore we obtain a state in the @xmath389algebra by a closed curve in the leaf which extends to a knot ( an embedded , closed curve ) in a submanifold of the 3-manifold defined up to the skein relation .
finally : @xmath399}\ ] ] we will later explain this correspondence as a deformation quantization . we will close this subsection by some remarks . every representation @xmath400 defines ( up to conjugacy ) a flat connection . at the same time it defines also a hyperbolic structure on @xmath198 ( for @xmath401 ) . by the argumentation above
, the quantized version of this geometry ( as defined by the @xmath389algebra of the foliation ) is given by the skein space ( see subsection [ sub : drinfeld - turaev - quantization ] for the definition of the skein space ) .
our previous work implied that the transition from the standard @xmath100 to a small exotic @xmath1 has much to do with quantum gravity ( qg ) .
therefore one would expect that a submanifold in the standard @xmath100 with an appropriated geometry represents a classical state . before we construct this state
, there is a lot to say about the wild embedded 3-sphere as a quantum state . to describe this wild 3-sphere
, we will construct the sequence of @xmath198 by using the example of @xcite which was already partly explained in subsection [ sub : small - exotic - r4-sequnce-3mf ] .
at first we remark that the interior of the handle body in figure [ fig : handle - picture - of - exotic - r4 ] is the @xmath1 .
the casson handle for this @xmath1 is given by the simplest tree @xmath402 , one positive self - intersection for each level .
the compact 4-manifold inside of @xmath1 can be seen in figure [ fig : link - picture - for - k ] as a handle body .
the 3-manifold @xmath198 surrounding this compact submanifold @xmath134 is given by surgery ( @xmath112framed ) along the link in figure [ fig : handle - picture - of - exotic - r4 ] with a casson handle of @xmath142levels .
in @xcite , this case is explicitly discussed .
@xmath198 is given by @xmath112framed surgery along the @xmath21-th untwisted whitehead double of the pretzel @xmath403 knot ( see figure [ fig : pretzel - knot ] ) .
obviously , there is a sequence of inclusions @xmath404 with the 3-manifold @xmath405 as limit .
let @xmath406 be the corresponding ( wild ) knot , i.e. the @xmath407-th untwisted whitehead double of the pretzel knot @xmath144 ( or the knot @xmath408 in rolfson notation ) .
the surgery description of @xmath405induces the decomposition @xmath409 where @xmath410 is the knot complement of @xmath406 . in @xcite , the splitting of knot complements was described .
let @xmath411 be the pretzel knot @xmath144 and let @xmath412 be the whitehead link ( with two components )
. then the complement @xmath413 has one torus boundary whereas the complement @xmath414 has two torus boundaries .
now according to @xcite , one obtains the splitting @xmath415 and we will describe each part separately ( see figure [ fig : splitting - of - knot - complement ] ) . at first
the knot @xmath411 is a hyperbolic knot , i.e. the interior of the 3-manifold @xmath413 admits a hyperbolic metric . by the work of gabai @xcite ,
@xmath413admits a codimension - one foliation .
the whitehead link is a hyperbolic link but we need more : the whitehead link is a fibered link of genus @xmath416 .
that is , there is a fibration of the link complement @xmath417 over the circle so that @xmath418 is a surface of genus @xmath416 ( seifert surface ) for all @xmath419 .
now we will also describe the changes for a general tree .
at first we will modify the whitehead link : we duplicate the linked circle , i.e. there are as many circles as branching in the tree to get the link @xmath420 with @xmath421 components .
then the complement of @xmath420 has also @xmath421 torus boundaries and it also fibers over @xmath119 . with the help of @xmath420 we can build every tree @xmath422 . now
the 3-manifold @xmath423 is given by @xmath112framed surgery along the @xmath407-th untwisted ramified ( usage of @xmath420 ) whitehead double of a knot @xmath110 , denoted by the link @xmath424 .
the tree @xmath422 has one root , then @xmath423 is given by @xmath425 and the complement @xmath426 splits like the tree into complements of @xmath420 and one copy of @xmath427 ( see figure [ fig : splitting - of - knot - complement ] ) . using a deep result of freedman @xcite
, we obtain : + _ @xmath423 is a wild embedded 3-sphere @xmath240 . _ our result about the existence of a codimension - one foliation for @xmath423 can be simply expressed : foliations are characterized by the holonomy properties of the leafs .
this principle is also the corner stone for the usage of non - commutative geometry as description of the leaf space . in the previous subsection
, we already characterized the state as an element of the kauffman skein module .
here we are interested in a reconstruction of the underlying space but now assuming a global foliation so that we will obtain the whole spatial space .
starting point is the state constructed in the subsection [ sub : construction - of - state ] . here , we got a relation between the state @xmath218 as linear functional over the algebra and the kauffman skein module . using this relation
, we consider a leaf @xmath428 $ ] and the 3-dimensional extension as solid torus @xmath117 .
the kauffman skein module @xmath429 is polynomial algebra with one generator ( the loop around @xmath119 ) .
now we consider one 3-manifold @xmath198 with the corresponding foliation .
using the splitting above , the kauffman skein module @xmath430 is determined by the skein module for the parts , i.e. by the knot complements .
therefore we have to consider the skein module for hyperbolic 3-manifolds .
hyperbolic 3-manifolds contain special surfaces , called essential or incompressible surfaces , see appendix c. it is known @xcite that the skein module of 3-manifolds containing essential surfaces is not finitely generated .
therefore , the state itself is not finitely generated .
if we use the leaf @xmath117 as a local model for one generator then we will obtain an infinitely complicated 3-manifold made from pieces @xmath117 so that the corresponding generators are not related to each other .
an example of this structure is the whitehead manifold having a non - finitely generated kauffman skein module @xcite . in general we will obtain a wild embedded 3-manifold by using this simple pieces . by the argumentation in the previous subsection
we know that this wild embedded 3-manifold is the wild embedded 3-sphere @xmath423 .
finally we obtain : @xmath431 the state @xmath218 is realized by some wild embedded 3-sphere .
following @xcite we will construct a @xmath233algebra from the wild embedded 3-sphere .
let @xmath432 be a wild embedding of codimension - one so that @xmath433 .
now we consider the complement @xmath434 which is non - trivial , i.e. @xmath435 .
now we define the @xmath233algebra @xmath436 ) associated to the complement @xmath437 with group @xmath438 .
if @xmath439 is non - trivial then this group is not finitely generated . from an abstract point of view
, we have a decomposition of @xmath440 by an infinite union @xmath441 of level sets @xmath442
. then every element @xmath443 lies ( up to homotopy ) in a finite union of levels .
the basic elements of the @xmath233algebra @xmath436 ) are smooth half - densities with compact supports on @xmath440 , @xmath444 , where @xmath354 for @xmath443 is the one - dimensional complex vector space of maps from the exterior power @xmath445 ( @xmath446 ) , of the union of levels @xmath262 representing @xmath225 , to @xmath283 such that @xmath447 for @xmath448 , the convolution product @xmath363 is given by the equality @xmath364 with the group operation @xmath449 in @xmath439 .
then we define via @xmath365 a @xmath315operation making @xmath450 into a @xmath315algebra . each level set @xmath442 consists of simple pieces ( in case of alexanders horned sphere ,
we will explain it below ) denoted by @xmath451 .
for these pieces , one has a natural representation of @xmath450 on the @xmath367 space over @xmath451 .
then one defines the representation @xmath452 the completion of @xmath450 with respect to the norm @xmath453 makes it into a @xmath266-algebra @xmath454 ) .
finally we are able to define the @xmath233algebra associated to the wild embedding . using a result in @xcite
, one can show that the corresponding von neumann algebra is the factor @xmath3 . among all elements of the @xmath266-algebra
, there are distinguished elements , idempotent operators or projectors having a geometric interpretation . for later use
, we will construct them explicitly ( we follow @xcite sec . @xmath455 closely ) .
let @xmath456 be the wild submanifold .
a small tubular neighborhood @xmath376 of @xmath423 in @xmath0 defines the corresponding @xmath266-algebra @xmath457 is isomorphic to @xmath458 with @xmath380 the @xmath266-algebra of compact operators .
in particular it contains an idempotent @xmath381 , @xmath459 , where @xmath12 is a minimal projection in @xmath380 .
it induces an idempotent in @xmath460 . by definition ,
this idempotent is given by a closed curve in the complement @xmath434 .
these projection operators form the basis in this algebra in this section we will describe a way from a ( classical ) poisson algebra to a quantum algebra by using deformation quantization .
therefore we will obtain a positive answer to the question : does the @xmath233algebra of the foliation ( as well of a wild ( specific ) embedding ) comes from a ( deformation ) quantization ?
of course , this question can not be answered in all generality , but for our example we will show that the enveloping von neumann algebra of foliation and of this wild embedding is the result of a deformation quantization using the classical poisson algebra ( of closed curves ) of the tame embedding .
this result shows two things : the wild embedding can be seen as a quantum state and the classical state is a tame embedding . in this section we will describe the formal structure of a classical theory coming from the algebra of observables using the concept of a poisson algebra . in quantum theory , an observable
is represented by an hermitean operator having the spectral decomposition via projectors or idempotent operators .
the coefficient of the projector is the eigenvalue of the observable or one possible result of a measurement .
at least one of these projectors represents ( via the gns representation ) a quasi - classical state .
thus , to construct the substitute of a classical observable algebra with poisson algebra structure , we have to concentrate on the idempotents in the @xmath266-algebra .
now we will see that the set of closed curves on a surface has the structure of a poisson algebra .
let us start with the definition of a poisson algebra .
let @xmath342 be a commutative algebra with unit over @xmath461 or @xmath283 .
poisson bracket _ on @xmath342 is a bilinearform @xmath462 fulfilling the following 3 conditions : 1 . anti - symmetry @xmath463 2 .
jacobi identity @xmath464 3 .
derivation @xmath465 .
then a _ poisson algebra _ is the algebra @xmath466 .
now we consider a surface @xmath467 together with a closed curve @xmath225 .
additionally we have a lie group @xmath268 given by the isometry group .
the closed curve is one element of the fundamental group @xmath468 .
from the theory of surfaces we know that @xmath468 is a free abelian group .
denote by @xmath469 the free @xmath470-module ( @xmath470 a ring with unit ) with the basis @xmath468 , i.e. @xmath469 is a freely generated @xmath470-module . recall goldman s definition of the lie bracket in @xmath469 ( see @xcite ) . for a loop @xmath471
we denote its class in @xmath468 by @xmath472 .
let @xmath473 be two loops on @xmath467 lying in general position .
denote the ( finite ) set @xmath474 by @xmath475 .
for @xmath476 denote by @xmath477 the intersection index of @xmath478 and @xmath479 in @xmath480 .
denote by @xmath481 the product of the loops @xmath473 based in @xmath480 .
up to homotopy the loop @xmath482 is obtained from @xmath483 by the orientation preserving smoothing of the crossing in the point @xmath480 . set @xmath484=\sum_{q\in\alpha\#\beta}\epsilon(q;\alpha,\beta)(\alpha_{q}\beta_{q})\quad.\label{eq : lie - bracket - loops}\ ] ] according to goldman @xcite ( theorem 5.2 ) , the bilinear pairing @xmath485:z\times z\to z$ ] given by ( [ eq : lie - bracket - loops ] ) on the generators is well defined and makes @xmath469 a lie algebra . the algebra @xmath486 of symmetric tensors is then a poisson algebra ( see turaev @xcite ) .
the whole approach seems natural for the construction of the lie algebra @xmath469 but the introduction of the poisson structure is an artificial act . from the physical point of view , the poisson structure is not the essential part of classical mechanics .
more important is the algebra of observables , i.e. functions over the configuration space forming the poisson algebra . for the foliation discussed above
, we already identified the observable algebra ( the holonomy along closed curves ) as well the corresponding group to be @xmath394 .
therefore for the following , we will set @xmath487 .
now we introduce a principal @xmath268 bundle on @xmath467 , representing a geometry on the surface .
this bundle is induced from a @xmath268 bundle over @xmath488 $ ] having always a flat connection .
alternatively one can consider a homomorphism @xmath489 represented as holonomy functional @xmath490 with the path ordering operator @xmath491 and @xmath218 as flat connection ( i.e. inducing a flat curvature @xmath492 ) .
this functional is unique up to conjugation induced by a gauge transformation of the connection .
thus we have to consider the conjugation classes of maps @xmath493 forming the space @xmath494 of gauge - invariant flat connections of principal @xmath268 bundles over @xmath467 . now
( see @xcite ) we can start with the construction of the poisson structure on @xmath494 , based on the cartan form as the unique bilinearform of a lie algebra . as discussed above we will use the lie group @xmath487 but the whole procedure works for every other group too .
now we consider the standard basis @xmath495 of the lie algebra @xmath496 with @xmath497=h,\,[h , x]=2x,\,[h , y]=-2y$ ] .
furthermore there is the bilinearform @xmath498 written in the standard basis as @xmath499 now we consider the holomorphic function @xmath500 and define the gradient @xmath501 along @xmath12 at the point @xmath149 as @xmath502 with @xmath503 and @xmath504 the calculation of the gradient @xmath505 for the trace @xmath506 along a matrix @xmath507 is given by @xmath508 given a representation @xmath509 of the fundamental group and an invariant function @xmath510 extendable to @xmath511 .
then we consider two conjugacy classes @xmath512 represented by two transversal intersecting loops @xmath513 and define the function @xmath514 by @xmath515 .
let @xmath516 be the intersection point of the loops @xmath513 and @xmath517 a path between the point @xmath278 and the fixed base point in @xmath468 .
then we define @xmath518 and @xmath519 .
finally we get the poisson bracket @xmath520 where @xmath521 is the sign of the intersection point @xmath278 . thus , _ the space @xmath511 has a natural poisson structure ( induced by the bilinear form ( [ eq : lie - bracket - loops ] ) on the group ) and the poisson algebra _ @xmath522 _ of complex functions over them is the algebra of observables .
_ now we introduce the ring @xmath523 $ ] of formal polynomials in @xmath188 with values in @xmath283 .
this ring has a topological structure , i.e. for a given power series @xmath524 $ ] the set @xmath525 $ ] forms a neighborhood .
now we define a _ quantization _ of a poisson algebra @xmath342 is a @xmath523 $ ] algebra @xmath526 together with the @xmath283-algebra isomorphism @xmath527 so that \1 .
the module @xmath526 is isomorphic to @xmath528 $ ] for a @xmath283 vector space @xmath79 \2 .
let @xmath529 and @xmath530 be @xmath531 , @xmath532 then @xmath533 one speaks of a deformation of the poisson algebra by using a deformation parameter @xmath188 to get a relation between the poisson bracket and the commutator .
therefore we have the problem to find the deformation of the poisson algebra @xmath534 . the solution to this problem
can be found via two steps : 1 .
at first find another description of the poisson algebra by a structure with one parameter at a special value and 2 .
secondly vary this parameter to get the deformation .
fortunately both problems were already solved ( see @xcite ) .
the solution of the first problem is expressed in the theorem : _ the skein module @xmath535)$ ] ( i.e. @xmath536 ) has the structure of an algebra isomorphic to the poisson algebra @xmath534 . _ _ ( see also @xcite ) _
then we have also the solution of the second problem : _ the skein algebra @xmath537)$ ] is the quantization of the poisson algebra @xmath534 with the deformation parameter @xmath538.(see also @xcite ) _ to understand these solutions we have to introduce the skein module @xmath539 of a 3-manifold @xmath127 ( see @xcite ) . for that purpose
we consider the set of links @xmath540 in @xmath127 up to isotopy and construct the vector space @xmath541 with basis @xmath540
. then one can define @xmath542 $ ] as ring of formal polynomials having coefficients in @xmath541 .
now we consider the link diagram of a link , i.e. the projection of the link to the @xmath543 having the crossings in mind . choosing a disk in @xmath543 so that one crossing is inside this disk .
if the three links differ by the three crossings @xmath544 ( see figure [ fig : skein - crossings ] ) inside of the disk then these links are skein - related .
.[fig : skein - crossings ] ] then in @xmath542 $ ] one writes the skein relation . ] @xmath545 .
furthermore let @xmath546 be the disjoint union of the link with a circle and one writes the framing relation @xmath547 .
let @xmath548 be the smallest submodule of @xmath542 $ ] containing both relations .
then we define the kauffman bracket skein module by @xmath549/s(m)$ ] .
we list the following general results about this module : * the module @xmath550 for @xmath536 is a commutative algebra .
* let @xmath467 be a surface , then @xmath537)$ ] carries the structure of an algebra . the algebraic structure of @xmath537)$ ]
can be simply seen by using the diffeomorphism between the sum @xmath488\cup_{s}s\times[0,1]$ ] along @xmath467 and @xmath488 $ ] .
then the product @xmath551 of two elements @xmath552)$ ] is a link in @xmath488\cup_{s}s\times[0,1]$ ] corresponding to a link in @xmath488 $ ] via the diffeomorphism .
the algebra @xmath537)$ ] is in general non - commutative for @xmath553 . for the following
we will omit the interval @xmath81 $ ] and denote the skein algebra by @xmath554 . in subsection [ sub : construction - of - state ] , we described the state as an element of the kauffman skein module @xmath555 of the leaf @xmath222 .
now we obtained also that the observable algebra is the kauffman skein module again .
how does this whole story fit into the description of the observable algebra for the foliation as factor @xmath3 ? in @xcite
, it was shown that the kauffman bracket skein module of a cylinder over the torus embeds as a subalgebra of the noncommutative torus .
however , the noncommutative torus can be seen as the leaf space of the kronecker foliation of the torus leading to the factor @xmath335 . then by using ( [ eq : crossed - product - factor - iii ] )
, we obtain the factor @xmath3 back .
we will use this relation in the next section to get the quantum action .
above , we used the foliation to get quantum states which agreed with the deformation quantization of a classical state .
central point in our argumentation is the construction of the @xmath233algebra with the corresponding von neumann algebra as observable algebra .
this von neumann algebra is a factor @xmath3 . by using the tomita - takesaki modular theory ,
there is a relation to the factor @xmath335 by using an action of the group @xmath349 by automorphisms of a lebesgue measure space leading to the decomposition of the factor @xmath3 .
this action is related to an invariant , the flow of weights mod(m ) .
the main property of the factor @xmath3 is the constant flow of weights mod(m ) .
connes @xcite described the flow of weights as a bundle of densities over the leaf space , i.e. the @xmath349 homogeneous space of nonzero maps . in case of foliation considered above ,
this density is constant and we can naturally identify this density with the volume of the submanifold defining the foliation . by definition , this volume is given by the godbillon - vey invariant ( see eqn .
( [ eq : gv - number - thurston - foliation-1 ] ) in appendix b , the circle in the fiber has unit size ) .
this invariant can be seen as an element of @xmath556 with the holonomy groupoid @xmath268 of the foliation . as shown by connes @xcite
, the godbillon - vey class @xmath557 can be expressed as a cyclic cohomology class ( the so - called flow of weights ) @xmath558 of the @xmath233algebra for the foliation .
then we define an expression @xmath559 uniquely associated to the foliation ( @xmath560 is the dixmier trace ) .
the expression @xmath467 generates the action on the factor by @xmath561 so that @xmath467 is the action or the hamiltonian multiplied by the time ( see ( [ eq : modular - generator ] ) ) .
it is an operator which defines the dynamics by acting on the states . for explicit calculations we have to evaluate this operator .
one way is the usage of the relation between the foliation and the wild embedding .
this wild embedding is determined by the fundamental group @xmath439 of its complement . in @xcite , we discussed the properties of this group @xmath439 .
it is a perfect group , i.e. every element is generated by a commutator .
then a representation of this group into some other group like @xmath562 ( the limit of @xmath563 for @xmath139 ) reduces to the representation of the maximal perfect subgroup . for that purpose
we consider the representation of the group @xmath439 into the group @xmath564 of elementary matrices , which is the perfect subgroup of @xmath562 .
then we obtain matrix - valued functions @xmath565 as the image of the generators of @xmath439 w.r.t .
the representation @xmath566 labeled by the dimension @xmath567 of the embedding space @xmath568 . via the representation @xmath569
, we obtain a cyclic cocycle in @xmath570 generated by a suitable fredholm operator @xmath341 . here
we use the standard choice @xmath571 with the dirac operator @xmath201 acting on the functions in @xmath572 .
then the cocycle in @xmath570 can be expressed by @xmath573[f , x^{\nu}]\ ] ] using a metric @xmath574 in @xmath1 via the pull - back using the representation @xmath569 .
finally we obtain the action @xmath575[f , x_{\mu}])=tr_{\omega}([d , x^{\mu}][d , x_{\mu}]|d|^{-2})\label{eq : quantum - action}\ ] ] which can be evaluated by using the heat - kernel of the dirac operator @xmath201 .
the appearance of the heat kernel is a sign for a relation to quantum field theory where the heat kernel is a very convenient tool for studying one - loop divergences , anomalies and various asymptotics of the effective action .
away from this operator expression for the godbillon - vey invariant , there are geometrical evaluations which are not defined on the leaf space but rather on the whole manifold .
as mentioned above , this invariant admits values in the real numbers and we can evaluate them according to the type of the number : for integer values one obtains the euler class and for rational numbers the pontrjagin class ( for the corresponding bundles ) . therefore using the ideas of section [ sec : action - induced - by - top ] , we obtain the einstein - hilbert and the holst action but also a correction given by irrational values of the godbillon - vey number .
a good test for the theory is the dependence of the action ( [ eq : quantum - action ] ) on the scale .
the theory has a strong geometrical flavor and therefore the scaling behavior can be understood by a geometrical construction using the exotic @xmath1 . as explained above
, the central point in the construction is the casson handle . from the scaling point of view
, the casson handle contains disks of any size ( with respect to the embedding @xmath576 ) .
the long scales are given by the first levels of the casson handle whereas the small scales are represented by the higher levels of the casson handle .
let us consider the small exotic @xmath1 . from the physics point of view
, the large scale is given by the first levels of the casson handle .
in the construction of the foliation of @xmath1 , the first levels describe a polygon in the hyperbolic space @xmath343 with a finite and small number of vertices .
the godbillon - vey number of this foliation is given by the volume of this polygon . in principle , it is also true for the inclusion of the higher levels ( and also for the whole casson handle ) but every higher level gives only a very small contribution to the godbillon - vey number .
therefore , the first levels of the casson handle can be simply characterized by the godbillon - vey number , i.e. by the size of the polygon in the scale @xmath577 .
then the godbillon - vey number is given by @xmath578 .
in @xcite we analyzed this situation and found the relation @xmath579 to the godbillon - vey number .
here we integrate over the disk ( equal to the polygon ) which is used to define the foliation .
this model is the non - linear sigma model ( for the embedding of the disk into @xmath198 with metric @xmath147 ) depending on the scale @xmath580 .
the scaling behavior of this model was studied in @xcite and one obtains the rg flow equation @xmath581 reducing to the ricci flow equations for large scales ( @xmath582 ) .
the fixed point of this flow are geometries of constant curvature ( used to prove the thurston geometrization conjecture ) .
therefore in the classical limit of large scales , we obtain a geometry of the 3-manifold of constant curvature whereas for small scales one has to take into account higher curvature corrections . on the spacetime , one has also flow equations from one 3-manifold of constant curvature to another 3-manifold of constant curvature .
this flow equation is equivalent to the ( anti-)self - dual curvature ( or instantons ) by using the gradient flow of the chern - simons functional @xcite .
this approach has much in common with the non - linear graviton of penrose @xcite .
we will explain these ideas in subsection [ sub : the - graviton - propagator ] .
for the short scale , we need the full power of the casson handle . as a first step we can evaluate the action ( [ eq : quantum - action ] ) so that the dirac operator @xmath201 acts on usual square - integrable functions , so that @xmath583=dx^{\mu}$ ] is finite
. then the action ( [ eq : quantum - action ] ) reduces to @xmath584 where @xmath585 is the index for the coordinates on @xmath1 and @xmath586 represents the index of the disk ( inside of the casson handle ) .
now we will choose a small fluctuation @xmath587 of a fixed embedding of the disk in the casson handle given by @xmath588 with @xmath589 .
then we obtain @xmath590 and we use a standard argument to neglect the terms linear in @xmath591 : fluctuations have no preferred direction and therefore only the square contributes
. then we have @xmath592 for the action . by using a result of @xcite one
obtains for the dixmier trace @xmath593 with the first coefficient @xmath594 of the heat kernel expansion @xcite @xmath595 and the action simplifies to @xmath596 for the main contributions where @xmath74 is the scalar curvature of the embedded disk @xmath125 . again , but now for small fluctuations , we obtain the flow equation ( [ eq : rg - flow - equation ] ) but we have to consider the small case @xmath597
. then we have to take arbitrary curvature contributions into account .
this short calculation showed that the short - scale behavior is given by a two - dimensional action . in the next section
we will understand this behavior geometrically . for small fluctuations we obtained a disk but what happens for larger fluctuations
? then we have to take even the higher levels of the casson handle into account .
these higher levels form a complicated surface with a fractal structure ( a generalization of the cantor set ) .
then the action ( [ eq : action - small - fluctuations ] ) has to be replaced by an integral over this fractal space . for the evaluation of the quantum action ( [ eq : quantum - action ] ) one can use the ideas of noncommutative geometry as used for fractals and quasi - fuchsian groups , see @xcite ( section iv.3 ) .
in this section we will present some properties of the theory . for an impression ,
it is enough to present the main ideas .
the details will be published separately . by using the large scale behavior in subsection [ sub : long - scale - behaviour ]
, we have to consider ricci - flat spaces and an easy calculation gives the well - known propagator in the linearized version , however , we are not interested in the linearized version .
grt is a highly non - linear theory and therefore one has to take this non - linearity into account .
the ricci - flatness of the spacetime goes over to the 3-manifold as the spatial component where it implies a 3-manifold of constant curvature ( as fixed point of the ricci flow ) .
then as shown by witten @xcite , the 3-dimensional einstein - hilbert action @xmath598 is related to the chern - simons action @xmath599 with respect to the ( levi - civita ) connection @xmath149 and the length @xmath262 . by using the stokes theorem
we obtain @xmath600)=\intop_{m_{t}}tr(f\wedge f ) \
, , \ ] ] i.e. the action for the 4-manifold @xmath601 $ ] ( as local spacetime ) with the curvature @xmath602 , i.e. the action is the ( topological ) pontrjagin class of the 4-manifold .
from the formal point of view , the curvature 2-form @xmath602 is generated by a @xmath215 connection @xmath149 in the frame bundle , which can be lifted uniquely to a @xmath394- ( spin- ) connection . according to the ambrose - singer theorem ,
the components of the curvature tensor are determined by the values of holonomy which is in general a subgroup of @xmath394 .
thus we start with a suitable curvature 2-form @xmath602 with values in the lie algebra @xmath603 of the lie group @xmath268 as subgroup of the @xmath394 .
the variation of the chern - simons action gets flat connections @xmath604 as solutions .
the flow of solutions @xmath605 in @xmath601 $ ] ( parametrized by the variable @xmath606 , the time ) between the flat connection @xmath607 in @xmath608 to the flat connection @xmath609 in @xmath610 will be given by the gradient flow equation ( see for instance @xcite ) @xmath611 where the coordinate @xmath606 is normal to @xmath336 .
therefore we are able to introduce a connection @xmath612 in @xmath601 $ ] so that the covariant derivative in @xmath606-direction agrees with @xmath613
. then we have for the curvature @xmath614 , where the fourth component is given by @xmath615
. thus we will get the instanton equation with ( anti- ) self - dual curvature @xmath616 it follows @xmath617\times n)=\intop_{n\times[0,1]}tr(\tilde{f}\wedge\tilde{f})=\pm\intop_{n\times[0,1]}tr(\ \tilde{f}\wedge*\tilde{f})\,,\ ] ] ( i.e. the macdowellmansouri action ) .
we remark the main point in this argumentation : we obtain a self - dual curvature as gradient flow between two 3-manifolds of constant curvature . of course
, ( anti-)self - dual curvatures are also solutions of einsteins equation ( but the reverse is not true ) .
following penrose @xcite , we call these solutions the nonlinear graviton . above we constructed the observable algebra from the foliation leading to the kauffman bracket skein module . in the subsection
we will discuss the relation to lattice gauge field theory .
main source for this discussion is the work of bullock , frohman and kania - bartoszyska @xcite . in this paper
the authors realize that gauge fields come from the restricted dual of the hopf algebra on which the theory is based .
this leads to a coordinate free formulation .
then they comultiply connections in a way that implies the usual exchange relations for fields while preserving their evaluability .
their new foundations allow them to compute wilson loops and many other operators using a simple extension of tangle functors .
then they analyzed the structure of the algebra of observables . in their viewpoint , the observables correspond to quantum groups seen as rings of invariants of n - tuples of matrices under conjugation .
the connection with lattice gauge field theory is that each n - tuple of matrices corresponds to a connection on a lattice with one vertex and n - edges , with the gauge fields based on a classical group .
the construction given in this paper leads to an algebra of characters of a surface group with respect to any ribbon hopf algebra .
the algebras are interesting from many points of view : they generalize objects studied in invariant theory ; they should provide tools for investigating the structure of the mapping class groups of surfaces ; and they should give a way of understanding quantum invariants of @xmath618-manifolds .
the algebra of observables based on the enveloped lie algebra @xmath619 is proved to be the ring of @xmath268-characters of the fundamental group of the associated surface .
then , given the ring of @xmath268-characters of a surface group , they showed that the observables based on the corresponding drinfeld - jimbo algebra form a quantization with respect to the usual poisson structure .
furthermore they proved for the classical groups that the algebra of observables is generated by wilson loops . finally , invoking a quantized cayley - hamilton identity
, they obtain a new proof , that the @xmath620-characters of a surface are exactly the kauffman bracket skein module of a cylinder over that surface .
the power of lattice gauge field theory is that it places the representation theory of the underlying manifold and the quantum invariants in the same setting .
ultimately the asymptotic analysis of the quantum invariants of a @xmath618-manifold in terms of the representations of its fundamental group should flow out of this setting .
the identification of the representation theory of a quantum group with that of a compact lie group leads to rigorous integral formulas for quantum invariants of @xmath618-manifolds .
this should in turn lead to a simple explication of the relationship between quantum invariants and more classical invariants of @xmath618-manifolds .
this relation to lattice gauge field theory seems to imply an underlying discrete structure of the space and/or spacetime , but the approach in the paper uncovers the reason @xcite : the kauffman bracket skein module is discrete structure containing only a finite amount of information .
therefore , any description has to be discrete as well including the approach via gauge fields .
this idea can be extended to the 4-manifold .
as explained above , every smooth 4-manifold can be effectively described by handles and one only needs a finite number to describe every compact 4-manifold .
then the handles can be simply triangulated by using simplices to end up with a piecewise - linear ( or pl ) structure
. the surprising result of cerf for manifolds of dimension smaller than 7 was simple : pl - structure ( or triangulations ) and smoothness structure are the same .
this implies that every pl - structure can be smoothed to a smoothness structure and vice versa .
therefore _ the discrete approach ( via triangulations ) and the smooth approach to defining a manifold are the same _ !
so , our spacetime admits a kind of duality : it contains discrete information in its handle structure but it is a continuous space at the same time . both approaches are interchangeable .
therefore the underlying structure of the spacetime is discrete but the spacetime itself is a smooth 4-manifold . or , the information contained in a smooth 4-manifold is finite . in @xcite we describe an exotic black hole by constructing a smooth metric for the interior .
here we will present the main argument shortly .
in @xcite the existence of an exotic black hole ( as exotic kruskal space ) using an exotic @xmath0 was suggested .
the idea was simply to consider the complement @xmath621 where @xmath622 was only understood topologically . in case of the exotic small @xmath0 given by a casson handle
, we can reproduce our construction of an exotic @xmath623 by using a casson handle .
therefore we will here concentrate on the representation of the exotic @xmath623 by using the casson handle @xmath124 to get @xmath624 in @xcite the analytical properties of the casson handle were discussed .
the main idea is the usage of the theory of end - periodic manifolds , i.e. an infinite periodic structure generated by @xmath59 glued along a compact set @xmath134 to get @xmath625 the end - periodic manifold .
the definition of an end - periodic manifold is very formal ( see @xcite ) and we omit it here . all casson handles generated by a balanced tree have the structure of end - periodic manifolds as shown in @xcite . by using the theory of taubes @xcite
one can construct a metric on @xmath626 by using the metric on @xmath59 .
then a metric @xmath147 in @xmath627 transforms to a periodic function @xmath628 on the infinite periodic manifold @xmath629 where @xmath630 is the building block @xmath59 at the @xmath631-th place . to reflect the number of the building block
, we have to extend @xmath628 to @xmath632by using a metric @xmath633holomorphic in @xmath634 with @xmath635 where @xmath631 identifies the two boundaries of @xmath59 . from the formal point of view
we have the _ generalized fourier - laplace transform ( or fourier - laplace transform for short ) _
@xmath636 where the coefficient @xmath637 represents the building block @xmath638 in @xmath639 . without loss of generality
we can choose the coordinates @xmath278 in @xmath127 so that the @xmath143-th component @xmath640 is related to the integer @xmath641 $ ] via its integer part @xmath642 $ ] .
using the inverse transformation we can construct a smooth metric @xmath147 in @xmath639 at the @xmath21-th building block via @xmath643 for @xmath644 , @xmath645 , @xmath641 $ ] with the projection @xmath646 ( mathematically : @xmath639 is the universal cover of @xmath647 like @xmath461 is the universal cover of @xmath119 ) . in the case of the kruskal space
we have the metric @xmath648 in the usual units with a singularity at @xmath649 used for the whole space @xmath623 .
the coordinates @xmath650 together with the relation @xmath651 represent @xmath543 and the angles @xmath652 the 2-sphere @xmath212 parametrized by the radius @xmath577 .
clearly this metric can be also used for each building block @xmath59 having the topological structure @xmath653 with two attaching regions topologically given by @xmath654 forming the boundary @xmath655 ( see the description of a casson handle above ) .
remember that the casson handle is topologically the subset @xmath656 .
now we consider the decomposition @xmath657 and the part @xmath658 will be later the @xmath543 part of the casson handle .
the size of the @xmath125 is parametrized by @xmath577 as above
. then we obtain the metric ( [ eq : kruskal - metric ] ) for the building block @xmath59 . our model of the black hole based on the implicit dependence of the two coordinates @xmath650 on the parameter @xmath577 , the radius of the 2-sphere .
therefore we choose for the coordinate @xmath659 the relation @xmath660 and obtain a metric @xmath628 on @xmath632 .
so , we make the assumptions : 1 . the coordinate @xmath661 is related to the radius by @xmath660 .
only the @xmath650 part of the metric is periodic and we do not change the other component @xmath662 of the metric .
the integer part @xmath663 $ ] of the coordinate @xmath664 gives the number of the building block @xmath638 in the casson handle ( seen as end - periodic manifold ) .
the metric on @xmath665 is given by a fourier transformation ( [ eq : fl - trafo ] ) of the @xmath650 part of the metric in the building block @xmath59 .
some more comments are in order .
the number @xmath663 $ ] is related to the coordinate @xmath664 as substitute of time .
the metric @xmath147 in @xmath639 is smooth with respect to @xmath664 and we obtain the number of the building block by @xmath663 $ ] . to express this property we have to identify @xmath650 with the coordinates of @xmath658 .
then we obtain the metric on @xmath665 by the generalized fourier - laplace transformation of the metric on @xmath635 using the metric of the building block @xmath59 and the coordinate @xmath661 similar to ( [ eq : fl - trafo ] ) @xmath666 especially the singular part of the metric ( i.e. the @xmath650 part ) on the building block @xmath59 @xmath667 transforms to the heaviside jump function @xmath668 using the relation ( [ eq : kruskal - relation ] ) , having no singularity .
the metric vanishes , however , for large values of @xmath664 in the interior of the black hole .
this sketch of some arguments gives a hint that the transformation of the smoothness to exotic smoothness could possibly smooth out some singularity in the black hole case .
this metric vanishes along two directions or _
one obtains a dimensional reduction from 4d to 2d .
_ but , is there a geometrical reason for this reduction ? a hyperbolic 3-manifold @xmath127 admits a hyperbolic structure by fixing a homomorphism @xmath400 ( up to conjugation ) . from the physics point of view , this homomorphism is given by the holonomy along a closed curve ( as element in @xmath669 ) for a flat connection .
a sequence of these holonomies does not converge but it is possible to compactify the space of flat @xmath394 connections .
this limit can be understood geometrically : the hyperbolic 3-manifold is triangulated by tetrahedrons . however , because of the hyperbolic geometry , the edge between two vertices is not the usual line but rather a geodesics in the hyperbolic geometry .
the curvature of this geodesics depends on the hyperbolic structure . in the limit ,
all geodesics of the tetrahedron meet and one obtains a tree instead of tetrahedrons .
therefore in the limit of large curvature , one obtains a reduction from 3d ( = tetrahedrons ) to 1d ( = tree ) .
[ fig : degen - hyp - tree ] visualizes the transition from 2d(triangle ) to 1d(tree ) . ]
in a purely geometrical theory , one has to answer this question . it can not be shifted to assume the appearance of quantum fluctuations .
instead we have to understand the root of these quantum fluctuations .
starting point of our approach is the foliation of the exotic @xmath1 by using the anosov flow .
main point in the argumentation above is the appearance of the hyperbolic geometry in 3- and 4-dimensional submanifolds .
the foliation can , however , be interpreted differently : a foliation defines a dynamics at a manifold leading to a splitting into leafs ( the integral curves of the dynamics ) .
therefore , a tiny variation in the initial conditions will lead to a strong variation of the corresponding integral curve .
this chaotic behavior is a natural consequence of the exotic smoothness structure ( leading to the non - trivial @xmath241foliation ) . for completeness we will describe this dynamics , called the anosov flow . for that purpose
we consider the standard basis @xmath670 of the lie algebra @xmath671 with @xmath672=x\quad\quad[j , y]=-y\quad\quad[x , y]=2j\ ] ] leading to the exponential maps @xmath673 defining right - invariant flows on the unit tangent bundle @xmath674 of the hyperbolic space .
the connection to the anosov flow comes from the realization that @xmath675 is the geodesic flow on @xmath676 . with
lie vector fields being ( by definition ) left invariant under the action of a group element , one has that these fields are left invariant under the specific elements @xmath675 of the geodesic flow .
this flow goes over to a surface @xmath677 defined by a subgroup @xmath678 with @xmath679 .
now the geodesic flow @xmath675 acts on the exponential maps @xmath680 so that the geodesic flow itself is invariant , @xmath681 , but the other two shrink and expand : @xmath682 and @xmath683 .
then the bundle @xmath684 splits into three subbundles @xmath685 where one bundle @xmath686 expands , one bundle @xmath687 contracts and one bundle @xmath688 is invariant w.r.t .
geodesic flow .
this property is crucial for the following discussion . because of the expanding behavior of one subbundle , the anosov flow is the generator of a chaotic dynamics .
therefore , two geodesics diverge exponentially in this foliation , but this behavior goes over to the holonomies characterizing the geometry .
the transport of a holonomy along two diverging geodesics can lead to totally different holonomies .
currently this dynamics is deterministic , i.e. if we choose exactly the same initial condition then we will end at the state ( seen as limit point ) .
this situation changes if we are unable to choose the initial condition exactly ( by choosing real numbers ) but instead we can only choose a rational number where this rational number is the characterizing property of the state
. then all initial conditions ( represented by all real numbers ) in this class represent the same state but have totally different limit points of the corresponding dynamics .
now we will describe this dynamics .
starting point is the observable algebra @xmath689 , i.e. the space of holonomies @xmath690 ( i.e. homomorphisms ) up to conjugation , see subsection [ sub : the - observable - algebra ] .
the deformation quantization ( see subsection [ sub : drinfeld - turaev - quantization ] ) is the kauffman bracket skein module .
here we made use of the identity @xmath691 between two elements @xmath692 ( w.r.t . a representation ) .
using the group commutator @xmath693=aba^{-1}b^{-1}$ ] one also obtains @xmath694)=(tr(a))^{2}+(tr(b))^{2}+(tr(ab))^{2}-tr(a)tr(b)tr(ab)\ ] ] according to deformation procedure , pairs of elements @xmath692 coming from closed curves via the holonomy and fulfilling @xmath695)=\pm2\ ] ] can be a canonical pair w.r.t . the symplectic structure .
the sign is purely convention and we choose @xmath696)=-2 $ ] . then the canonical pair has to fulfill the equation @xmath697 which can be written in a more familiar form @xmath698 by using @xmath699 .
because of the discreteness of @xmath468 , we have to look for rational solutions of this equation ( diophantine equation ) .
the solutions of the equation are markoff triples forming a binary tree ( see fig .
[ fig : binary - tree - of - markoff ] ) . binary tree of markoff numbers as solution of equation ( [ eq : markoff - equation ] ) . ] the set of these elements @xmath700 corresponding to discrete groups is known to be fractal in nature @xcite .
it is the large class of quasi - fuchsian groups having a fractal curve ( julia set ) as limit set .
then we have the desired behavior : + _ the set of canonical pairs ( as measurable variables in the theory ) forms a fractal subset of the space of all holonomies .
then we can only determine the initial condition up to discrete value ( given by the canonical pair ) and the chaotic behavior of the foliation ( i.e. the anosov flow ) makes the limit not predictable . _
+ at the end of this section , one remark about the role of the canonical pair .
it is always possible to construct a classical continuous random field that has the same probability density as the quantum vacuum state .
furthermore it is known that a random field can be generated by a chaotic dynamics .
there is , however , a large difference between the classical random field and a quantum field : there are pairs of not equally accurate measurable observables ( mostly the canonical pairs ) for quantum fields impossible for the classical random fields . with our approach , we showed the same behavior for the canonical pairs .
our geometrical approach should also lead to a description of the measurement process ( including the collapse of the wave function ) . in section
[ sec : wild - embeddings ] , we constructed the geometrical expression for a quantum state given by a wild embedding ( the wild @xmath60 ) .
the reduction of the quantum state ( as linear combination ) to an eigenstate ( or the collapse of the wave function ) is equivalent to a reduction of the wild embedding to a tame embedding .
therefore we need a mechanism to reduce the wild embedding to a tame one .
the construction of the wild @xmath60 is strongly related to the casson handle .
the exoticness of the smooth structure of @xmath1 and the wildness of the @xmath60 depend both on the self - intersection of some disk .
if we are able to remove these self - intersections then we will obtain the desired reduction . according to the discussion in subsection [ sub : small - exotic ]
, one needs a casson handle for the cancellation .
how many levels of the casson handle are needed to cancel the self - intersection ?
this question was answered by freedman @xcite : one needs three levels ( a three - level casson tower ) ! at the same time , however , one produces more self - intersections in the higher levels . therefore one needs a little bit more : a casson tower where a complete casson handle can be embedded .
then this casson handle is able to cancel the self - intersection and we will obtain a tame embedding or a classical state . as shown by freedman @xcite and gompf / singh @xcite , one needs a 5-stage casson tower so that a casson handle with the same attaching circle can be embedded into this 5-stage tower .
we obtain a process which is the collapse of the wave function . what is the cause of this collapse ? as explained above , we can not choose a single disk to remove the self - intersections . instead we have to choose a casson tower where each stage is a boundary - connected sum of @xmath701 , i.e. its boundary is the sum @xmath702 where the number of components is equal to the number of self - intersections .
so , every piece @xmath703 of the boundary is given by the identification of the two boundary components for @xmath207 $ ] . in section [ sec :
action - induced - by - top ] , we identified this 3-manifold with the graviton or _ the collapse of the wave function is caused by a gravitational interaction_. the corresponding process is known as decoherence . in the following we will calculate the minimal decoherence time for the gravitational interaction . the 5-stage casson tower can be also understood as a cobordism between the 3-manifold @xmath704 ( the @xmath119 defines the attaching circle ) and a 3-manifold having the same homology . in case of the simplest casson tower ,
it is given by five complements of the whitehead link @xmath705 closed by two solid tori , i.e. @xmath706 and this manifold can be very complicated for more complex towers .
now we will add some geometry to calculate the decoherence time .
as shown by witten @xcite , the action @xmath707 for every 3-manifold ( in particular for @xmath708 and @xmath709 denoted by @xmath710 ) is related to the chern - simons action @xmath711 .
the scaling factor @xmath262 is related to the volume by @xmath712{vol(\sigma_{0,1})}$ ] and we obtain formally @xmath713 by using @xmath714 with the ( unit ) volume @xmath715 . if @xmath710 is a hyperbolic 3-manifold then the ( unit ) volume is a topological invariant which can not be normalized to 1 . together with @xmath716 one
can compare the kernels of the integrals of ( [ eq : witten - relation ] ) and ( [ eq : cs - integral - relation ] ) to get for a fixed time @xmath717 this gives the scaling factor @xmath718 where we set @xmath719 in the following .
the hyperbolic geometry of the cobordism is best expressed by the metric @xmath720 also called the friedmann - robertson - walker metric ( frw metric ) with the scaling function @xmath721 for the ( spatial ) 3-manifold .
mostow rigidity enforces us to choose @xmath722 in the length scale @xmath262 of the hyperbolic structure . in the following
we will switch to quadratic expressions because we will determine the expectation value of the area .
a second reason for the consideration of quadratic expressions is again the hyperbolic structure of @xmath343 .
we needed this structure for the construction of the foliation which is given by a polygon in @xmath343 .
this polygon defines a compact surface of genus @xmath723 .
then the foliation of the polygon induces a foliation of the small exotic @xmath1 .
the area of the polygon is mainly the godbillon - vey invariant of the foliation .
it is known that foliations of surfaces are given by quadratic differentials of the form defined below . here
, there are deep connections to trees and @xmath394 flat connections , i.e. a tree defines a quadratic differential and vice versa @xcite .
using the previous equation , we obtain @xmath724 with respect to the scale @xmath725 . by using the tree of the casson handle , we obtain a countable infinite sum of contributions for ( [ eq : scale - quadratic - expansion ] ) .
before we start we will clarify the geometry of the casson handle . a casson handle
admits a hyperbolic geometry . therefore the tree corresponding to the casson handle must be interpreted as a metric tree with hyperbolic structure in @xmath343 and metric @xmath726
the embedding of the casson handle in the cobordism is given by the rules 1 .
the direction of the increasing levels @xmath727 is identified with @xmath728 and @xmath729 is the number of edges for a fixed level with scaling parameter @xmath725 .
the contribution of every level in the tree is determined by the previous level best expressed in the scaling parameter @xmath725 .
3 . an immersed disk at level @xmath21 needs at least one disk to resolve the self - intersection point .
this disk forms the level @xmath421 but this disk is connected to the previous disk .
so we obtain for @xmath730 at level @xmath421 @xmath731 up to a constant . by using the metric @xmath726 with the embedding ( @xmath732 , @xmath733 ) we obtain for the change @xmath734 along the @xmath735direction ( i.e. for a fixed @xmath279 ) @xmath736 .
this change determines the scaling from the level @xmath21 to @xmath421 , i.e. @xmath737 and after the whole summation ( as substitute for an integral in case of discrete values ) we obtain for the relative scaling @xmath738 with @xmath739 .
the chern - simons invariant for @xmath708 vanishes and we are left with @xmath740 and the complements @xmath705 are hyperbolic 3-manifolds with @xmath741 by using the software snapea
. finally for the scaling we obtain @xmath742 and for the time we have to choose @xmath743 using the well - known relation @xmath744 between length and time , i.e. we see one coordinate along the casson handle as time axis .
the time @xmath745 has to be identified with the planck time @xmath746 ( see section [ sec : action - induced - by - top ] ) so that @xmath747 is the minimal decoherence time for the gravitational interaction .
now we also discuss the entanglement which has to be also geometrically expressed .
a quantum state is an element of the skein algebra @xmath554 for @xmath488 $ ] . for two disjoint surfaces
@xmath748 one has @xmath749 now let us choose a knot @xmath750 in @xmath751 $ ] as element of @xmath752 as well a knot @xmath753 in @xmath754 .
then @xmath755 is an element of @xmath756 .
furthermore we can assume that the knots @xmath753 and @xmath750 can be also an element of @xmath752 and @xmath754 , respectively .
then the element @xmath757 exists but now as an element of @xmath554 with @xmath758 .
using the skein relations in @xmath554 , see fig .
[ fig : skein - crossings ] , we obtain a linking between the corresponding knots , i.e. @xmath750 and @xmath753 forming a link . fig . [
fig : entangled - circle ] visualizes the transition from disjoint circles ( = disjoint states ) to linked circles ( = entangled states ) . ]
then entanglement is reduced to a linking ! next we have to think about the measurement which reduces the entangled state to one product state . here
we will only present some rough ideas for the description of the measurement process , but at first we have to define a measurement device . in this proposal , it is a union of casson handles which can be used to unlink two linked components . at the level of skein algebras ,
the casson handle is also given by elements of a skein algebra ( given by closed , knotted curves at the levels ) .
the particular structure of the casson handle is not determined ( see also section [ sec : where - does - fluctuation ] ) .
now a given quantum state is linked to this casson handle .
the limit point of the casson handle ( i.e. the leafs of the tree ) give the result of the unlinking .
all limit points of the casson handle have a fractal structure ( a cantor set ) expressing our inability to know the outcome of the measurement .
the tree structure of the casson handle has also another effect : the limit points are exponentially separated from each other and can be seen as classical states . with these speculations
, we will close this section .
in the last section we will collect some implications for a cosmological model .
let us assume the topology @xmath4 for the spacetime but with an exotic smoothness structure @xmath759 .
one can construct this spacetime from the exotic @xmath1 by @xmath760 . from previous work , we know : * cosmological anomalies like dark matter and dark energy are ( conjecturally ) rooted in exotic smoothness @xcite .
* the initial state of the cosmos must be a wild 3-sphere representing a quantum state @xcite .
* then there is an inflationary phase @xcite driven by a decoherence which can be described by the starobinsky model .
+ in this model , we have a topological transition from a 3-manifold @xmath708 to another 3-manifold @xmath709 . both 3-manifolds are homology 3-spheres .
therefore let us describe this change ( a so - called homology cobordism ) between two homology 3-spheres @xmath708 and @xmath709 .
the situation can be described by a diagram @xmath761 which commutes .
the two functions @xmath199 and @xmath762 are the morse function of @xmath708 and @xmath709 , respectively , with @xmath763 .
the morse function over @xmath710 is a function @xmath764 having only isolated , non - degenerated , critical points ( i.e. with vanishing first derivatives at these points ) .
a homology 3-sphere has two critical points ( located at the two poles ) .
the morse function looks like @xmath765 at these critical points .
the transition @xmath766 represented by the ( homology ) cobordism @xmath767 maps the morse function @xmath768 on @xmath708 to the morse function @xmath769 on @xmath709 .
the function @xmath770 represents also the critical point of the cobordism @xmath767 . as we learned above
, this cobordism has a hyperbolic geometry and we have to interpret the function @xmath771 not as an euclidean form but change it to the hyperbolic geometry so that @xmath772 i.e. we have a preferred direction represented by a single scalar field @xmath773 .
therefore , the transition @xmath774 is represented by a single scalar field @xmath773 and we identify this field as the moduli .
finally we interpret this morse function in the interior of the cobordism @xmath767 as the potential ( shifted away from the point @xmath143 ) of the scalar field @xmath775 with lagrangian @xmath776 with two free constants @xmath777 and @xmath778 .
for the value @xmath779 and @xmath780 we obtain the starobinski model @xcite ( by a conformal transformation using @xmath775 and a redefinition of the scalar field @xcite ) @xmath781 with the mass scale @xmath782 much smaller than the planck mass . from our discussion above
, the appearance of this model is not totally surprising .
it favors a surface to be incompressible ( which is compatible with the properties of hyperbolic manifolds ) . *
this inflationary phase is followed by another exponentially increasing phase leading to a hyperbolic 4-manifold with constant curvature which is rigid by mostow rigidity @xcite . here
, we obtained the global geometry of the spacetime : it is a de sitter space @xmath232 with a cosmological constant which is the curvature of the spacetime .
* this constant curvature can be identified with the cosmological constant in good agreement with the planck satellite results @xcite .
the cosmological constant is constant by mostow rigidity ( but now for the 4-manifold ) . *
the topology of the spatial component ( seen as 3-manifold ) is strongly restricted @xcite by the smoothness of the spacetime . *
the inclusion of matter can be done naturally as direct consequence of exotic smoothness @xcite . *
the interior of black holes can be described by exotic smoothness where the singularity is smoothed out @xcite .
smooth quantum gravity , the usage of exotic smoothness structures on 4-manifolds , are the attempt to obtain a consistent theory of quantum gravity without any further assumptions . for us , the change of the smoothness structure is the next step in extending general relativity , where non - euclidean geometry was used to describe gravity and all accelerations .
then , two different smoothness structures represent two different physical systems .
in particular i think that the standard smoothness structure represents the case of a spacetime without matter and non - gravitational fields . in this paper
we are going a more radical way to construct a quantum theory without quantization but by using purely geometrical ideas from mathematical topics like differential and geometric topology .
the flow of ideas can be simply described by the following points : * an exotic @xmath0 is given by an infinite handlebody ( so one needs infinitely many charts ) and one finds also the description by an infinite sequence of 3-manifolds together with 4-dimensional cobordisms connecting them . * every 3-manifold admits a codimension - one foliation which goes over to the 4-dimensional cobordisms .
the leaf space of this foliation is an operator algebra with a strong connection to algebraic quantum field theory . * the states ( as linear functionals in the algebra ) depend on knotted curves and are elements of the kauffman bracket skein algebra .
the reconstruction of the spatial space gives a wild embedded 3-sphere which is therefore related to the state , or the quantum state can be identified with the wild embedding .
the classical state is a tame ( i.e. usual ) embedding where the deformation quantization of a tame embedding is a wild embedding . *
the structure of the operator algebra can be analyzed by the tomita - takesaki modular theory .
then it is possible to construct the quantum action by using the quantized calculus of connes . * for large scales ,
one gets the einstein - hilbert action . whereas for small scales , one obtains a dimensional reduced action . *
the foliation is given by a hyperbolic dynamics having a chaotic behavior .
for our states , one gets an unpredictable behavior so that the dynamics can generate the quantum fluctuations .
this list shows the current state but there are many open points , where we list only the most important here : * what is the hamiltonian of the theory ? in principle
we constructed this operator but have a problem connecting to loop quantum gravity .
* what are the states seen as knots ?
the states are knots but the skein and mandelstam identities give a class of knots : the states are conjecturally the concordance class of knots .
* is the state a solution of the hamiltonian ?
here we conjecture that the concordance class of the knot lies already in the kernel of the hamiltonian ( therefore it is a solution of the hamiltonian constraint ) a lot is done but there are also many open problems .
+ * happy birthday carl ! *
i have to thank carl for 20 years of friendship and collaboration as well numerous discussions .
special thanks to jerzy krl for our work and many discussions about fundamental problems in math and physics .
now i understand the importance of model theory .
let us now consider the basic construction of the casson handle @xmath124 .
let @xmath127 be a smooth , compact , simply - connected 4-manifold and @xmath783 a ( codimension-2 ) mapping . by using diffeomorphisms of @xmath125 and @xmath127
, one can deform the mapping @xmath12 to get an immersion ( i.e. injective differential ) generically with only double points ( i.e. @xmath784 ) as singularities @xcite .
but to incorporate the generic location of the disk , one is rather interesting in the mapping of a 2-handle @xmath115 induced by @xmath785 from @xmath12 . then every double point ( or self - intersection ) of @xmath786 leads to self - plumbings of the 2-handle @xmath115 .
a self - plumbing is an identification of @xmath787 with @xmath788 where @xmath789 are disjoint sub - disks of the first factor disk . in complex
coordinates the plumbing may be written as @xmath790 or @xmath791 creating either a positive or negative ( respectively ) double point on the disk @xmath792 .
consider the pair @xmath793 and produce finitely many self - plumbings away from the attaching region @xmath794 to get a kinky handle @xmath795 where @xmath796 denotes the attaching region of the kinky handle
. a kinky handle @xmath795 is a one - stage tower @xmath797 and an @xmath798-stage tower @xmath799 is an @xmath21-stage tower union of kinky handles @xmath800 where two towers are attached along @xmath801 .
let @xmath802 be @xmath803 and the casson handle @xmath804 is the union of towers ( with direct limit topology induced from the inclusions @xmath805 ) .
a casson handle is specified up to ( orientation preserving ) diffeomorphism ( of pairs ) by a labeled finitely - branching tree with base - point * , having all edge paths infinitely extendable away from * . each edge should be given a label @xmath133 or @xmath806 and each vertex corresponds to a kinky handle ; the self - plumbing number of that kinky handle equals the number of branches leaving the vertex .
the sign on each branch corresponds to the sign of the associated self plumbing .
the whole process generates a tree with infinite many levels . in principle
, every tree with a finite number of branches per level realizes a corresponding casson handle .
the simplest non - trivial casson handle is represented by the tree @xmath807 : each level has one branching point with positive sign @xmath133 .
the reverse construction of a casson handle @xmath808 by using a labeled tree @xmath422 can be found in the appendix a. let @xmath809 and @xmath810 be two trees with @xmath811 ( it is the subtree ) then @xmath812.given a labeled based tree @xmath813 , let us describe a subset @xmath814 of @xmath115 .
now we will construct a @xmath815 which is diffeomorphic to the casson handle associated to @xmath813 . in @xmath115 embed a ramified whitehead link with one whitehead link component for every edge labeled by @xmath133 leaving * and one mirror image whitehead link component for every edge labeled by @xmath806(minus ) leaving * .
corresponding to each first level node of @xmath813 we have already found a ( normally framed ) solid torus embedded in @xmath816 . in each of these solid tori embed a ramified whitehead link ,
ramified according to the number of @xmath133 and @xmath806 labeled branches leaving that node .
we can do that process for every level of @xmath813 .
let the disjoint union of the ( closed ) solid tori in the @xmath21-th family ( one solid torus for each branch at level @xmath21 in @xmath813 ) be denoted by @xmath817 .
@xmath813 tells us how to construct an infinite chain of inclusions : @xmath818 and we define the whitehead decomposition @xmath819 of @xmath813 .
@xmath820 is the whitehead continuum @xcite for the simplest unbranched tree .
we define @xmath814 to be @xmath821 alternatively one can also write @xmath822 where @xmath823 is the cone of a space @xmath824/(x,0)\sim(x',0)\qquad\forall x , x'\in a\ ] ] over the point @xmath825 . as freedman ( see @xcite theorem 2.2 ) showed @xmath814 is diffeomorphic to the casson handle @xmath826 given by the tree @xmath813 .
in @xcite thurston constructed a foliation of the 3-sphere @xmath60 which depends on a polygon @xmath342 in the hyperbolic plane @xmath343 so that two foliations are non - cobordant if the corresponding polygons have different areas . for later usage
, we will present the main ideas of this construction only ( see also the book @xcite chapter viii for the details ) . starting point is the hyperbolic plane @xmath343 with a convex polygon @xmath827 having @xmath110 sides @xmath828 . assuming the upper half plane model of @xmath343
then the sides are circular arcs .
the construction of the foliation depends mainly on the isometry group @xmath255 of @xmath343 realized as rational transformations ( and this group can be lifted to @xmath829 ) .
the followings steps are needed in the construction : 1 .
the polygon @xmath134 is doubled along one side , say @xmath830 , to get a polygon @xmath831 .
the sides are identified by ( isometric ) transformations @xmath832 ( as elements of @xmath829 ) .
2 . take @xmath833-neighborhoods @xmath834 with @xmath835 sufficient small and set @xmath836 having the topology of @xmath837 and we set @xmath838 .
3 . now consider the unit tangent bundle @xmath839 , i.e. a @xmath840bundle over @xmath343 ( or the tangent bundle where every vector has norm one ) .
the restricted bundle over @xmath841 is trivial so that @xmath842 .
let @xmath843 be circular arcs ( geodesics ) in @xmath343 ( invariant w.r.t .
@xmath829 ) starting at a common point which define parallel tangent vectors w.r.t .
the metrics of the upper half plane model .
the foliation of @xmath841 is given by geodesics transverse to the boundary and we obtain a foliation of @xmath844 ( as unit tangent bundle ) .
this foliation is given by a @xmath829-invariant smooth 1-form @xmath218 ( so that @xmath845defines the leaves ) which is integrable @xmath243 .
( @xmath846invariant foliation @xmath847 ) 4 . with the relation @xmath848 ,
we obtain @xmath849 or the gluing of @xmath850 solid tori to @xmath844 gives a solid tori .
every glued solid torus will be foliated by a reeb foliation . finally using @xmath851 ( the heegard decomposition of the 3-sphere ) again with a solid torus with reeb foliation
, we obtain a foliation on the 3-sphere . the construction of this foliation @xmath852 ( thurston foliation ) will be also work for any 3-manifold .
thurston @xcite obtains for the godbillon - vey number @xmath853 and @xmath854 so that _ any real number can be realized by a suitable foliation of this type_. furthermore , two cobordant foliations have the same godbillon - vey number ( but the reverse is in general wrong ) .
let @xmath855\in h^{3}(s^{3},\mathbb{r})$ ] be the dual of the fundamental class @xmath856 $ ] defined by the volume form , then the godbillon - vey class can be represented by @xmath857\label{eq : godbillon - vey - class - thurston - foliation}\ ] ] the godbillon - vey class is an element of the derham cohomology @xmath858 .
now we will discuss the general case of a compact 3-manifold carrying a foliation of the same type like the 3-sphere above .
the main idea of the construction is very simple and uses a general representation of all compact 3-manifolds by dehn surgery . here
we will use an alternative representation of surgery by using the dehn - lickorish theorem ( @xcite corollary 12.4 at page 84 ) .
let @xmath859 be a compact 3-manifold without boundary .
there is now a natural number @xmath860 so that any orientable 3-manifold can be obtained by cutting out @xmath110 solid tori from the 3-sphere @xmath60 and then pasting them back in , but along different diffeomorphisms of their boundaries .
moreover , it can be assumed that all these solid tori in @xmath60 are unknotted . then any 3-manifold @xmath859 can be written as @xmath861 where @xmath862 is the gluing map from each boundary component of @xmath863 to the boundary of @xmath864 .
this gluing map is a diffeomorphism of tori @xmath865 ( where @xmath866 ) .
the dehn - lickorish theorem describes all diffeomorphisms of a surface : every diffeomorphism of a surface is the composition of dehn twists and coordinate transformations ( or small diffeomorphisms ) .
the decomposition @xmath867 of the 3-sphere can be used to get a decomposition of @xmath859 by @xmath868 which will guide us to the construction of a foliation on @xmath859 : * construct a foliation @xmath869 on @xmath844 using a polygon @xmath342 ( see above ) and * glue in @xmath110 reeb foliations of the solid tori using the diffeomorphisms @xmath870 .
finally we get a foliation @xmath871 on @xmath859 . according to the rules above
, we are able to calculate the godbillon - vey number @xmath872 therefore for any foliation of @xmath60 , we can construct a foliation on any compact 3-manifold @xmath859 with the same godbillon - vey number .
both foliations @xmath852 and @xmath871 agree for the common submanifold @xmath844 or there is a foliated cobordism between @xmath873 and @xmath874 .
of course , @xmath60 and @xmath859 differ by the gluing of the solid tori but every solid torus carries a reeb foliation which does not contribute to the godbillon - vey number .
a connected 3-manifold @xmath336 is prime if it can not be obtained as a connected sum of two manifolds @xmath875 neither of which is the 3-sphere @xmath60 ( or , equivalently , neither of which is the homeomorphic to @xmath336 ) .
examples are the 3-torus @xmath876 and @xmath703 but also the poincare sphere . according to @xcite , any compact ,
oriented 3-manifold is the connected sum of a unique ( up to homeomorphism ) collection of prime 3-manifolds ( prime decomposition ) .
a subset of prime manifolds are the irreducible 3-manifolds .
a connected 3-manifold is irreducible if every differentiable submanifold @xmath467 homeomorphic to a sphere @xmath212 bounds a subset @xmath201 ( i.e. @xmath877 ) which is homeomorphic to the closed ball @xmath123 .
the only prime but reducible 3-manifold is @xmath703 . for the geometric properties ( to meet thurstons geometrization theorem )
we need a finer decomposition induced by incompressible tori . a properly embedded connected surface @xmath878 is called 2-sided then correspond to the components of the complement of @xmath467 in a tubular neighborhood @xmath488\subset n$ ] .
] if its normal bundle is trivial , and 1-sided if its normal bundle is nontrivial
. a 2-sided connected surface @xmath467 other than @xmath212 or @xmath125 is called incompressible if for each disk @xmath879 with @xmath880 there is a disk @xmath881 with @xmath882 .
the boundary of a 3-manifold is an incompressible surface .
most importantly , the 3-sphere @xmath60 , @xmath883 and the 3-manifolds
@xmath884 with @xmath885 a finite subgroup do not contain incompressible surfaces .
the class of 3-manifolds @xmath884 ( the spherical 3-manifolds ) include cases like the poincare sphere ( @xmath886 the binary icosaeder group ) or lens spaces ( @xmath887 the cyclic group ) .
let @xmath888 be irreducible 3-manifolds containing incompressible surfaces then we can @xmath336 split into pieces ( along embedded @xmath212 ) @xmath889 where @xmath890 denotes the @xmath21-fold connected sum and @xmath885 is a finite subgroup . the decomposition of @xmath336 is unique up to the order of the factors .
the irreducible 3-manifolds @xmath891 are able to contain incompressible tori and one can split @xmath888 along the tori into simpler pieces @xmath892 @xcite ( called the jsj decomposition ) .
the two classes @xmath268 and @xmath187 are the graph manifold @xmath268 and the hyperbolic 3-manifold @xmath187 ( see figure [ fig : torus - decomposition ] ) .
the hyperbolic 3-manifold @xmath187 has a torus boundary @xmath895 , i.e. @xmath187 admits a hyperbolic structure in the interior only . in this paper
we need the splitting of the link / knot complement .
as shown in @xcite , the whitehead double of a knot leads to jsj decomposition of the complement into the knot complement and the complement of the whitehead link ( along one torus boundary of the whitehead link complement ) .
one property of hyperbolic 3-manifolds is central : mostow rigidity . as shown by mostow @xcite ,
every hyperbolic @xmath142manifold @xmath896 with finite volume has this property : _ every diffeomorphism ( especially every conformal transformation ) of a hyperbolic @xmath142manifold with finite volume is induced by an isometry .
_ therefore one can not scale a hyperbolic 3-manifold and the volume is a topological invariant . together with the prime and jsj decomposition @xmath897 we can discuss the geometric properties central to thurstons geometrization theorem : _ every oriented closed prime 3-manifold can be cut along tori ( jsj decomposition ) , so that the interior of each of the resulting manifolds has a geometric structure with finite volume .
_ now , we have to clarify the term geometric structure s .
a model geometry is a simply connected smooth manifold @xmath37 together with a transitive action of a lie group @xmath268 on @xmath37 with compact stabilizers .
a geometric structure on a manifold @xmath336 is a diffeomorphism from @xmath336 to @xmath898 for some model geometry @xmath37 , where @xmath899 is a discrete subgroup of @xmath268 acting freely on @xmath37 .
t is a surprising fact that there are also a finite number of three - dimensional model geometries , i.e. 8 geometries with the following models : spherical @xmath900 , euclidean @xmath901 , hyperbolic @xmath902 , mixed spherical - euclidean @xmath903 , mixed hyperbolic - euclidean @xmath904 and 3 exceptional cases called @xmath905 ( twisted version of @xmath906 ) , nil ( geometry of the heisenberg group as twisted version of @xmath907 ) , sol ( split extension of @xmath543 by @xmath461 , i.e. the lie algebra of the group of isometries of 2-dimensional minkowski space ) .
we refer to @xcite for the details .
h. abchir .
invariants at infinity for the whitehead manifold . in c.
gordan , v.f.r .
jones , l. kauffman , s. lambropoulou , and j.h .
przytycki , editors , _ knots in hellas 98 _ , pages 117 , singapore , 1998 .
world scientific .
t. asselmeyer - maluga and j. krl .
small exotic smooth @xmath909 and string theory . in _ international congress of mathematicians icm 2010 short communications abstracts book _ , ed .
r. bathia , page 400 , hindustan book agency , 2010 . t. asselmeyer - maluga and r. mader .
exotic @xmath909 and quantum field theory . in c. burdik et .
al . , editor , _
7th international conference on quantum theory and symmetries ( qts7 ) _ , page 012011 , bristol , uk , 2012 .
iop publishing .
arxiv:1112.4885 , doi:10.1088/1742 - 6596/343/1/012011 . | over the last two decades , many unexpected relations between exotic smoothness , e.g. exotic @xmath0 , and quantum field theory were found . some of these relations are rooted in a relation to superstring theory and quantum gravity .
therefore one would expect that exotic smoothness is directly related to the quantization of general relativity . in this article
we will support this conjecture and develop a new approach to quantum gravity called _ smooth quantum gravity _ by using smooth 4-manifolds with an exotic smoothness structure .
in particular we discuss the appearance of a wildly embedded 3-manifold which we identify with a quantum state .
furthermore , we analyze this quantum state by using foliation theory and relate it to an element in an operator algebra . then we describe a set of geometric , non - commutative operators , the skein algebra , which can be used to determine the geometry of a 3-manifold .
this operator algebra can be understood as a deformation quantization of the classical poisson algebra of observables given by holonomies .
the structure of this operator algebra induces an action by using the quantized calculus of connes .
the scaling behavior of this action is analyzed to obtain the classical theory of general relativity ( grt ) for large scales .
this approach has some obvious properties : there are non - linear gravitons , a connection to lattice gauge field theory and a dimensional reduction from 4d to 2d .
some cosmological consequences like the appearance of an inflationary phase are also discussed . at the end
we will get the simple picture that the change from the standard @xmath0 to the exotic @xmath1 is a quantization of geometry .
_ on the occasion of the 80-th birthday of carl h. brans _ |
the past decade has been marked by the emergence of electron quantum optics . in the spirit of quantum optics with photons , it aims to generate and manipulate single electronic excitations in quantum coherent circuits for fundamental and applied science . as a first step to achieve this goal ,
several single electron sources have been implemented in sub - micron cavities , such as the so called quantum capacitor @xcite and others @xcite , or by applying a periodic sequence of lorentzian voltage pulses to an electronic reservoir in order to generate a clean and coherent train of electronic excitations @xcite . once injected into quantum circuits ,
such excitations can be used to study fundamental aspects of quantum mechanics such as entanglement @xcite , interference effects @xcite and quantum correlations @xcite or interaction effects @xcite and coherence properties @xcite which would be of great interest when it comes to applications in quantum electronics or information processing .
however , due to quantum effects , it is now well established that charge transport at the nanoscale is a statistical process @xcite .
going beyond the knowledge of average currents is then unavoidable and extremely useful at the same time as pointed out by r. landauer in his famous quote `` the noise is the signal '' .
therefore , many efforts have been made in this direction in the past two decades using noise measurements @xcite and full counting statistics ( fcs ) @xcite , namely the second moment of current fluctuations and the statistics of charges transferred during a long time interval .
recently , other tools have been introduced to characterize current fluctuations in the time domain like the finite - frequency noise @xcite and fcs @xcite , the wigner function @xcite or the waiting time distribution ( wtd ) @xcite .
indeed , in such quantum devices , the time between the detection of two consecutive electrons is random because of their quantum nature and the knowledge of its probability distribution provides an original point of view on quantum correlations and current fluctuations . .
the outgoing many - body state is made of transmitted and missing ( reflected ) electrons represented by filled curves and dashed lines respectively .
the quantity of interest is the distribution of delay times @xmath0 between the detection of two electrons far away from the scattering region.[fig_setup ] ] we study here the wtd of non - interacting electrons emitted in a sequence of lorentzian pulses as proposed theoretically @xcite and demonstrated experimentally recently @xcite .
we focus here on the regime where electrons are emitted one by one when applying a periodic voltage to an electronic reservoir in a clean fashion ( no spurious electron - hole pairs ) .
these electrons are then propagating along a one - dimensional quantum channel and may scatter onto a quantum point contact ( qpc ) before being detected ( see fig .
[ fig_setup ] ) .
this situation is physically different from the one studied in ref . where free electrons were injected directly from the reservoir subjected to a constant bias @xmath1 into a single channel .
the ratio @xmath2 between the width of the wave packets @xmath3 and the interparticle distance @xmath4 is a new control parameter that drives the many - body quantum state from a solid - like state , when the wave packets weakly overlap , to a liquid - like behavior at strong overlap . in the latter limit
, the constant bias case is recovered .
we use a wave packet approach @xcite to derive general formulas for the wtd that we evaluate both numerically and analytically in some limiting cases .
these formulas are equivalent to those derived in @xcite from a second - quantized formulation .
in addition we show that the decay rate of the wtd corresponds to the one of a binomial process and describe analytically the quantum fluctuations around this classical result .
although we focus on lorentzian pulses here , our theory is valid for arbitrary shapes of the wave packets .
this paper is organized as follows . in sec .
ii we describe the setup under investigation and show how to compute the wtd from the many - body state and the scattering matrix of the quantum point contact .
we then turn to the results obtained with this formalism , starting from the two - particle case for pedagogical reasons in sec .
iii before discussing the periodic case in sec .
section v is devoted to a brief discussion of experimental measurements of the wtd .
conclusion and outlook are given in sec .
vi and several technical details are available in appendices .
we consider a qpc connected to two one - dimensional electronic leads .
a periodic sequence of lorentzian pulses of the form @xmath5 with @xmath6 , is locally applied to the left reservoir and generates a train of single electron excitations without electron - hole pairs as demonstrated theoretically @xcite and experimentally @xcite recently . here
@xmath3 and @xmath4 will be respectively the width and the spatial period of the wave packets .
we focus here on the special case where each pulse contains exactly one electron .
these electrons are moving to the right and scatter on the qpc as depicted in fig .
[ fig_setup ] .
in addition we consider zero temperature , therefore only electrons emitted from the left reservoir are involved in the transport process .
the incoming train of electrons , generated by the periodic voltage , is built up from a set of @xmath7 single particle lorentzian wave packets separated by time interval @xmath8 , @xmath9 @xmath10 . at low energy ,
the dispersion relation is linear and the fermi velocity @xmath11 is supposed to be a constant ( independent of energy ) .
each wave packet is a superposition of plane waves of energy @xmath12 but the amplitude in the superposition is exponentially decaying with energy ( or wave number since @xmath13 ) .
a crucial quantity is the overlap between wave functions @xmath14 , for @xmath15 that controls the correlations between the different electrons .
taking into account the fermionic statistics , the many - body wave - function has to be antisymmetrized and as far as interactions are negligible , it will be given by a slater determinant of all the possible orbitals @xmath16 .
now , the pulses coming from the left of the scattering regions ( see fig.[fig_setup ] ) are impinging on a quantum point contact of transparency @xmath17 , so that @xmath16 needs to be changed to @xmath18 with @xmath19 , for @xmath20 far away to the right of the qpc , and will be measured by a detector located in this region .
for the sake of simplicity , we only consider energy independent scattering with therefore @xmath21 . for negligible overlap , the normalization is trivial apart from an overall @xmath22 factor .
however , for finite overlap , this statistical coefficient has to be completed with the determinant of the overlap matrix elements between different wave packets @xcite @xmath23 .
as mentioned in the introduction , the quantity of interest is the wtd @xmath24 , namely the probability distribution of delay times between the detection of two consecutive electrons .
the detector is located at @xmath25 somewhere far away from the scattering region between the qpc and the right reservoir . in that case , the wtd depends on both the delay time @xmath0 and a second time @xmath26 which is chosen here to be the time when the first electron is measured .
this situation is more general than the one of a stationary flow of particles described in ref . where it only depends on the delay time .
to calculate this quantity , we will either refer to the joint probability of measuring an electron at time @xmath26 and nothing until @xmath27 , @xmath28 or the probability of not detecting anything between @xmath26 and @xmath29 , @xmath30 even if they are related quantities .
following these definitions , it is straightforward to show the following useful relations @xmath31 @xmath32 and @xmath33 where @xmath34 denotes the probability density to detect an electron at time @xmath26 and is simply proportional to the average current . although not fundamentally different , we will use either @xmath35 or @xmath36 to compute the wtd depending on mathematical convenience . in a real experiment ,
the time of the first detection is random .
however , for periodic systems ( @xmath37 ) , one can construct an average wtd that only depends on the time delay @xmath0 .
such a quantity is constructed from the time integration of the wtd over a period with weight @xmath34 . using this definition and ( [ wtaumethod2_2 ] ) yields @xmath38 with @xmath39 , @xmath40 and @xmath41 the mean waiting time .
this is the exact analog of the formula for a stationary process @xcite and the same formula as the one proposed by dasenbrook _
_ @xcite in a recent related work . for pedagogical reasons ,
we will start to explain the calculations in the case @xmath42 before describing the physics of a periodic state made of @xmath37 electrons .
the normalized wave - function for the two electrons is simply @xmath43 with @xmath44)^{1/2}. \end{split}\ ] ] the detector is located at @xmath45 , in a very small interval .
suppose we measure an electron at @xmath45 , at time @xmath26 .
the new wave function is obtained by acting the operator @xmath46 , where @xmath47 is a very small time , much smaller than both @xmath8 , the interval between the pulses and @xmath48 .
after the measurement , the packet is reduced , so that the wave - function of one electron which has been detected is now @xmath49 , where @xmath49 is very peaked around the detector and almost zero everywhere else .
we can take @xmath50 , so that @xmath51 . after a few algebraic manipulations , the many - body wave function after the measurement @xmath52 assumes the form @xmath53 therefore , the probability of detecting nothing before time @xmath0 , having detected an electron at time @xmath26 , is @xmath54 and the wtd , through ( [ eq_wtd_p ] ) , reads @xmath55 as a consequence , if the single particle wave functions @xmath16 are differentiable , then @xmath56 vanishes as @xmath57 for small waiting times , in accordance with the pauli principle . a more expanded discussion of this result is given in sec .
iii . in order to mimic an infinite train of lorentzian pulses ,
we generalize the previous calculation to an arbitrary number of electrons @xmath7 .
we then analyze the asymptotic properties of the wtd for large @xmath7 .
two different physical situations will be treated separately for mathematical convenience . in sec .
ii b 1 , we consider the situation where the single particle wave functions weakly overlap in real space .
such a situation is more easily tackled in the basis of wave - functions @xmath16 in real space . however , for large overlap , the matrices that appear in the calculation have very small determinants and are thus very ill - conditioned for numerical calculations . therefore , it is more convenient mathematically to use another basis . such a basis is constructed from the fourier transforms of the original localized wave packets @xcite as explained in sec .
ii b 2 .
we derive here a general formula for the wtd of a train of @xmath7 electrons in terms of the ratio of different determinants .
this formula is general but not convenient in the limit of large overlap between the single particle wave functions , as we will discuss later .
the first step of the derivation consists in computing the many - body wave function after the measurement of the first particle . before the detection ,
the system is in state @xmath58 , with @xmath59 the matrix of elements @xmath60 and @xmath61 is the overlap matrix @xmath62 .
immediately after the measurement , the state collapses through the application of a projector onto the state @xmath63 , i.e. one electron is now confined in the detector . after normalization and a little algebra , the new wave function can be cast as @xmath64 where @xmath65 is the same matrix as @xmath66 but the first line @xmath67 has been replaced by @xmath68 .
@xmath69 is the same matrix as @xmath61 except that the first line has been replaced by @xmath70 , @xmath71 to @xmath7 .
@xmath72 is the operator @xmath73 , namely the projector on the detector .
the probability of not detecting anything before @xmath0 is thus @xmath74 where the subscript @xmath75 means operation on coordinate @xmath76 .
detailed calculations can be found in appendix a. expanding the wave function @xmath77 , which is a determinant , as a sum over permutations , we arrive at @xmath78 where @xmath79 and @xmath80 are two matrices defined as follows .
@xmath81 is the matrix @xmath82 .
@xmath79 is obtained from @xmath81 by substitution of the @xmath83 line by the line vector @xmath84 .
@xmath85 is the overlap matrix @xmath86 and @xmath80 is again obtained from @xmath85 by substituting the @xmath87line by @xmath88 .
equation [ generalformula ] is the central result of this section and is equivalent to the determinant formula derived in @xcite from a second - quantized formulation . for large overlaps , in order to perform efficient and reliable numerical simulations , we need another line of attack .
this is the reason why we adapt in this case the more traditional methods used previously for qpc @xcite .
it has also the benefit of providing other methods for analytical calculations , as explained in the following .
the central quantity is the idle time probability @xmath30 .
it is the probability of not detecting anything in the time interval @xmath89 , irrespective of what happens at time @xmath26 .
we stress that @xmath30 is a different quantity from @xmath28 , as explained before .
when the @xmath7 electrons are in states @xmath90 , the formula giving @xmath30 reads @xmath91 where @xmath92 is @xmath93 .
however , as it stands , in real space basis , @xmath94 is very small and leads to numerical problems for large overlaps @xmath2 .
we thus switch to fourier representation and define @xmath95 with @xmath96 $ ] . for large @xmath7 and for @xmath20 in the bulk of the train , the @xmath97 s assume the form @xmath98 where @xmath99 is the heaviside function .
since the @xmath97 s are delocalized , within the span @xmath100 of the whole train , supposed to be the largest distance ( much larger than @xmath3 ) , the @xmath97 s are not normalized .
thus we defined normalized functions @xmath101 s by @xmath102 .
the @xmath103 s have essentially the same form as the @xmath104 s , see eq .
( [ formulaphik ] ) , except that the real exponential factor @xmath105 is now absent . in full analogy to eq .
( [ formulapitau2 ] ) , we have the formula @xmath106 where the matrices @xmath107 and @xmath108 are the exact analogs of @xmath109 and @xmath110 previously defined ; the @xmath111 s just need to be replaced by the @xmath101 s . in the limit of large @xmath7 ( only ) , the elements of the matrices @xmath107 and @xmath108 depend solely on @xmath112 , they are toeplitz matrices .
this leads to great simplifications , both analytically and numerically .
although mathematically trivial , it is instructive to discuss the behaviors of @xmath24 for two pulses in limiting cases .
we focus here on the results at perfect transmission , probing only the fluctuations encoded in the many - body state . from now
, we assume that the first electron has been detected at @xmath113 at time @xmath114 .
the position of the center of the first and the second wave - packet are @xmath115 and @xmath116 respectively .
using ( [ spwf ] ) and ( [ wtd2e1 ] ) yields the following compact formula @xmath117 where @xmath118 ( @xmath119 ) , @xmath120 , @xmath121 and @xmath122 these formulas are not particularly illuminating , thus , we consider two limiting cases . in the weak overlap case , @xmath123 , and for small times
we get @xmath124 which indeed vanishes as @xmath57 , as expected from the pauli principle .
in contrast , for large times , we find @xmath125 which decays algebraically with @xmath0 .
this is a direct consequence of the shape of the wave function of the emitted electrons .
since there are only two electrons , the last one is not correlated with any other following electron and keeps its lorentzian tail .
physically , it means that the fluctuations of the waiting time can be arbitrary large and are not characterized by a finite second moment . such a behavior is not possible in the bulk of a train of @xmath37 electrons . indeed
, all the electrons have to be contained in a time window of order @xmath100 which prevents large waiting times fluctuations .
in addition we find that the maximum of @xmath126 occurs for a time of the order @xmath8 . in the opposite case of strongly overlapping pulses , @xmath127
, we get @xmath128 again , @xmath56 starts as @xmath57 for small times and decays as @xmath129 at large times . however , now , the only relevant time is @xmath48 , which is related to the span of wave - packets .
time @xmath8 has completely scaled out of the problem .
we now discuss the results obtained for an infinite train of electrons by looking at the large @xmath7 limit .
this section starts with the results for weakly overlapping wave packets where the wtd presents a very clear internal structure made from periodic peaks every @xmath8 . as the overlap
is increased , such strong solid - like correlations smoothly disappear until becoming small wiggles as expected for free fermions propagating in a single quantum channel @xcite .
we recall that the overlap parameter is defined as @xmath130 . for @xmath131 ( solid red line ) , and @xmath132 ( dashed - dotted black line ) . @xmath133 and the number of pulses is always @xmath134 .
inset : enlargement of the small time region compared to quadratic fitted functions ( thin solid lines ) to show the @xmath57 dependence.[fig_wtd1 ] ] in this regime , the train of particles resembles a series of well separated thin lorentzian wave packets .
these wave packets are almost independent and periodically distributed in time .
we therefore expect to observe well defined peaks in the wtd , reflecting the shape of a single particle wave function . at perfect transmission , there must be ( approximately ) only one peak centered around @xmath8 .
however , for finite transmission the wtd exhibits a double structure made of an envelope function containing several similar peaks . roughly speaking , these peaks are the hallmark of the fluctuations encoded in the quantum state
whereas the envelope is coming from the random scattering through the quantum point contact .
such a situation is similar to what has been observed for the wtd distribution of a single electron source in the phase noise regime @xcite .
we perform numerical calculations using eq .
( [ generalformula ] ) which is more convenient in this regime .
figure [ fig_wtd1 ] shows @xmath135 for two values of @xmath2 and confirms our expectations .
moreover , the small time behavior is shown to correspond to the @xmath57 prediction imposed by pauli principle .
such a behavior is indeed observed for times shorter than @xmath136 .
note that we show @xmath56 for @xmath137 but the results are qualitatively similar for any value of this parameter .
we now show how to obtain an approximant of the wtd in the limit of extremely small overlap and long waiting times . in that case ,
( [ generalformula ] ) reduces for large times to @xmath138 with @xmath139 @xmath140 and @xmath141 . for vanishing overlap
we throw out the non diagonal elements because only @xmath142 is non negligible .
such a naive computation gives ( see appendix b for details ) @xmath143 where `` @xmath144 '' denotes the integer part .
this result is physically illuminating and allows us to extract the decay rate of the wtd . at this level of approximation
, the wtd reduces to a periodic series of peaks contained in an exponentially decaying envelope .
the presence of several peaks is due to the imperfect transmission through the qpc that allows electrons to be reflected .
actually , this is the wtd of a classical binomial process .
the only random process is the rate of success @xmath17 every @xmath8 and the wtd provides only information about the scatterer .
quantum corrections would give a finite width to the peaks as a hallmark of quantum jittering . such information
are encoded in the many - body state and in that case would lead to a lorentzian shape even though we have not proved it explicitly .
however , when @xmath145 we can assume that the electrons are uncorrelated and the shape of the peaks of the wtd just reflects the one of the wave packets .
this yields @xmath146 with @xmath147 which is in very good agreement with numerical calculations for @xmath148 ( see fig . [ fig_trans ] for @xmath149 and @xmath150 ) .
the decay rate extracted numerically for @xmath133 and @xmath131 is @xmath151 , in very good agreement with the prediction @xmath152 .
figure [ fig_decay ] presents a more detailed comparison for different transmissions .
however , as @xmath2 becomes larger , the peaks continue to spread and eventually overlap . as explained in the next subsection , this modifies the correlations between particles and breaks the uncorrelated electrons picture . at perfect transmission , the crude approximation leading to eq .
( [ simpleapproximation ] ) is not valid anymore .
since there is no scattering , a particle is almost surely found every @xmath8 .
therefore , the wtd consists in a single peak centered around @xmath8 plus a very small satellite around @xmath153 as depicted on fig .
[ fig_trans ] .
other satellites exist at @xmath154 , with @xmath155 integer , but have negligible amplitudes .
moreover , fig .
[ fig_trans ] illustrates the agreement with ( [ wtd_ap2 ] ) for @xmath149 and @xmath150 .
however , this agreement is only valid for times around @xmath8 and the asymptotic decay of the wtd is not lorentzian but rather gaussian .
we now turn to the discussion of intermediate and large values of the overlap parameter @xmath121 .
we start with numerical results before giving analytical arguments in the infinite @xmath2 and in the finite but large @xmath2 limits .
we then try to explain how the finiteness of @xmath3 alters the results . in this case
, it is convenient to switch to the formulation in fourier space described in the previous section .
we then use ( [ eq_pi_fourier ] ) and ( [ wtaumethod2_2 ] ) to compute the wtd numerically .
figure [ fig_wtd1 ] shows the result for @xmath132 and @xmath133 .
we clearly see that the correlations are strongly reduced since the peaks melt down to small wiggles .
these small oscillations are still the hallmark of periodic correlations encoded in the periodic train of electrons .
the situation is comparable to a liquid where correlations on the scale of the inter - particle distance are much weaker and decay rapidly .
then , as @xmath2 is tuned to large values , we recover the case @xcite of a single quantum channel subjected to an effective constant bias @xmath156 as pointed out in ref . .
indeed , we already see for @xmath157 and @xmath149 on fig .
[ fig_trans ] that the finite overlap case approaches the infinite case .
however , a closer look at the asymptotic properties shows that for times of the order or larger than @xmath136 the decay rate , although still gaussian , is different . to get a better understanding of this ,
we study analytically the decay rate and the wiggles for infinite and finite @xmath2 using the theory of toeplitz matrices .
with @xmath137 for @xmath150 ( red solid line ) , and @xmath157 ( blue solid line ) .
@xmath149 and the number of pulses is always @xmath158 .
black dashed lines represent ( [ wtd_ap2 ] ) .
inset : enlargement to illustrate the agreement with the wigner - dyson distribution ( [ eq_wd ] ) , ( dashed line ) for @xmath157 and the presence of a second peak in the wtd for @xmath150.[fig_trans ] ] as mentioned above , the limit of infinite @xmath3 is equivalent to a biased single quantum channel filled with free fermions @xcite .
numerical calculations , as well as a few limiting cases such as qpc close to pinch - off , have been presented in ref . .
however , using szeg s theorem , we can evaluate the envelope and the wiggles in the long time limit with good accuracy for @xmath159 .
the case of perfect transmission has to be tackled aside .
we recall that , in this limit , @xmath160 that appears in the determinant formula ( [ eq_pi_fourier ] ) simplifies to @xmath161 moreover , in real space @xmath162 is a lorentzian function decaying on scale @xmath3 , therefore in fourier space , the magnitude of the elements of @xmath107 decays exponentially on scale @xmath163 . for infinite overlap
, i.e. @xmath164 but still @xmath165 , @xmath107 goes to the identity matrix .
it is convenient to introduce the symbols @xmath166 of the matrices @xmath107 defined as @xmath167 with @xmath168 and @xmath169 taken in @xmath170 and analogously , @xmath171 the symbol of @xmath108 . for infinite overlap and setting @xmath172 @xcite , we find @xmath173 if @xmath174 and @xmath175 otherwise , with @xmath176 and @xmath166 is @xmath177 for any @xmath169 . to obtain the long time behavior of @xmath30 using eq .
( [ eq_pi_fourier ] ) , we use szeg s theorem which gives the long time asymptotic behavior of toeplitz determinants .
this reads as @xmath178 \
, \frac{d \theta}{2 \pi}\right\},\ ] ] which yields the main behavior of the wtd for large times .
@xmath179 again , we find that for @xmath180 , the wtd decays exponentially with a rate @xmath181 and a comparison to numerical results on fig .
[ fig_decay ] shows a very good agreement with this prediction .
this is related to the fact that the physics of this system , on a long time scale , is a binomial process .
note that , in this regime , @xmath26 scales out of the problem since the system becomes stationary .
whenever applicable , szeg s theorem not only gives the main exponential behavior but also the next correction .
let @xmath182 be the fourier transform of @xmath183 , at wave vector @xmath184 , for @xmath155 integer between @xmath185 and @xmath186 .
the correction is of the form @xmath187 , whenever the sum converges .
however , in our case , szeg s theorem is not directly applicable , because the sum on @xmath188 diverges logarithmically .
a way around it would be to apply a theorem for matrices having fisher - hartwig singularities @xcite but it yields only results at very large times , in terms of a power law correction to the exponential decay of the envelope , missing the wiggles at intermediate times .
we thus apply an ultraviolet cutoff on @xmath188 , called @xmath189 , which is of order @xmath7 .
if we choose @xmath190 with @xmath191 a numerical constant of order one , we obtain @xmath192 with @xmath193 and @xmath194 . in the limit of very large @xmath7
, @xmath195 can be expressed in terms of the cosine integral and @xmath195 reduces to @xmath196 .
@xmath197 is always a strictly decreasing function of @xmath0 but its second derivative , @xmath198 exhibits , for moderate @xmath199 , weak wiggles with period @xmath8 , due to the cosine integral .
although not shown , this prediction is in good agreement with numerical results in the long time limit . for @xmath149 ,
( but @xmath3 still infinite ) the behavior is known exactly @xcite from an analogy with random matrix theory .
the decay at large times is mainly gaussian with power law and higher order corrections but is well approximated by the wigner distribution @xmath200 ( @xmath201 ) , @xmath132 ( @xmath202 ) and @xmath203 ( @xmath204 ) as a function of @xmath205 in agreement with ( [ wtd_ap1 ] ) and ( [ wtd_ap3 ] ) .
dashed - dotted black line is the poissonian result which states that for small transparency the decay rate is simply @xmath206.[fig_decay ] ] for the sake of simplicity , we only consider the main term in szeg s theorem and look at the decay of the wtd at large times .
following the same method as above , eq .
( [ eq_pi_fourier ] ) translates into a formula similar to eq .
( [ pi_ap ] ) , but we find more convenient to write it as a riemann sum @xmath207 with @xmath208 , taking discrete values .
it is thus necessary to see how the finiteness of @xmath3 alters @xmath171 and @xmath166 .
let us begin with @xmath166 .
the overlap between @xmath101 and @xmath209 is still @xmath177 ( by definition ) for @xmath210 but will no longer be zero for @xmath211 .
however , its magnitude falls to order @xmath212 as soon as @xmath112 is @xmath177 or larger .
it is mainly of order @xmath212 up to @xmath112 of the order @xmath163 . for large values of @xmath112 , it is still smaller and we shall neglect it .
thus , @xmath213 will no longer be @xmath177 for all @xmath169 s but will be slightly modified for @xmath214 smaller than @xmath215 .
the change is of order @xmath212 .
we now turn to @xmath171 . for @xmath112
smaller than @xmath163 , @xmath216 will assume the same form as for @xmath2 infinite .
the reason is that , small @xmath112 mean large distance and the finiteness of @xmath3 just changes the short distance cutoff in real space . in fourier space , small @xmath217
s do not see modifications on small scales . for @xmath169
s larger than @xmath215 , @xmath218 keeps its infinite @xmath3 form , which is @xmath177 for @xmath219 and zero otherwise . only @xmath169 s of absolute value smaller than @xmath215
deserve much more care .
it is natural to single out the @xmath220 s for which @xmath218 and @xmath213 retain their infinite @xmath2 values .
we call theses values @xmath221 and @xmath222 .
as @xmath222 is always one , we omit it .
times much larger than @xmath136 correspond to @xmath223 . in this case , we split the sum in eq .
( [ riemann ] ) into two parts , @xmath224 for @xmath223 , the second sum will make the main contribution and the first one is much smaller .
thus , we expect that the rate of decay for large times will be approximately the same as the one observed for infinite overlap @xmath2 .
this is indeed what is observed numerically .
already for @xmath157 , the rate of decay is very similar to the one observed for infinite @xmath2 .
this argument does not work for @xmath149 because szeg s theorem is not applicable .
numerically , however , we always observe gaussian behavior for perfect transmission at times larger than @xmath136 , whether @xmath2 is finite or not . for times , smaller than @xmath136 , only @xmath225 smaller than @xmath2
are involved and both @xmath226 and @xmath227 do get perturbed by the finiteness of @xmath3 .
we did not attempt to make analytical predictions in this time range , @xmath228 .
we now address the question of measurability of wtd . detecting single electron events in a quantum coherent conductors
is one of the most challenging goal of electron quantum optics nowadays .
most of the single electron sources operate at the ghz range ( namely they inject coherent electrons at a ghz frequency ) while standard electronic technology limits the measurement precision to a few mhz bandwidth .
an obvious solution to circumvent this problem is to lower the emission rate to the khz regime where single electron tunneling events have been shown to be accessible and measured with high accuracy @xcite . in that case
the wtd is fully accessible but coherent effects are washed out and the physics is dominated by interactions ( coulomb blockade regime ) .
an alternative idea to get information in the time domain is to follow the approach developed by hong , ou and mandel in quantum optics @xcite .
such an interference experiment between indistinguishable particles created by two identical sources gives access to the second order coherence @xmath229 which has been measured in two recent ground breaking experiments @xcite .
in fact , the wtd and the second order correlation function have the same short time behavior and for a large class of systems ( systems with renewal properties which is the case here only for weak overlap ) they contain the same amount of information and are related to each other @xcite . for more general systems ,
the wtd is a mixture of all the correlation functions but could be reconstructed from their measurement .
unfortunately , only a few of the first ones are measurable with current technology .
however , high frequency measurements are progressing very fast @xcite and bring some hope to observe electrons one by one in a near future .
for instance , the setup described in @xcite enables us to measure electrons on time scales of fractions of nanoseconds .
although the magnetic field in theses experiments is still too low to reach the single quantum channel limit , they have access to the ghz frequencies needed for wtd measurements .
the last possibility would be to consider alternative systems to electrons where the typical coherence time is much larger .
therefore , coherent effects could be measurable on time scale accessible with present technology .
this is the case with cold atoms where the bosonic wtd and fcs have already been measured in atomic lasers @xcite .
fermionic atoms have been a bit more challenging but many progresses have been done in this direction recently @xcite . in that case
there is , to our knowledge , no available protocol to generate lorentzian pulses yet .
however , continuous sources are already available @xcite and combined with extremely efficient observation tools are very promising setups to investigate wtd .
moreover , such systems present many advantages compared to electronic systems and one of them is the amazing control over interactions and decoherence .
we have presented a theory of wtd for periodic trains of quantized pulses impinging on a qpc , with transmission @xmath17 .
this generalizes wtd for a qpc subjected to a constant voltage @xcite . when the pulses weakly overlap , the wtd exhibits strong oscillations on the scale of the period and the decay , at large times
, is exponential with a decay rate proportional to @xmath230 .
the internal structure reveals correlations encoded in the many - body state whereas the overall envelope is mainly controlled by the scattering matrix of the qpc . as the overlap between wave packets
is increased , the oscillations are reduced to small wiggles . for very large overlaps ,
the wtd is analogous to that for a qpc subjected to a finite voltage @xmath231 but with the period @xmath8 playing the role of @xmath232 . for @xmath17 not close to @xmath177 , regardless of overlap , the envelope of the wtd is like the one that would be obtained for a binomial process .
electrons cross the qpc randomly with probability @xmath17 , every @xmath8 .
this looks surprising since the quantum system seems to behave like a classical one . for perfect transmission
however , quantum correlations are stronger and cause the wtd to show gaussian decay at large times , reminiscent to what happens for random matrix models .
finally , we have briefly explained how the short time behavior of the wtd could be extracted from hong - ou - mandel experiments .
extension of this type of approach to other physical situations such as klein tunneling , dynamical coulomb blockade or to the case of several channels could be envisaged .
we thank d. dasenbrook , p. degiovanni , g. fve , c. flindt , g. haack , t. jonckheere , m. moskalets and k. h. thomas for useful discussions and remarks .
in addition , we thank c. flindt for letting us know about a related work @xcite .
in this section , we derive eq .
( [ generalformula ] ) of the main text , giving the idle time probability for arbitrary @xmath7 . before detecting an electron at @xmath114 ,
the system is in state @xmath233 summation is over all permutations @xmath234 of @xmath235 and @xmath236 is the signature of @xmath234 .
applying @xmath237 , @xmath238 the wave function after measurement can be recast as a determinant @xmath239 the probability of measuring nothing before @xmath0 is @xmath240 with @xmath241 for @xmath242 and @xmath243 , for @xmath244 . for @xmath244
, we measure an electron and for @xmath242 , we measure no electron . as usual , the composition of @xmath234 and @xmath245 has to be considered but , contrary to what usually happens , the operator @xmath246 is not the same for all @xmath247 .
the case @xmath244 is different from the other @xmath247 s .
we set @xmath248 where `` @xmath249 '' means composition of applications and @xmath250 the inverse of @xmath234 .
@xmath251 among all permutations @xmath252 of @xmath235 , only @xmath253 will give @xmath254 .
finally , @xmath255 where @xmath256 means @xmath257 everywhere except for @xmath258 , where @xmath257 has to be replaced by @xmath72 . as an example , here is the term for @xmath259 , @xmath260 the term for @xmath261 is , @xmath262 normalization of @xmath30 requires that one has to divide by the same sum of determinants , but now , the @xmath263 have to be replaced by @xmath177 .
this gives eq .
( [ generalformula ] ) of the main text .
in this section , we justify why , in the case of small overlap , the approximation by diagonal matrices works so well . if we forget the logarithmic terms , for @xmath264 , @xmath265 becomes @xmath266 with @xmath267 and @xmath268 .
thus , in the limit @xmath269 , @xmath270 factorizes as @xmath271 , where @xmath272 is the same as @xmath273 except that all elements with row _ or _ column index smaller than @xmath274 have to be replaced by the elements of the identity matrix .
so far , we have neglected the logarithmic terms .
they vanish on the diagonal .
they also vanish for @xmath275 or @xmath276 much smaller than @xmath199 and also when @xmath275 and @xmath276 much larger than @xmath199 . neglecting them should amount to neglecting some correlations and the decay rate including them may be smaller than the one given in the diagonal approximation , but we do not have clear evidence for that .
calculations of the first non - zero order in @xmath2 would require retaining all the terms in eq .
( [ generalformula ] ) of the main text .
j. gabelli , g. fve , j .-
berroir , b. plaais , a. cavanna , b. etienne , y. jin , and d. c. glattli , science * 313 * , 499 ( 2006 ) . g. fve , a. mah , j .-
berroir , t. kontos , b. plaais , d. c. glattli , a. cavanna , b. etienne , and y. jin , science * 316 * , 1169 ( 2007 ) . m. bttiker , h. thomas , and a. prtre , phys .
a * 180 * , 364 ( 1993 ) .
m. d. blumenthal , b. kaestner , l. li , s. giblin , t. j. b. m. janssen , m. pepper , d. anderson , g. jones , and d. a. ritchie , nat .
phys . * 3 * , 343 ( 2007 ) .
j. p. pekola , j. j. vartiainen , m. mttnen , o .- p .
saira , m. meschke , and d. v. averin , nat . phys . * 4 * , 120 ( 2008 ) .
f. giazotto , p. spathis , s. roddaro , s. biswas , f. taddei , m. governale , and l. sorba , nat . phys .
* 7 * , 857 ( 2011 ) . c. leicht , p. mirovsky , b. kaestner , f. hohls , v. kashcheyevs , e. v. kurganova , u. zeitler , t. weimann , and k. pierz , and h. w. schumacher , semicond .
* 26 * , 055010 ( 2011 ) .
s. hermelin , s. takada , m. yamamoto , s. tarucha , a. d. wieck , l. saminadayar , c. buerle , and t. meunier , nature ( london ) * 477 * , 435 ( 2011 ) .
r. p. g. mcneil , m. kataoka , c. j. b. ford , c. h. w. barnes , d. anderson , g. a. c. jones , i. farrer , and d. a. ritchie , nature ( london ) * 477 * , 439 ( 2011 ) .
g. c. tettamanzi , r. wacquez , and s. rogge , arxiv:1401.3080 ( 2014 ) .
l. s. levitov , h. lee , and g. b. lesovik , j. math . phys . *
37 * , 4845 ( 1996 ) .
d. a. ivanov , h. lee , and l. s. levitov , phys .
b * 56 * , 6839 ( 1997 ) .
j. keeling , i. klich , and l. levitov , phys . rev
. lett . * 97 * , 116403 ( 2006 ) .
j. dubois , t. jullien , f. portier , p. roche , a. cavanna , y. jin , w. wegscheider , p. roulleau , and d. c. glattli , nature * 502 * , 659 ( 2013 ) .
j. splettstoesser , m. moskalets , and m. bttiker , phys .
lett . * 103 * , 076804 ( 2009 ) .
s. olkhovskaya , j. splettstoesser , m. moskalets , and m. bttiker , phys .
* 101 * , 166802 ( 2008 ) .
s. juergens , j. splettstoesser , and m. moskalets , europhys .
lett . * 96 * , 37011 ( 2011 ) .
t. jonckheere , j. rech , c. wahl , and t. martin , phys .
b * 86 * , 125425 ( 2012 ) .
e. bocquillon , v. freulon , j .-
m berroir , p. degiovanni , b. plaais , a. cavanna , y. jin , and g. fve , science * 339 * , 1169 ( 2013 ) .
e. bocquillon , f. d. parmentier , c. grenier , j .-
berroir , p. degiovanni , d. c. glattli , b. plaais , a. cavanna , y. jin , and g. fve , phys .
lett . * 108 * , 196803 ( 2012 ) .
e. bocquillon , v. freulon , j .-
berroir , p. degiovanni , b. plaais , a. cavanna , y. jin , and g. fve , nature communications * 4 * , 1839 ( 2013 ) .
p. degiovanni , c. grenier , and g. fve , phys .
b * 80 * , 241307(r ) ( 2009 ) .
g. haack , m. moskalets , j. splettstoesser , and m. bttiker , phys .
b * 84 * , 081303(r ) ( 2011 ) . ya
. m. blanter and m. bttiker , phys
* 336 * , 1 ( 2000 ) .
l. s. levitov and g. b. lesovik , jetp lett . * 58 * , 230 ( 1993 ) , l. s. levitov , h .- w .
lee , and g. b. lesovik , j. math . phys . *
37 * , 4845 ( 1996 ) .
v. nazarov , ed . , _ quantum noise in mesoscopic physics _
( nato science series , kluwer , dordrecht , 2003 ) .
a. andreev and a. kamenev , phys .
lett . * 85 * , 1294 ( 2000 ) .
makhlin and a. d. mirlin , phys .
* 87 * , 276803 ( 2001 ) .
m. albert , c. flindt , and m. bttiker , phys .
b * 82 * , 041407 ( 2010 ) . j. basset , h. bouchiat , and r. deblock , phys . rev .
b * 85 * , 085435 ( 2012 ) .
a. mah , f. d. parmentier , e. bocquillon , j .-
berroir , d. c. glattli , t. kontos , b. plaais , g. fve , a. cavanna , and y. jin , phys .
b * 82 * , 201309(r ) ( 2010 ) .
f. d. parmentier , e. bocquillon , j .-
berroir , d. c. glattli , b. plaais , g. fve , m. albert , c. flindt , m. bttiker , phys .
b * 85 * , 165438 ( 2012 ) .
r. zamoum , a. crpieux , and i. safi , phys .
b * 85 * , 125421 ( 2012 ) . c. emary , d. marcos , r. aguado , and t. brandes , phys . rev .
b * 76 * , 161404 ( 2007 ) . c. flindt , t. novotn , a. braggio , m. sassetti , and a. p. jauho , phys . rev .
* 100 * , 150601 ( 2008 ) .
d. marcos , c. emary , t. brandes , and r. aguado , new j. phys .
* 12 * , 123009 ( 2010 ) .
d. marcos , c. emary , t. brandes , and r. aguado , phys .
b * 83 * , 125426 ( 2011 ) .
n. ubbelohde _
et al . _ , nat
. commun . * 3 * , 612 ( 2012 ) .
d. ferraro , a. feller , a. ghibaudo , e. thibierge , e. bocquillon , g. fve , c. grenier , and p. degiovanni , phys .
b * 88 * , 205303 ( 2013 ) .
j. schriefl , m. clusel , d. carpentier , and p. degiovanni , phys .
b * 72 * , 035328 ( 2005 ) .
j. koch , m. e. raikh , and f. von oppen , phys .
lett . * 95 * , 056801 ( 2005 ) .
s. welack , m. esposito , u. harbola , and s. mukamel , phys .
b * 77 * , 195315 ( 2008 ) , s. welack , s. mukamel , and y. yan , europhys .
* 85 * , 57008 ( 2009 ) .
t. brandes , ann .
( berlin ) * 17 * , 477 ( 2008 ) .
m. albert , c. flindt , and m. bttiker , phys .
* 107 * , 086805 ( 2011 ) .
m. albert , g. haack , c. flindt and m. bttiker , phys .
lett . * 108 * , 186806 ( 2012 ) .
k. h. thomas and c. flindt , phys .
b * 87 * , 121405(r ) ( 2013 ) .
d. dasenbrook , c. flindt , and m. bttiker , phys .
* 112 * , 146801 ( 2014 ) .
l. rajabi , c. pltl and m. governale , phys .
lett . * 111 * , 067002 ( 2013 ) .
k. h. thomas and c. flindt , phys .
b * 89 * , 245420 ( 2014 ) .
tang , f. xu , j. wang , phys .
b * 89 * , 205310 ( 2014 ) .
r. landauer , physica d * 38 * , 226 ( 1989 ) , t. martin and r. landauer , phys .
b * 45 * , 1742 ( 1992 ) .
f. hassler , m. v. suslov , g. m. graf , m. v. lebedev , g. b. lesovik , and g. blatter , phys .
b * 78 * , 165330 , ( 2008 ) , g. b. lesovik and i. a. sadovskyy , phys .- usp . *
54 * , 1007 ( 2011 ) . in this limit
, one can easily show that @xmath45 and @xmath26 scale out of the problem .
g. abanov , d. a. ivanov , and y. qian , j. phys . a , math .
theor * 44 * , 485001 ( 2011 ) .
m. e. fisher and r. e. hartwig , adv .
phys . , * 15 * , 333 ( 1968 ) .
i. v. krasovsky , int .
notices 2004 , 1249 ( 2004 )
. also arxiv.math 0401258v2 , ( 2004 ) . c. a. tracy and h. widom , in _ geometric and quantum aspects of integrable systems _ , ed
. g. f. helminck , lecture notes in physics , vol .
424 , springer verlag , berlin 1993 , pp .
103 - 130 . c. flindt , c. fricke , f. hohls , t. novotny , k. netocny , t. brandes , and r. j. haug , proc .
usa * 106 * , 10116 ( 2009 ) .
c. k. hong , z. y. ou , and l. mandel , phys .
lett . * 59 * , 2044 ( 1987 ) .
d. r. cox , _ renewal theory _
( chapman and hall , london , 1962 ) .
r. thalineau , a. d. wieck , c. buerle , and t. meunier , arxiv cond / mat 1403.7770 ( 2014 ) .
a. ttl , s. ritter , m. khl , and t. esslinger , phys .
lett . * 95 * , 090404 ( 2005 ) .
m. khl , a. ttl , s. ritter , t. donner , t. bourdel , and t. esslinger , applied physics b * 86 * , 391 ( 2007 ) .
j. h. thywissen , r. m. westervelt , and m. prentiss , phys .
lett . * 83 * 3762 ( 1999 ) .
j. p. brantut , j. meineke , d. stadler , s. krinner , and t. esslinger , science * 337 * , 1069 ( 2012 ) . s. krinner , d. stadler , d. husmann , j. p. brantut , and t. esslinger , arxiv cond / mat 1404.6400 ( 2014 ) . | we consider a sequence of quantized lorentzian pulses of non - interacting electrons impinging on a quantum point contact ( qpc ) and study the waiting time distribution ( wtd ) , for any transmission and any number of pulses . as the degree of overlap between the electronic wave functions
is tuned , the wtd reveals how the correlations between particles are modified . in the weak overlap regime ,
the wtd is made of several equidistant peaks , separated by the same period as the incoming pulses , contained in an almost exponentially decaying envelope . in the other limit
, the wtd of a single quantum channel subjected to a constant voltage is recovered . in both cases ,
the wtd stresses the difference between the fluctuations induced by the scatterer and the ones encoded in the incoming quantum state .
a clear cross - over between these two situations is studied with numerical and analytical calculations based on scattering theory . |
the goal of the development of the code was to have a simple and efficient tool for the computation of adiabatic oscillation frequencies and eigenfunctions for general stellar models , emphasizing also the accuracy of the results .
not surprisingly , given the long development period , the simplicity is now less evident .
however , the code offers considerable flexibility in the choice of integration method as well as ability to determine all frequencies of a given model , in a given range of degree and frequency .
the choice of variables describing the equilibrium model and oscillations was to a large extent inspired by @xcite .
as discussed in section [ sec : eqmodel ] the equilibrium model is defined in terms of a minimal set of dimensionless variables , as well as by mass and radius of the model .
fairly extensive documentation of the code , on which the present paper in part is based , is provided with the distribution packagejcd / adipack.n ] .
@xcite provided an extensive review of adiabatic stellar oscillations , emphasizing applications to helioseismology , and discussed many aspects and tests of the aarhus package , whereas @xcite carried out careful tests and comparisons of results on polytropic models ; this includes extensive tables of frequencies which can be used for comparison with other codes .
the equilibrium model is defined in terms of the following dimensionless variables : @xmath0 here @xmath1 is distance to the centre , @xmath2 is the mass interior to @xmath1 , @xmath3 is the photospheric radius of the model and @xmath4 is its mass ; also , @xmath5 is the gravitational constant , @xmath6 is pressure , @xmath7 is density , and @xmath8 , the derivative being at constant specific entropy .
in addition , the model file defines @xmath4 and @xmath3 , as well as central pressure and density , in dimensional units , and scaled second derivatives of @xmath6 and @xmath7 at the centre ( required from the expansions in the central boundary condition ) ; finally , for models with vanishing surface pressure , assuming a polytropic relation between @xmath6 and @xmath7 in the near - surface region , the polytropic index is specified .
the following relations between the variables defined here and more `` physical '' variables are often useful : @xmath9 we may also express the characteristic frequencies for adiabatic oscillations in terms of these variables . thus if @xmath10 is the buoyancy frequency , @xmath11 is the lamb frequency at degree @xmath12 and @xmath13 is the acoustical cut - off frequency for an isothermal atmosphere , we have @xmath14 where @xmath15 is the adiabatic sound speed , and @xmath16 is the pressure scale height , @xmath17 being the gravitational acceleration .
finally it may be noted that the squared sound speed is given by @xmath18 these equations also define the dimensionless characteristic frequencies @xmath19 , @xmath20 and @xmath21 as well as the dimensionless sound speed @xmath22 , which are often useful . as is well known the displacement vector of nonradial ( spheroidal ) modes
can be written in terms of polar coordinates @xmath23 as @xmath24 \exp ( - { { \rm i}}\omega t ) \right\ } \ ; . \nonumber\end{aligned}\ ] ] here @xmath25 is a spherical harmonic of degree @xmath12 and azimuthal order @xmath2 , @xmath26 being co - latitude and @xmath27 longitude ; @xmath28 is an associated legendre function , and @xmath29 is a suitable normalization constant .
also , @xmath30 , @xmath31 , and @xmath32 are unit vectors in the @xmath1 , @xmath26 , and @xmath27 directions . finally , @xmath33 is time and @xmath34 is the angular frequency of the mode .
similarly , e.g. , the eulerian perturbation to pressure may be written @xmath35 \ ; . \label{eq : e2.2}\ ] ] as the oscillations are adiabatic ( and only conservative boundary conditions are considered ) @xmath34 is real , and the amplitude functions @xmath36 ,
@xmath37 , @xmath38 , etc . can be chosen to be real .
the equations of adiabatic stellar oscillations , in the nonradial case , are expressed in terms of the following variables : , @xmath39 results from the earlier use of an unconventional sign convention for @xmath40 ; now , as usual , @xmath40 is defined such that the perturbed poisson equation has the form @xmath41 , where @xmath42 is the eulerian density perturbation . ]
@xmath43 here @xmath40 is the perturbation to the gravitational potential . also , we introduce the dimensionless frequency @xmath44 by @xmath45 corresponding to eqs [ eq : buoy ] [ eq : cutoff ] . these quantities satisfy the following equations : @xmath46 y_1 + ( a - 1 ) y_2 + \eta a y_3 \ ; , \\ \label{eq : ea.3 } x { { { \rm d}}y_3 \over { { \rm d}}x } & = & y_3 + y_4 \ ; , \\ \label{eq : ea.4 } x { { { \rm d}}y_4 \over { { \rm d}}x } & = & - a u y_1 - u { v_g \over \eta } y_2 \\ & & + [ l ( l + 1 ) + u(a - 2 ) + u v_g ] y_3 + 2(1 - u ) y_4 \ ; . \nonumber\end{aligned}\ ] ] here @xmath47 , and the notation is otherwise as defined in eq .
[ eq : fivea ] . in the @xcite approximation , where the perturbation to the gravitational potential is neglected ,
the terms in @xmath48 are neglected in eqs [ eq : ea.1 ] and [ eq : ea.2 ] and eqs [ eq : ea.3 ] and [ eq : ea.4 ] are not used .
the dependent variables @xmath49 in the nonradial case have been chosen in such a way that for @xmath50 they all vary as @xmath51 for @xmath52 . for large @xmath12 a considerable ( and fundamentally unnecessary ) computational effort
would be needed to represent this variation sufficiently accurately with , e.g. , a finite difference technique , if these variables were to be used in the numerical integration .
instead i introduce a new set of dependent variables by @xmath53 these variables are then @xmath54 in @xmath55 near the centre .
they are used in the region where the variation in the @xmath56 is dominated by the @xmath51 behaviour , for @xmath57 , say , where @xmath58 is determined on the basis of the asymptotic properties of the solution . this transformation permits calculating modes of arbitrarily high degree in a complete model . for radial oscillations only @xmath59 and @xmath60
are used , where @xmath59 is defined as above , and @xmath61 here the equations become @xmath62 y_1 + a y_2 \ ; . \label{eq : ea.6}\end{aligned}\ ] ] the equations are solved on the interval @xmath63 $ ] in @xmath55 . here ,
in the most common case involving a complete stellar model @xmath64 , where @xmath65 is a suitably small number such that the series expansion around @xmath66 is sufficiently accurate ; however , the code can also deal with envelope models with arbitrary @xmath67 , typically imposing @xmath68 at the bottom of the envelope .
the outermost point is defined by @xmath69 where @xmath70 is the surface radius , including the atmosphere ; thus , typically , @xmath71 .
the centre of the star , @xmath72 , is obviously a singular point of the equations .
as discussed , e.g. , by @xcite boundary conditions at this point are obtained from a series expansion , in the present code to second significant order . in the general case
this defines two conditions at the innermost non - zero point in the model . for radial oscillations , or nonradial oscillations in the cowling approximation
, one condition is obtained .
the surface in a realistic model is typically defined at a suitable point in the stellar atmosphere , with non - zero pressure and density . here
the simple condition of vanishing lagrangian pressure perturbation is implemented and sometimes used .
however , more commonly a condition between pressure perturbation and displacement is established by matching continuously to the solution in an isothermal atmosphere extending continuously from the uppermost point in the model .
a very similar condition was presented by @xcite . in addition , in the full nonradial case a condition is obtained from the continuous match of @xmath40 and its derivative to the vacuum solution outside the star . in full polytropic models , or other models with vanishing surface pressure , the surface is also a singular point . in this case
a boundary condition at the outermost non - singular point is obtained from a series expansion , assuming a near - surface polytropic behaviour ( see * ? ? ? * for details ) .
the code also has the option of considering truncated ( e.g. , envelope ) models although at the moment only in the cowling approximation or for radial oscillations . in this case
the innermost boundary condition is typically the vanishing of the radial displacement @xmath73 although other options are available .
the numerical problem can be formulated generally as that of solving @xmath74 with the boundary conditions @xmath75 @xmath76 here the order @xmath77 of the system is 4 for the full nonradial case , and 2 for radial oscillations or nonradial oscillations in the cowling approximation .
this system only allows non - trivial solutions for selected values of @xmath78 which is thus an eigenvalue of the problem .
the programme permits solving these equations with two basically different techniques , each with some variants .
the first is a shooting method , where solutions satisfying the boundary conditions are integrated separately from the inner and outer boundary , and the eigenvalue is found by matching these solutions at a suitable inner fitting point @xmath79 .
the second technique is to solve the equations together with a normalization condition and all boundary conditions using a relaxation technique ; the eigenvalue is then found by requiring continuity of one of the eigenfunctions at an interior matching point .
for simplicity i do not distinguish between @xmath80 and @xmath56 ( cf . section [ sec : eq ] ) in this section .
it is implicitly understood that the dependent variable ( which is denoted @xmath56 ) is @xmath80 for @xmath57 and @xmath56 for @xmath81 .
the numerical treatment of the transition between @xmath80 and @xmath56 has required a little care in the coding .
it is convenient here to distinguish between @xmath77 = 2 and @xmath77 = 4 . for @xmath77 = 2 the differential eqs [ eq : e3.1 ]
have a unique ( apart from normalization ) solution @xmath82 satisfying the inner boundary conditions [ eq : e3.2 ] , and a unique solution @xmath83 satisfying the outer boundary conditions [ eq : e3.3 ] .
these are obtained by numerical integration of the equations .
the final solution can then be represented as @xmath84 .
the eigenvalue is obtained by requiring that the solutions agree at a suitable matching point @xmath85 , say .
thus @xmath86 these equations clearly have a non - trivial solution @xmath87 only when their determinant vanishes , i.e. , when @xmath88 equation [ eq : e3.5 ] is therefore the eigenvalue equation .
for @xmath77 = 4 there are two linearly independent solutions satisfying the inner boundary conditions , and two linearly independent solutions satisfying the outer boundary conditions .
the former set @xmath89 is chosen by setting @xmath90 and the latter set @xmath91 is chosen by setting @xmath92 the inner and outer boundary conditions are such that , given @xmath59 and @xmath48 , @xmath60 and @xmath39 may be calculated from them ; thus eqs [ eq : e3.6 ] and [ eq : e3.7 ] completely specify the solutions , which are obtained by integrating from the inner or outer boundary .
the final solution can then be represented as @xmath93 at the fitting point @xmath79 continuity of the solution requires that @xmath94 this set of equations only has a non - trivial solution if @xmath95 where , e.g. , @xmath96 .
thus eq .
[ eq : e3.9 ] is the eigenvalue equation in this case .
clearly @xmath97 as defined in either eq .
[ eq : e3.5 ] or eq .
[ eq : e3.9 ] is a smooth function of @xmath78 , and the eigenfrequencies are found as the zeros of this function .
this is done in the programme using a standard secant technique .
however , the programme also has the option for scanning through a given interval in @xmath78 to look for changes of sign of @xmath97 , possibly iterating for the eigenfrequency at each change of sign .
thus it is possible to search a given region of the spectrum completely automatically .
the programme allows the use of two different techniques for solving the differential equations .
one is the standard second - order centred difference technique , where the differential equations are replaced by the difference equations @xmath98
, \quad i = 1 , \ldots , i \ ; .
\label{eq : e3.11}\ ] ] here i have introduced a mesh @xmath99 in @xmath55 , where @xmath10 is the total number of mesh points ; @xmath100 , and @xmath101 .
these equations allow the solution at @xmath102 to be determined from the solution at @xmath103 .
the second technique was proposed by @xcite ; here on each mesh interval @xmath104 we consider the equations @xmath105 with constant coefficients , where @xmath106 .
these equations may be solved analytically on the mesh intervals , and the complete solution is obtained by continuous matching at the mesh points .
this technique clearly permits the computation of solutions varying arbitrarily rapidly , i.e. , the calculation of modes of arbitrarily high order . on the other hand solving eqs [ eq : e3.12 ] involves finding the eigenvalues and eigenvectors of the coefficient matrix , and therefore becomes very complex and time consuming for higher - order systems .
thus in practice it has only been implemented for systems of order 2 , i.e. , radial oscillations or nonradial oscillations in the cowling approximation .
if one of the boundary conditions is dropped , the difference equations , with the remaining boundary condition and a normalization condition , constitute a set of linear equations for the @xmath107 which can be solved for any value of @xmath44 ; this set may be solved efficiently by forward elimination and backsubstitution ( e.g. , * ? ? ?
* ) , with a procedure very similar to the so - called henyey technique ( e.g. , * ? ?
* see also christensen - dalsgaard 2007 ) used in stellar modelling .
the eigenvalue is then found by requiring that the remaining boundary condition , which effectively takes the role of @xmath108 , be satisfied .
however , as both boundaries , at least in a complete model , are either singular or very nearly singular , the removal of one of the boundary conditions tends to produce solutions that are somewhat ill - behaved , in particular for modes of high degree .
this in turn is reflected in the behaviour of @xmath97 as a function of @xmath44 .
this problem is avoided in a variant of the relaxation technique where the difference equations are solved separately for @xmath109 and @xmath110 , by introducing a double point @xmath111 in the mesh . the solution is furthermore required to satisfy the boundary conditions [ eq : e3.2 ] and [ eq : e3.3 ] , a suitable normalization condition ( e.g. @xmath112 ) , and continuity of all but one of the variables at @xmath85 , e.g. , @xmath113 ( when @xmath77 = 2 clearly only the first continuity condition is used ) we then set @xmath114 and the eigenvalues are found as the zeros of @xmath97 , regarded as a function of @xmath78 . with this definition
, @xmath97 may have singularities with discontinuous sign changes that are not associated with an eigenvalue , and hence a little care is required in the search for eigenvalues .
however , close to an eigenvalue @xmath97 is generally well - behaved , and the secant iteration may be used without problems . as implemented here
the shooting technique is considerably faster than the relaxation technique , and so should be used whenever possible ( notice that both techniques may use the difference eqs [ eq : e3.11 ] and so they are numerically equivalent , in regions of the spectrum where they both work ) . for _ second - order systems _ the shooting technique can probably always be used ; the integrations of the inner and outer solutions should cause no problems , and the matching determinant in eq .
[ eq : e3.5 ] is well - behaved . for _ fourth - order systems _
, however , this needs not be the case . for modes where the perturbation to the gravitational potential has little effect on the solution , the two solutions @xmath115 and @xmath116 , and similarly the two solutions @xmath117 and @xmath118 , are almost linearly dependent , and so the matching determinant nearly vanishes for any value of @xmath78 .
this is therefore the situation where the relaxation technique may be used with advantage .
this applies , in particular , to the calculation of modes of moderate and high degree which are essential to helioseismology . to make full use of the increasingly accurate observed frequencies the computed frequencies should clearly at the very least match the observational accuracy , for a given model . only in this way
do the frequencies provide a faithful representation of the properties of the model , in comparisons with the observations
. however , since the numerical errors in the computed frequencies are typically highly systematic , they may affect the asteroseismic inferences even if they are smaller than the random errors in the observations , and hence more stringent requirements should be imposed on the computations .
also , the fact that solar - like oscillations , and several other types of asteroseismically interesting modes , tend to be of high radial order complicates reaching the required precision .
the numerical techniques discussed so far are generally of second order .
this yields insufficient precision in the evaluation of the eigenfrequencies , unless a very dense mesh is used in the computation ( see also * ? ? ?
the code may apply two techniques to improve the precision .
one technique ( cf .
* ) uses the fact that the frequency approximately satisfies a variational principle @xcite .
the variational expression may formally be written as @xmath119 where @xmath120 and @xmath121 are integrals over the equilibrium model depending on the eigenfunction , here represented by @xmath122 .
the variational property implies that any error @xmath123 in @xmath122 induces an error in @xmath124 that is @xmath125 .
thus by substituting the computed eigenfunction into the variational expression a more precise determination of @xmath78 should result .
this has indeed been confirmed @xcite .
the second technique uses explicitly that the difference scheme [ eq : e3.11 ] , which is used by one version of the shooting technique , and the relaxation technique , is of second order . consequently the truncation errors in the eigenfrequency and eigenfunction scale as @xmath126 .
if @xmath127 and @xmath128 are the eigenfrequencies obtained from solutions with @xmath129 and @xmath10 meshpoints , the leading - order error term therefore cancels in @xmath130 \ ; . \label{eq : e3.18}\ ] ] this procedure , known as _ richardson extrapolation _ , was used by @xcite .
it provides an estimate of the eigenfrequency that is substantially more accurate than @xmath131 , although of course at some added computational expense .
indeed , since the error in the representation [ eq : e3.11 ] depends only on even powers of @xmath132 , the leading term of the error in @xmath133 is @xmath134 . even with these techniques
the precision of the computed frequencies may be inadequate if the mesh used in stellar - evolution calculations is used also for the computation of the oscillations .
the number of meshpoints is typically relatively modest and the distribution may not reflect the requirement to resolve properly the eigenfunctions of the modes .
@xcite discussed techniques to redistribute the mesh in a way that takes into account the asymptotic behaviour of the eigenfunctions ; a code to do so , based on four - point lagrangian interpolation , is included in the adipls distribution package .
on the other hand , for computing low - order modes ( as are typically relevant for , say , @xmath135 scuti or @xmath136 cephei stars ) , the original mesh of the evolution calculation may be adequate .
it is difficult to provide general recommendations concerning the required number of points or the need for redistribution , since this depends strongly on the types of modes and the properties of the stellar model .
it is recommended to carry out experiments varying the number and distribution of points to obtain estimates of the intrinsic precision of the computation ( e.g. , * ? ? ?
* ; * ? ? ?
in the latter case , considering simple polytropic models , it was found that 4801 points yielded a relative precision substantially better than @xmath137 for high - order p modes , when richardson extrapolation was used . in the discussion of the frequency calculation it is important to distinguish between _ precision _ and _ accuracy _ ,
the latter obviously referring to the extent to which the computed frequencies represent what might be considered the ` true ' frequencies of the model .
in particular , the manipulations required to derive eq . [ eq : varprinc ] and to demonstrate its variational property depend on the equation of hydrostatic support being satisfied .
if this is not the case , as might well happen in an insufficiently careful stellar model calculation , the value determined from the variational principle may be quite precise , in the sense of numerically stable , but still unacceptably far from the correct value .
indeed , a comparison between @xmath138 and @xmath133 provides some measure of the reliability of the computed frequencies ( e.g. * ? ? ?
the programme finds the order of the mode according to the definition developed by @xcite and @xcite , based on earlier work by @xcite .
specifically , the order is defined by @xmath139 here the sum is over the zeros @xmath140 in @xmath59 ( excluding the centre ) , and @xmath141 is the sign function , @xmath142 if @xmath143 and @xmath144 if @xmath145 . for a complete model that includes the centre @xmath146 for radial oscillations and @xmath147 for nonradial oscillations .
thus the lowest - order radial oscillation has order @xmath148 .
although this is contrary to the commonly used convention of assigning order 0 to the fundamental radial oscillation , the convention used here is in fact the more reasonable , in the sense that it ensures that @xmath149 is invariant under a continuous variation of @xmath12 from 0 to 1 . with this definition @xmath150 for p modes , @xmath151 for f modes , and @xmath152 for g modes , at least in simple models .
it has been found that this procedure has serious problems for dipolar modes in centrally condensed models ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the eigenfunctions @xmath59 are shifted such that nodes disappear or otherwise provide spurious results when eq .
[ eq : e4.1 ] is used to determine the mode order . a procedure that does not suffer from this difficulty has recently been developed by @xcite
; i discuss it further in section [ sec : develop ] . a powerful measure of the characteristics of a mode is provided by the _ normalized inertia_. the code calculates this as @xmath153 \rho r^2 { { \rm d}}r \over m [ \xi_r ( r_{\rm phot } ) ^2 + l(l+1 ) { \xi_{\rm h}}(r_{\rm phot } ) ^2 ] } \nonumber \\ & = & { \int_{x_1}^{{x_{\rm s } } } \left [ y_1 ^ 2 + y_2 ^ 2 / l ( l + 1 ) \right ] q u { { \rm d}}x / x \over 4 \pi [ y_1 ( x_{\rm phot } ) ^2 + y_2 ( x_{\rm phot } ) ^2/l(l+1 ) ] } \ ; .\end{aligned}\ ] ] ( for radial modes the terms in @xmath60 are not included . ) here @xmath154 and @xmath155 are the distance of the innermost mesh point from the centre and the surface radius , respectively , and @xmath156 is the fractional photospheric radius .
the normalization at the photosphere is to some extent arbitrary , of course , but reflects the fact that many radial - velocity observations use lines formed relatively deep in the atmosphere .
a more common definition of the inertia is @xmath157 where @xmath158 is the so - called _
mode mass_. the code has the option to output the eigenfunctions , in the form of @xmath159 .
in addition ( or instead ) the displacement eigenfunctions can be output in a form indicating the region where the mode predominantly resides , in an energetical sense , as @xmath160 ( for radial modes only @xmath161 is found ) .
these are defined in such a way that @xmath162 { { \rm d}}x / x \over 4 \pi [ y_1 ( x_{\rm phot } ) ^2 + y_2 ( x_{\rm phot } ) ^2/l(l+1 ) ] } \ ; . \label{eq : e4.4}\ ] ] the form provided by the @xmath163 is also convenient , e.g. , for computing rotational splittings @xmath164 ( e.g. , * ? ? ? * ) , where @xmath165 is the frequency of a mode of radial order @xmath149 , degree @xmath12 and azimuthal order @xmath2 . for slow rotation
the splittings are obtained from first - order perturbation analysis as @xmath166 characterized by _ kernels _
@xmath167 , where in general the angular velocity @xmath168 depends on both @xmath1 and @xmath26 .
the code has built in the option to compute kernels for first - order rotational splitting in the special case where @xmath168 depends only on @xmath1 .
several revisions of the code have been implemented in preliminary form or are under development .
a substantial improvement in the numerical solution of the oscillation equations , particularly for high - order modes , is the installation of a fourth - order integration scheme , based on the algorithm of @xcite .
this is essentially operational but has so far not been carefully tested .
comparisons with the results of the variational expression and the use of richardson extrapolation , of the same formal order , will be particularly interesting . as discussed by @xcite the use of @xmath169 ( or , as here , @xmath170 ) as one of the integration variables has the disadvantage that the quantity @xmath171 enters into the oscillation equations .
in models with a density discontinuity , such as results if the model has a growing convective core and diffusion is neglected , @xmath171 has a delta - function singularity at the point of the discontinuity . in the adipls calculations this
is dealt with by replacing the discontinuity by a very steep and well - resolved slope .
however , it would obviously be an advantage to avoid this problem altogether . this can be achieved by using instead the lagrangian pressure perturbation @xmath172 as one of the variables .
implementing this option would be a relatively straightforward modification to the code and is under consideration .
the proper classification of dipolar modes of low order in centrally condensed models has been a long - standing problem in the theory of stellar pulsations , as discussed in section [ sec : results ] .
such a scheme must provide a unique order for each mode , such that the order is invariant under continuous changes of the equilibrium model , e.g. , as a result of stellar evolution . as a major breakthrough , takata in a series of papers has elucidated important properties of these modes and defined a new classification scheme satisfying this requirement @xcite .
a preliminary version of this scheme has been implemented and tested ; however , the latest and most convenient form of the takata classification still needs to be installed .
a version of the code has been established which computes the first - order rotational splitting for a given rotation profile @xmath173 , in addition to setting up the corresponding kernels .
this is being extended by k. burke , sheffield , to cover also second - order effects of rotation , based on the formalism of @xcite .
an important motivation for this is the integration , discussed by @xcite , of the pulsation calculation with the astec evolution code to allow full calculation of oscillation frequencies for a model of specified parameters ( mass , age , initial rotation rate , etc . ) as the result of a single subroutine call .
i am very grateful to w. dziembowski and d. o. gough for illuminating discussions of the properties of stellar oscillations , and to a. moya and m. j. p. f. g. monteiro for organizing the comparisons of stellar oscillation and model calculations within the esta collaboration .
i thank the referee for useful comments which , i hope , have helped improving the presentation .
this project is being supported by the danish natural science research council and by the european helio- and asteroseismology network ( helas ) , a major international collaboration funded by the european commission s sixth framework programme .
christensen - dalsgaard , j. , berthomieu , g. : theory of solar oscillations . in : cox , a. n. , livingston , w. c. , matthews , m. ( eds ) , _ solar interior and atmosphere _ , p. 401
space science series , university of arizona press ( 1991 ) moya , a. , christensen - dalsgaard , j. , charpinet , s. , lebreton , y. , miglio , a. , montalbn , j. , monteiro , m. j. p. f. g. , provost , j. , roxburgh , i. , scuflaire , r. , surez , j. c. , suran , m. : inter - comparison of the g- , f- and p - modes calculated using different oscillation codes for a given stellar model .
apss , this volume ( 2007 ) | development of the aarhus adiabatic pulsation code started around 1978 .
although the main features have been stable for more than a decade , development of the code is continuing , concerning numerical properties and output .
the code has been provided as a generally available package and has seen substantial use at a number of installations .
further development of the package , including bringing the documentation closer to being up to date , is planned as part of the helas coordination action . |
the knowledge of the in - medium hadronic properties at supra - normal densities is of major importance for the understanding of ( nuclear ) astrophysics such as the physical mechanism of supernovae explosions and the physics of neutron stars @xcite .
strangeness production at intermediate energy heavy - ion collisions has been a very helpful tool in studying the hadronic equation of state ( eos ) and the properties of strangeness matter under extreme conditions of baryon density and temperature @xcite .
it is well established that ( @xmath0 ) kaons feel a weak repulsive potential of the order of @xmath1 mev at saturation density @xmath2 , as it has been verified by studying strangeness flows in heavy - ion collisions @xcite .
the determination of the high density behavior of the nuclear eos has been successfully investigated in this context @xcite .
it has been realized that strangeness production could also serve as an important messenger for the isovector part of the nuclear eos at high densities , i.e. for the symmetry energy @xcite .
a reliable determination of the symmetry energy at supra - normal densities from heavy - ion collisions is still an object of current debates and investigations @xcite .
however , a reliable study of the high density nuclear eos and , in particular , its isovector part involves a detailed analysis of the role of the isospin dependent part of the kaon - nucleon interaction on strangeness production in intermediate energy nucleus - nucleus collisions , which has not been studied so far .
theoretical studies on static hadronic matter based on the chiral perturbation theory ( chpt ) @xcite and on the one - boson - exchange approach ( obe ) @xcite predict a splitting of self energies between positive charged ( @xmath3 ) and neutral ( @xmath4 ) kaons .
this isospin effect increases with baryon density and isospin asymmetry , and it may thus influence the production and propagation of particles with strangeness in dense hadronic matter during a nucleus - nucleus collision .
it may thus be a sensitive observable to determine the symmetry energy at high densities .
this work is an extension of former theoretical investigations of kaon medium effects in heavy - ion collisions @xcite to asymmetric hadronic matter .
the main topic of our study is therefore the influence of the isovector sector of the kaon - nucleon interaction on kaon dynamics at incident energies near the strangeness production threshold . in extension to our previous work @xcite , two different models for the kaon - nucleon interaction and their particular isovector properties
are now discussed in detail : one based on the chiral perturbation theory ( chpt ) @xcite and a second one based on the one - boson - exchange ( obe ) approach for the effective meson - nucleon field theory @xcite .
we use a fully relativistic transport model , in which isospin effects are included in the mean - field and in the collision integral @xcite , and also in the different realizations of the kaon potential .
these models are applied in heavy - ion collisions at intermediate energy below the threshold for strangeness production . the production of the @xmath5 isospin states and , in particular , their ratio is extensively investigated within the different scenarios for the isovector sector of the nuclear mean - field and the kaon - nucleon potential .
the antikaon field @xmath6 is not considered here .
a discussion of the stronger and more complex @xmath7-nucleon interaction would go beyond the scope of this work , e.g. , we refer to refs .
in this section we describe the theoretical model used in this work . after a brief introduction of the transport equation we focus on the different treatments of the kaon - nucleon potential in isospin - asymmetric hadronic matter , which is the major focus of the present work .
the theoretical description of a heavy ion collision is based on the relativistic boltzmann - uehling - uhlenbeck ( rbuu ) dynamical approach , as introduced in refs .
@xcite and then extended for isospin and strangeness degrees of freedom @xcite : @xmath8
f_i(x , k^ { * } ) = { \cal i}_{\rm coll } \label{rbuu } \quad .\end{aligned}\ ] ] the rbuu transport equation describes the phase space evolution of the one - particle density distribution functions @xmath9 for the different particle species under consideration under the influence of a mean - field ( lhs of eq.([rbuu ] ) ) and binary collisions ( rhs of eq.([rbuu ] ) ) which couple the different fields . here
we consider the propagation of all baryons up to the @xmath10-resonance and hyperons ( @xmath11 ) .
the production of pions results through the resonance decay and kaons ( @xmath0 ) are produced in baryon - baryon and baryon - pion binary collisions ( see for details ref .
@xcite ) . the mean - field or the nuclear eos enters via the effective mass @xmath12 , the kinetic momenta @xmath13 and the field tensor @xmath14 , where @xmath15 and @xmath16 are the lorentz - scalar and lorentz - vector components of the baryon self energy .
pions are propagated under the influence of the coulomb field only .
mesons with strangeness obey the same transport equation ( [ rbuu ] ) , however , with different self energies with respect to the baryon ones , as it will be discussed below .
the collision integral is determined by the total elastic and inelastic in - medium cross sections and includes all important channels on pion and kaon production , without the inclusion of antikaons .
isospin effects on the threshold conditions are explicitly accounted for to respect energy conservation @xcite .
more details can be found in ref .
furthermore , the quasi - elastic @xmath17 channel with isospin exchange is included , however , it does not essentially influence the strangeness ratios .
kaplan and nelson were the first to apply the chiral lagrangian approach to the properties of kaons in dense nuclear matter @xcite .
starting from the effective chiral lagrangian an in - medium klein - gordon equations for kaons at the mean field level follow can be derived @xmath19 \phi_{k}(x)= 0 \label{kaonmean } \quad , \ ] ] where @xmath20 and @xmath21 are the baryon current and the scalar density , respectively .
( [ kaonmean ] ) can be rewritten in a more compact form as @xmath22\phi_{k}(x)= 0 \quad .
\label{kaonmean2}\ ] ] in isospin asymmetric matter the kaon effective mass is given by @xcite @xmath23 where @xmath24 are the total and isospin scalar baryon densities .
the vector potential is defined as @xmath25 with @xmath26 isoscalar and isovector baryon currents , respectively , and @xmath27 for @xmath3 and @xmath4 .
@xmath28 @xmath29 is the free kaon mass , @xmath30 @xmath29 the pion decay constant and @xmath31 the kaon - nucleon sigma term , usually chosen in the range @xmath32 @xmath29 .
@xmath33 is the in - medium pion decay constant @xcite which appears only in the vector field . according to brown and rho @xcite this accounts for the fact that the enhancement of the the scalar part is compensated by higher order corrections in the chiral expansion .
finally , the parameter @xmath34 is fixed to the value of @xmath35 @xcite . in an alternative approach kaon potentials can also be derived within an effective meson field @xmath36 approach @xcite . in this picture
the kaons interact through the exchange of mesons with different lorentz properties ( @xmath37 ) .
these mesons mediate the isoscalar - scalar , isoscalar - vector , isovector - vector and isovector - scalar parts of the kaon - kaon interaction . in the mean - field approximation
the kaon equation of motion reads @xmath38\phi_k(x)=0 \label{kaoneq2 } \quad .\end{aligned}\ ] ] as in the chpt case ( see eq .
( [ kaonmean2 ] ) ) eq .
( [ kaoneq2 ] ) can be written more compact as : @xmath39 \phi_k(x)=0 \label{kaoneq}\ ] ] the vector potential @xmath40 and the effective mass @xmath41 are now defined differently relative to the chpt approach @xmath42 where @xmath43 and @xmath44 , @xmath45 and @xmath46 are the isoscalar , isovector baryon currents , the isoscalar - scalar field and isovector - scalar baryon density , respectively .
we will use a simple constituent quark - counting prescription to relate the kaon - meson couplings to the nucleon - meson couplings , i.e. a reduction by a factor @xmath47 ; @xmath48 ( @xmath49 ) .
thus , in the isoscalar - vector sector we have @xmath50 due to the enhanced kaon - scalar / vector coupling . in the isovector - vector sector
one obtains @xmath51 .
the scalar kaon field finally includes the isoscalar - scalar and isovector - scalar couplings @xmath52 and @xmath53 , respectively .
for the parameters of the nucleon - meson couplings we refer to ref .
we note that the picture of a simple constituent quark - counting may not be consistent with @xmath54 scattering in free space . however , this is not of importance here , since @xmath54 scattering may influence kaon spectra , but not absolute kaon yields and their ratio , which is the main topic of this work .
( 0.5,0.0 ) comparing both models for the kaon - nucleon interaction , eqs .
( [ vpot2],[emass ] ) versus ( [ kaoneffiso],[kapotiso ] ) , one sees that in the @xmath36 model the nucleon - meson couplings enter explicitly in the kaon self energies , while in the @xmath18 approach the kaon mean field is determined by the parameters @xmath31 and @xmath34 .
it is therefore helpful to study the isospin dependence of the @xmath36 model for various parametrizations of the isovector eos , i.e. , the symmetry energy . as in previous works @xcite
, we use the @xmath55 , @xmath56 and @xmath57 prescriptions for the isovector mean - field . in the first one ( @xmath55 )
no isospin potential is included .
@xmath56 contains only the @xmath58 meson , while @xmath57 accounts for both , the @xmath58 and the @xmath59 mesons in the isovector potential .
the symmetry energy becomes stiffer in the direction @xmath60 , since one has to increase the @xmath58-meson coupling when the @xmath59 meson field is taken into account in the potential part of the symmetry energy .
the crucial quantity for the interpretation of the transport results shown later is the kaon in - medium energy , given by the dispersion relation : @xmath61 the density dependence of the in - medium kaon energy @xmath62 for @xmath0 is shown in fig .
[ kmass ] for both models for the kaon - nucleon potential . in general
the @xmath18 approach leads to a stiffer density dependence of the in - medium kaon energy relative to the @xmath36 model .
this feature is more pronounced at high densities @xmath63 , which are reached in intermediate energy heavy - ion collisions near the kaon production threshold .
thus , one expects observable effects of the model of the symmetry energy in the produced kaons ( @xmath0 ) from heavy - ion collisions . the isospin effect for the considered asymmetry parameter of @xmath64 ( @xmath65 are the neutron and proton numbers ) , which corresponds approximately to @xmath66-nuclei , is pronounced at the baryon densities of @xmath63 , see again fig .
[ kmass ] . according to eqs .
( [ kaoneffiso],[emass ] ) the isovector contributions to the kaon - nucleon @xmath18 and @xmath36 scalar self energies lead to an effective mass splitting between positive ( @xmath3 ) and neutral ( @xmath4 ) kaons .
also the vector self energies between @xmath3 and @xmath4 differ according eqs .
( [ kapotiso],[vpot2 ] ) .
these isospin effects finally lead to the pronounced isospin dependence of the in - medium kaon energy .
we should note here the different origin of the isospin part of the kaon - nucleon potential between the two approaches . in the @xmath18 case the isospin dependence enters only through the isospin - currents , while in the @xmath36 model
the effective isovector meson - nucleon couplings are explicitly included .
this leads to an isospin splitting between @xmath4 and @xmath3 , which additionally depends on the model for the symmetry energy . the non - trivial isospin dependence of the two models for the kaon - nucleon potential may have interesting effects on the absolute @xmath4 and @xmath3 yields and their collective flows and , in particular , on the strangeness ratio @xmath67 in heavy - ion collisions at intermediate energies near the kaon production threshold , which is the topic of the next section .
the results of the following dynamical calculations are based on the transport equation ( [ rbuu ] ) , in which the isospin effects of the kaon - nucleon potentials are of particular interest , since they are investigated for the first time and may affect absolute kaon yields and strangeness ratios .
we first discuss in subsection @xmath68 the qualitative features of the two models of the kaon - nucleon potential and compare with experiments . for a transparent discussion on the the isospin effects
the following strategy has been chosen : the isoscalar part of the nuclear mean - field is the same for all different cases of transport calculations , while the isovector channel includes only the @xmath58 meson . in the strangeness sector following cases of different transport calculations will be discussed : 1 .
@xmath18 refers to transport results with only the isospin independent part of the chiral model for the kaon potential .
2 . @xmath36 refers to transport results with only the isospin independent part of the effective mean - field model for the kaon potential .
3 . @xmath69 refers to transport results with the full isospin dependent chiral model for the kaon potential .
4 . @xmath70 refers to transport results with the full isospin dependent effective mean - field model for the kaon potential .
thus in the first two cases isospin effects are included only in the baryonic sector ( @xmath58 meson ) , while in the last two ones they are taken into account also in the strangeness sector .
we then investigate in subsection @xmath71 in detail the role of the different isospin - dependent kaon - nucleon potentials on the strangeness @xmath72-ratio in @xmath73 central collisions . here
we particularly distinguish for each case listed above between the @xmath55 , @xmath56 and @xmath57 models , in which the symmetry energy becomes more stiffer in the sequence @xmath74 .
we start the discussion with the influence of the different kaon potentials on absolute @xmath75 meson yields .
this is shown in fig .
[ kpotrap ] for heavy - ion collisions between heavy systems ( au+au ) , in terms of their rapidity distributions for the four cases .
consistent with fig .
[ kmass ] the transport calculations with the @xmath18 model with the larger repulsion leads to overall smaller kaon yields with respect to those in the @xmath36 approach .
the inclusion of the isospin dependent part in the kaon potentials ( compare the curves @xmath69 and @xmath18 as well as those between @xmath70 and @xmath36 in fig . [ kpotrap ] ) changes only very little the absolute yields .
this is again consistent with fig .
[ kmass ] , where the isospin effects on the two isovector states are small at baryon densities around @xmath76 .
furthermore , in fig . [ kmass ] the asymmetry parameter was fixed to a value of @xmath77 corresponding to the asymmetry of a @xmath66-nucleus ) , while in heavy - ion collisions the local isospin - asymmetry , in the stage of kaon production , is reduced with respect to that of the initial situation @xcite .
this is because of partial isospin equilibration due to stopping and inelastic processes with associated isospin exchange .
( 2,0.0 ) it is interesting to study the role of the different kaon potentials in lighter colliding systems , such as @xmath78 , in which also experimental information is available @xcite .
[ knini ] shows the comparison between the data and the transport calculations obtained with both sets of the kaon potential without and with the isospin dependence ( left and right panel , respectively ) .
the isospin dependence is again not of relevance here .
the different kaon potentials influence the kaon yields in a similar manner as in the previous figure .
these in - medium dependencies are less pronounced now , since less compression is achieved in the stage of kaon production for the lighter @xmath78-system .
both sets of calculations are close to the experimental data , with the tendency of an improvement in the @xmath36 and @xmath70 cases . however
, this result does not neccesarily mean that the @xmath36 and @xmath70 kaon - nucleon potentials are the most realistic ones , because other physical quantities may influence the transport calculations .
a general and non - trivial feature in heavy - ion collisions concerns the treatment of the in - medium effects on the elastic and inelastic cross sections , which are usually parametrized to empirical free nn - scattering .
microscopic studies on elastic nn - scattering at finite baryon density exist @xcite , but for inelastic processes , e.g. , @xmath79 , theoretical calculations are very rare and often limited in baryon density and momentum @xcite .
attempts have been recently done in studying this non - trivial problem in heavy - ion collisions @xcite using a simple parametrization from microscopic dirac - brueckner calculations @xcite . for the strangeness sector , e.g. , @xmath80 and @xmath81 ( @xmath82 , @xmath83 and @xmath84 stand for nucleons or higher resonances , hyperons and kaons , respectively ) , the in - medium effects are trivially included only in the phase - space factors and the threshold conditions , but not in the matrix elements , i.e. , here free parametrizations are used @xcite .
a selfconsistent treatment of kaon - nucleon scattering in the hadronic environment would be desirable in this context , as it has been already investigated for the antikaonic sector @xcite .
furthermore , a more realistic density and , in particular , energy dependence of the hadronic relativistic mean field would be important @xcite . however , in order to keep the discussions transparent , we do not expand on these topics here , and use the same parametrization as in @xcite .
( 2.,0.0 ) in fig .
[ kpotcentr ] the centrality dependence of the @xmath3 yields for @xmath85-collisions is shown .
first of all , all sets of data understimate the data due to the strong density dependence of the inelastic cross section for @xmath10-production , see also ref . @xcite and the discussion above .
the in - medium dependence of the kaon potentials is now more pronounced for the most central @xmath85-collisions , see left panel in fig .
[ kpotcentr ] .
isospin effects of both kaon potentials appear due to the asymmetry of the colliding @xmath86-systems , see right panel of the same figure .
the effect of the inclusion of the isovector potential turns out to be stronger in the @xmath69 case compared to the @xmath70 one , since the in - medium energy for @xmath3 is stronger affected in the @xmath18 model , see again fig .
[ kmass ] .
thus , the kaon yield is enhanced more strongly in the @xmath69 transport calculations , and comes closer to the @xmath70 calculations , which are less sensitive to isospin effects .
collective kaon ( @xmath3 ) flow ( not shown here ) has been also studied in the framework of the different kaon potentials .
all the calculations reproduce the experimental flow @xcite reasonable well , however , the isospin effects were only moderate
. it would be more helpful to study _ isospin _ strangeness flows , i.e. , the difference of collective flows between @xmath4 and @xmath3 , since the isovector part of the kaon potential affects @xmath3 and @xmath4 differently . in conclusion
, the evaluation scheme of the kaon - nucleon potential within the effective @xmath36 approach seems to lead to a better description of the experimental data .
however , a definitive conclusion on the question , which potential would be more realistic , can not be drawn at this level of study . a more systematic study on in - medium effects also on the level of @xmath87-scattering is in this context neccesary . ( 8.,7.0 ) ( 2.0,0.0 ) thus at this level of the comparison between the theoretical calculations and the data the @xmath36 framework describes the experiment better than the @xmath18 model .
however , for a transparent discussion specifically of the isospin effects on strangeness ratios both models of the kaon - nucleon interaction will be investigated . a more promising observable to study isovector effects
may be given by the ratio of the two isospin kaon states , e.g. , the strangeness ratio @xmath67 , since ratios are expected to be independent on possible in - medium modifications of the isoscalar medium .
e.g. , the influence of in - medium inelastic @xmath88 cross sections has been studied in @xcite with the conclusion of a very moderate dependence of the strangeness ratio on the density dependence of the inelastic cross section .
a crucial question is whether particle yield _ ratios _ are influenced by the in - medium modifications of kaon potentials .
this is important particularly for kaons , since ratios of particles with strangeness have been widely used to determine the nuclear eos at supra - normal density @xcite .
more recently , the @xmath89- and @xmath72-ratios have been proposed as a good probe to determine the high density behaviour of the isovector part of the nuclear mean field @xcite .
we thus extend our previous studies @xcite by explicitly including both schemes of the kaon potential in the transport calculations ( @xmath18 and @xmath36 models ) .
in particular , we distinguish between the different scenarios for the symmetry energy in the baryon sector , i.e. , by neglecting the isovector potential ( @xmath55 case ) , by including only the @xmath58-meson ( @xmath56 case ) , and both , the @xmath58 and @xmath59 mesons ( @xmath57 ) in the isovector baryon and kaon potentials ( @xmath69 and @xmath70 cases ) .
( 8.,7.0 ) ( 2.0,0.0 ) fig .
[ ratios ] shows the strangeness ratio for the different parametrisations of the isovector part of the nuclear @xmath90 .
calculations are also performed for the two approaches of the kaon potential ( @xmath69 and @xmath70 model ) .
first of all , an overall linear increase of the ratio with respect to the stiffness of the symmetry energy is observed .
this feature is not trivial and reflects the interplay of isospin effects between mean - field and collision dynamics at the threshold . as discussed in detail in ref .
@xcite , isovector mean - field and threshold effects act differently on the isospin kaon states @xmath3 and @xmath4 , where the later feature governs the strangeness dynamics at low intermediate energies near the kaon production threshold .
the sensitivity of the strangeness ratio on the different evaluations of the isospin _ independent _ kaon potential ( @xmath18 and @xmath36 cases ) is very moderate , as expected . if the isovector part is neglected , the density behavior of the kaon potential is the same for both isospin @xmath4 and @xmath3 states .
therefore the ratio should be rather independent on the evaluation scheme of the kaon potential , as it is the case in the transport calculations of fig .
[ ratios ] .
taking explicitly account of the isospin effects in the kaon potentials ( @xmath69 and @xmath70 cases ) the dynamical situation changes . the isospin effect in the @xmath69 model
strongly decreases the strangeness ratio , however , the relative difference between the various choices of the symmetry energy ( @xmath55 , @xmath56 and @xmath57 ) remains stable .
this is due to the fact that the isovector sector of the @xmath69 kaon self energies contain only the isospin lorentz - scalar and lorentz - vector densities without additional parameters , e.g. , meson - nucleon couplings @xcite .
this feature is also shown in fig .
[ kmass ] , in which the isospin splitting is independent on the stiffness choice of the symmetry energy .
the decrease of the ratio in the @xmath69 calculations is related to the fact that the in - medium energy becomes more stiffer ( softer ) for @xmath4 ( @xmath3 ) .
the situation in the @xmath70 calculations is similar , but the isospin effects are less pronounced .
the suppression of the kaon ratio in the transport calculations due to the isospin effects is now weak , due to the softer character of the @xmath91 kaon potential , compared to the @xmath92 model .
however , the relative difference between the different models of the symmetry energy is slightly modified . when isospin effects are accounted for
, the @xmath70 self energies explicitly contain the meson - nucleon couplings , which are different for each model of the symmetry energy , see also again fig .
[ kmass ] .
this leads to the observed different slope of the kaon ratio in the @xmath70 transport calculations , however , the effect is rather moderate .
it is interesting to note that the situation in a chemical and thermal equilibrated ( closed ) system , e.g. , in studying strangeness production in a box with periodic boundary conditions , is rather different .
the isospin effects on the strangeness ratio are very pronounced in box calculations ( see fig .
@xmath93 in ref .
however , the isospin effect is considerably reduced in an open system in heavy - ion collisions due to , e.g. , isospin transparency which decreases the local asymmetry of the kaon production source . in summary
, the consideration of the isospin effects on the level of the kaon - nucleon interaction leads to modifications of the strangeness ratio on the absolute value , but not to significant changes with respect to the stiffness of the symmetry energy . in the @xmath69 and @xmath70 approaches a sensitivity to the symmetry energy on the level of @xmath94 and @xmath95 survives .
we thus conclude that the strangeness ratio is a very promising observable to determine the isospin dependence of the underlying kaon - nucleon interaction and particularly the isospin part of the nuclear eos .
we have investigated the role of different kaon potentials and , in particular , their explicit isospin dependence in asymmetric nuclear matter .
the kaon potentials have been evaluated in two schemes , one following a chiral perturbation approach and a second one within the framework of an effective one - boson - exchange picture .
we have studied in detail the effects of various isospin dependent kaon potentials on kaon dynamics in heavy - ion collisions within a covariant transport equation .
the isospin effects turns out to be less sensitive to absolute kaon yields , but more pronounced to the strangeness @xmath67 ratio .
we thus focused our attention to the @xmath67 ratio in collisions between heavy and isospin - asymmetric systems ( @xmath85 ) at incident energies below the kaon production threshold .
different evaluation schemes of the kaon potential and its isovector part lead to different results on the strangeness ratio .
the isospin dependence changes this ratio on an absolute level . as an interesting result ,
however , the relative dependence of the strangeness ratio to the stiffness of the isovector eos remains stable . at this level of investigation
we conclude that the strangeness ratio is an important observable to determine the high density behavior of the symmetry energy .
future experiments with exotic radioactive beams , as they are planned in new experimental facilities at msu and at gsi ( fair project ) , will be very helpful to resolve the theoretical uncertainties of the symmetry energy at high baryon densities .
* acknowledgements * + we acknowledge very fruitful discussions with a.b .
larionov , m. kaskulov , h. lenske .
we would like to particularly thank m. kaskulov and h. lenske for a carefully reading of the manuscript and further suggestions .
this work is supported by bmbf and by the dfg cluster of excellence _ origin and structure of the universe_. one of the authors ( v.p . )
acknowledges support by the hellenic state scholarships foundation ( i.k.y . ) .
v. thorsson , m. prakash , j.m .
lattimer , nucl . phys .
* a572 * ( 1994 ) 693 ; + r. knorren , m. prakash , p.j .
ellis , phys . rev .
* c52 * ( 1995 ) 3470 ; + g.e .
brown , k. kubodera , m. rho , v. thorsson , phys . lett .
* b291 * ( 1992 ) 355 ; + v. thorsson , j.p .
ellis , phys . rev .
* d55 * ( 1997 ) 5177 ; + t. klhn et al .
, phys . rev .
* c74 * ( 2006 ) 035802 . | kaon properties are studied within the framework of a fully covariant transport approach .
the kaon - nucleon potential is evaluated in two schemes , a chiral perturbative approach and an effective one - boson - exchange model .
isospin effects are explicitly accounted for in both models .
the transport calculations indicate a significant sensitivity of momentum distributions and total yields of @xmath0 isospin states on the choice of the kaon - nucleon interaction .
furthermore , isospin effects are rather moderate on absolute kaon yields , but appear on strangeness ratios .
this is an important issue in determining the high density symmetry energy from studies of strangeness production in heavy - ion collisions .
asymmetric nuclear matter , symmetry energy , relativistic heavy ion collisions , kaon - nucleon potential , strangeness production .
+ pacs numbers : * 25.75.-q * , * 21.65.+f * , 21.30.fe , 25.75.dw . |
the detection and characterization of astronomical sources becomes increasingly difficult as we attempt to observe these sources in the euv , x - ray , and gamma - ray spectral regimes .
there are several reasons for this .
first , in these high - energy regimes , source data may consist of only a few counts , so that we must rely on the poisson distribution when making statistical inferences rather than using gaussian statistics that are considerably easier to apply , but that are strictly applicable only in the high - counts limit .
second , spatially extended sources , such as supernova remnants and galaxy clusters , exhibit bright diffuse emission at high energies which may overlap with point sources , rendering the latter more difficult both to detect and characterize . and third , the present generation of broad - band high - energy telescopes , unlike optical telescopes , have spatially non - uniform point spread functions ( psfs ) as an unavoidable by - product of their design .
for instance , the psf of the wolter i - type high resolution mirror assembly ( hrma ) on the _ chandra x - ray observatory _
( cxo ) has a 50% encircled energy radius that varies in width from @xmath0 0.3@xmath1 on - axis to @xmath2 10@xmath1 near the outer edges of an acis - i chip ( @xmath0 10@xmath3 off - axis ) . a standard method for the analysis of poisson count data involves the application of the so - called `` sliding cell '' ( harnden et al . 1984 )
. in sliding - cell analysis , two co - aligned but differently sized square cells are placed at each image pixel , with the number of counts in the annular region between cells providing an estimate of the local background amplitude at the pixel .
this amplitude is then used to compute the poisson significance of the observed number of counts in the inner cell .
the usefulness of a sliding - cell algorithm is limited both when the field - of - view ( fov ) is crowded , since overlapping sources can not be handled and/or nearby sources contributing counts to the estimated background may decrease detection sensitivity , and when sources are observed off - axis , since uncertainties in the model of the psf , which generally increase with off - axis angle , can greatly affect source property estimates ( see , e.g. , kashyap et al .
. the sliding cell also may not provide accurate source property estimates for extended sources .
source characterization can be improved by fitting the detected sources using maximum likelihood methods ( as in the algorithm of hasinger , johnston , & verbunt 1994 ) , but the accuracy of this method is still limited by uncertainties in the psf . within the last decade , astronomers have begun to apply wavelet functions to the problem of source detection .
( for an introduction to the theory of wavelet functions , see , e.g. , mallat 1998 and references therein . ) these functions are scaleable , oscillatory , have a finite support ( i.e. , are non - zero within a limited spatial regime ) , and have an average value of zero ; they can be used to define a set of basis functions that act as highly localized filters in both the spatial and frequency domains , and thus are superior source characterizers .
certain wavelet functions also exhibit the property of vanishing moments , which is important for source detection : the integral of the product of a wavelet with @xmath4 vanishing moments and a polynomial of degree @xmath5 is zero .
thus the correlation of a suitably chosen wavelet function with a photon counts image will yield correlation coefficients which are significantly large only in the vicinity of localized high - order variations in the data , e.g. in the vicinity of astronomical sources , which appear as psf - broadened bumps with infinite power - series representation .
wavelet - based source detection thus boils down to the statistical problem of identifying image pixels with significantly large " correlation coefficients .
damiani et al .
( 1997a ) were the first to present a wavelet - based generalized method for source detection and characterization , which they apply to _ rosat _ pspc data in a subsequent work ( damiani et al .
1997b ) . among its features , this method , unlike the others , uses exposure maps to handle situations in which , e.g. , support rib shadows or the edge of the fov lie within the wavelet support , allowing the analysis of the entire fov .
their algorithm is , however , instrument - dependent in that they must make significant changes to it in order to account for qualitative differences between exposure maps of different detectors ( e.g. between the exposure maps of the _ rosat _ pspc and _ rosat _ hri ; micela et al .
1999 , provide a short description of the changes that are necessary to perform hri image analysis ) .
in addition , while one can apply the damiani et al . method to data from other photon - counting detectors , the psfs must be very nearly gaussian .
last , they publish a function for the computation of source detection thresholds which is strictly valid only when the mean background amplitude per image pixel is @xmath6 , where @xmath7 is the wavelet scale size . in this paper ( and in freeman
et al . 2001 ) we describe our own algorithm for wavelet source detection and characterization that has been developed for a generic detector and implemented in the _ chandra interactive analysis of observations ( ciao ) _ routine wavdetect . in subsequent papers ( e.g. kashyap et al .
2001 , in preparation ) we will discuss the application of this algorithm to specific scientific problems .
our algorithm is considerably more general and flexible than others that have been developed , in that it can : ( 1 ) operate effectively in the low background counts regime , which is crucial because of the expected low particle and cosmic background count rates for the _ chandra _ detectors ( the overall rate being @xmath8 10@xmath9 and 10@xmath10 ct sec@xmath11 pix@xmath11 for the _ chandra _ acis and hrc detectors , respectively ) ; and ( 2 ) operate effectively regardless of the psf _ shape _ , also crucial because of the ( non - gaussian ) nature of the off - axis _ chandra _ psfs .
it also ( 3 ) corrects for the effect of exposure variations in a general , non - detector - specific manner .
thus our algorithm may be immediately adapted for the analysis of data from virtually any other photon - counting detector .
in [ sect : wave ] , we provide a brief description of wavelet functions , and define the marr , or
mexican hat " ( mh ) wavelet function which we use in our algorithm .
we then present a simple example in which we apply the mh function to idealized data , in order to build the reader s intuition about how to interpret the results .
the mh function has been used often in astronomical wavelet analyses , in for example : the analysis of @xmath12co spectral maps of the molecular cloud l1551 ( gill & henrikson 1990 ; see also langer , wilson , & anderson 1993 , who apply laplacian pyramid transforms to a @xmath12co image map of barnard 5 ) ; the analysis of galaxy cluster structure in optically derived catalogs ( slezak , bijaoui , & mars 1990 ) ; the detection and localization of features in optical ccd images of galaxies ( coupinot et al . 1992 ) ; the analysis of substructures observed in _ rosat _ pspc images of the coma cluster ( vikhlinin , forman , & jones 1994 ) ; the examination of _ einstein _ hri and _ rosat _ pspc images of abell 1367 ( grebenev et al .
1995 ) ; the detection of serendipitous x - ray clusters in archival _ rosat _ pspc data ( sharc ; cf .
ulmer et al .
1995 , freeman et al .
1996 , nichol et al . 1997 ) ; and the analysis and modeling of x - ray emission in abell clusters ( lazzati & chincarini 1998 ; see also lazzati et al . 1998 ) and stellar clusters ( damiani et al .
1997b ) .
in [ sect : alg ] we describe the basic steps of our algorithm , in which sources are detected and characterized .
the basic steps of _ source detection _ include : the correlation of the wavelet function with the data to create the correlation image ; the estimation of the local background ( if necessary ) in each pixel ; the computation of the source detection threshold in each pixel ; accounting for the effects of exposure variations within the fov on the correlation value ; and the identification and cleansing " of source counts from the image .
source count cleansing is done iteratively , first by analyzing the raw data , then the first version of the cleansed data , etc .
, until the background is estimated .
this final background is used to compute source detection thresholds , which are subsequently used to identify putative sources .
_ source characterization _ involves the combination of outputs from a number of wavelet scales , and the creation of source cells on the initial image within which source properties ( positions , count rates , etc . ) are estimated .
in [ sect : ver ] we demonstrate the efficacy of the algorithm by applying it to a large variety of cases , such as : an idealized image containing simulated extended and point sources , with spatially invariant gaussian psf and an exposure map similar to that of the _ einstein _ ipc ; a 32-ksec _ rosat _ pspc observation of the pleiades cluster ( cf .
micela et al . 1996 ) ; and a simulated 30-ksec _
acis - i observation of the lockman hole region ( t. gaetz , private communication ) .
we then describe the differences between our method and previously published methods , especially that of damiani et al .
, in [ sect : comp ] , and provide our summary and conclusions in
[ sect : sum ] . in future works
, our algorithm will be applied to the analysis of specific scientific problems , e.g. observed spatial variations in the diffuse x - ray background of the pleiades cluster ( see kashyap et al .
1996 and kashyap et al .
2001 , in preparation ) .
wavelets can be used to filter an image at a given length - scale .
this can be seen by considering the difference between two smoothed functions : one formed by convolving an ( arbitrary ) function @xmath13 with a real - valued , non - negative , infinitely differentiable smoothing function @xmath14 whose size is characterized by a scale @xmath7 and satisfies the condition @xmath15 1 ( e.g. , the gaussian function ) , and one formed by the convolution of @xmath13 with the same smoothing function with scale size @xmath16 ( holschneider 1995 ) . because all structure at length - scales smaller than the smoothing scales
would be suppressed , the difference of these two smoothed functions will provide information about the details of @xmath13 that are introduced at the scale @xmath7 itself .
in the situation of interest to us , the analysis of two - dimensional images , the relationship between @xmath17 and the wavelet function @xmath18 is , in the limits @xmath19 0 , @xmath20 ( the reason for our use of a prime symbol will become apparent below . ) this is an analyzing wavelet , or mother wavelet ( hence the subscript @xmath21 ) .
other members of the same wavelet family ( so - called `` atoms '' ) can be generated from this mother wavelet via dilations and translations : @xmath22 @xmath23 and @xmath24 are the dilation parameters , and @xmath25 and @xmath26 are the translation parameters .
there are many functions @xmath17 that can be used to create a mother wavelet .
one example is the two - dimensional gaussian function : @xmath27 from which is created the marr , or
mexican hat " ( mh ) , wavelet function : @xmath28e^{-\frac{x^2}{2\sigma_x^2}-\frac{y^2}{2\sigma_y^2 } } \label{eqn : wave1 } \\ & = & \frac{1}{2{\pi}\sigma_x\sigma_y } w_m(\frac{x}{\sigma_x},\frac{y}{\sigma_y } ) \ , .
\label{eqn : wave}\end{aligned}\ ] ] this is the wavelet function we use in our source detection algorithm .
( note that we use the function @xmath29 instead of @xmath30 , to be consistent with damiani et al .
the mh function has a positive kernel with shape similar to a canonical psf ( @xmath31 ) , surrounded by a negative annulus ( @xmath32 ; figure [ fig : wave ] ) .
ellipses with semi - axes of length @xmath33 , 2@xmath34 , and 5@xmath34 describe the boundary between @xmath31 and @xmath32 , the minimum of the mh function , and the effective support of the mh function , respectively .
the mh function offers several advantages which motivate its use for source detection : ( 1 ) it has two vanishing moments ( i.e. the correlation of the mh function with constant or linear functions is zero ) , so that it acts to suppress the contribution of the ( generally spatially constant ) background to the correlation coefficients ; ( 2 ) while it does not have compact support , and can not be used to construct a set of orthogonal basis functions , a dyadic sequence of mh functions ( i.e. , with mh functions with scales separated by factors of two ) is sufficient to sample the entire frequency domain , because of the limited extent of an mh function s fourier transform ; ( 3 ) its limited extent in both spatial and fourier domains helps to minimize effects of aliasing ; and ( 4 ) it is analytically manipulable , so that many numerical operations may be performed using analytically derivable functions ( see appendix [ sect : mh ] ) , which can reduce computation time significantly . before describing our algorithm in detail
, we present a simplified example to help build the reader s intuition .
we assume that we are analyzing a subset of an evenly exposed image with the mh wavelet function , far from any edge of the fov . within this image ,
the counts in each pixel , @xmath35 , are sampled from a function which has a constant , relatively large ( @xmath36 1 ) amplitude , which we denote @xmath37 .
these assumptions allow us to build intuition with a minimum of unnecessary detail ( such as correcting for exposure variations , etc . ) , and should not be construed as being reflective of limitations on the applicability of our algorithm . as will be described in
[ sect : alg ] , the first step in our algorithm is to correlate the wavelet function @xmath38 and the binned image data @xmath39 .
we denote correlation using the symbol @xmath40 ; in this case , we would write @xmath41 , or for a particular pixel , @xmath42 .
would be written @xmath43 $ ] ; however , we feel our notation makes complicated expressions in the remainder of this work more easily interpretable . ]
because the average value of the wavelet is zero and the data are sampled from a constant amplitude function , the mean correlation value will tend asymptotically to zero , with statistical sampling causing positively and negatively valued deviations from zero in individual pixels .
indeed , for our simple situation , the resulting probability sampling distribution ( psd ) of correlation values , denoted @xmath44 , will tend asymptotically to a zero - mean gaussian , with width @xmath45 ( freeman et al .
1996 , damiani et al . 1996 ,
damiani et al .
1997a ) .
we now assume that there is a tightly bunched clump of counts in our otherwise source - free subset image , which could be caused by a poisson sampling fluctuation or by an astronomical source .
we assume the clump has gaussian shape with amplitude @xmath46 and width @xmath47 , which is comparable to the size of the psf of the instrument .
the correlation of the mh function with a gaussian yields a mh function that has its centroid at the gaussian centroid , and has centroid amplitude @xmath48 proportional to @xmath46 : @xmath49 .
\label{eqn : cmax}\ ] ] for our simple example , @xmath50 0 , i.e. @xmath51 is always greater than the mean of the psd . to determine whether we should associated the clump with an astronomical source
, we would compute the integral of the psd from @xmath51 to @xmath52 ; if this quantity is smaller than a predefined significance threshold ( e.g. @xmath53 10@xmath9 , or @xmath54 ) , then we associate the clump with a source . because @xmath55 , source detectability will vary as a function of wavelet scale size . in the limits
@xmath56 0 and @xmath57 , the ratio @xmath58 goes to zero , i.e. the source becomes undetectable .
if we apply symmetric wavelets in our analysis , the maximum value of this ratio occurs at @xmath59 .
hence sources are most easily detected when one analyzes the image with a wavelet function that has a size similar to that of the source .
since such a function acts to filter , " or selectively enhance , structures of similar scale , this behavior is to be expected .
if the clump is associated with a previously known source , but is not detected , we can use eq .
( [ eqn : cmax ] ) to determine the upper limit on source counts , by substituting the source detection threshold for @xmath51 , and solving for @xmath46 ( see , e.g. ,
[ sect : ver1 ] ) .
( because the gaussian is normalized , @xmath46 is identically the number of source counts . )
we stress that the use of this method to place upper limits on source counts is limited to cases where the psf shape is that of a two - dimensional gaussian function , and should not be used in the place of simulations if the psf has arbitrary shape .
our first objective is _ source detection _ : the identification of putative source pixels in binned , two - dimensional image data .
this identification is normally done by carrying out the steps in the algorithm described below separately with each of a number of wavelet functions ( see also figure [ fig : flow ] ) .
the basic steps in this algorithm include ( 1 ) the correlation of the data with the given wavelet function ( [ sect : wcor ] ) , and ( 2 ) the identification of image pixels with correlation values larger than pre - defined thresholds for source detection ( [ sect : thresh ] ) .
the second step requires knowledge of the local background in each pixel ( due to unresolved point sources , diffuse astrophysical emission , particle background , etc . ) .
if the background has not been determined previously , then it must be estimated , e.g. via the method described in
[ sect : back ] .
also , instrumental artifacts such as support rib structures , hot spots , and the edge of the fov can adversely affect the second step , resulting in the detection of instrumental features in the image , which are astrophysically uninteresting . in
[ sect : wexp ] , we describe a method for rejecting such
instrumental sources . " finally , in
[ sect : werror ] we describe how the variances of the correlation and local background amplitudes are estimated .
the first step in the source detection process at a given scale is to compute the correlation of the wavelet function @xmath60 with the binned , two - dimensional image data @xmath39 .
the translation parameters @xmath25 and @xmath26 shown in eq .
( [ eqn : trans ] ) correspond to image pixel indices @xmath61 and @xmath62 ; for a given pixel , the correlation value is @xmath63 ( henceforth , if a quantity is given with subscripts , we are referring to its value at pixel @xmath64 ; otherwise , we are referring to the quantity s array of values . ) the interested reader may find details about how @xmath65 is computed in appendix [ sect : mh ] . to determine whether pixel @xmath64 should be associated with a source , we compute the probability of observing the correlation value @xmath66 if there are only background counts present within the support of @xmath38 : @xmath67 @xmath68 is dubbed the significance ( or the type i error ; see , e.g. , eadie et al .
1971 , pp . 215 - 216 ) , and it is the estimated probability that we would erroneously identify pixel @xmath64 with a source . in [ sect : simp ] , we indicated how @xmath68 could be determined analytically in the high - counts limit , where @xmath69 tends asymptotically to a zero - mean gaussian of width @xmath70 .
in general , however , @xmath69 must be determined via simulations ( see appendix b ) . because the number of simulations we carried out is only sufficient to directly determine significances @xmath71 10@xmath10 , and because strong sources may have much greater significances ( in a qualitative sense )
, we instead compute a source detection threshold @xmath72 via the equation @xmath73 where @xmath74 is the user - specified threshold significance .
, where @xmath75 is the number of analyzed pixels in the image ; with this choice , the average number of false detections is one per image . ] if @xmath76 , we associate the pixel @xmath64 with a source . in this section ,
we describe how the local background counts amplitude @xmath77 is estimated if it is unknown _ a priori _ ( see also figure [ fig : bgflow ] ) . while there is no unique way to make this estimate , we seek a method that does not depend on a detector s psf , both for increased generality and computational speed , and also because we want our algorithm to be able to detect and analyze sources of arbitrary size , not just point sources .
we can fulfill this condition by creating background maps _ at each wavelet scale _ , . ] using a localized function that is wavelet - scale , and not psf - scale , dependent : the wavelet negative annulus @xmath78 ( i.e. , @xmath38 with positive values reset to zero ) . while the function @xmath32 is related to a wavelet function , _ we stress that it itself is not a wavelet function , and that its use in background estimation does not constitute a transform .
_ if the exposure within the support of @xmath32 is constant , and if @xmath23 and @xmath24 are sufficiently large such that the integral of the source counts distribution ( i.e. the psf for a point source ) over the @xmath32 is insignificant with respect to the integrated background , then one can estimate the background using the formula @xmath79 where @xmath80 is the integrated volume of @xmath78 ( see appendix [ sect : negann ] for derivation ) .
if on the other hand the exposure is not constant , the exposure map can be used as a weighting function : @xmath81 where @xmath82 is the normalized ( i.e. flat - fielded ) number of expected background counts at pixel @xmath64 .
we ignore the distinction between vignetted and non - vignetted components of the background ( e.g. , the particle background ) in our estimate , because of degeneracy .
we note that if the amplitude of the non - vignetted background component @xmath83 is known , then one could in principle estimate the background using a variation of eq .
( [ eqn : back ] ) : @xmath84 there are three situations in which counts from sources may bias the local background estimate : ( 1 ) if the estimate is being made within diffuse extended emission ; ( 2 ) if the pixel in which the estimate is being made is a source pixel , but @xmath23 and/or @xmath24 is smaller than the source size @xmath85 ; or ( 3 ) if sources are located within the @xmath32 .
the first situation is a non - issue , because if the analysis goal is detection , say of a source within a supernova remnant , then the diffuse emission should be treated as a local background component : for instance , should a clump of counts be associated with a source , or with poisson fluctuations in the background _ and _ the diffuse emission ? in the second situation
, the background map will exhibit a bump " at the location of the source , whose amplitude is greatest where the number of source counts within the support of the @xmath32 is maximized , generally at the source centroid ( figure [ fig : bgbump ] ) . within the bump
, the source detection threshold is overestimated , and thus this perhaps - otherwise - detectable source may remain undetected _ at the given scale_. however , this is not a critical problem , since if the source is detectable , it will be detected when ( @xmath23,@xmath24 ) @xmath86 , a scale regime where the bump is minimized , and the issue becomes moot .
note that these bumps do not adversely affect source property estimation , since they are eliminated when the background map that is used for source property estimates is computed ( [ sect : nbkg ] ) .
the third situation can be more problematic for source detection .
if there are sources in the fov , then there will _ always _ be some image pixels for which the background amplitude is overestimated : for instance , for a symmetric wavelet function , these pixels will surround sources in circular rings " of radius @xmath0 2@xmath7 ( the radius at which @xmath32 achieves its minimum value ; see figure [ fig : bgring ] ) .
in these rings , @xmath87 is overestimated , so that otherwise - detectable sources whose locations coincide with these rings may go undetected .
rings can appear regardless of the scale or source size , and thus can impede source detection at all scales .
they can also adversely affect the computation of source properties , as a background that is overestimated in the vicinity of a source can lead to underestimated count rates , etc .
, in the final source list .
( this last point is demonstrated below in figure [ fig : spropest ] , which shows the effect of rings on the estimation of pleiades cluster source properties .
[ sect : ver3 ] . )
one way to remove the rings is to remove source counts iteratively from the raw data , via the following algorithm : 1 .
identify pixels to be cleansed " using @xmath44 and the initial background map , which we will dub @xmath88 ; 2 .
mask out these pixels or replace their data with other values , creating a new image we denote @xmath89 ; 3 .
estimate @xmath90 , where @xmath91 is the iteration number ( @xmath92 2 ) ; and 4 .
if the background map is to be refined yet again , determine @xmath93 , identify pixels to be cleansed using @xmath94 , and return to step ( 2 ) .
then , when the final background map @xmath95 is determined , compare the original values @xmath96 with @xmath97 to create a final list of putative source pixels .
regarding steps ( 1 ) and ( 4 ) , since the goal of this iterative approach is to remove as many source counts as possible from the raw data , we advocate an aggressive approach to identifying the pixels to be cleansed : the threshold significance should be set high , e.g. to @xmath98 ( although never higher than 0.05 , which corresponds to the oft - used 95% rejection level of statistics ) . regarding step ( 2 ) , while masking is used by damiani et al .
( 1997a ) in their two - iteration approach to source detection , it would preclude us from using ffts to calculate @xmath99 .
thus we replace the data in cleansed pixels with the inferred background amplitude .
there are no rigorous quantitative rules governing how one should specify the number of iterations , as that can depend on the crowdedness of the field , the source distribution , the source strengths , and the wavelet scale size .
we do note that iterative cleansing will cease if the background map does not change from one iteration to the next , i.e. , if no new pixels are marked for cleansing .
in [ sect : thresh ] , we describe how we use probability sampling distributions @xmath69 to identify sources .
these distributions are derived assuming a spatially constant exposure .
if , for instance , the exposure map exhibits localized high - order variations , then the list of detected sources may contain a mixture of astrophysically interesting sources and `` instrumental sources '' aligned near support rib shadows , near the edge of the fov , at the location of hot pixels , etc .
e.g. _ , figure 1 of damiani et al . 1997b . )
thus , an efficacious source detection algorithm should include additional calculations that act to decrease the detectability of instrumental sources , while leaving the detection efficiency of astrophysical sources unchanged .
are also derived assuming a spatially constant background map , from which simulated data are sampled .
thus an efficacious source detection algorithm should also include calculations that mitigate the effect of background variations caused by , _
e.g. _ , x - ray shadows .
the current algorithm does not take such variations into account ; anecdotal evidence ( e.g. in [ sect : ver3 ] ) indicates that they have little effect , possibly because they are far less sharp " than variations induced by support rib shadows , etc .
note that if the background is known _ a priori _ , one can in principle remove the effect of high - order background variations on correlation values using a transformation similar to the one described below.]@xmath100 one possibility is to construct new sampling distributions @xmath101 for each observation , taking into account all the exposure variations that can appear within the wavelet support ; however , this is not computationally practical .
instead , we estimate the systematic effect that exposure variations have on the correlation coefficients .
assuming the null hypothesis , we may write the observed correlation coefficient as @xmath102 where @xmath77 is the estimated ( noise - free ) background intensity and @xmath103 is the noise ( and possibly source count ) contribution to @xmath66 .
we ignore the latter term ( see the caveats below ) and rewrite the former so that its dependence on exposure variations is explicit : @xmath104 the quantity @xmath105 encapsulates exposure variability within the wavelet support , and thus the last term in eq .
( [ eqn : expcor_deriv ] ) encapsulates the effect of exposure variability upon @xmath66 ; subtracting this term from @xmath66 yields an exposure - corrected " quantity that contains only information of astrophysical value : @xmath106 it is this quantity that is compared with the distribution @xmath69 to determine whether @xmath64 is a source pixel . if @xmath107 is constant ( or linear ) within the wavelet support , then eq .
( [ eqn : fullcor ] ) reduces to @xmath108 because @xmath109 is computed only once , we dub this the fast " exposure correction , as opposed to the full " exposure correction of eq .
( [ eqn : fullcor ] ) .
one should not use the fast " correction if non - linear structures ( caused , e.g. , by x - ray shadows ) are apparent in the background map .
one should keep the following caveats in mind : 1 . strictly speaking ,
@xmath110 still can not be directly compared with the probability sampling distribution @xmath69 because the noise term @xmath103 is itself uncorrected .
if we concentrate on the issue of false positives ( _ i.e. _ assume that there is no source count contribution to @xmath103 ) , the important question is : is the asymptotic width of the distribution from which @xmath111 is sampled _ smaller _ than the width of the distribution from which @xmath103 is sampled ?
if so , then the rate of false detections will still be greater than expected .
this is a problem if and only if for a given pixel , @xmath112 is smaller than the average exposure @xmath113 over the wavelet support @xmath64 , i.e. this is only a problem within troughs or beyond the edge of the fov . to see this , harken back to the simple example of [ sect : simp ] : what would happen to the width of @xmath114 if we were to reduce the exposure ? fewer counts would be detected , so @xmath77 would decrease , and the width of the noise distribution , which is @xmath115 , would also decrease . thus we suggest that one should carefully scrutinize all sources detected in low - exposure regions ( @xmath116 ) .
2 . note the distinction between the correlation maps @xmath96 and @xmath117 , especially if the analysis goal is not just source detection , but also image decomposition ( the scale - by - scale characterization of sources in correlation space ) . ) .
thus this caveat is only an issue if the user desires to analyze correlation maps _ outside _ of wavdetect . ]
the quantity @xmath117 represents a _
pixel - by - pixel _ wavelet filtering not of the raw data @xmath39 , but of the quantity @xmath118 ( eq . [ eqn : fullcor ] ) . because @xmath107 may be estimated using the @xmath32 , which is most sensitive to low - frequency components of the data ( see figure [ fig : pow ] ) , these modified data may be contaminated " with low - frequency information ( although wavelet filtering [ eq .
[ eqn : fullcor ] ] mitigates the effect of the contamination ) .
we note that while systematic overestimates of @xmath107 ( caused for reasons discussed in [ sect : back ] ) adversely affect the computation of @xmath117 , they will not lead to an increased number of false detections .
this is because the only situation where such a systematic overestimate affects _ non - source _ pixels in the _ final _ , refined background image is when the background is computed in regions of extended diffuse emission .
in this situation , the algorithm treats the sum of the real background and the diffuse emission as the background , " so the rate of false detections ( which is independent of background amplitude ) will be unchanged .
we estimate the variances of @xmath77 ( if one does not provide a background map ) , and of @xmath66 ( or @xmath110 or @xmath119 ) , using the standard formula ( eadie et al .
1971 , p. 23 ) @xmath120~&=&~v[\sum_i \sum_j a_{i , j } x_{i , j } ] \nonumber \\ & = & ~\sum_i \sum_j a_{i , j}^2 v[x_{i , j } ] + 2 \sum_i \sum_{i ' > i } \sum_j \sum_{j ' > j } a_{i , j } a_{i',j ' } { \rm cov}[x_{i , j},x_{i'j ' } ] \label{eqn : covar } % % % & \approx&~\sum_i \sum_j a_{i , j}^2 v[x_{i , j } ] \nonumber \ , , \label{eqn : var}\end{aligned}\ ] ] where @xmath121 is quantity of interest and @xmath122 are random variables ( either @xmath35 or functions of @xmath35 ) .
for instance , @xmath123~&=&~v[\sum_{i ' } \sum_{j ' } w_{i - i',j - j'}d_{i',j ' } ] \nonumber \\ & = & ~\sum_{i ' } \sum_{j ' } w_{i - i',j - j'}^2 v[d_{i',j ' } ] \nonumber \\ & = & ~\sum_{i ' } \sum_{j ' } w_{i - i',j - j'}^2 d_{i',j'}\label{eqn : corerr2 } \\ & = & ~<w^2{\star}d>_{i , j } \ , .
\label{eqn : corerr3}\nonumber\end{aligned}\ ] ] note that we make two assumptions when deriving this formula : ( 1 ) the datum @xmath124 is sampled from a poisson distribution with variance @xmath124 ; and ( 2 ) each pixel s raw datum is independently sampled ( so covariance terms do not contribute to @xmath125 $ ] ) .
we list variance formulae related to source detection in table [ tab : sderr ] .
the reader should keep in mind two important caveats about them : 1 .
these formulae ignore the contribution of the covariance terms , which are non - zero for @xmath126 $ ] , @xmath127 $ ] , and @xmath128 $ ] if the data are iteratively cleansed , i.e. if @xmath37 is a not just a function of the raw data @xmath39 only .
we ignore these terms because even the simplest covariance computation , that for a two - iteration background map ( see appendix [ sect : covar ] ) , has a staggeringly high computational cost : we find that the cpu time needed to compute the variance increases by a factor @xmath129 , where @xmath130 and @xmath131 are the x- and y - axis lengths in pixels , respectively .
also , additional arrays containing information needed to compute the covariance terms must be kept in memory , so there is a resource cost as well .
we find that including covariance terms increases the variance by a median value of @xmath0 7% , and at most by only @xmath0 30% adjacent to strong sources , although this is a source - strength- and source - geometry - dependent result that obviously can not be blindly applied to all fields .
ultimately it is up to the user to judge whether adding the computation of covariance to our base algorithm is worthwhile .
when computing @xmath132 $ ] for the final background map , we make the simplifying assumption that the variance of a cleansed datum is equal to the cleansed datum itself ; for instance , for a two - iteration background map , @xmath133~&=&~\sum_{i'}\sum_{j ' } a_{i , i',j , j'}^2 v[d_{2,i',j ' } ] \nonumber \\ & = & ~\sum_{i'}\sum_{j ' } a_{i , i',j , j'}^2 d_{2,i',j ' } \label{eqn : bvar } \,,\end{aligned}\ ] ] where @xmath134 we make this assumption because if the data are a mixture of raw data and background estimates , the variance estimates become increasingly complicated : for one iteration , @xmath135~&=&~d_{i , j } \ , , \nonumber\end{aligned}\ ] ] for two iterations , @xmath136~&=&~\cases { d_{i , j } & { \rm uncleansed pixel } \cr \sum_{i'}\sum_{j ' } a_{i , i',j , j'}^2 d_{i',j ' } & { \rm cleansed pixel } } \label{eqn : twoiter}\end{aligned}\ ] ] etc .
once we have identified putative source pixels at each of a number of wavelet scale size pairs @xmath137 , our next objective is _ source characterization _ , wherein we combine information derived at each scale pair to generate a final source list and to estimate source properties ( see figure [ fig : wrflow ] ) . unlike source detection
, source characterization algorithms can be arbitrarily complex , depending , for instance , upon whether one wishes to use detailed psf information .
our method is particularly simple , in that we use only the characteristic psf size at a given pixel , @xmath138 .
this size may be associated with , e.g. , 50% encircled energy ; we find that smaller values , such as 39.3% ( which corresponds to the integral of a symmetric normalized two - dimensional gaussian to radius @xmath47 ) , work better than large values .
while using the detailed psf shape may allow for more accurate estimates of source properties , the simplicity of our scheme makes it more immediately applicable to images from virtually any counts detector ( including those for which calibration is on - going or is for other reasons incomplete ) .
in [ sect : back ] , we describe how we calculate a scale - dependent normalized local background estimate @xmath107 , by assuming that there are no source counts in the negative annulus , @xmath32 . source detection itself is not markedly affected if this assumption is violated and the background overestimated , for reasons given in [ sect : back ] , but an overestimated background will adversely affect the computation of source properties . we create a new , corrected , background estimate by combining information across scales ( denoted with a subscript @xmath139 ) , noting that the assumption that there no source counts in @xmath32 is always violated around sources if either @xmath140 or @xmath141 is less than the source size : @xmath142 @xmath4 is the number of scale pairs used , and @xmath143 @xmath21 is a multiplicative factor , set to one when estimating the properties of point sources , and to larger values when extended sources are analyzed .
we use eq .
( [ eqn : covar ] ) to estimate the variance of @xmath144 , with covariance terms ignored : @xmath145 \ , .
\label{eqn : corbvar}\ ] ] @xmath146 $ ] is estimated using the approximate equation listed in table [ tab : sderr ] .
the next step in source characterization is to determine which pixels of the original image @xmath39 are to be associated with each detected source .
we term contiguous pixels which are associated with a particular source a _ source cell _ , and as we describe below in [ sect : sprop ] , we use the raw data in a source cell to determine a source s properties .
we need to create source cells because , as noted above , we do not use psf shape information in our algorithm , and because we want an algorithm that is applicable to both point and extended sources .
source cells are not computed in algorithms such as the sliding cell , where the integrated psf volume and sum of ( point ) source counts for a given user - defined cell can be used to estimate the total number of ( point ) source counts , etc . to create source cells ,
we first compute source count images , smoothing the raw data with the positive kernel of the wavelet function , @xmath31 , at user - specified scales , and then subtracting the corrected background image , @xmath147 : @xmath148 ( we denote these images @xmath149 to avoid confusion with either the significance @xmath150 or the correlation image @xmath96 . )
we use @xmath31 as the smoothing function because it has the desirable properties of being localized , and , for the particular case of a symmetric mh function , of mimicking the shape of a canonical gaussian psf . in regions where there are no sources , source count image values are either zero , or positive and nearly zero
; only in the vicinity of sources do the values deviate markedly from zero .
thus the source count images appear to contain numerous islands " of non - zero flux in a sea of zero values , with their relative size increasing with size of the smoothing function @xmath31 ( see figure [ fig : scounts ] ) .
each island contains one or more peaks , and sub - islands may be defined using each peak , with saddle points providing the boundaries between them .
( sub-)islands observed in selected source count images define the source cells . to define a source cell , we must select a source counts image and then determine to which ( sub-)island the putative source belongs .
the selection proceeds as follows .
the location of a putative source in correlation - space is assumed to be the location of the correlation maximum , @xmath151 . at this location
, we compute the psf size , in pixels , @xmath152 .
( this introduces a bias towards point sources ; we return to this point below . )
we then select the source count image with smoothing scale
closest " to @xmath152 by minimizing @xmath153 , where @xmath154 is defined for each scale pair : @xmath155 \ , .
\nonumber\end{aligned}\ ] ] on the selected source count image , we examine pixel @xmath151 : the ( sub-)island to which this pixel belongs defines the source cell . a source cell defined in this manner
has advantageous properties : ( 1 ) if @xmath156 , then nearly all isolated point source counts should lie within a source cell ( figure [ fig : scex1 ] ) ; ( 2 ) exposure variations are taken into account via the use of @xmath157 in eq .
( [ eqn : flux ] ) , so that source cells are not truncated near , e.g. , support rib shadows ; and ( 3 ) , as noted above , saddle points in the source count images provide natural boundaries between sources in crowded fields ( figure [ fig : scex2 ] ) .
we note two situations where care must be exercised when interpreting results .
first , the source cell for an extended source may be too small if the smoothing scale is @xmath158 .
the steps one must take to deal with this situation will vary depending upon analysis circumstances , but one possible step is to create only one source count image , with @xmath159 , and to use this image to define the cell , while being careful to note whether previously detected point sources are located within it .
( see , e.g. ,
[ sect : ver1 ] . )
another situation for which care must be exercised is when the psf is bimodal or otherwise strangely behaved ( such as the off - axis _
chandra _ psf ) ; two or more source cells could be created for one detected source . the necessary steps to deal with
this situation depend upon the details of the detector itself , and thus we will not discuss this particular situation further here . because the same source will generally be detected at multiple scales , and to further decrease the possibility of finding false sources , it is necessary to reject sources from the lists of correlation maxima generated at each scale .
a maximum observed at @xmath151 , for the scale pair @xmath160 , is rejected from further consideration if any of the following conditions are met : ( 1 ) @xmath151 lies in a previously defined source cell ; ( 2 ) @xmath161 0 ; ( 3 ) the ellipse defined by @xmath162 contains one or more previously defined sources detected at smaller scales ( this can occur when @xmath140 or @xmath141 is @xmath163 , since previously identified sources will eventually merge if the field is crowded , creating new " sources at new locations ) ; and ( 4 ) if , after the source cell is defined for a particular scale , it is found not to contain any correlation maxima _ at that scale_. this last check is aimed at rejecting small - scale poisson fluctuations that may be observed in the background data .
we present the formulae we use to estimate source properties and their variances in tables [ tab : prop ] and [ tab : error ] respectively .
the summations performed when making these estimates are carried out over the pixels within the source cell .
we use the raw counts data , @xmath35 , as a weighting function , instead of the source fluence @xmath164 , because the use of the latter can greatly complicate the estimation of variances . using the data rather than
the source fluence will lead to similar estimates when the background amplitude is small relative to source amplitude .
we can use the information present in the correlation images and the source cell image to create a
noise - free " rendering of the observed source data , @xmath165 , with the effect of exposure variations removed : @xmath166 dividing the correlation value by @xmath140 and @xmath141 restores the normalization contained in eq .
( [ eqn : wave ] ) , allowing a scale - by - scale summation .
the quantity @xmath167 = 1 if ( 1 ) @xmath168 0 , ( 2 ) the local maximum corresponding to @xmath64 has been identified as a source pixel , and ( 3 ) the associated local correlation maximum is contained within a source cell ( the second condition ensures that random maxima which are not associated with a source but which happen to lie within a source cell are not included in the source image ; the last condition ensures that rejected sources are not included ) ; otherwise , @xmath167 = 0 .
to verify its source detection and characterization capabilities , we apply our algorithm to : ( 1 ) 1 and 10 ksec observations by an idealized detector with a spatially invariant psf ; ( 2 ) a 32-ksec _ rosat _ pspc observation of the pleiades cluster ; and ( 3 ) a simulated 30-ksec _ chandra _
acis - i observation of the lockman hole region .
these tests allow us to demonstrate that our algorithm can efficiently detect and accurately describe well - sampled sources in an uncrowded field , and can effectively analyze crowded fields , even in the low - background limit of the _ chandra _ detectors .
we first demonstrate that our algorithm can efficiently detect , and accurately describe , well - sampled sources in an uncrowded field .
we apply it to two 512@xmath169512 images , hereafter images a and b , that represent 1 and 10 ksec observations by an idealized detector with an effective area 1000 @xmath170 and a spatially invariant gaussian psf of width @xmath171 2.56 pixels ( figure [ fig : simdata ] ) .
the exposure map for this detector is similar to that of the _ einstein _ ipc . within each image
we randomly place 42 point sources , and 4 extended sources with elliptical shape .
the fluxes of the point sources were sampled from a @xmath172 distribution with slope @xmath173 , above 10@xmath174 erg @xmath175 sec@xmath11 .
we also simulate a locally variable background ( amplitude @xmath8 10@xmath176 ct sec@xmath11 pix@xmath11 ) by setting the background amplitude at five reference points , performing minimum - curvature - surface interpolation , and sampling background data in each pixel .
we analyze the images assuming the input parameters listed for test 1 in table [ tab : numdet ] . in figure
[ fig : counts ] , we plot the source counts for detected sources , and the upper limits for undetected sources , against the number of predicted counts .
upper limits are defined using the source detection threshold values at the correlation maxima nearest the location of the undetected sources , and are computed using eq .
( [ eqn : cmax ] ) , with @xmath177 4.43 pixels .
we conclude that our algorithm efficiently detects and describes point sources with @xmath2 10 counts .
this is not an absolute quantity : the minimum number of counts needed for source detection varies as a function of the background amplitude and source size ( see [ sect : simp ] ) .
hence while we may conclude that a 1 ksec observation is sufficient for the detection by _
chandra _ of nearly on - axis point sources with fluxes @xmath2 10@xmath174 erg @xmath175 sec@xmath11 , since it has a similar effective area as , and lower expected background count - rates than , our idealized detector , it may not be sufficient far off - axis , or for detectors that have higher rates of background accumulation . in figure
[ fig : cell ] , we show the cells for the detected sources , along with the input source locations
. we use a wavelet scale of 2@xmath178 pixels to create the source counts image that is in turn used to delineate the source cells .
this scale is the closest
standard " scale to the assumed psf size ( if we assume scale sizes separated by factors of @xmath178 rather than 2 for greater source detection efficiency ) . in both images ,
our algorithm detects one false source , which is consistent with the assumption of @xmath179 for a 512@xmath169512 image .
creating source cells by using information derived at a wavelet scale close to the psf size is not optimal when one wishes to analyze and describe extended sources , as discussed in
[ sect : cell ] , and indeed by examining figure [ fig : cell ] we can see that the extended source cells are undersized .
there are many ways by which an analyst may wish to treat extended sources ; here , we show how one could derive an image showing the ( normalized ) number of counts per pixel within the extended source .
we use as our example the largest extended object in image b. first , we must expand the source cell so that it just encloses the extended source . to do this
, we increase the minimum scale size at which a source counts image is to be computed ( in this example , from 2@xmath178 pixels to 8@xmath178 pixels ; see figure [ fig : ext]a - b ) .
( note that the parameter @xmath21 in eq .
[ eqn : bsumm ] must also be increased so that the background is not overestimated within the source . )
second , we would use the new source cell as a spatial filter , applying it to the original source counts image created at the psf scale ( e.g. figure [ fig : ext]c ) , or to the data , etc .
( if one outputs source count image data , one can then , in principle , fit directly to them .
for instance , the image may be of a galaxy cluster , and one may wish to assess the detectability of its constituent galaxies .
however , care must be exercised since the data in contiguous pixels are not independent . taking into account the width of the @xmath31 used to smooth the raw image data , we can state that data @xmath180 pixels apart are independent .
thus simple statistical fitting can be done to a sparse grid of data .
this process of fitting would be essentially equivalent to the decimation " method described by lazzati et al .
1998 , except that they fit to correlation image data . )
next , we demonstrate that our algorithm outperforms the sliding cell in efficiently detecting sources in a crowded field , by applying it to the deep ( 32 ksec ) _ rosat _ pspc observation of the core of the pleiades cluster ( rp200068 , cf
. micela et al . 1996 ) .
these data were obtained in two segments separated by roughly six months , and the slight boresight offset between the two segments has been corrected using a method described by micela et al .
the data were also filtered to exclude times of high background contamination , and to exclude pulse - height - invariant ( pi ) channels at both the low - energy ( pi@xmath5320 , to avoid the so - called `` ghost image '' problem ; nousek & lesser 1993 ) and high - energy ( pi@xmath181201 , where no instrument map is available to determine exposure variations ) ends of the spectral response .
we have computed exposure maps taking into account these changes using software developed by snowden & kuntz ( 1998 ) . in table [ tab : numdet ] , we show how the number of detected sources varies as a function of the number of iterations , the spacing of scales , the exposure correction method , and the source detection significance . comparing tests 1 through 6 ,
for which @xmath74 is constant , we find that the number of detected sources changes little if the number of iterations is increased beyond two , or if the full exposure correction method ( eq . [ eqn : fullcor ] ) is used instead of the fast one ( eq . [ eqn : fastcor ] ) . however , there is an @xmath0 5% increase in source yield by analyzing the image with the scale sizes spaced by factors of @xmath178 instead of @xmath182 .
the extra sources are relatively weak sources whose probability of detection is maximized around the scales @xmath178 pixels , @xmath183 pixels , etc .
it is the decision of the user as to whether the increase in weak - source detection efficiency is worth nearly doubling the computation time .
we compare the source detection results for our test 8 with those shown in the wgacat and rosatsrc catalogs ( white , giommi , & angelini 1994 , and voges et al . 1994 , respectively ) , as well as those shown in micela et al . and damiani et al . ( 1997b ) .
( the wgacat and rosatsrc catalog teams , and micela et al . ,
use variants of the sliding - cell algorithm . ) we choose test 8 because it most closely resembles the analysis of damiani et al .
, who perform a two - iteration analysis of the pleiades image with wavelet functions of scale sizes 1 , @xmath178 , ... , 16 pixels , and with threshold significance @xmath184 ( 4.2@xmath7 ) .
our result is nearly equal to that of damiani et al .
, as we detect 148 sources , while they detect 150 .
we can not directly compare our numerical results with those of damiani et al .
, who only publish figures showing the performance of their algorithm on a subset of the pleiades field .
however , we can emulate damiani et al . by showing how our results compare with those of the wgacat and rosatsrc catalogs ( table [ tab : compare ] , cf .
damiani et al .
table 1 ) .
our results are virtually identical to those of damiani et al . ,
in , e.g. , how many sources are detected only by our algorithm , etc .
we thus may conclude that our algorithm and that of damiani et al .
generate a largely similar source list .
a large fraction of the sources that we detect , but that are not included in the wgacat or rosatsrc catalogs , are located either near the inner telescope support ring ( @xmath0 20@xmath3 off - axis ) or near the precipitous drop - offs in exposure caused by telescope vignetting near the edge of the fov ( figure [ fig : ple ] ; see also figure [ fig : spropest ] ) .
a visual examination of these sources indicates that they are not spurious .
we further examine in detail the two sources near the edge of the fov that were not included in either the wgacat or rosatsrc catalogs , but are in regions covered by other _ rosat _
pspc pointings described by stauffer et al . (
we find that these sources lie within @xmath185 pixels of the stauffer et al .
sources 159 and 197 ( see their table 2 ) , which have reported count rates @xmath0 0.023 and 0.016 ct sec@xmath11 respectively .
these rates are consistent with our derived count rates .
this demonstrates the robustness of our simple exposure correction method .
we also note the intriguing result that the local background map computed by our algorithm indicates the presence of an x - ray shadow in the core of the cluster ( figure [ fig : ple_bkg ] ; kashyap et al .
2001 , in preparation ) .
because the pleiades cluster is located beyond the edge of the local bubble of hot gas ( frisch 1995 ) , this shadow , which is most pronounced at low energies , is of more distant sources of the diffuse x - ray background ( dxbg ) , such as the extragalactic component and obscured stars in the pleiades cluster itself .
the depth of the shadow places constraints on the nature of the stellar mass - function at low masses , and in particular rules out models where the mass - function is extrapolated at a constant slope from higher masses , thus providing independent , x - ray observational support for optical observations that report drops in the mass - function at low masses ( @xmath186 ; see , e.g. , tinney , mould , & reid 1992 , bahcall et al .
1994 ) .
finally , we demonstrate that our algorithm can handle the low background amplitudes which are characteristic of _ chandra _ observations , while continuing to outperform the sliding cell , by applying it to a simulated 30 ksec _
acis - i image of the lockman hole ( t. gaetz , private communication ; figure [ fig : lh_data ] ) .
the acis - i is comprised of four 1024@xmath1691024 pixel ccds configured in a 2@xmath1692 square , with fov @xmath0 17@xmath18717@xmath3 ( @xmath0 50 times smaller than that of the _ rosat _ pspc ) . within the acis - i field
are placed : ( 1 ) 12 optically identified _
pspc x - ray sources catalogued by schmidt et al .
( 1998 ) , including one extended cluster source ; ( 2 ) @xmath0 6000 point sources sampled from the hasinger et al .
( 1998 ) @xmath172 distribution between 10@xmath188 and 5@xmath16910@xmath189 erg @xmath175 sec@xmath11 ; and ( 3 ) 19500 particle background counts ( corresponding to a particle background rate of 1.5@xmath16910@xmath10 ct pix@xmath11 sec@xmath11 ) .
if we assume the input parameters listed for test 1 in table [ tab : numdet ] , and do not use an exposure map , we detect 171 sources in the full acis - i field , of which four are directly associated with the extended galactic cluster ( figure [ fig : lh_data ] ) . in figure
[ fig : logn ] , we show differential @xmath172 distributions for both detected , and all , points sources in the fov . on the basis of this figure , we may conclude that our algorithm will efficiently detect sources in acis - i images with fluxes @xmath2 10@xmath189 erg @xmath175 sec@xmath11 ( 0.5 - 2 kev ) , and has the ability to detect poisson sampling fluctuations for sources with fluxes @xmath190 10@xmath191 erg @xmath175 sec@xmath11 .
our result compares very favorably with the performance of celldetect , which detects 51 sources with @xmath192 3 .
( further comparison of the source detection efficiencies of wavdetect and celldetect is provided by kim et al .
2001 , in preparation . )
we conclude that the relative superiority of our wavelet detection algorithm , with respect to the sliding cell , is inversely proportional to the background amplitude . in figure
[ fig : offax ] , we plot the offsets of the locations of detected sources from their actual locations within the fov .
( these offsets are caused by the asymmetry of the _ chandra _ psf , and their magnitude increases with off - axis angle . )
we note two characteristics of these offsets .
first , the variation in the offsets does not depend on source strength , signifying that the source - locating process is insensitive to the strength of the source .
second , the observed offsets are significantly smaller than the expected mean separation between sources ( @xmath193 pix ) , implying that the observed offsets are due to the asymmetries inherent in the psf and not due to source misidentifications .
we thus find that the weak sources are as `` well - behaved '' as strong sources ( about whose detection and identification there can be little doubt ) , and hence infer that even the weakest detected sources are real .
the source detection algorithm which we present in this work resembles algorithms published previously by vikhlinin et al .
( 1994 ) , rosati et al .
( 1995 ) , grebenev et al .
( 1995 ) , damiani et al .
( 1997a ) , and lazzati et al .
( 1998 ) . in this section ,
we highlight important differences between our algorithm and these other algorithms , all of which , having been developed for analyzing _
pspc data , suffer from inherent limitations that do not allow them to be directly applied to data from , e.g. , _
chandra_. we do not discuss how our method of source characterization differs from those described previously , because these other methods are built upon the premise that the psf has gaussian shape .
thus they are simply not directly applicable in situations where the psf has a more complex shape .
our basic method for computing the correlation images is the same as that used by vikhlinin et al . , rosati et al . , grebenev et al . , and lazzati et al . , with the exception that rosati et al .
use the symmetric morlet wavelet function @xmath194\ ] ] instead of the mh function .
however , these authors do not attempt to correct for exposure variations , as they focus their attention upon the center of the _ rosat _ pspc fov , and they use a fundamentally different method to determine source detection thresholds ( see below )
. the method by which damiani et al . correct for exposure and
detect sources differs substantially from ours .
they first divide the raw data image ( which they refer to as the photon image " ) by an exposure map on a pixel - by - pixel basis , to create the so - called count - rate image . "
pixels with relative exposure less than a certain amount ( e.g. 0.2 ) are not included , which introduces a sharp edge in the count - rate image where telescope vignetting becomes important .
damiani et al .
correlate the wavelet function with this image , applying an analytic correction to correlation values near the edge ( as given in their eq .
12 ) . because the data in the count - rate image are not poisson - distributed , damiani et al .
must convert source detection threshold values , derived from their background map in the photon image , to values appropriate for the count - rate image .
they accomplish this by dividing the photon - image detection thresholds by an effective exposure time @xmath195 , which is not a source exposure time that can be used to convert estimated source counts to count rates . because damiani et al .
derive the equation with which they compute effective exposure time in the gaussian limit ( see their appendix b ) , their exposure correction method is effectively limited to the high - counts regime : it can not be applied as is to , e.g. , typical _ chandra _ data . as described in
[ sect : alg ] , we estimate the local background counts amplitude in each pixel by assuming that within the negative annulus of the wavelet , @xmath32 , there are no source counts ( eq .
[ eqn : back2 ] ) .
we then use that inferred amplitude to determine source detection thresholds .
this is a generalization of the approach used by vikhlinin et al .
, rosati et al . , grebenev et al . , and lazzati et al . , in which the correlation variance @xmath196 is used to determine thresholds ( e.g. , @xmath197 3.5 ) . this approach will work if the background is locally flat or has a locally constant gradient , in which case the correlation of wavelet and background is zero .
such an approach is obviously insufficient for use with the whole fov of x - ray detector , where support rib shadows and vignetting will affect the background , and/or in situations where the non - instrumental background varies markedly ( such as in the pleiades ; see figure [ fig : ple_bkg ] ) .
damiani et al .
use the same basic approach to source detection as we do , in that they compute a local background amplitude , and use it to compute source detection thresholds .
damiani et al .
compute the background map by first smoothing the raw data with a gaussian with width @xmath198 .
they then interpolate background estimates within support rib shadows . at each scale , they compute the median value of the smoothed background within a square region with side - length @xmath199 centered on the given pixel .
( we explicitly use @xmath200 instead of @xmath201 to highlight their assumption of gaussian psf shape . )
after point sources are detected ( using a source cleansing significance that is 0.2@xmath7 smaller than their source detection significance ) , regions around them containing 95% of the counts are masked out , and a refined background estimate is made by interpolating over the masked regions .
in this work , we present a generalized wavelet - based source detection algorithm that , in principle , can be applied immediately to image data collected by any photon counts detector , although it was developed specifically for the analysis of _ chandra x - ray observatory _ image data .
we exclusively use the marr , or mexican hat , wavelet function in this paper , but the basic details of our algorithm would be unchanged if we use other wavelet functions , such as the haar or morlet wavelet functions
. aspects of our algorithm include : ( 1 ) the computation of the correlation of the wavelet function and the data image using either analytic or fft methods ; ( 2 ) the computation of a local , exposure - corrected normalized ( i.e. flat - fielded ) background in each pixel ; ( 3 ) its applicability within the low - counts regime , as it does not require a minimum number of background counts per pixel for the accurate computation of source detection thresholds ; ( 4 ) the correction of those correlation values which are affected by large exposure variations within the wavelet support ( due to , e.g. , telescope support ribs or the edge of the field of view ) , using either one of two methods given by eq .
[ eqn : fullcor ] and eq .
[ eqn : fastcor ] ; ( 5 ) the generation of a source list in a manner that does not depend upon the details of the psf shape , including the creation of general , data - dependent source cells for the estimation of source properties and filtering of extended source data ; and ( 6 ) full error analysis . in these respects ,
our algorithm is considerably more general than the similar wavelet - based methods developed by vikhlinin et al .
( 1994 ) , rosati et al .
( 1995 ) , grebenev et al .
( 1995 ) , damiani et al .
( 1997a , b ) , starck & pierre ( 1998 ) , and lazzati et al .
nearly all of these methods were developed specifically for treating data collected by the _
rosat _ pspc ( starck & pierre 1998 apply their method to data from the _ rosat _ hri ) ; none except damiani et al . attempt to correct for variations in exposure within the fov ; and all except for damiani et al .
assume a flat background across the region of interest .
the relatively more general method of damiani et al . is limited to analyzing data from detectors which have psfs with gaussian shape , and which have high rates of background accumulation .
these limitations make the damiani et al .
approach , as published , inappropriate for use with , e.g. , _ chandra _ data .
in [ sect : ver ] , we demonstrate the robustness of our algorithm by applying it , _ without algorithmic changes _ , to data collected by an idealized detector with a spatially invariant gaussian psf ; to _ rosat _ pspc data of the crowded field of the pleiades cluster ; and to a simulated _ chandra _ acis - i image of the lockman hole region .
collectively , these test cases indicate that our algorithm : ( 1 ) effectively detects and describes point sources , and can be applied to the study of extended sources ; ( 2 ) does not detect more spurious sources than expected ; ( 3 ) is more sensitive than sliding - cell methods ; and ( 4 ) has equal sensitivity to the damiani et al .
method for the specific case of _ rosat _ pspc data .
we find that while we can use the algorithm as presented to analyze extended sources , such analysis requires careful monitoring on the part of the analyst .
work on generalizing this analysis is on - going , and will be reported in a future work .
the authors acknowledge the support of the cxc beta test site grant and nasa grants nag5 - 3173 , nag5 - 3189 , nag5 - 3195 , nag5 - 3196 , nag5 - 3831 , nag5 - 6755 , and nag5 - 7226 .
we are grateful to t. gaetz and a. vikhlinin for making available the simulated acis - i image of the lockman hole , analyzed in [ sect : ver2 ] .
we would like to thank r. c. nichol , b. holden , j. flanagan , t. calderwood , and a. dobryzcki for their assistance with coding issues , e. kolaczyk for useful discussions regarding the computation of covariance , and f. damiani , s. sciortino , and a. bijaoui for useful discussions .
finally , we would like to thank the anonymous referees for their detailed comments and suggestions , which greatly helped the preparation of this manuscript .
in particular , our description of the vanishing moments property of wavelets and its relationship to source detection owes much to one of these referees .
the source detection algorithm that we have presented does not explicitly depend on the details of wavelet function itself , and thus should be applicable with other , simple , wavelet functions which have one central positive mode .
hence we have deferred to this appendix and the next the derivation of various formulae that we use in our algorithm which are valid specifically when using the marr , or mexican hat ( mh ) , wavelet function . before carrying out source detection at given scales ( @xmath23,@xmath24 ) , we must determine the grid of values @xmath202 ( see [ eqn : coranal ] ) .
this is accomplished by integrating the function @xmath203 , defined in eqs .
( [ eqn : wave1]-[eqn : wave ] ) , on a pixel - by - pixel basis within an ellipse with axes with full lengths 10@xmath23 and 10@xmath24 , with values farther from the origin set to zero : @xmath204 \exp\left(-\frac{x^2}{2\sigma_{x}^2}\right ) \exp\left(-\frac{y^2}{2\sigma_{y}^2}\right ) \ , .
\label{eqn : anal}\end{aligned}\ ] ] @xmath205 represent coordinates relative to the center of pixel @xmath64 , and and @xmath206 and @xmath207 denote the limits of integration over pixel @xmath208 ( figure [ fig : coord ] ) : @xmath209 we expand eq .
( [ eqn : anal ] ) and perform each integral separately .
the first integral in the expansion is computed using @xmath210 \nonumber \\ & = & \sqrt{\frac{\pi}{2 } } \sigma_x x_{\rm erf } \ , , \nonumber\end{aligned}\ ] ] where the substitution @xmath211 is made , and @xmath212 is the error function . performing the same integration over @xmath213 , we determine that the first term is @xmath214 . to compute the second and third integrals , we must determine , e.g. , @xmath215 this integral can be solved by parts , by taking the derivative of @xmath216 and the integral of @xmath217 .
we present the solution : @xmath218 the final solution is @xmath219 analytic computation of correlation values ( eq . [ eqn : coranal ] ) may be too computationally intensive if the image and/or wavelet scale sizes are too large .
so in wavdetect , for instance , fast fourier transforms ( ffts ) are used if scale sizes are @xmath220 2 pixels : @xmath221 .\ ] ] to mitigate the effect of the wrap - around , " any image that is to be transformed is padded with zeros ; the minimum width of the padding is 10@xmath222 . in our own fortran code , we use the analytic fourier transform of the mh function , which we denote @xmath223 : @xmath224 \int_{-\infty}^{+\infty } dx w(x , y ) \cos(2{\pi}k_{x}x ) \label{eqn : evodd2 } \nonumber \\ & = & ~\int_{-\infty}^{+\infty } dy \cos(2{\pi}k_{y}y ) \int_{-\infty}^{+\infty } dx w(x , y ) \cos(2{\pi}k_{x}x ) \nonumber \\ & = & ~\int_{-\infty}^{+\infty } dy \cos(2{\pi}k_{y}y ) \times c_{\rm w } \label{eqn : cw } \,.\end{aligned}\ ] ] the wave - number @xmath139 equals @xmath225 , where @xmath226 is the pixel number in fourier space , @xmath227 is the nyquist wave - number , and @xmath228 is half the length of the relevant axis in the padded image .
the fourth integral in eq .
( [ eqn : evodd ] ) and the second integral in eq .
( [ eqn : evodd2 ] ) are zero , as the integrands are products of even and odd functions .
@xmath229 \exp\left(-\frac{x^2}{2\sigma_x^2}\right)\cos(2{\pi}k_xx ) \nonumber \\ & = & ~\exp\left(-\frac{y^2}{2\sigma_y^2}\right ) \left [ \left(2 - \frac{y^2}{\sigma_y^2 } \right ) \psi_{\rm 0,c}(2{\pi}k_x,\frac{1}{\sqrt{2}\sigma_x},0 ) - \frac{1}{\sigma_x^2 } \psi_{\rm 2,c}(2{\pi}k_x,\frac{1}{\sqrt{2}\sigma_x } ) \right ] \label{eqn : c2 } \,.\end{aligned}\ ] ] @xmath230 and @xmath231 represent two integral solutions which one may find , e.g. , in gradshteyn & ryzhik ( 1980 ; formulae 3.896 - 2 and 3.952 - 4 respectively ) : @xmath232 and @xmath233 substituting these quantities into eq .
( [ eqn : c2 ] ) , one finds that @xmath234 \label{eqn : cw2}\ ] ] substituting eq .
( [ eqn : cw2 ] ) into eq .
( [ eqn : cw ] ) and solving , we find : @xmath235 \nonumber \\ & = & ~\exp(-2{\pi}^2k_{x}^2\sigma_{x}^2 ) \sigma_{x } \sqrt{2{\pi } } \left [ ( 4{\pi}^2k_{x}^2\sigma_{x}^2 + 1)\psi_{\rm 0,c}(2{\pi}k_{y},\frac{1}{\sqrt{2}\sigma_{y}},0 ) - \frac{1}{\sigma_{y}^2}\psi_{\rm 2,c}(2{\pi}k_{y},\frac{1}{\sqrt{2}\sigma_{y}})\right ] \nonumber \\ & = & ~(2{\pi})^3\sigma_{x}\sigma_{y}(k_{x}^2\sigma_{x}^2 + k_{y}^2\sigma_{y}^2)\exp[-2{\pi}^2(k_{x}^2\sigma_{x}^2+k_{y}^2\sigma_{y}^2 ) ] \ , .
\label{eqn : ftw}\end{aligned}\ ] ] by using @xmath223 instead of @xmath236 , we reduce computation time , but numerical estimates are less accurate ( with respect to estimates derived analytically ) .
we quantify the discrepancy between any two of our three methods ( @xmath223 , @xmath236 , or analytic ) using @xmath237 we use the analytically derived detection threshold @xmath238 in the denominator because the expectation value of @xmath96 is zero .
it also provides an intuitive way to describe the discrepancy at the detection threshold .
we find @xmath239 , while @xmath240 , increasing with @xmath241 .
( see figure [ fig : ftan ] . )
the discrepancy results from the fact that we analyze binned images . because the data are binned , we should actually compute @xmath223 using the equation @xmath242
if @xmath243 exhibits significant curvature in a bin , then @xmath244 within that bin ; the effect of this discrepancy in the final correlation value is proportional to the number of counts in the bin .
thus @xmath245 will increase with source strength , as shown in the right panel of figure [ fig : ftan ] .
if we wish to compute the error of the correlation value in a pixel , we calculate @xmath246 , saving computation time by using the analytic fourier transform @xmath247 .
the derivation of this function is similar to the derivation of @xmath223 . a new integral which appears is : @xmath248 one may solve this integral by parts , differentiating the term @xmath249 and integrating the term @xmath250 .
the integral @xmath251 has solution ( see , e.g. , gradshteyn & ryzhik 1980 , formula 3.952 - 5 ) : @xmath252 the final solution is @xmath253\exp[-{\pi}^2(k_{x}^2\sigma_{x}^2+k_{y}^2\sigma_{y}^2 ) ] .\ ] ] use of @xmath247 instead of @xmath254 leads to a less accurate accounting of the errors as derived using purely analytic methods , for reasons given above .
we find the magnitude of the discrepancy in error estimates , as a fraction of the detection threshold , to be similar to that seen above ( @xmath8 10@xmath255 ) ; if we use the analytic error in the denominator instead of the detection threshold , we find that the discrepancy is constant as a function of @xmath256 , at @xmath8 10@xmath255 . if exposure variations and the fov edge may be ignored in the computation of the background , then we may replace @xmath257 in eq .
( [ eqn : back2 ] ) with a background normalization factor , @xmath258 , which we derive here .
the average value of the mh function is zero .
hence @xmath258 may be derived by integrating the mh function over either @xmath31 or @xmath32 .
we choose the former : @xmath259^{\frac{1}{2}}\sigma_y}^{[2-\frac{x^2}{\sigma_{x}^2}]^{\frac{1}{2}}\sigma_y } dy \left [ 2 - \frac{x^2}{\sigma_{x}^2 } - \frac{y^2}{\sigma_{y}^2 } \right ] \exp\left(-\frac{x^2}{2\sigma_{x}^2}\right ) \exp\left(-\frac{y^2}{2\sigma_{y}^2 } \right ) .\ ] ] the limits of integration reflect that the core extends over an ellipse with axes @xmath260 and @xmath261 .
we reparametrize the integral using polar coordinates @xmath262 , after first mapping the boundary ellipse to a boundary circle using the transformation @xmath263 : @xmath264 \exp\left(-\frac{x^2+y'^2}{2\sigma_{x}^2}\right ) \\ & = & ~\frac{\sigma_{y}}{\sigma_{x}}\int_0^{2{\pi } } d\theta \int_0^{\sqrt{2}\sigma_{x } } dr r \left[2 - \frac{r^2}{\sigma_{x}^2}\right ] \exp\left(-\frac{r^2}{2\sigma_{x}^2}\right ) \\ & = & ~2{\pi}\frac{\sigma_{y}}{\sigma_{x } } \left[\int_0^{\sqrt{2}\sigma_{x } } dr 2 r \exp\left(-\frac{r^2}{2\sigma_{x}^2}\right ) - \int_0^{\sqrt{2}\sigma_{x } } dr \frac{r^3}{\sigma_{x}^2 } \exp\left(-\frac{r^2}{2\sigma_{x}^2}\right)\right ] \ , .
\label{eqn:1back}\end{aligned}\ ] ] the determinant of the jacobian of the transformation from @xmath265 is @xmath266 .
we evaluate the second integral using integration by parts : @xmath267 + \int_0^{\sqrt{2}\sigma_{x } } dr 2 r \exp\left(-\frac{r^2}{2\sigma_{x}^2}\right ) .
\label{eqn:2back}\ ] ] the second integral in eq .
( [ eqn:2back ] ) cancels with the first integral in eq .
( [ eqn:1back ] ) , leaving : @xmath268 \nonumber \\ & = & ~2{\pi}\frac{\sigma_{y}}{\sigma_{x } } \frac{2\sigma_{x}^2}{\exp(1 ) } \nonumber \\ & = & ~\frac{4{\pi}\sigma_{x}\sigma_{y}}{\exp(1 ) } \,.\end{aligned}\ ] ]
we associate an image pixel @xmath64 with a source if the significance @xmath68 of its correlation value @xmath66 is greater than a user - defined detection threshold significance @xmath74 : @xmath269 where @xmath77 is either @xmath270 , the convolution of the wavelet negative annulus @xmath32 with ( possibly cleansed ) image data @xmath39 , or the amplitude in an input background map .
@xmath69 is the probability sampling distribution ( psd ) for @xmath96 given @xmath77 .
( we are assuming here that the field of view is evenly exposed , which is true for the simulations we describe below . in a more general expression
, we would replace @xmath270 with @xmath271 . ) by defining the significance in this manner , we are making the assumption that the expected background counts amplitude is constant throughout the wavelet support , and that if we are computing the background , there are no source counts in the negative annulus . since the psd is dimensionless , it is independent of scale size ( e.g. the psd for @xmath77 would be the same if we were to double @xmath23 and @xmath24 and reduce the count rate by a factor of four ) . we replace @xmath77 with the equivalent quantity @xmath272 , the expected number of background counts within the spatial region spanned by the positive kernel of the wavelet , @xmath31 .
( @xmath272 is related to @xmath77 by multiplicative factors . )
the psd does not have analytic form when the total number of expected counts within the positive kernel of the wavelet , @xmath272 , is small ( @xmath190 1 ) ; this is demonstrated by damiani et al .
( 1997a ) .
like damiani et al . , we use simulations to estimate significances and detection thresholds ; we carry these simulations out both within the regime they examine and also for lower expected counts values . because this is a computationally intensive problem , we can not accurately estimate significances for each pixel if @xmath273 .
instead , we compare the value of @xmath66 in each pixel with the detection threshold @xmath274 , defined by @xmath275 here , we recast the equation using the variable @xmath272 , the number of expected counts in @xmath31 , as its use clarifies our description below of how we estimate @xmath276 .
to determine @xmath277 , we simulated over 50,000 1024@xmath1691024 flat - field images .
for each image , we : * randomly selected values @xmath278 from the range -10 @xmath279 3.25 ( we describe why we chose this upper limit below ) ; * determined the expected background amplitude @xmath280 in each image pixel ( @xmath281 pixels ) ; * sampled data in each pixel from the appropriate poisson distribution given rate @xmath282 ; * computed @xmath66 and @xmath272 for each image pixel ; * and recorded @xmath66 in bins of size @xmath283 = 0.2 , for -6.9 @xmath284 3.1 , with one bin being used for all values of @xmath285 -6.9 . from these distributions
@xmath286 , we can determine @xmath276 . because there is an inverse correlation between observed values of @xmath66 and @xmath272
, it is important to record @xmath286 and _ not _ @xmath287 .
use of the latter distribution leads to underestimated detection thresholds , and thus to larger numbers of false source detections than one would expect , given @xmath74 .
we determined 25 values of @xmath288 in each @xmath289 bin , for values of @xmath290 10@xmath10 , using the central 68% ( 17 values ) to estimate
one-@xmath7 " errors on @xmath87 .
we then fit these data with simple functions , minimizing the @xmath291 statistic .
these functions we use describe the observed detection thresholds well , except in the regime @xmath292 and @xmath293 , where we use a look - up table instead .
these fits allow us in principle to compute detection threshold for significances below our computational lower limit @xmath294 10@xmath10 ( such as for @xmath294 10@xmath295 , the significance corresponding to one false source pixel in an _ chandra _ hrc image ) . in the regime @xmath292 and @xmath296 , we compute @xmath297 using the function @xmath298 where @xmath299 in the regime @xmath300 , we compute @xmath301 using the function @xmath302 where @xmath303 this function is also used by damiani et al . to fit detection thresholds in this
counts regime , though their derived coefficients differ from ours . for values @xmath304
, we find that we must use the formula @xmath305 to compute the detection threshold , with @xmath306 and @xmath307 for our simulations , we chose @xmath308 3.25 as a upper limit because of the contention of damiani et al . that if @xmath309 3 , the psd probability sampling distribution is analytically representable as a gaussian with width @xmath310 .
( note that the value of @xmath272 that we use in this work is larger than that used by damiani et al . by a factor of @xmath311 . )
if this is the case , the significance is given by : @xmath312 \ , .
\label{eqn : anl}\end{aligned}\ ] ] we find , however , that if we use this formula in the regime @xmath309 3 , the derived values of @xmath276 are smaller than those predicted by eq .
( [ eqn : hi ] ) above .
thus , because it is more conservative , we use eq .
( [ eqn : hi ] ) to compute detection thresholds for all values @xmath309 0 .
if the data are cleansed ( see [ sect : back ] ) , then an exact calculation of @xmath132 $ ] will include non - zero covariance terms . in this section
, we derive these terms assuming that the data are cleansed only once , i.e. we compute the variance for @xmath313 , a background estimate made by convolving the wavelet negative annulus with data @xmath89 that are a mixture of raw data @xmath39 and first - iteration background estimates @xmath88 : @xmath314 where @xmath315 = @xmath316 . using eq .
( [ eqn : covar ] ) , @xmath317~&=&~\sum_{i ' } \sum_{j ' } ( n_{i , j}nw_{i - i',j - j'})^2 v[d_{i',j ' } ' ] \nonumber \\ & & + 2 \sum_{i ' } \sum_{i '' > i ' } \sum_{j ' } \sum_{j '' > j ' } ( n_{i , j}nw_{i - i',j - j ' } ) ( n_{i , j}nw_{i - i'',j - j '' } ) { \rm cov}[d_{i',j'}',d_{i'',j '' } ' ] \ , .
\nonumber\end{aligned}\ ] ] to calculate the variance , we must estimate both @xmath318 $ ] and @xmath319 $ ] .
the estimate of the former is given in eq .
( [ eqn : twoiter ] ) , derived assuming that the raw datum @xmath124 is sampled from a poisson distribution with variance @xmath124 , and that each pixel s raw datum is independently sampled .
estimation of @xmath319 $ ] is more complicated .
for the two - iteration background case , there are three possibilities : @xmath320 and @xmath321 ; @xmath320 and @xmath322 ( or vice - versa ) ; or @xmath323 and @xmath322 . assuming that the raw data are independently sampled , in the first case the covariance is zero . in the second case
, we can estimate the covariance using the approximation ( eadie et al .
p. 27 ) : @xmath324~&=&~\sum_k \sum_l \sum_{k ' } \sum_{l ' } \left(\frac{{\partial}d_{i',j'}}{{\partial}\mu_{k , l}}\right )
\left(\frac{{\partial}b_{1,i'',j''}}{{\partial}\mu_{k',l'}}\right ) cov[d_{k , l},d_{k',l ' } ] \nonumber \\ & = & ~\sum_k \sum_l \left(\frac{{\partial}d_{i',j'}}{{\partial}\mu_{k , l}}\right ) \left(\frac{{\partial}b_{1,i'',j''}}{{\partial}\mu_{k , l}}\right ) v[d_{k , l } ] \nonumber \\ & = & ~\sum_k \sum_l \left(\frac{{\partial}d_{i',j'}}{{\partial}d_{k , l}}\right ) \left(\frac{{\partial}b_{1,i'',j''}}{{\partial}d_{k , l}}\right ) v[d_{k , l } ] \nonumber \\ & = & ~\left(\frac{{\partial}b_{1,i'',j''}}{{\partial}d_{i',j'}}\right ) v[d_{i',j ' } ] \nonumber \\ & = & ~v[d_{i',j ' } ] \frac{\partial}{{\partial}d_{i',j ' } } n_{i'',j '' } \sum_{i ' '' } \sum_{j ' '' } nw_{i''-i''',j''-j ' '' } d_{i''',j ' '' } \nonumber \\ & = & ~v[d_{i',j ' } ] n_{i'',j '' } nw_{i''-i',j''-j ' } \nonumber \\ & = & ~d_{i',j ' } n_{i'',j '' } nw_{i''-i',j''-j ' } \nonumber \,.\end{aligned}\ ] ] in the above equations , @xmath325 represents the expectation value of the sampling distribution for @xmath39 ; for a poisson distribution with variance equal to the datum , @xmath325 will be equal to the datum . in the third case ,
we find ( while skipping some steps ) : @xmath326~&=&~\sum_k \sum_l \left(\frac{{\partial}b_{1,i',j'}}{{\partial}d_{k , l}}\right ) \left(\frac{{\partial}b_{1,i'',j''}}{{\partial}d_{k , l}}\right ) v[d_{k , l } ] \nonumber \\ & = & ~\sum_k \sum_l n_{i',j ' } nw_{i'-k , j'-l } n_{i'',j '' } nw_{i''-k , j''-l } v[d_{k , l } ] \nonumber \\ & = & ~\sum_k \sum_l n_{i',j ' } nw_{i'-k , j'-l } n_{i'',j '' } nw_{i''-k , j''-l } d_{k , l } \nonumber \,.\end{aligned}\ ] ] to demonstrate the effect of including covariance terms , we compute @xmath327 $ ] for a 50@xmath16950 subfield of the _ rosat _ pspc pleiades image , with @xmath23 = @xmath24 = 4 pixels . in figure
[ fig : covar ] , we show the ratio @xmath327_{\rm covar}{\slash}v[b_2]_{\rm no covar}$ ] .
we infer the following from this computation : ( 1 ) it is quantitatively unimportant at the location of sources and within voids , where the change in variance is @xmath190 1% ; ( 2 ) in the vicinity of a strong sources , the change in variance is @xmath8 10% ( with maximum
@xmath0 30% for this particular field , in the vicinity of an 800-count source ) , and is affected by source crowding ; and ( 3 ) including covariance terms increases the cpu time needed to compute @xmath327 $ ] by a factor @xmath129 , where @xmath130 and @xmath131 are the x- and y - axis lengths , respectively , in pixels .
snowden , s. l. , & kuntz , k. d. 1998 , cookbook for analysis procedures for rosat xrt observations of extended objects and the diffuse background , part i : individual observations ( nasa / gsfc : us rosat science data center ) , as carried out at selected scales @xmath137 .
parantheses indicate optional input or computations .
the local background map may be estimated or provided by the user ; a flow chart illustrating how it can be iteratively estimated using the input data ( and exposure map ) is given in figure [ fig : bgflow ] . ] ) , causing bumps " if the wavelet scale sizes @xmath328 , the characteristic psf size at a given pixel .
we use a 50@xmath16950 subfield of the _ rosat _ pspc pleiades cluster image in which @xmath329 3 pixels .
top : image and surface plot of the background estimate for @xmath23 = @xmath24 = 1 pixel .
middle : same as top , but for @xmath23 = @xmath24 = 2@xmath178 pixels .
bottom : same as top , but for @xmath23 = @xmath24 = 8 pixels . ] ) , causing rings " if sources are located within the negative annulus of the wavelet .
we use a 50@xmath16950 subfield of the _ rosat _ pspc pleiades cluster image .
top : image and surface plot of the background estimate after one iteration ( i.e. no cleansing " ) .
middle : same as top , but after two iterations .
bottom : same at top , but after three iterations . ] ) , created by first smoothing the data in a 50@xmath16950 subfield of the _ rosat _ pspc pleiades cluster image with a @xmath31 function of size @xmath23 = @xmath24 = 2 pixels , then subtracting the estimated background .
pixels associated with the strongest local maxima comprise the source cells that are used for source property estimation .
see figures [ fig : scex1 ] and [ fig : scex2 ] . ]
function of size @xmath23 = @xmath24 = 2 pixels , then subtracting the estimated background .
bottom : the source cell defined using the source counts image data ( [ sect : cell ] ) .
this cell , used in the estimation of source properties , contains nearly all , if not all , of the counts from this source . ]
function of size @xmath23 = @xmath24 = 2 pixels , then subtracting the estimated background . bottom : the source cells defined using the source counts image data .
the saddle point seen in the middle image defines the boundary between the cells ( [ sect : cell ] ) . ]
512 pixel image ( image a ) showing a 1-ksec observation by an idealized detector with effective area 1000 @xmath170 , a spatially invariant gaussian psf of width @xmath330 = 2.56 pixels , and an exposure map similar to that of the _ einstein _ ipc ( [ sect : ver1 ] ) . within this image
were placed 42 point and 4 extended sources .
the background is assumed to be locally variable , with amplitude @xmath8 10@xmath176 ct sec@xmath11 pix@xmath11 .
bottom : same as top , except the observation time is 10 ksec ( image b ) . ] , showing the cells generated in the vicinity of the largest extended source in image b. middle : same as top , except that now the source cells are defined using a source counts image with smoothing size 8@xmath178 pixels instead of 2@xmath178 pixels .
bottom : the normalized count - rate image for the extended source ( and three point sources ) , generated by creating a source counts image with smoothing size 2 pixels , then filtering out all pixels not contained in the source cell shown immediately above .
if this source were a galaxy cluster , the analyst could proceed to fit to these data ( respecting the caveats discussed in
[ sect : ver1 ] ) . ] principal axes and rotation angle for each source .
middle left and right : same as top left and right , with only the central 1@xmath3311@xmath332 portion of the image shown .
bottom left and right : same as top left and right , with only the central 30@xmath33330@xmath3 portion of the image shown . note : this is a bitmap image .
the original image may be downloaded from http://hea-www.harvard.edu/~pfreeman/f15.eps.gz . ] ) via the method of [ sect : nbkg ] .
the contour levels are 0 , 0.9 , 1.05 , 1.2 , 1.35 , and 1.5 counts , with darker areas having more counts .
since the estimated error in each pixel is @xmath8 0.01 counts , the perceived structure is real and is indicative of x - ray shadowing ( kashyap et al .
2001 , in preparation ) .
note : this is a bitmap image .
the original image may be downloaded from http://hea-www.harvard.edu/~pfreeman/f17.eps.gz . ]
15 pixels .
bottom : ellipses representing the 171 sources detected by our algorithm .
the ellipses , whose sizes are normally set by deriving the 1@xmath7 principal axes and rotation angle for each source , have minimum axis lengths of 10 pixels for greater visual clarity . ] and @xmath213 are continuous variables describing the wavelet function centered in pixel @xmath64 . the correlation value for pixel @xmath64 is computed by summing the product of the data in pixels @xmath208 and the integral of the wavelet function in those pixels .
the dotted line shows the boundary between the positive kernel @xmath31 and negative annulus @xmath32 of a mexican hat wavelet with @xmath334 1 pixel , while the solid line shows the extent over which the summation is carried out ; beyond this line , the amplitude of wavelet function is @xmath0 0 . ] for details ) . left : ratio of variances @xmath327_{\rm
covar}{\slash}v[b_2]_{\rm no~covar}$ ] , for a 50@xmath16950 subfield of the _ rosat _ pspc pleiades cluster image , analyzed with a wavelet with scale sizes @xmath23 = @xmath24 = 4 pixels .
light regions have values @xmath0 1 , while dark regions have values up to @xmath0 1.3 .
right : the 50@xmath16950 subfield of the _ rosat _ pspc pleiades image for which the ratio of variances shown at left was computed .
sources in this field correspond with light ( low - amplitude ) regions in the ratio image . ]
ll correlation ( @xmath66 ) & @xmath246 + + exposure - corrected & @xmath125i + $ ] + correlation ( @xmath110 ) & @xmath335)>_{i , j } + $ ] + & 2 @xmath336)>_{i , j}$ ] + + & @xmath337)>_{i , j}$ ] + + approximate & @xmath125 + b_{{\rm norm},i , j } < w^2{\star}e^2>_{i , j}$ ] + exposure - corrected & + correlation ( @xmath119 ) & + + normalized background ( @xmath82 ) & @xmath338 + [ tab : sderr ] lll x location & pix & @xmath339 + + y location & pix & @xmath340 + + bkg counts & ct & @xmath341 + + counts & ct & @xmath342 + + exposure & varies & @xmath343 + + count rate & ct @xmath344 exp unit & @xmath345 + + bkg count rate & ct @xmath344 exp unit & @xmath346 + [ tab : prop ] lll x location ( @xmath347 ) & pix & @xmath348 + + y location ( @xmath349 ) & pix & @xmath350 + + bkg counts ( @xmath351 ) & ct & @xmath352 $ ] + + counts ( @xmath353 ) & ct & @xmath354 $ ] + + exposure ( @xmath355 ) & sec & @xmath356 + + count rate ( @xmath357 ) & ct @xmath344 exp unit & @xmath358 + r_{s , c}^2 v[t_s])$ ] + + bkg count rate ( @xmath359 ) & ct @xmath344 exp unit & @xmath360 + r_{s , b}^2 v[t_s])$ ] + [ tab : error ] cccccc 1 & 2 & 10@xmath9 & 2 & fast ( eq . [ eqn : fastcor ] ) & 129 + 2 & 2 & 10@xmath9 & @xmath178 & fast & 136 + 3 & 3 & 10@xmath9 & 2 & fast & 130 + 4 & 3 & 10@xmath9 & @xmath178 & fast & 137 + 5 & 3 & 10@xmath9 & 2 & full ( eq . [ eqn : fullcor ] ) & 132 + 6 & 3 & 10@xmath9 & @xmath178 & full & 138 + 7 & 2 & 10@xmath176 & 2 & fast & 144 + 8 & 2 & 10@xmath176 & @xmath178 & fast & 148 + [ tab : numdet ] lc our algorithm ( test
8) & 148 + damiani et al .
( 1997b ) & 150 + micela et al .
( 1996 ) & 99 + wgacat & 127 + mpe & 100 + + common to : + all methods & 81 + our algorithm & wgacat & 27 + our algorithm & mpe & 12 + wgacat & mpe & 3 + our algorithm only & 28 + wgacat only & 16 + mpe only & 4 + [ tab : compare ] | wavelets are scaleable , oscillatory functions that deviate from zero only within a limited spatial regime and have average value zero , and thus may be used to simultaneously characterize the shape , location , and strength of astronomical sources .
but in addition to their use as source characterizers , wavelet functions are rapidly gaining currency within the source detection field .
wavelet - based source detection involves the correlation of scaled wavelet functions with binned , two - dimensional image data .
if the chosen wavelet function exhibits the property of vanishing moments , significantly non - zero correlation coefficients will be observed only where there are high - order variations in the data ; e.g. , they will be observed in the vicinity of sources .
source pixels are identified by comparing each correlation coefficient with its probability sampling distribution , which is a function of the ( estimated or _ a priori_-known ) background amplitude . in this paper
, we describe the mission - independent , wavelet - based source detection algorithm wavdetect , part of the freely available _ chandra interactive analysis of observations ( ciao ) _ software package .
our algorithm uses the marr , or
mexican hat " wavelet function , but may be adapted for use with other wavelet functions .
aspects of our algorithm include : ( 1 ) the computation of local , exposure - corrected normalized ( i.e. flat - fielded ) background maps ; ( 2 ) the correction for exposure variations within the field - of - view ( due to , e.g. , telescope support ribs or the edge of the field ) ; ( 3 ) its applicability within the low - counts regime , as it does not require a minimum number of background counts per pixel for the accurate computation of source detection thresholds ; ( 4 ) the generation of a source list in a manner that does not depend upon a detailed knowledge of the point spread function ( psf ) shape ; and ( 5 ) error analysis .
these features make our algorithm considerably more general than previous methods developed for the analysis of x - ray image data , especially in the low count regime .
we demonstrate the robustness of wavdetect by applying it to an image from an idealized detector with a spatially invariant gaussian psf and an exposure map similar to that of the _ einstein _ ipc ; to pleiades cluster data collected by the _ rosat _
pspc ; and to simulated _ chandra _ acis - i image of the lockman hole region . |
at present cp - violation is observed only in the @xmath4 and @xmath5 decays , and first indication of the effect in b - decays was recently reported @xcite .
another possible domain for cp - violation studies are still unseeing @xmath6 and @xmath7 decays , of which the latter must be a pure cp - violating process @xcite , because for three neutral pions only cp - odd states exist .
cp - violation in the @xmath7 decay can be parameterized in terms of @xmath8 parameter , which is defined as : @xmath9 one can estimate the decay branching ratio @xcite : @xmath10 .
the lowest existing experimental upper limit of @xmath11 was reported by cplear collaboration @xcite . in this
paper results of the study of the @xmath0 decay with the snd detector are presented .
the experiment was performed in 19961998 at vepp-2 m collider with snd detector @xcite .
the snd is a general purpose nonmagnetic detector .
its main part is a spherical electromagnetic calorimeter , consisting of 1632 nai(tl ) crystals .
the calorimeter energy resolution for photons is @xmath12{e(\mathrm{gev})}$ ] , the angular resolution is @xmath13 and solid angle coverage is close to @xmath14 of @xmath15 steradian @xcite .
presented analysis is based on experimental data collected in the center - of - mass energy region 9801040 @xmath16 , with most of the data , taken in the close vicinity of the @xmath17 peak .
the @xmath18 decays were used as a source of @xmath2 mesons .
experimental data corresponds to about @xmath19 produced @xmath20 mesons or @xmath21 @xmath22 decays .
the search for the @xmath0 decay was performed using the process @xmath23 in this process the @xmath24 having momentum of about @xmath25 and decay length of @xmath26 may either produce signals in the detector due to nuclear interaction in the calorimeter or decay in flight , or it can punch through the detector unseen . the detection efficiency of @xmath24 mesons in the calorimeter material was studied in the process @xmath27 with the probability of @xmath28 the @xmath24 produces a single cluster of hit crystals in the calorimeter , and with the @xmath29 probability more than one cluster .
the clusters produced by @xmath24 mesons are interpreted as `` photons '' by event reconstruction program .
the rest @xmath30 of @xmath24 mesons produce no signal in the calorimeter .
events with 6 or 7 reconstructed photons were used in the search for @xmath0 decay . in order to reject background caused by stray particles from the accelerator and cosmic events ,
constraints were imposed on total energy deposition ( @xmath31 ) and total momentum of events ( @xmath32 ) , where @xmath33 is a beam energy . to suppress cosmic background even further , the events , where the most of hit crystals could be fitted by a single straight track , were rejected . due to worse energy resolution near
the calorimeter edges the polar angle of all reconstructed photons was limited to @xmath34 .
remaining background comes mainly from two processes : @xmath35 and from ( [ ks3eq2 ] ) with @xmath24 producing extra 2 or 3 `` photons '' due to nuclear interaction or decay .
kinematic fitting based on @xmath36-method was performed for the events satisfying selection criteria described above . for each event
two hypotheses were checked : * @xmath37 : an event is due to the process ( [ ks3eq1 ] ) , i.e. there are 3 @xmath38-s from @xmath2 decay in the event ; * @xmath39 : an event is due to the process ( [ ks3eq2 ] ) , i.e. it contains 2 @xmath38-s from @xmath2 decay . as result of kinematic fitting the following parameters were evaluated : * @xmath40 and @xmath41 the chi - square values for the two hypotheses ; * @xmath42 and @xmath43 the momentum of reconstructed @xmath2 ; * @xmath44 and @xmath45 the polar angle of reconstructed @xmath2 ; * @xmath46 and @xmath47 raw invariant masses of photon pairs , attributed to pions during kinematic fitting . in order to isolate events of the process ( [ ks3eq1 ] )
, the following cuts were applied : * @xmath48 * @xmath49 * @xmath50 as a result significant part of background events of the processes ( [ ks3eq3 ] ) and ( [ ks3eq2 ] ) was rejected .
the total of 19 6-photon and 15 7-photon events survived the cuts .
the process ( [ ks3eq2 ] ) was used as a reference for the detection efficiency monitoring .
the events with 4 and 5 reconstructed photons were selected using the same primary cuts as for the process ( [ ks3eq1 ] ) .
after kinematic fitting the following additional cuts were applied : * @xmath51 * @xmath52 * @xmath53 the @xmath41 and @xmath43 distributions for experimental and simulated events of the process ( [ ks3eq2 ] ) are shown in fig .
[ ks3fig1 ] and fig .
[ ks3fig2 ] . further analysis of the process ( [ ks3eq1 ] ) candidates was performed separately for events with 6 and 7 photons .
the reference process ( [ ks3eq2 ] ) was studied in 4- and 5-photon classes respectively .
in the events of the process ( [ ks3eq1 ] ) with 7 reconstructed photons , one of the photons must be in fact a @xmath24 meson , thus additional cuts on its parameters can be imposed . in the analysis
it was required that energy deposition of @xmath24 cluster is at least @xmath54 , and spatial angle between the cluster and reconstructed @xmath2 direction is more than @xmath55 degrees .
the following cuts were based on specific for @xmath24 profiles of energy deposition in the calorimeter .
the parameters @xmath56 and @xmath57 were introduced to quantitatively describe the differences between the energy deposition profiles for photons and @xmath24 mesons .
the parameter @xmath56 @xcite represents the likelihood of a hypothesis that the transverse profile of energy deposition in a cluster was produced by a single photon .
the parameter @xmath57 has the same meaning , but for the longitudinal profile .
both parameters were studied in the process ( [ ks3eq2 ] ) , their distributions are shown in figs .
[ ks3fig3 ] and [ ks3fig4 ] .
the figures show that the requirement of either @xmath58 or @xmath59 reliably identifies @xmath24 meson .
the same requirement was applied to @xmath24 meson in 5-photon events of the reference process ( [ ks3eq2 ] ) . as a result
the number of selected events was @xmath60 for the process ( [ ks3eq1 ] ) and @xmath61 for the process ( [ ks3eq2 ] ) . the detection efficiencies @xmath62 and @xmath63 for the processes ( [ ks3eq1 ] ) and ( [ ks3eq2 ] ) were calculated by monte carlo simulation using unimod2 package @xcite
. the branching ratio of the @xmath0 decay was calculated as follows : @xmath64 an upper limit was obtained at @xmath14 confidence level : @xmath65
in 6-photon ( [ ks3eq1 ] ) candidates all detected particle must be photons .
thus the remaining ( [ ks3eq2 ] ) background can be suppressed by the following cuts on @xmath56 and @xmath57 : @xmath66 and @xmath67 for all six particles .
it was required also that raw invariant masses of photon pairs , reconstructed as pions are restricted to the range @xmath68 .
the same cuts were used in parallel analysis of 4-photon events of the reference process ( [ ks3eq2 ] ) .
the detection efficiencies of processes ( [ ks3eq1 ] ) and ( [ ks3eq2 ] ) were obtained by monte carlo simulation to be @xmath69 and @xmath70 respectively .
@xmath71 events of the process ( [ ks3eq1 ] ) and @xmath72 events of process ( [ ks3eq2 ] ) survived the cuts .
the upper limit at the confidence level of @xmath14 is : @xmath73
by using the following relation between numbers of found events of the processes ( [ ks3eq1 ] ) and ( [ ks3eq2 ] ) : @xmath74 the results of both analyses can be combined : @xmath75 the resulting upper limit according to eq.([ks3eq9 ] ) amounts to @xmath76 at the confidence level of @xmath14 .
the systematic error of the detection efficiency is determined mainly by imprecise simulation of @xmath24 nuclear interaction .
its estimated value is @xmath30 .
since the ( [ ks3eq1 ] ) branching ratio depends on the ratio of the detection efficiencies for the processes ( [ ks3eq1 ] ) and ( [ ks3eq2 ] ) , the common systematic error in the @xmath24 simulation cancels . the remaining systematic error in the efficiency ratio
is determined mainly by the accuracy of simulation of electromagnetic showers . in order to estimate this error , the process ( [ ks3eq3 ] )
was studied with the cuts similar to those in the analysis of process ( [ ks3eq1 ] ) .
resulting branching ratio of @xmath77 is @xmath78 lower than its world averaged value @xcite .
this difference was taken as an estimate of the systematic error of the ratio of the detection efficiencies .
the final result for the upper limit is then : @xmath79
the experiment was performed with snd detector at vepp-2 m @xmath1 collider .
the total statistics of @xmath19 @xmath20 mesons was analyzed . as a result , no candidate events of the @xmath0 decay were found .
the upper limit of the branching ratio of @xmath0 @xmath80 at the confidence level of @xmath14 was placed .
this work is supported in part by the russian fund for basic researches ( grant 96 - 15 - 96327 ) and stp `` integration '' ( no.274 ) .
99 manfred paulini , b lifetimes , mixing and cp violation at cdf , fermilab - pub-99/014-e , hep - ex/9903002 . l.maiani and n.paver , the second da@xmath81ne physics handbook , vol.i ( 1995 ) , p. 51 .
cplear collaboration , phys .
b 425 ( 1998 ) pp.391 - 398 . m.n.achasov et al . , experiments at vepp-2 m with snd detector , preprint binp 98 - 65 , ( 1998 ) .
m.n.aulchenko et al . , spherical nonmagnetic detector ( snd ) for electron - positron collider vepp-2 m , preprint binp 99 - 16 , ( 1999 ) , in russian .
v.m.aulchenko et al . , proc .
workshop on physics and detectors for da@xmath81ne , frascati , italy , april 9 - 12 ( 1991 ) , p.605 .
a.v.bozhenok , v.n.ivanchenko , z.k.silagadze , transverse energy profile of electromagnetic shower , nucl .
inst . and meth .
, a379 ( 1996 ) pp.507 - 508 .
a.d.bukin et al . , in proc .
workshop on detector and event simulation in high energy physics , the netherlands , amsterdam , 8 - 12 april 1991 , nikhef , p.79 .
c.caso et al .
( particle data group ) europ . phys .
* c 3 * 1 ( 1998 ) .
1 . the distribution of @xmath36 of kinematic fit for @xmath82 events .
the histogram represents the simulation , points experimental data .
2 . the distribution of the momentum of reconstructed @xmath2 meson in @xmath82 events . the histogram represents the simulation , points data .
3 . the @xmath56 distributions for photons ( clear histogram ) and for @xmath24 mesons ( shaded histogram ) in the process ( [ ks3eq2 ] ) .
4 . the @xmath57 distributions for photons ( clear histogram ) and @xmath24 mesons ( shaded histogram ) in the process ( [ ks3eq2 ] ) . | the search for cp - violating @xmath0 decay was performed with snd detector at vepp-2 m @xmath1 collider .
the total amount of data corresponding to 7 million produced @xmath2 mesons was analyzed .
no candidate events were found , giving an upper limit of the decay branching ratio : @xmath3 at 90% confidence level .
+ 13.25.es ; 11.30.er ; 13.65.+i ; 13.75.-n ; 14.40.aq + _ keywords : _
@xmath1 collisions ; kaon decays ; cp - violation ; upper limit |
an important manifestation of the activity inside the disks of gas - rich galaxies is their highly structured hi distribution , marked by cavities , shells and supershells .
first discovered in the milky way ( heiles 1979 , 1984 ) , such features are now known to exist in a number of spiral galaxies ( e.g. lehnert & heckman 1996 , irwin & seaquist 1990 , puche et al .
1992 , brinks & bajaja 1986 ) .
exceptionally huge hi arcs and loops extending across several kiloparsecs have been identified with greater clarity in the hi images of a number of edge - on spirals , such as ngc 5775 ( irwin 1994 ) , ngc 4631 ( rand & van der hulst 1993 ) , ngc 3044 ( lee & irwin 1997 , hereafter li97 ) and ngc 3556 ( m 108 , king & irwin 1997 , hereafter ki97 ) . these have been interpreted as expanding supershells because of a loop - like or circular appearance in projection and either a persistence over a wide velocity range or , in a few cases , as some evidence for expansion in position - velocity space .
two main classes of explanations for the supershells posit the source of their kinetic energy to be , respectively , internal and external to the parent galaxy .
the internal source model involves starbursts , driving stellar winds ( or superwinds ) and subsequent supernova explosions ( e.g. lehnert & heckman 1996 ) .
the chimney model " ( norman & ikeuchi 1989 ) , for example , attempts to explain disk - halo features and other halo gas via processes related to underlying star formation .
the association between extraplanar h@xmath1 filaments and star forming regions in the disk of ngc 891 and other correlations between halo emission and in - disk tracers of star formation ( dahlem et al .
1995 ; rand 1997 ) argue in favour of such models . if the presence of hi supershells is found to correlate with the existence of other halo gas , as might be expected in the chimney model , then stellar winds and supernovae are expected to be responsible for the hi supershells as well .
the main difficulty with the starburst model for hi supershells lies in the required input energies for the largest shells . using standard assumptions that the expanding supershells are in the post - sedov phase following an ` instantaneous ' injection of energy ( cf .
chevalier 1974 ) , hi supershells often require energy input from staggering numbers of spatially correlated supernova events .
this was realized early on for our own galaxy ( heiles 1979 , 1984 ) . for external edge - on galaxies ,
since we are selectively observing only the largest shells , the energy deficit problem is exacerbated . in some cases ,
hundreds of thousands of clustered supernovae are required ( e.g. ki97 , li97 ) , a conclusion which is not changed significantly if the energy is injected continuously over the lifetime of the shells . other evidence against star formation processes creating the hi shells is also emerging .
rhode et al . ( 1999 ) find no optical evidence for recent star formation in the numerous lower energy hi holes of holmberg ii and note that x - ray and fuv emission are also absent .
they conclude that supernovae have not played a part in the formation of the hi shells .
efremov et al .
( 1998 ) outline numerous other examples in which there appears to be no relation between hi shells and star formation .
they , as well as loeb & perna ( 1998 ) , propose that the hi shells are produced , instead , by gamma ray bursts .
the alternative external source hypothesis invokes infall of massive gas clouds on to the galactic plane , as a result of gravitational interaction with neighbouring galaxies ( see tenorio - tagle & bodenheimer 1988 ) .
this resolves the energy problem since input energy is then a function of the mass and velocity of the infalling cloud .
evidence in favour of this hypothesis comes from observations of high velocity clouds ( hvcs ) around our own milky way and the signatures of interaction in m 101 ( van der hulst & sancisi 1988 ) and ngc 4631 ( rand & stone 1996 ) .
it does , however , require that the galaxy be in some way interacting with a companion or , at least , that sufficiently massive clouds be in the vicinity .
recent observations are revealing galaxies which are apparently isolated , yet harbour extremely large hi supershells .
two striking examples are the nearby , sb(s)cd galaxy , ngc 3556 ( ki97 ) and the sbc galaxy , ngc 3044 ( li97 ) .
both of these galaxies exhibit radio continuum halos extending to @xmath2 kpc from the galactic plane and have a number of supershells requiring energies up to a few @xmath3 10@xmath4 ergs .
these supershells are too large and energetic to have been produced by conventional clustered supernovae . at the same time
, there appears to be no evidence for interaction or nearby companions , either .
we propose here a new explanation for hi supershells .
that is , that they have been formed by radio jets which plow through the interstellar medium ( ism ) , accreting ism gas and sometimes inflating bubbles .
this allows for an internal energy source for the hi shells , provides a natural explanation for any spatial symmetries seen in the hi features , and also resolves the energy problem . in sect . 2
, we provide arguments in favour of jet inflated hi bubbles , sect . 3 presents the model , and sect .
4 discusses the implications of this scenario .
seyferts are one class of disk galaxy for which several examples of the nucleus ejecting a radio jet pair have been found ( e.g. ulvestad & wilson 1984a , 1984b , kukula et al .
1995 , aoki et al .
likewise , several cases of jets occurring in normal spiral galaxies have been reported ( e.g. hummel et al . 1983 ) .
prominent examples include ngc 3079 ( de bruyn 1977 ) , ngc 5548 ( ulvestad et al .
1999 ) and circinus ( elmouttie et al . 1998 ) .
the total energy output from such nuclear activity can approach @xmath5 erg , assuming that the nuclear activity lasts @xmath0 years , and arises from accretion onto a supermassive black hole of @xmath6 at 10% of the eddington limit with the canonical 10% efficiency . in
normal " spiral galaxies , the central mass may be lower ; for example , the bipolar outflow from the sab(rs)cd galaxy , m 101 may be produced by a central @xmath7 10@xmath8 m@xmath9 black hole ( moody et al .
while jets are not always observed directly , the growing number of supermassive black holes inferred to be present in the nuclei of normal disk galaxies ( kormendy & richstone 1995 , van der marel 1999 ) , including late - type disks ( ho et al .
1997 ) , the milky way itself ( e.g. genzel et al .
1997 ) and gas - rich , large low - surface - brightness galaxies ( schombert 1998 ) , lend weight to the idea that many such galaxies may have undergone phases of nuclear activity accompanied with a bi - directional ejection of relativistic plasma jets , before entering a dormant nuclear phase .
recent studies have also revealed that the ejection of jets can take place at small angles to the large - scale galactic disk ( e.g. nagar & wilson 1999 ; kinney et al 2000 ) , plausibly leading to clear signatures of their dynamical interaction with the ism .
a classic case of such an interaction is the bow - shaped morphology of the radio lobe in the well known seyfert galaxy , ngc 1068 , which strongly suggests that the kiloparsec - scale radio jets are ploughing through the disk medium ( wilson & ulvestad 1987 , axon et al .
a dynamical interaction of the jets with the disk is also evident in the case of the spiral galaxy m 51 which is at the low end of nuclear activity scale . here
, the jets seem to have created a pair of ` radio lobes ' on opposite sides of the nucleus ; whereas one of the lobes is bow - shaped , the other one is actually resolved into a ring which is also detected in h@xmath1 emission ( ford et al .
another example is ngc 4258 whose vlbi jet has been ejected close to the galaxy disk ( see cecil et al .
1995 , herrnstein et al .
1997 , cecil et al .
large scale ( 14 kpc length ) anomalous arms " are also seen _ within _ the galactic disk and have been interpreted as manifestations of a larger - scale jet ( e.g. martin et al .
1989 , but see cox & downes 1996 ) . in m 101 ,
the bipolar outflow is also roughly confined to the disk , with the outflow not expected to extend beyond a height of 400 pc ( moody et al .
additional support for the jet - disk interaction hypothesis in these disk galaxies comes from the detection of shock - excited optical emission lines associated with the radio lobes , as discussed in some of the references cited above .
it is further interesting to note that radio bubbles and shells of synchrotron plasma have been discovered within the lobes of a few radio galaxies .
two spectacular examples are 3c310 ( van breugel & fomalont , 1984 ) and hercules a ( dreher & feigelson 1984 ) . in hercules a , in addition , peculiar dark shells , about 3 kpc in size , have been found straddling the nucleus along the radio outflow axes ( baum et al .
for the dark shells , baum et al . prefer an explanation in terms of expanding bubbles of hot gas ejected from the active nucleus along the radio axis , in agreement with plasmon - like ejection from the core , though they do not rule out other possibilities .
these cases provide a clue that the radio bubbles / shells may have been ` puffed ' up due to localized instabilities in the radio jets , occurring at distances of a few kiloparsecs from the active nucleus ( e.g. morrison & sadun 1996 ) .
should such instabilities arise also in the jets ejected within disk galaxies ( sect .
2.1 ) , it is quite plausible that the resulting increase in pressure at those locations would inflate bubbles and shells out of the ambient ism material which is hi rich , in the case of spirals .
such features , puffed out of the disk during the brief phase of jet instability , would remain visible past the radiative lifetime of the synchrotron plasma against the radiative and expansion losses , which is typically of order @xmath10 years . in ngc 3556 ,
the two most prominent supershells are located symmetrically near the opposite extremities of the galaxy .
ki97 have carried out a detailed mapping of the hi and radio continuum emission from this nearly edge - on galaxy located 11.6 mpc away .
the superimposition of 2 channels of hi emission , symmetrically placed in velocity about systemic , are shown in fig .
1 a. the most obvious features are two loops of hi , producing extensions to the east and west along the major axis . fig .
1b shows the hi emission from these velocities superimposed on the radio continuum map , the latter taken from entirely independent observations .
the abrupt density drop - off in the radio continuum map occurs roughly where the optical galaxy ends ( fig .
1a inset ) .
the salient features of this system most relevant to the present study ( see fig . 1 ; ki97 ) are : \(a ) two symmetrically placed giant hi loops are seen originating at a projected galactocentric radius of about 12 kpc on the east and west major axis . they appear to originate right at the opposite edges of the optical disk and extend , not perpendicular to the disk plane , but rather parallel to it and outwards .
\(b ) the giant hi loops or supershells are slightly bent into a z - symmetry .
\(c ) these hi loops are most obvious in velocity channels which are equally spaced about the systemic velocity of the galaxy ( though they extend over a larger velocity range , see ki97 ) .
\(d ) from their velocity dependent morphology , each of the loops shows evidence for shell expansion , but with only half of the shell present . in both cases , the receding side ( with respect to galaxy rotation ) is open .
\(e ) other smaller hi loops and extensions do exist at smaller galactocentric radii .
\(f ) the parameters of these two hi supershells are as follows : the eastern supershell has a mass , m = 5.9 @xmath3 10@xmath11 m@xmath9 , a radius , r = 3.2 kpc , an expansion velocity , @xmath12 = 51.7 km s@xmath13 , a kinematic age , @xmath14 = 6.0 @xmath3 10@xmath11 yr , a kinetic energy , e@xmath15 = 1.6 @xmath3 10@xmath16 ergs , and an implied input energy ( assuming instantaneous input from supernovae ) of e = 2.6 @xmath3 10@xmath4 ergs
. the parameters of the western supershell are : m = 2.3 @xmath3 10@xmath11 m@xmath9 , r = 1.85 kpc , @xmath12 = 41.4 km s@xmath13 , @xmath14 = 4.4 @xmath3 10@xmath11 yr , e@xmath15 = 3.9 @xmath3 10@xmath17 ergs , and e = 3.4 @xmath3 10@xmath18 ergs .
\(g ) the energy needed to create the eastern supershell alone would require a star cluster populated by @xmath19 ob stars , if supernovae and stellar winds are the drivers .
\(h ) the brighter and more spectacular eastern supershell has associated hi gas which extends as far as 30 kpc beyond the eastern edge of the optical disk and possibly much farther ( see fig .
1a inset ) .
the huge eastern hi extension appears to open up from one side of the eastern shell and extends fairly straight outwards in the radial , rather than the vertical direction as if formed in a short period of time in comparison to galactic rotation .
\(i ) the galaxy contains a large nonthermal radio halo with a scale height of @xmath2 kpc above the plane ( see also de bruyn & hummel 1979 , bloemen et al .
1993 ) .
\(j ) radio continuum emission appears to be associated with the hi loops , but is not spatially coincident with them .
the radio continuum emission extends farther out than the hi supershells , as if funnelling " along or through the hi features ( ki97 ; fig .
\(k ) ngc 3556 has no obvious interacting companions .
the brightest galaxy in the vicinity is the 16th magnitude , mcg+09 - 19 - 001 , 25@xmath20 in size , undisturbed in appearance and @xmath7 12@xmath21 to the east of ngc 3556 .
this is most likely a background source .
ngc 5775 , an interacting galaxy , shows symmetrically placed hi features ( lee et al .
2000 ) in the sense that hi extensions occur on opposite sides of the major axis in which there appears to be an underlying disturbance ; these features occur at similar galactocentric radii , in projection ( see also fig . 3 of irwin 1994 which shows 3 of the extensions ) .
ngc 2613 ( chaves & irwin 2000 ) also shows six hi extensions which occur in pairs symmetrically on either side of the major axis ; two of the pairs occur near the ends of the optical disk .
another potentially very interesting example is m 31 .
blitz et al . ( 1999 ) show that two massive hi clouds exist symmetrically placed on opposite sides of this galaxy .
they bear a remarkable resemblance to extragalactic radio jets .
the two clouds are redshifted by @xmath22 km s@xmath13 with respect to the systemic velocity of m 31 .
conceivably , this could be explained by ram pressure sweeping as m 31 falls through the igm towards the milky way .
as mentioned above , the galaxy ngc 3556 puts to a severe test both of the standard explanations for the supershell phenomenon , namely : ( i ) localized starburst ( generating intense stellar winds , followed by multiple supernova explosions ) , and ( ii ) infall of massive gas clouds on to the galactic disk .
the major difficulties faced by these models , as highlighted in ki97 , are : \(a ) * localized starbursts : * the energy needed to create the eastern supershell alone would require a star cluster populated by @xmath23 ob stars .
even for the recently discovered ` super star clusters ( ssc ) ' in some starburst galaxies , not more than @xmath24 ob stars are implied ( meurer et al . 1995 ) .
there is no indication of an ob association even remotely approaching the level of sscs at the locations of the supershells in ngc 3556 .
since the kinematic ages of the supershells are @xmath7 5 @xmath3 10@xmath11 yr , the starburst would have to have occurred within this time frame . yet
starburst durations and ob association ages are typically also of this order , so some evidence of the starburst might be expected to have survived .
no such evidence is presently found , however .
ngc 3556 does not show a markedly high supernova rate globally ( irwin et al . 1999 ) and independent low and high resolution radio continuum observations show no evidence for a source of energy at the bases of the supershells ( ki97 , irwin et al .
2000 ) .
\(b ) * infalling clouds : * several arguments have been advanced against this possibility ( ki97 ) .
firstly , this galaxy appears isolated , making the source of the putative clouds puzzlesome . secondly , to impart sufficient kinetic energy for the creation of the eastern supershell , an infalling cloud of mass @xmath25 would be required . a cloud this massive
would easily have been detected on the sensitive hi maps , yet no such cloud or clouds are seen .
( an exception would be if the clouds had very narrow line widths , low fractal dimensions and were optically thick , in which case they could have been missed due to low filling factors . )
thirdly , if an encounter with intergalactic clouds has occurred in the past , it would be necessary to postulate that two very massive clouds just happened to fall in parallel to the major axis at opposite ends of the galaxy , roughly at the same time , a scenario which seems implausible .
fourthly , we could instead suggest that a rain " of hi clouds is infalling ( since other high latitude hi structures are also seen in this galaxy ) , including two which fell in along the major axis at either end .
however , in order for the rain to be no longer visible , the infall would have to be completed over a timescale which is less than or equal to the age of the supershells , i.e. a few 10@xmath11 years .
this is unlikely since infall timescales are of order a dynamical timescale which is typically a few 10@xmath26 yr . while these arguments have been applied to ngc 3556 alone , the difficulty with internal energy from starbursting applies to a number of galaxies , as outlined in sect . 1 .
infalling clouds are certainly plausible , but run into difficulties for isolated galaxies such as ngc 3556 and ngc 3044 . barring new information coming to light on these galaxies , we therefore argue that a new model should be considered . as mentioned earlier , the duality of the expanding hi supershells in ngc 3556 , together with their locations at the opposite edges of the disk marked by steep density gradients , their z - symmetric deviations from spherical symmetry , as well as their anomalously large energy requirements ( sect .
2.3.1 ) , together lead us to argue that the origin of these two supershells is linked to the jet phenomenon . in sect .
2 we have noted several manifestations of jet instabilities giving rise to bubble - shaped features , or shells , in the form of optical nebulosities or nonthermal radio emission in normal spirals and radio galaxies . in the present context of hi supershells , we examine the possibility of the supershell being inflated due to localised instabilities in the radio jets as they plough through the ism of the disk of ngc 3556 .
the present discussion should only be viewed as a feasibility check based on energy considerations , and not as a detailed quantitative model for the supershells . even in the case of ngc 3556 ,
the standard mechanisms for the supershells ( sect .
1 ) may well have contributed at some level . according to our proposal ,
the two radio jets ejected close to the plane of the disk , during the active phase of the nucleus , undergo a rapid _ flaring _ as they encounter the region of large scale density decline near the outskirts of the galactic disk . consequently , both synchrotron plasma and the entrained thermal plasma deposited by the jets in these two regions get heated and the resulting high - pressure bubble of hot plasma undergoes a rapid expansion , sweeping the gas - rich ambient medium into the shape of the hi supershells .
an idealization to such a situation is the flaring of a jet crossing an ism / igm interface , which was first discussed in the context of radio galaxies analytically by gopal - krishna & wiita ( 1987 ) , followed by two - dimensional ( norman et al .
1988 , wiita et al .
1990 , zhang et al .
1999 ) and three - dimensional ( loken et al .
1995 , hooda & wiita 1996 , 1998 ) numerical simulations . the simulations by loken et al .
showed that intermediate power radio jets associated with wide - angle - tail ( wat ) sources undergo a rapid flaring upon crossing the ism / igm density discontinuity where a moderately supersonic jet becomes subsonic .
as shown by the numerical simulations , a density drop of a factor of just a few can cause such a flaring ( e.g. hooda & wiita 1996 ) .
observational evidence for the jet flaring comes from wide - angle - tail ( wat ) radio sources which are associated with the dominant members of rich clusters of galaxies ( e.g. odonoghue et al .
further evidence is provided , e.g. , by the recent radio / optical / x - ray study of the wat radio galaxy 3c465 in the abell cluster a2634 ( sakelliou & merrifield 1999 ) . particularly instructive for the present study
is the case of the active disk galaxy iras 04210 + 0400 ; the two @xmath27 kpc long radio jets associated with this disk galaxy are seen to flare up near the opposite optical boundaries of the galaxy ( holloway et al 1996 ) .
it may be noted that during their passage through the inner parts of the galactic disk , the jets are likely to encounter large ism density fluctuations .
however , these are less prone to disrupt the jets because of the higher jet thrust there , combined with the expected small spatial scale of the density fluctuations compared with the jets cross - section .
since the postulated radio jets in ngc 3556 are currently too weak for detection , a point which we re - address below , we shall adopt here the jet parameters inferred for another massive disk galaxy , ngc 4258 ( sect .
2.1 ) , and apply them to known conditions in ngc 3556 .
the jet velocity , as measured from the emission line gas , is _ at least _
v@xmath28 = 2000 km s@xmath13 which , together with a density of @xmath29 = 0.02 @xmath30 g @xmath31 , where @xmath30 is the mass of the proton , correspond to a kinetic luminosity of l@xmath28 @xmath7 10@xmath32 erg s@xmath13 ( falcke & biermann 1999 ) .
assuming that the ram pressure in the jet dominates over the internal thermal pressure , then across the shock front we require @xmath33 , where @xmath34 is the ambient gas density , and @xmath35 is the shock velocity . taking @xmath34 = 0.26 @xmath30 g @xmath31 at the positions of the supershells ( ki97 ) , we obtain @xmath35 = 435 km s@xmath13 .
this is slightly higher than the value of @xmath7 300 km s@xmath13 estimated for ngc 4258 .
the shock velocity puts a minimum timescale on the duration of the jet in this model , since there must be sufficient time for the jet to propagate out to the locations of the supershells at 12 kpc galactocentric radius .
thus , the minimum duration of the jets is 2.7 @xmath3 10@xmath11 yr for the parameters in this illustration .
the mechanical power of the jet would then be @xmath36 = @xmath37 = 5 @xmath3 10@xmath38 erg s@xmath13 , where the effective jet radius , r has a typical value of 1 kpc ( e.g. cecil et al 1995 ) .
this is similar to the jet power estimated for the well known spiral galaxy m 51 , which is at the low end of the scale of nuclear activity ( cf .
ford et al 1985 ) . in ngc 3556
, the integrated mechanical luminosity of the jet over its minimum lifetime is then @xmath7 4 @xmath3 10@xmath4 erg .
this lower limit is adequate to account for the observed kinetic energy associated with the larger eastern supershell ( sect .
here we have assumed that the efficiency for converting the input energy into the kinetic energy of the shell is of order 1% as usually taken for the multiple supernova model . for a higher efficiency , the required jet power can be lower .
the expansion of a typical bubble in the ism should then proceed similarly to previously modelled scenarios , the difference being the source of input energy .
since the input energy is much higher than conventional supernovae , the bubbles are more likely to achieve blowout , providing a natural conduit through which cosmic rays ( including those supplied by the jets ) can escape into the halo .
thus , the presence of supershells is expected to correlate with the existence of a nonthermal radio halo .
the locations of the anchor points of the two supershells in ngc 3556 suggest that the postulated blasting of each supershell would have occurred at a galactocentric distance which is just past the peak of the hi rotation curve ( see fig .
6 of ki97 ) .
the radially outward expansion of each supershell would then expose it to the velocity shear in the medium , exerting a side - ways wind pressure counter to the galactic rotation .
due to this , the half of the supershell towards the direction of the galactic rotation would be compressed and hence brightened , whilst the other half would be dragged out due to the velocity shear in the external medium and , consequently , dimmed .
such a deformation of the two supershells from spherical symmetry , in course of their expansion , would give rise to a z - symmetry , which is consistent with the observations ( sect . 2.3.1 ; ki97 ) .
the cooling time for the x - ray emitting sheaths around the jets of ngc 4258 , assuming unity filling factor , is only 4 @xmath3 10@xmath8 yr ( cecil et al . 1995 ) and therefore , once the jets have turned off , such a signature of their existence would disappear rapidly .
the lifetime of the synchrotron emitting particles in the jets is longer , typically from 10@xmath8 to 10@xmath11 yr . for ngc 4258
, it has been estimated to be between 1 to 5 @xmath3 10@xmath11 yr ( martin et al . 1989 ) .
however , shorter timescales are possible and will depend on local magnetic field strength and spectral index .
for instance , the magnetic field in the north - east radio jet in ngc 1068 , which extends to @xmath7 450 pc from the nucleus , is @xmath7 5 @xmath3 10@xmath39 gauss and the spectral index is -1 ( wilson & ulvestad 1987 ) .
this gives a synchrotron lifetime of only 1.5 @xmath3 10@xmath40 yr .
the particle lifetime is expected to be shorter in shocks where the magnetic field is compressed .
thus , we expect the radio jet in ngc 3556 to fade over a timescale less than the ages of the supershells .
similar arguments apply to the radio core which is expected to decay faster than the jets , once the nuclear activity has switched off , given the likelihood of a stronger magnetic field and flatter spectral index in the core .
even if some radio emission from a core were to persist after the activity ceases , it may be below the rms noise level of the available maps .
chary & becklin ( 1997 ) , for example , estimate the radio luminosity of the core of ngc 4258 , which is known to have a supermassive central object of mass 3.6 @xmath3 10@xmath11 m@xmath9 and a vlbi jet , to be l@xmath41 = 100 l@xmath9 .
the 2@xmath42 noise level of the radio maps of irwin et al .
( 2000 ) corresponds to a radio power of 6.8 @xmath3 10@xmath43 w hz@xmath13 .
integrating over 10@xmath44 hz , this yields a radio luminosity of 180 l@xmath9 .
thus , even if ngc 3556 harbours a radio core of the same luminosity as ngc 4258 , it would not have been detected in the observations mentioned above .
in this study we have sketched a scenario whereby radio jets ejected during an active nuclear phase in the life of a large spiral galaxy could inflate out of the disk exceptionally large shells , i.e. supershells of neutral hydrogen gas . when applied to the scd galaxy , ngc 3556
, this model can account for each of the several features enumerated in sect .
2.3.1 . in the context of the hi supershells seen in this galaxy , this scheme appears to fare distinctly better than either of the two standard models for supershells which invoke either a starburst induced superwind or an infall of external gas clouds on to the galaxy disk .
thus , the jet - ism interaction scenario for disk galaxies could effectively supplement the other two mechanisms already proposed for the supershell phenomenon , with a greater relevance whenever the shells are extraordinarily large as well as energetic and occur at large galactocentric distances .
the energy requirement is no longer a major issue in the present model .
even a modest energy input from a 10@xmath11 m@xmath9 compact object at the galactic nucleus , such as that observed in ngc 4258 , is more than adequate to supply the input energy required for the supershells of ngc 3556 .
this model can therefore account for hi supershells which are difficult to form via conventional processes involving massive star formation , or those occurring in galaxies lacking potentially interacting companions .
larger supershells ( or parts thereof , given the probable development of instabilities ) may also be predicted , since known central masses and/or accretion rates may well surpass those considered here .
furthermore , the present model can account for the symmetric hi features in galaxies .
for instance , it is more likely that bubbles will be inflated by the propagating jets at galactocentric radii where the ism density has declined to sufficiently low values for blow - out . at the same time
, symmetry is not a generic outcome our model , since local shocks , dictated by the interaction of the jet with local density perturbations , can be important in determining where the bubbles will be inflated .
for example , known bubbles in the lobes of radio galaxies ( sect .
2.2 ) are not always seen to be equidistant on either side of the nucleus . in our sample illustration ( sect .
3.2 ) , the jets must exist for long enough to reach large galactocentric radii ( 3 @xmath3 10@xmath11 yr ) but need to inject energy over only a fraction of the lifetime of the supershells ( 2 @xmath3 10@xmath11 yr , or even less , if the efficiency exceeds 1% ) .
the jets presumably switch off thereafter , with the synchrotron emitting particles decaying in 10@xmath8 - 10@xmath11 yr .
it is difficult to predict whether jets and hi supershells should co - exist in spirals .
nuclear activity in galaxies is commonly believed to be a transient phenomenon , lasting for @xmath45 years ( e.g. eilek 1996 , scheuer 1995 ; soltan 1982 ; haehnelt & rees 1993 ; ho et al 1997 ) .
if agn activity occurs on timescales longer than a few @xmath3 10@xmath11 yr , this would suggest that there might be disk galaxies in which both phenomena should be observed at the same time .
however , such nuclear activity timescales are likely to be more representative of radio galaxies ; agn activity in spirals is probably shorter lived .
an important next step is to consider what jet parameters are required to reach large , or at least kpc - scale radii , in traversing the typical ism of a spiral galaxy . if jets in spirals both recur and precess , it may be possible to inflate bubbles over a variety of galactocentric radii and azimuthal angles
. however , it is unlikely that the frothy " nature of the hi seen in galaxies such as holmberg ii or the holes in m 31 could be produced by such jets directly without seeing some lingering evidence for the jets as well .
also , the jet phenomenon can not directly account for features which correlate with the star forming disk , such as some known radio and h@xmath1 halos .
if there is a connection , it is more likely to be indirect .
for example , once the expansion of the postulated jet - blown supershells is halted due to the gravitational pull of the host galaxy , their segments ( i.e. hi clumps ) would eventually shower back on to the galactic disk , giving rise to secondary shells , bubbles and cavities in the hi component of the disk .
the impact could also trigger sporadic bursts of star formation , especially in the outer disk , as seen , e.g. , in gas - rich low - surface - brightness galaxies ( cf .
oneil et al .
1997 ) which are thought to be the present - day remnants of quasars ( e.g. schombert 1998 ) .
thus , much after its cessation , the nuclear activity in gas - rich galaxies may continue to influence the evolution of their disks .
aoki k. , kosugi g. , wilson a. s. , yoshida m. 1999 , apj 521 , 565 axon d. j. , marconi a. , capetti a. , et al .
1998 , apj 496 , 75 baum s. a. , odea c. p. , de koff s. , et al .
1996 , apj , 465 , l5 blitz l. , spergel d. n. , teuben p. j. , hartmann d. , burton w. b. 1999 , apj 514 , 818 bloemen h. , duric n. , irwin j. a. 1993 , in 23rd international cosmic ray conference , eds .
d. a. leahy , r. b. hicks , d. venkatesan ( new jersey , world scientific ) , p. 279 brinks e. , bajaja e. 1986 , a&a 169 , 14 cecil g. , wilson a. s. , de pree c. 1995 , apj 440 , 181 cecil g. , de pree c. g. , greenhill l. j. , moran j. m. , dopita m. a. 1998 , aas 193.0710 chary r. , becklin e. e. 1997 , apj 485 , l75 chaves t. , irwin j. a. , in gas and galaxy evolution , asp conf .
ser . , ed .
j. e. hibbard , m. p. rupen , j. h. van gorkom , in press chevalier , r. a. 1974 , apj 188 , 501 cox p. , downes d. 1996 , apj 473 , 219 dahlem m. , lisenfeld u. , golla g. 1995 , apj 444 , 119 de bruyn a. g. 1977 , a&a , 58 221 de bruyn a. g. , hummel e. 1979 , a&a 73 , 196 dreher j. w. , feigelson e. d. 1984 , nature 308 , 43 efremov y. n. , elmegreen b. g. , hodge p. w. 1998 , apj 501 , l163 eilek j. a. 1996 , in energy transport in radio galaxies and quasars , asp conf .
100 , ed . p.e .
hardee , a.h .
bridel , j.a .
zensus , ( san francisco : astr .
pacific ) , 281 elmouttie m. , haynes r. f. , jones k. l. , sadler e. m. , ehle m. 1998 , mnras 297 , 1202 falcke h. , biermann p. l. 1999 , a&a 342 , 49 ford h. c. , crane p. c. , jacoby g. h. , lawrie d. g. , van der hulst j m. 1985 , apj 293 , 132 genzel r. , eckart a. , ott t. , eisenhauer f. 1997 , mnras 291 , 219 gopal - krishna , wiita p. j. 1987 , mnras 226 , 531 haehnelt m. , rees m. 1993 , mnras 263 , 168 heiles c. 1979 , apj 229 , 533 heiles c. 1984 , apjs 55 , 585 herrnstein j. r. , moran j. m. , greenhill l. j. , et al .
1997 , apj 475 , l17 , erratum , apj , 482 , l113 ho l. c. , filippenko a. v. , sargent w. l. w. , 1997 , apj 487 , 568 holloway a. j. , steffen w. , pedlar a. , et al .
1996 , mnras 279 , 171 hooda j. s. , wiita p. j. 1996 , apj 470 , 211 hooda j. s. , wiita p. j. 1998 , apj 493 , 81 hummel e. , van gorkom j. h. , kotanyi c. g. 1983 , apj 267 , l5 irwin j. a. 1994 , apj 429 , 618 irwin j. a. , seaquist e. r. 1990 , apj 353 , 469 irwin j. a. , english j. , sorathia b. 1999 , aj 117 , 2102 irwin j. a. , saikia d. j. , english j. 2000 , aj 119 , 1592 king d. , irwin j. a. 1997 , new astr . 2 , 251 ( ki97 ) kinney a. l. , 2000 , apj
( july 2000 issue ) kormendy j. richstone d. 1995 , araa 33 , 581 kukula m. j. , pedlar a. , baum s. a. , odea , c. p. 1995 , mnras 276 , 1262 lee s .- w . , irwin j. a. 1997 , apj 490 , 247 ( li97 ) lee s .- w .
, irwin j. a. , dettmar r .- j . ,
2000 , in prep .
lehnert m. d. , heckman t. m. 1996 , apj 462 , 651 loeb a. , perna r. , 1998 , apj 503 , l35
loken c. , roettiger k. , burns j. o. , norman m. 1995 , apj 445 , 80 martin p. , roy j .-
, noreau l. , lo k. y. 1989 , apj 345 , 707 meurer g. r. , heckman t. m. , leitherer c. , et al .
1995 , aj 110 , 2665 moody j. w. , roming p. w. a. , joner m. d. , et al .
1995 , aj 110 , 2088 morganti r. , killeen n. e. b. , ekers r. d. , oosterloo t. a. 1999 , mnras 307 , 750 morrison p. , sadun a. 1996 , mnras 278 , 265 nagar n. m. , wilson a. s. 1999 , apj 516 , 97 norman m. l. , burns j. o. , sulkanen m. 1988 , nature 335 , 146 norman c. a. , ikeuchi s. 1989 , apj 345 , 372 odonoghue a. a. , eilek j. a. , jones j. m. 1993 , apj 408 , 428 oneil k. , bothun g. d. , schombert j. , cornell m. e. , impey c. d. 1997 , aj 114 , 2448 puche d. , westpfahl d. , brinks e. , roy j .- r .
1992 , aj 103 , 1841 rand r. j. 1997 , in the interstellar medium in galaxies , ed .
j. m. van der hulst ( dordrecht : kluwer ) , 105 rand r. j. , stone j. m. 1996 , aj 111 , 190 rand r. j. , van der hulst j. m. 1993 , aj 105 , 2098 , erratum , aj , 107 , 392 rhode k. l. , salzer j. j. , westpfahl d. j. , radice l. a. 1999 , aj 118 , 323 sakelliou i. , merrifield m. r. 1999 , mnras 305 , 417 scheuer p. a. g. 1995 , mnras 277 , 331 schombert j. 1998 , aj 116 , 1650 soltan a. 1982 , mnras 200 115 tenorio - tagle g. , bodenheimer p. , 1988
, ara&a 26 , 145 ulvestad j. s. , wilson a. s. 1984a , apj 278 , 544 ulvestad j. s. , wilson a. s. 1984b , apj 285 , 439 ulvestad j. s. , wrobel j. m. , roy a. l. , et al . 1999 , apj 517 , l81 van breugel w. , fomalont e. b. 1984 , apj 282 , l55 van der marel r. p. , 1999
, aj 117 , 744 van der hulst t. , sancisi r. 1988 , aj 95 , 1354 wiita p. j. , rosen a. , norman m. l. 1990 , apj 350 , 545 wilson a. s. , ulvestad j. s. 1987 , apj 319 , 105 zhang h .- m . , koide s. , sakai j .-
1999 , pasj 51 , 449 | taking a clue from the pair of hi supershells found in the scd galaxy ngc 3556 ( m 108 ) , we propose a new mechanism for the origin of hi supershells in gas - rich massive galaxies . in this scenario ,
the two supershells were inflated out of the neutral hydrogen disk due to the localised flaring of a pair of radio lobes formed by the jets ejected from the nucleus during an active phase about @xmath0 years ago , but have faded away by now .
it is shown that the salient features of this supershell pair , such as their symmetrical locations about the galactic centre , the anomalously large energy requirements , the large galacto - centric distances , as well as the z - symmetric hemispherical shapes , find a more natural explanation in terms of this scenario , as compared to the standard models which postulate either a massive starburst , or the infall of external gas clouds .
other possible implications of this hypothesis are briefly discussed . |
an intriguing phenomenon in diluted magnetic semiconductors ( dms ) is the formation of bound magnetic polarons @xcite .
the bound magnetic polaron ( bmp ) is a local ordering of magnetic moments induced by the exchange interaction with a localized carrier .
most bmp studies , both experimental and theoretical , have been performed so far in a regime of the mean - field approach ( when @xmath4 ) @xcite . here
, @xmath5 is the concentration of magnetic ions , @xmath6 is the concentration of cations , and @xmath7 is the localization radius .
this approach implies that the localized carrier interacts with a infinite number of @xmath8 ions , and hence it is typically applicable for high and moderate manganese concentrations ( @xmath9 ) . at present systems with a countable number of magnetic moments
attract growing interest @xcite .
this corresponds to the very dilute regime @xmath10 , i.e. when on average one carrier interacts with one ( or few ) mn ion(s ) .
clearly , at this condition the mean - field approach is not an appropriate model . up to now bound magnetic polarons
have not been observed in a very dilute regime .
this is because their observation requires the polaron formation time @xmath11 to be shorter than the life time @xmath12 of nonequilibrium carriers , @xmath13 . in dms with @xmath14 this condition is fulfilled .
however , in samples with lower mn concentrations @xmath15 the polaron formation time is of the order @xmath16 s @xcite .
therefore , the bmp formation process is interrupted by exciton recombination ( with characteristic time @xmath17
s ) @xcite . in this paper
we study bmp in bulk @xmath0 samples with very low mn concentration ( @xmath1 ) , i.e. at the condition @xmath10 . in order to avoid the above problem of bmp detection in the very dilute regime we use resonant ( selective ) excitation of donor - acceptor ( da ) pairs with transition energies @xcite @xmath18 here , @xmath19 and @xmath20 are the binding energies of donors and acceptors respectively
, @xmath21 is the band gap and @xmath22 is the dielectric constant in @xmath0 . as follows from eq .
( [ eq_da ] ) by changing the excitation energy @xmath23 one can selectively excite pairs with a given da separation @xmath24 . because of the weak overlap [ @xmath25 of the donor and acceptor wavefunctions , the pairs with large @xmath24 are characterized by a long recombination time @xmath26 . as a result the bound magnetic polaron forms before the recombination occurs . without polaronic effects such resonantly excited pairs
should manifest themselves in emission spectra as a narrow line of width within either the inverse recombination time ( @xmath27 ) or the laser line width .
the bulk @xmath0 samples with mn concentrations @xmath28 have been grown by the bridgman technique .
the crystals are nominally undoped .
the concentration of residual impurities ( both donors and acceptors ) is about @xmath29 @xmath30 , which is much lower than the mn concentration @xmath31 . as grown crystals are usually p - type due to excess of te or cd - vacancies .
note , for acceptors with the bohr radius @xmath32 and @xmath33 ( @xmath34 @xmath30 ) the relationship is @xmath35 . in the experiments we used @xmath36 mm pieces cleaved out along the ( 110 ) plane from the massive monocrystals .
all measurements were carried out at a temperature @xmath37 k. external magnetic fields were applied either perpendicular to the sample plane ( faraday geometry ) or in the sample plane ( voigt geometry ) .
the photoluminescence ( pl ) was excited by a he - ne laser or a tunable ti - sapphire laser pumped by an ar - ion laser .
the excitation density was about 5 @xmath38 .
the laser beam was directed onto the samples at the angle close to the normal ( axis [ 110 ] ) , and the emission was registered in the backscattering geometry .
a polarizer ( in combination with a quater - wave plate ) was ( were ) used to linearly ( circularly ) polarize the excitation .
the degree of pl polarization was detected with use of a photo - elastic modulator and a two - channel photon counter .
sample excited by a he - ne laser ( @xmath39 ev ) .
( b ) pl spectra under quasiresonant excitation , the excitation energies ( @xmath40 ) being shown in the panel .
arrows indicate emission bands which are related to the bmp formation .
pl in the same spectral range but under nonresonant excitation [ as in ( a ) ] is given in the upper part for comparison .
( c ) pl spectra of da pairs recorded in @xmath41 and @xmath42 circular polarizations in an external magnetic field @xmath43 t applied in faraday geometry .
the excitation conditions are the same as in ( a ) .
( c ) degree of circular polarization vs magnetic field detected at the da band . ]
we found qualitatively very similar behavior in all samples studied . in the following we present experimental results for one piece cleaved out from the bulk @xmath44 sample .
figure [ fig1](a ) shows the pl spectrum excited above the cdte band gap @xmath21 ( he - ne laser with @xmath39 ev ) .
the high - energy line ( x ) is attributed to the bound exciton . at low - energy part of the pl spectrum we observe a broad band associated with recombination of da pairs and its phonon replicas .
these replicas are separated by 20.5 mev , which corresponds to the lo - phonon energy in cdte @xcite .
application of external magnetic fields in faraday geometry results in in the circularly polarized emission already in low fields [ fig .
[ fig1](c ) ] .
such a behavior is typical for magnetic semiconductors .
figure [ fig1](d ) demonstrates the degree of circular polarization vs magnetic field detected at the da band .
it saturates rapidly on a 75% level . in order to study bound magnetic polarons
we use the site selective spectroscopy @xcite .
we resonantly excite da pairs and record bmp spectra shifted towards lower energies with respect to the excitation energy @xmath40 ( the stokes shift ) .
a set of pl spectra for different @xmath40 is presented in fig .
[ fig1](b ) .
a new emission band ( indicated by arrows ) , being observed on the top of the da photoluminescence , clearly follows the laser line which is a fingerprint of the bmp formation .
the most pronounced results have been obtained for @xmath45 ev , as also shown in fig .
[ fig2 ] . in order to avoid scattered light from the laser we detect such bmp spectra at the first phonon replica . in zero magnetic field
the bmp spectrum represents itself as an asymmetric band with the 5-mev - tail towards low energies [ fig .
[ fig2](a ) ] .
our estimations ( given below ) show that this value is in agreement with the bmp energy as well as with its root - mean - square fluctuations .
such a behavior is an essential deviation from the case of high mn concentrations ( @xmath4 ) , where the stokes shift is larger than the line width , giving rise to a distinct emission line shifted from the excitation by the bmp energy @xcite .
ev ) in zero magnetic field .
( b ) bmp and bmp - polarization spectra obtained under the same excitation conditions as in ( a ) but in magnetic field @xmath43 t applied in faraday geometry .
( c ) degree of circular polarization vs magnetic field detected at different energies .
( d ) bmp spectrum obtained in magnetic field @xmath43 t applied in voigt geometry when the polarization axis of the excitation ( @xmath46 ) is parallel to the field direction ( @xmath47(a ) ] .
a new line appears at the resonant energy ( i.e. , there is no stokes shift at all ) @xcite . with growing magnetic field the amplitude of this line increases and the width reduces greatly .
in strong enough magnetic fields ( @xmath48 t ) the width saturates on a @xmath2-@xmath3-level [ see also fig .
[ fig2](b ) ] .
the circular polarization of this line is about 65% , and its dependence on the magnetic field [ fig .
[ fig2](c ) ] follows that for the da emission under nonresonant excitation [ fig .
[ fig1](d ) ] . one more lineament of this new line : its intensity reduces under excitation by the light being linearly polarized along the magnetic field direction applied in voigt geometry [ see fig .
[ fig2](d ) ] .
however , a deviation of the field direction from the sample plane or the polarization axis from the field direction restores the amplitude of this line . as we show in sec .
[ sec3 ] , the evolution of pl spectra presented in fig .
[ fig3](a ) can not be described in terms of the mean - field theory , and hence this requires an alternative approach .
first , we discuss the asymmetric band observed in zero magnetic field .
resonant excitation generates da pairs with a certain energy @xmath49 depending on donor - acceptor separation @xmath24 [ pairs ( i ) and ( iii ) in fig .
[ fig4](a ) ] . da pairs with transition energies different from the excitation energy @xmath50 are not excited [ pair ( ii ) in fig .
[ fig4](a ) ] .
the observation of the asymmetric band in fig .
[ fig2](a ) we attribute to the bmp formation accompanied by relaxation to the ground state .
the dominant contribution to the polaron shift arises from the hole .
this is because the absolute value of the @xmath51-@xmath52 exchange constant is larger than the absolute value of the @xmath53-@xmath52 exchange , and the bohr radius of the acceptor is smaller than that of the donor @xcite . in this approach
the line width is explained by the dispersion of the polaron shifts induced by static fluctuations of mn concentration .
for instance , pair ( iii ) in fig .
[ fig4](a ) shows a larger polaron shift than pair ( i ) .
because on average less than one mn falls into the volume restricted by the acceptor bohr radius , the root - mean - square fluctuations appear to be comparable with the polaron shift .
its value can be estimated from the line width in fig .
[ fig2](a ) , @xmath54 mev .
strong enough magnetic fields suppress the polaron formation @xcite , and one would expect the disappearance of the low - energy tail .
however , our experiments show the opposite behavior : this tail preserves and a new narrow intense line appears .
we suggest the following explanation .
acceptor ( donor ) states are split in an external magnetic field .
the photon absorption by the fixed da pair leads to the population of one of the zeeman sublevels [ fig .
[ fig4](b ) ] . relaxation from the higher - lying sublevels to the lowest sublevel [ pairs ( v ) and ( vi ) in fig .
[ fig4](b ) ] results in the low - energy tail .
the characteristic width of this tail does not change with magnetic fields because the zeeman splittings disperse in the same manner as the polaron shifts , i.e. due to the static fluctuations of the mn concentration . in case when the lowest zeeman sublevel of da pairs is excited [ pair ( iv ) in fig .
[ fig4](b ) ] a further relaxation is not possible , and the narrow line is observed at the excitation energy . with growing magnetic fields the width of this line should decrease and saturate in fields when the magnetization of magnetic ions saturates ( @xmath48 t ) .
it is now clear why the narrow peak is suppressed when the polarization axis of the excitation light is exact parallel to the magnetic field direction : due to optical selection rules the excitation to the ground state [ pair ( iv ) in fig .
[ fig4](b ) ] is forbidden in this configuration .
however , even a small deviation from this geometry restores the perpendicular - to - field component of the polarization vector , and the peak rises up .
in order to calculate the pl spectra under selective excitation one needs to find the distribution function of the stokes shift induced by the polaron formation in zero magnetic field . on the other hand , in case of strong magnetic fields the distribution function is determined by the zeeman splitting dispersion . the hole ( electron ) spins interact with mn ions via the @xmath51-@xmath52 ( @xmath53-@xmath52 ) exchange interaction . in case of localized carriers
this interaction can be written as @xmath55 where @xmath56 and @xmath57 are the electron and hole spin operators , @xmath58 and @xmath59 are their wavefunctions , respectively .
@xmath60 is the spin operator of the @xmath61-th mn ion , and @xmath62 is a radius - vector assigning mn location in the crystal .
the exchange constants in @xmath0 are @xmath63 @xmath64 and @xmath65 @xmath66 .
the donor and acceptor bohr radii are equal approximately 60 and 10 , respectively .
this implies that the spin splitting of the mn ion removed by the bohr radius from the acceptor ( donor ) is 1.28 mev ( 1.5 @xmath3 ) . in other words , at a temperature @xmath37 k ( @xmath67 mev ) the electron contribution can be neglected , and all mn spins in the vicinity of the acceptor are aligned along the exchange field of the hole .
the hole exchange energy in the ground state is @xmath68 .
its mean value averaged over all mn configurations is @xmath69 , where @xmath70 @xmath71 is a concentration of cations .
for @xmath72 it is equal @xmath73 mev . the root - mean - square fluctuation @xmath74 is equal @xmath75 mev . and the mean value of the hole energy in the fluctuating fields of nonpolarized mn ions @xmath76 is equal @xmath77 mev .
it follows from these estimations that the fluctuations of the exchange energy and its mean value are nearly the same . therefore the gaussian statistics and , as consequence , the mean - field approach is not applicable . in case of low mn concentrations
it is difficult to obtain an analytical solution for the stokes shift distribution function .
therefore , in order to calculate the pl spectra under selective excitation we use numerical computing based on the monte - carlo method .
first , a random distribution of magnetic ions in a sphere with a radius equal to ten acceptor bohr radii @xmath78 is generated [ see also fig .
[ fig5](b ) ] @xcite .
each mn ion has a random spin orientation in accordance with the equilibrium distribution for a given temperature and magnetic field strength .
then , the total exchange field acting on the hole bound to the acceptor placed in the center of the sphere is calculated .
after that , the eigenstates of the hole in this total exchange field and the transition probabilities on these states are evaluated .
polaron energies are determined for each generated mn distribution by adding the exchange filed of the hole acting on mn ions to the external magnetic field .
the stokes shift is calculated as an energy difference between the initial and final states .
finally the distribution function of the stokes shift is obtained after summation over 50000 realizations .
the calculated pl spectra for different magnetic fields are shown in fig .
[ fig3](b ) .
not only qualitative but also quantitative agreement with experimental data [ compare with fig .
[ fig3](a ) ] is achieved .
note , no fitting parameters are used in the calculation . because of low mn - ion count in the vicinity of the acceptor ( on average
there are 0.3 mn ions inside the sphere of radius @xmath7 ) it is natural to assume that the nearest mn ions give dominant contributions to the stokes shifts [ fig .
[ fig5](b ) ] .
we now consider this approach in details . in an external magnetic field
the hole states are split in four sublevels characterized by the different projection @xmath79 of the hole angular momentum on the field direction .
one of them with @xmath80 is a ground state .
the contribution to the pl spectra from this state is a delta function @xmath81 ( if neglecting native homogeneous broadening ) as shown in fig .
[ fig4](b ) [ pair ( iv ) ] . excitation in one of three other states is followed by relaxation to the ground state [ pairs ( v ) and ( vi ) in fig .
[ fig4](b ) ] .
therefore , the shift of the emission energy with respect to the excitation energy is described by the distribution function of the nearest ( to the hole ) mn ion , @xmath82 . then , under unpolarized excitation the distribution function @xmath83 of the skokes shift @xmath84 is written as @xmath85 where @xmath86 and @xmath87 are the probabilities to excite the hole with @xmath88 and @xmath89 , respectively . and the characteristic exchange energy is @xmath90 mev . the probability @xmath91 that the nearest mn ion is located in the spherical layer ( @xmath92,@xmath93 ) [ see fig .
[ fig5](b ) ] is given by the probability to find one mn ion in this spherical layer being equal to @xmath94 , and the probability that there are no mn ions inside the sphere of radius @xmath92 being equal to @xmath95 .
here , @xmath95 is the poisson distribution @xmath96 with @xmath97 and @xmath98
. then one writes @xmath99 recall , the energy of the exchange interaction between the hole and mn ion depends on a distance @xmath92 as @xmath100 , i.e. @xmath101 and @xmath102 . finally , one obtains the distribution function vs energy @xmath84 as @xmath103 @xmath104 \ , , \label{eq5}\ ] ] with @xmath105 .
figure [ fig5](a ) shows that the nearest - mn approach describes well the pl spectra in high magnetic fields .
however , the agreement becomes worse for small @xmath84 in zero magnetic field , while the high - energy tail is still described quite well .
the reason is that for the complex hole + nearest - mn ion the relaxation is not possible at the excitation to the ground state , and the stokes shift is absent .
this results in a @xmath106-function peak appearing in zero as well as in high magnetic fields [ first term in eq .
( [ eq3 ] ) ] . in real the contribution to
the stokes shift for @xmath107 arises from several distant ( @xmath108 ) mn ions . in weak magnetic fields
the spins of these distant ions are disordered , and their orientation induced by the exchange field of the hole results in the energy lowering . in strong magnetic fields
the spins are already oriented at the moment of excitation , further relaxation does not occur , which manifests itself as the appearance of the narrow peak at @xmath109 .
with use of the site selective spectroscopy of residual da pairs in bulk @xmath0 samples we are able to observe the formation of bound magnetic polarons in a very dilute ( @xmath110 ) regime . under selective excitation
the pl spectra show an asymmetric band with low - energy tail and the additional very narrow peak inflaming in external magnetic fields .
this behavior can not be described in terms of the mean - field approach , frequently applied to diluted magnetic semiconductors ( dms ) .
our findings suggest that the low - energy tail is formed due to the exchange interaction of the hole ( bound to the acceptor ) with the nearest mn ion , and is described by the poisson distribution .
unexpectedly , the dominant contribution to the narrow peak arises from the interaction with several mn ions remote from the hole .
this research was supported by the dfg ( 436 rus 113/843 and spp 1285 ) as well as the rfbr .
we thank prof .
r. r. gaazka ( institute of physics , warsaw , poland ) for providing us with high quality samples .
p. a. wolff , in _ semiconductors and semimetals _ , edited by j. k. furdyna and j. kossut , vol . 25 ( academic , london , 1988 ) .
d. heiman , j. warnock , p. a. wolff , r. kershaw , r. ridgley , k. dwight , and a. wold , solid state commun .
* 52 * , 909 ( 1984 ) .
j. warnock , r. n. kershaw , d. ridgely , k. dwight , a. wold , and r. r. galazka , j. lumin . * 34 * , 25 ( 1985 ) ; t. itoh and e. komatsu , j. lumin . * 38 * , 266 ( 1987 ) ; b. p. zakharchenya and yu .
g. kusrayev , jetp lett . * 50 * , 199 ( 1989 ) . in continuous wave ( cw ) experiments it is not possible to separate the emission from the laser light at the energy of excitation . but
this can be easily done at the phonon replicas of pl lines , where the contribution of scattered light is negligible . | we study bound magnetic polarons ( bmp ) in a very diluted magnetic semiconductor @xmath0 ( @xmath1 ) by means of site selective spectroscopy . in zero magnetic field
we detect a broad and asymmetric band with a characteristic spectral width of about 5 mev .
when external magnetic fields are applied a new line appears in the emission spectrum .
remarkably , the spectral width of this line is reduced greatly down to @xmath2 @xmath3 .
we attribute such unusual behavior to the formation of bmp , effected by sizable fluctuations of local magnetic moments .
the modifications of the optical spectra have been simulated by the monte - carlo method and calculated within an approach considering the nearest mn ion . a quantitative agreement with the experiment
is achieved without use of fitting parameters .
it is demonstrated that the low - energy part of the emission spectra originates from the energetic relaxation of a complex consisting of a hole and its nearest mn ion .
it is also shown that the contribution to the narrow line arises from the remote mn ions . |
ltude des groupes de kac - moody sur un corps local a t initie par howard garland @xcite pour certains groupes de lacets .
dans @xcite on a construit un immeuble `` microaffine '' pour tous les groupes de kac - moody ( minimaux au sens de j. tits ) sur un corps muni dune valuation relle .
cest un immeuble ( en gnral non discret ) avec les bonnes proprits habituelles des immeubles .
cependant cet immeuble microaffine nest pas lanalogue des immeubles de f. bruhat et j. tits pour les groupes rductifs .
il correspond plutt leur frontire dans la compactification de satake ou compactification polydrique .
de plus il ne traduit pas les dcompositions de cartan de @xcite .
une autre construction est envisageable pour un groupe de kac - moody sur un corps rellement valu .
elle est la gnralisation directe de celle de bruhat - tits ( @xcite et @xcite ) et traduit les dcompositions de cartan de @xcite dans le cas des groupes de lacets .
cependant , comme ces dcompositions ne sont vrifies quaprs torsion , lespace @xmath0 ainsi construit la bruhat - tits est tel que deux points quelconques ne sont pas toujours dans un mme appartement .
il ne mrite donc pas le nom dimmeuble .
il sest avr cependant utile et a donc t considr sous le nom de masure ( hovel ) .
la construction de cette masure a t effectue dans @xcite et on a pu lutiliser pour des rsultats en thorie des reprsentations .
le corps valu intressant dans ce cadre est le corps des sries de laurent complexes @xmath1 . on sest
donc plac dans le cas dun corps @xmath2 muni dune valuation discrte avec un corps rsiduel contenant @xmath3 .
cette situation dgale caractristique @xmath4 simplifie les raisonnements et permet , en particulier , dutiliser les rsultats du livre de s. kumar @xcite . on a cependant d
faire quelques hypothses restrictives sur le groupe de kac - moody ( en particulier la symtrisabilit ) .
par ailleurs dans @xcite on a labor une dfinition abstraite de masure affine ( inspire de la dfinition abstraite des immeubles affines de @xcite ) et on a montr quelle est satisfaite par la plupart des masures dfinies prcdemment . de cette dfinition dcoulent des proprits intressantes : les rsidus en chaque point sont des immeubles jumels , linfini on trouve des immeubles jumels et deux immeubles microaffines , il existe un prordre invariant sur la masure .
le but du prsent article est de construire la masure affine dun groupe de kac - moody dploy sur un corps muni dune valuation relle non triviale et de montrer quelle satisfait aux axiomes abstraits de masure affine ordonne de @xcite .
ceci est ralis sans aucune restriction sur le groupe de kac - moody ni sur le corps valu .
pour cela on a essentiellement remplac dans @xcite les groupes de kac - moody la kumar par ceux dolivier mathieu @xcite et la reprsentation adjointe dans lalgbre de lie par celle dans lalgbre enveloppante entire , puisquon va considrer aussi de la caractristique ( rsiduelle ) positive .
il y a en fait beaucoup de choix possibles pour les groupes de kac - moody , @xcite . on considre ici les groupes dploys `` minimaux '' tels que dfinis par jacques tits @xcite ; leurs proprits essentielles sont expliques dans @xcite et aussi @xcite . on a rsum celles - ci et prouv quelques complments ( essentiellement sur les morphismes )
dans la premire partie de cet article .
la seconde partie est consacre lalgbre enveloppante entire introduite par j. tits pour construire ses groupes et la reprsentation adjointe , dj largement utilise par b. rmy . on y montre un thorme de poincar - birkhoff - witt ( avec des
puissances divises tordues ) et on y construit des exponentielles tordues .
ces rsultats tirent leur origine dans un travail de gnralisation du thorme de simplicit de r. moody @xcite , qui est expliqu dans lappendice .
les relations de commutation dans un groupe de kac - moody minimal @xmath5 sont compliques ( voire inexistantes ) .
pour mener bien des calculs on va raisonner dans un groupe maximal qui ici sera celui ( @xmath6 ou @xmath7 ) dfini par o. mathieu . on a cependant besoin dune connaissance plus
concrte de celui - ci , cest le but de la troisime partie .
il est en partie accompli grce aux exponentielles tordues de la partie prcdente .
dans ces trois premires parties on tudie les groupes de kac - moody de manire gnrale sur un anneau quelconque , ou sur un corps si on veut plus de rsultats de structure . on les considre ensuite sur un corps valu ( mais aussi sur son anneau des entiers ) .
dans la quatrime partie on construit lappartement tmoin et les sous - groupes parahoriques ( ou assimils ) associs aux sous - ensembles ou filtres de cet appartement tmoin .
cest la partie la plus technique .
comme dans @xcite , ces sous - groupes parahoriques du groupe de kac - moody @xmath5 sont construits en plusieurs tapes en utilisant les groupes maximaux @xmath6 et @xmath7 contenant @xmath5 . on a cependant
d apporter des changements substantiels aux raisonnements de @xcite . on rcolte enfin
dans la cinquime partie les fruits du travail prcdent .
par un procd classique utilisant lappartement tmoin et les sous - groupes parahoriques , on y construit la masure affine ( ordonne ) dun groupe dploy sur un corps rellement valu .
et on montre les mmes rsultats quen @xcite ou @xcite .
les raisonnements sont sans changement substantiel , mais , grce aux parties prcdentes , il ny a plus dhypothse technique superflue .
lappendice rassemble des rsultats sur le groupe de kac - moody maximal la mathieu @xmath6 ( sur un corps @xmath8 quelconque ) qui sont consquences des rsultats des parties [ s2 ] et [ s3 ] . on y
compare @xmath6 avec dautres groupes de kac - moody maximaux dfinis par m. ronan et b. rmy @xcite ou l. carbone et h. garland @xcite . on y gnralise aussi le thorme de simplicit dun sous - quotient de @xmath6 d r. moody @xcite en caractristique @xmath4 ; on lobtient ici en caractristique @xmath9 assez grande , pour un corps non algbrique sur @xmath10 .
je remercie bertrand rmy pour mavoir suggr dtudier larticle @xcite de r. moody et olivier mathieu pour mavoir signal son article @xcite et pour quelques claircissements sur la construction de son groupe de kac - moody ou la simplicit des algbres de kac - moody .
on introduit ici le groupe de kac - moody objet principal dtude de cet article et on y tudie en particulier sa fonctorialit .
\1 ) une _ matrice de kac - moody _
( ou matrice de cartan gnralise ) est une matrice carre @xmath11 , coefficients entiers , indexe par un ensemble fini @xmath12 et qui vrifie : \(i ) @xmath13 , \(ii ) @xmath14 , \(iii ) @xmath15 .
\2 ) un _ systme gnrateur de racines _
( en abrg sgr ) @xcite est un quadruplet @xmath16 form dune matrice de kac - moody @xmath17 indexe par @xmath12 , dun @xmath18module libre @xmath19 de rang fini @xmath20 , dune famille @xmath21 dans son dual @xmath22 et dune famille @xmath23 dans @xmath19 .
ces donnes sont soumises la condition de compatibilit suivante : @xmath24 @xmath25 .
on dit que @xmath26 est le _ rang _ du sgr @xmath27 et @xmath20 sa _
dimension_. on dit
que le sgr @xmath27 est _ libre _ ( ou _ adjoint _ ) ( resp . _ colibre _ ( ou _ coadjoint _ ) ) si @xmath21 ( resp .
@xmath23 ) est libre dans ( ou engendre ) @xmath28 ( resp .
@xmath19 ) . par exemple le sgr _ simplement connexe _ @xmath29 avec @xmath30 de base les @xmath31 est le seul sgr colibre et coadjoint associ @xmath17 .
\3 ) on introduit le @xmath32module libre @xmath33 .
il y a donc un homomorphisme de groupes @xmath34 , @xmath35 tel que @xmath36 . on note @xmath37 et @xmath38 .
quand @xmath27 est libre on identifie @xmath39 un sous - module de @xmath28 et on ne fait pas de diffrence entre @xmath40 et @xmath41 .
le sgr _ adjoint minimal _
@xmath42 est libre et adjoint . on note @xmath43 .
on dit que le sgr @xmath27 est _ sans cotorsion _
si @xmath44 est facteur direct dans @xmath19 @xmath45 est sans torsion .
cest le cas de @xmath46 pour lequel @xmath47 .
\4 ) on note @xmath48 et @xmath49 ; pour @xmath50 et @xmath51 , on crit @xmath52 .
l_algbre de kac - moody complexe _
@xmath53 associe @xmath54 est engendre par @xmath55 et des lments @xmath56 avec les relations suivantes ( pour @xmath57 et @xmath58 ) : ( kmt1 ) @xmath59=0$];@xmath60=\alpha_i(h)e_i$];@xmath61=-\alpha_i(h)f_i$];@xmath62=-\alpha_i^{\vee}$ ] .
( kmt2 ) @xmath63=0$];@xmath64 . on a alors
une graduation de lalgbre de lie @xmath65 par @xmath39 : @xmath66 o @xmath67 est le _ systme de racines _ de @xmath65 . on a @xmath68 ,
@xmath69 et @xmath70 , de plus @xmath71={\overline\alpha}(x)\;\forall h\in{\mathfrak{h}}\}$ ] est non nul pour @xmath72 . on sait
que @xmath73 et les @xmath74 ne dpendent que de @xmath17 et non de @xmath75 . on note
@xmath76 .
les racines @xmath77 sont dites _
simples_. les racines de @xmath78 ( resp .
@xmath79 ) sont dites _ positives _ ( resp .
_ ngatives _ ) . on a @xmath80 .
la sous - algbre _ nilpotente _ ( resp .
_ de borel _ ) _ positive _ est @xmath81 ( indpendante de @xmath75 et donc note aussi @xmath82 ) ( resp .
@xmath83 ) . on
a de mme des sous - algbres ngatives @xmath84 et @xmath85 avec la dcomposition triangulaire @xmath86 .
\5 ) le _ groupe de weyl ( vectoriel ) _
@xmath87 associ @xmath17 est un groupe de coxeter de systme de gnrateurs lensemble @xmath88 des automorphismes de @xmath39 dfinis par @xmath89 .
il stabilise @xmath73 et agit aussi sur @xmath28 et @xmath19 par des formules semblables .
on note @xmath90 lensemble des _ racines relles _
cest dire des lments de @xmath73 ou @xmath39 de la forme @xmath91 avec @xmath92 et @xmath93 .
si @xmath94 , alors @xmath95 est bien dtermin par @xmath40 , indpendamment du choix de @xmath96 et de @xmath97 tels que @xmath91 .
pour @xmath98 on a @xmath99 pour un @xmath100 avec @xmath101 .
si @xmath102 et @xmath103 , on a @xmath104 . les racines de
@xmath105 sont dites _
si on a deux parties @xmath106 de @xmath107 , on dit que @xmath108 est _ close _ ( resp .
un _ idal _ de @xmath109 ) si : @xmath110 , ( resp .
@xmath111 ) , @xmath112 .
la partie @xmath108 est dite _ prnilpotente _ sil existe @xmath113 tels que @xmath114 et @xmath115 , alors @xmath108 est finie et contenue dans la partie @xmath116 de @xmath117 qui est _ nilpotente _ ( prnilpotente et close ) .
\6 ) un _ morphisme de sgr _ , @xmath118 est une application linaire
@xmath119 et une injection @xmath120 , @xmath121 telles que @xmath122 , @xmath123 et @xmath124 @xmath125 . on note
@xmath126 lapplication duale cette dfinition de morphisme est duale de celle de @xcite qui est valable dans un cadre plus gnral .
pour un morphisme de sgr @xmath127 , on dfinit de manire vidente un morphisme @xmath128 entre les algbres de lie correspondantes .
\7 ) soit @xmath127 un morphisme de sgr .
on dit que : @xmath75 est une _ extension centrale torique _
de @xmath129 si @xmath130 est surjective et @xmath131 ; elle est dite de plus _ scinde _
si ker@xmath132 a un supplmentaire contenant @xmath44 , @xmath75 est une _ extension centrale finie
_ de @xmath129 si @xmath130 est injective , @xmath131 et les dimensions sont gales , plus gnralement @xmath75 est _
extension centrale _
de @xmath129 si @xmath133 est injective et @xmath131 , @xmath75 est un _ sous - sgr _ de @xmath129 si @xmath130 est injective et @xmath134 est sans torsion , @xmath129 est une _ extension semi - directe _ ( torique )
de @xmath75 si @xmath75 est un sous - sgr de @xmath129 et si @xmath131 .
elle est dite _ directe _ si , de plus , il existe un supplmentaire de @xmath135 dans @xmath136 contenu dans ker@xmath137 , @xmath138 . @xmath130 est une _ extension commutative _ si @xmath131 ( alors @xmath139 et @xmath140 ) .
un sgr @xmath75 , on associe un @xmath32schma en groupe @xmath141 ( vu comme foncteur en groupe sur la catgorie des anneaux ) dfini par @xmath142 ( pour tout anneau @xmath8 ) ou par @xmath143)$ ] .
cest un tore ( isomorphe @xmath144 ) de groupe de cocaractres ( resp .
caractres ) @xmath19 ( resp .
@xmath28 ) , voir @xcite .
un morphisme @xmath127 de sgr donne un homomorphisme @xmath145 de tores .
le groupe de weyl @xmath87 agit sur @xmath146 via ses actions sur @xmath19 et @xmath28
. si @xmath130 est extension centrale , @xmath147 est surjectif ( au sens schmatique ) de noyau le groupe multiplicatif @xmath148 de groupe des caractres @xmath149 .
si @xmath130 est extension centrale torique , @xmath147 est surjectif ( au sens fonctoriel ) de noyau le tore @xmath150 .
si @xmath130 est extension centrale finie , alors le thorme des diviseurs lmentaires appliqu @xmath151 donne une description explicite de @xmath152 comme produit de groupes de racines de lunit et lhomomorphisme @xmath153 nest pas surjectif pour tout anneau @xmath8 . si @xmath75 est un _ sous - sgr _ de @xmath129 , @xmath147 identifie @xmath146 un sous - tore facteur direct de @xmath154 .
[ 1.3 ] soit @xmath155 un sgr .
\a ) il existe une extension centrale finie @xmath156 de @xmath75 qui est sans cotorsion .
le sgr @xmath75 est extension centrale de @xmath157 qui est adjoint .
si @xmath75 est libre alors @xmath156 et @xmath158 aussi .
\b ) le sgr @xmath159 nest en gnral pas libre .
le sgr @xmath156 ( resp .
@xmath160 ) en est extension semi - directe ( resp .
centrale torique ) .
\c ) il existe une extension centrale torique @xmath161 de @xmath75 qui est sans cotorsion et colibre .
si @xmath75 est colibre et sans cotorsion , cette extension est scinde .
si @xmath75 est libre ou si @xmath39 est facteur direct dans @xmath28 , alors cest galement vrai pour @xmath161 .
\d ) il existe une extension semi - directe @xmath162 de @xmath75 qui est libre ( et avec @xmath163 facteur direct de @xmath164 ) .
si @xmath75 est libre et @xmath39 facteur direct de @xmath28 , cette extension est directe .
si @xmath75 est colibre ou si @xmath44 est facteur direct dans @xmath19 , alors cest galement vrai pour @xmath162 .
\e ) si @xmath75 est libre , colibre et sans cotorsion , alors @xmath75 est extension semi - directe dun sgr @xmath165 libre , colibre , sans cotorsion et de dimension @xmath166 o @xmath167 est le rang de @xmath75 et @xmath168 ( @xmath169 ) le rang de la matrice @xmath17 .
\1 ) on a @xmath170 et les constructions de c ) et d ) sont fonctorielles en @xmath75 .
\2 ) dans le cas e ) le sgr @xmath165 vrifie donc les hypothses de ( * ? ? ?
nest pas toujours extension directe de @xmath165 .
\3 ) quand la matrice @xmath17 est inversible ( par exemple dans le cas classique dune matrice de cartan ) tout sgr est libre , colibre et extension semi - directe ( ou aussi centrale torique ) dun sgr de dimension gale au rang ( semi - simple dans le cas classique ) . on considre un supplmentaire @xmath171 de @xmath172 dans @xmath19 et on pose @xmath173 .
les assertions a ) et b ) sont alors videntes .
\c ) on dfinit @xmath174 , son dual est donc @xmath175 o les @xmath176 forment une base duale des @xmath177 . on note
@xmath178 , @xmath179 et @xmath130 est la projection de @xmath180 sur @xmath19 .
@xcite .
\d ) on dfinit @xmath181 , son dual est donc @xmath182 o les @xmath183 forment une base duale des @xmath184 . on note
@xmath185 , @xmath186 et @xmath130 est linjection canonique de @xmath19 dans @xmath187 .
\e ) soit @xmath188 , @xmath189 .
comme @xmath75 est libre , @xmath190 est de rang @xmath167 , tandis que @xmath191 est de rang @xmath168 .
le module @xmath192 a un supplmentaire dans @xmath190 de rang @xmath193 ; on suppose que la base de ce supplmentaire est @xmath194 pour des @xmath195 .
comme @xmath75 est colibre
@xmath196 est de rang @xmath166 .
comme le quotient de @xmath19 par @xmath44 est sans torsion , il en est de mme pour @xmath197 . on note @xmath198 et @xmath199 .
linjection de @xmath197 dans @xmath19 fait bien de @xmath75 une extension semi - directe de @xmath165 et @xmath165 est colibre , sans cotorsion , de dimension @xmath166 .
comme @xmath200 est de rang @xmath167 , les @xmath201 sont bien libres dans @xmath202 .
* et 3.6 ) pour @xmath204 , @xmath205 et @xmath206 sont localement nilpotents dans @xmath207 ou @xmath65 , @xmath208 , @xmath209 sont des automorphismes de @xmath207 ou @xmath65 et lon a @xmath210 @xmath211 , que lon note @xmath212 ou @xmath213 .
lapplication @xmath214 stend en un homomorphisme surjectif @xmath215 du sous - groupe @xmath216 de @xmath203 engendr par les @xmath212 sur le groupe de weyl @xmath87 .
le groupe @xmath216 stabilise @xmath217 et permute les @xmath74 pour @xmath218 ; les actions induites sur @xmath217 ou @xmath73 se dduisent des actions de @xmath87 via lhomomorphisme prcdent . en particulier @xmath216 est le mme quil soit
dfini sur @xmath207 ou @xmath65 .
pour @xmath219 , lespace @xmath74 est de dimension @xmath220 et on peut choisir les lments de base @xmath221 de faon que @xmath222 , @xmath223=-\alpha^\vee$ ] et @xmath224 pour @xmath219 , @xmath204 et @xmath225 .
si @xmath226 forment une paire prnilpotente de racines , lensemble @xmath227=\{\,p\alpha+q\beta\in\phi\,\vert\,p , q\in\n\}$ ] est fini .
si @xmath40 et @xmath228 sont non colinaires , on note @xmath229\alpha,\beta[=[\alpha,\beta]\setminus\{\alpha,\beta\}$ ] , et on lordonne , par exemple de faon que @xmath230 soit croissant ; on a alors la formule suivante pour les commutateurs ( pour @xmath231 ) : @xmath232 o @xmath233 parcourt @xmath229\alpha,\beta[$ ] et les @xmath234 sont des entiers bien dfinis .
@xcite ce foncteur @xmath235 de la catgorie des anneaux dans celle des groupes est engendr par @xmath236 exemplaires du groupe additif @xmath237 .
plus prcisment , pour @xmath219 , @xmath8 un anneau et @xmath238 , on introduit le symbole @xmath239 et le groupe @xmath240 est engendr par les lments @xmath239 pour @xmath219 , @xmath238 soumis aux relations @xmath241 et aux relations de commutation : ( kmt3)@xmath242 pour @xmath243 , @xmath244 prnilpotente et @xmath245 , @xmath234 comme ci - dessus en [ 1.4 ] . si @xmath108 est une partie nilpotente de @xmath117 et
si on remplace ci - dessus @xmath117 par @xmath108 , on dfinit un foncteur en groupes @xmath246 .
cest un groupe algbrique unipotent isomorphe comme schma @xmath247 et qui ne dpend que de @xmath108 et @xmath17 ( donc @xmath117 ) mais pas de @xmath75 .
si @xmath219 et @xmath248 , @xmath249 est un sous - foncteur en groupes de @xmath235 isomorphe @xmath237 par @xmath250 .
en fait @xmath235 est lamalgame des groupes @xmath251 et @xmath252}$ ] pour @xmath219 et @xmath244 prnilpotente . pour @xmath219 , @xmath8 un anneau et @xmath253 , on dfinit dans @xmath240 , @xmath254 , @xmath255 et @xmath256 . on a @xmath257 . en particulier @xmath258 .
si jamais le quotient @xmath259 de @xmath207 par son idal gradu maximal dintersection triviale avec @xmath260 est diffrent de @xmath207 , la diffrence naffecte que les espaces radiciels imaginaires .
le remplacement de @xmath207 par @xmath259 ne change donc pas @xmath235 ( ni @xmath261 dfini ci - dessous ) ; ce nest par contre pas vrai pour le groupe @xmath262 du 3 .
@xcite , @xcite pour un anneau @xmath8 , on dfinit le groupe @xmath263 comme le quotient du produit libre @xmath264 par les 4 relations suivantes , pour @xmath40 racine simple , @xmath265 , @xmath238 et @xmath266 : ( kmt4 ) @xmath267 , ( kmt5 ) @xmath268 , ( kmt6 ) @xmath269 , ( kmt7 ) @xmath270 , si @xmath271 et @xmath272 ( avec @xmath273 )
. les relations ( kmt5 ) et ( kmt7 ) permettent aussitt
de gnraliser ( kmt4 ) au cas o @xmath219 nest pas simple .
proprits \1 ) daprs ( * ? ? ?
* b ) il existe des foncteurs en groupes @xmath274 et @xmath275 satisfaisant aux axiomes ( kmg 1 9 ) de . daprs le thorme 1 ( i ) p. 553
de , il en rsulte que , pour tout anneau @xmath8 , lhomomorphisme canonique de @xmath276 dans @xmath263 est injectif .
\2 ) pour tout @xmath219 , lhomomorphisme @xmath277 est injectif , voir la dmonstration de @xcite si @xmath278 est sans torsion ( @xmath125 ) ou si @xmath8 na pas dlments nilpotents ; le cas gnral se traite comme dans la partie b ) de la dmonstration du lemme [ 4.11 ] ci - dessous . plus
gnralement ces mmes raisonnements montrent que , pour toute partie nilpotente @xmath108 de @xmath117 , lhomomorphisme canonique de @xmath279 dans @xmath263 est injectif et aussi que , pour @xmath280 et @xmath167 ou @xmath281 non nul @xmath282 .
\3 ) on identifie donc @xmath146 , @xmath283 et @xmath246 des sous - foncteurs en groupes de @xmath261 ; si @xmath109 est un idal de @xmath108 , @xmath284 est distingu dans @xmath246
. de mme on note @xmath285 ( resp .
@xmath286 ) le sous - foncteur en groupe de @xmath261 tel que , pour un anneau @xmath8 , @xmath287 ( resp .
@xmath288 ) est le
sous - groupe engendr par les @xmath289 pour @xmath290 ( resp .
et par @xmath276 ) .
\4 ) on note @xmath291 le sous - foncteur de @xmath261 tel que @xmath292 soit le sous - groupe engendr par @xmath276 et les @xmath293 pour @xmath40 racine simple . on a un homomorphisme @xmath294 de @xmath291 sur @xmath87 trivial sur @xmath146 tel que
@xmath295 et @xmath292 agit sur @xmath146 et @xmath117 via @xmath294 , ( kmt4 ) , ( kmt5 ) , ( kmt7 ) et 2 ) ci - dessus .
@xcite les @xmath296 satisfont aux relations de tresse , daprs ( kmt6 ) @xmath297 , le noyau de @xmath294 est donc gal @xmath276 .
si @xmath8 est un corps avec au moins quatre lments , @xmath292 est le normalisateur de @xmath276 dans @xmath263 @xcite .
si @xmath8 est un corps infini , on sait de plus que tous les sous - tores @xmath298dploys maximaux de @xmath261 sont conjugus @xmath146 par @xmath263 @xcite .
\5 ) les thormes 1 et 1 p. 553
de @xcite montrent que , sur les corps , @xmath261 vrifie les axiomes ( kmg 1 9 ) de : il existe un homomorphisme @xmath299 de foncteurs en groupes @xmath300 qui est un isomorphisme sur les corps et vrifie @xmath301 .
en particulier ( kmg4 ) dit que , pour une extension de corps @xmath302 , @xmath263 sinjecte dans @xmath303 .
si @xmath8 est un corps , @xmath304 est une donne radicielle de type @xmath117 , au sens de @xcite , ou donne radicielle jumele entire @xcite ( cest plus prcis que les donnes radicielles jumeles de , les rgd - systems de @xcite ) . en particulier ,
pour @xmath305 , il existe @xmath306 tels que @xmath307 conjugue @xmath308 en @xmath309 , @xmath310 , donc @xmath311 ; on a @xmath312 .
on montre que @xmath288 est produit semi - direct de @xmath276 et @xmath287 @xcite .
pour une partie nilpotente de la forme @xmath313 , on a @xmath314 3.5.4 .
le centre de @xmath263 est @xmath315 8.4.3 .
toujours si @xmath8 est un corps , la donne radicielle ci - dessus permet de construire des immeubles combinatoires jumels @xmath316 et @xmath317 , agits par @xmath263 : limmeuble @xmath318 de @xmath261 sur @xmath8 est associ au systme de tits @xmath319 .
pour un anneau k , on dfinit @xmath320 comme le quotient de @xmath321 par les relations suivantes , pour @xmath322 racines simples , @xmath323 et @xmath243 : ( @xmath3245 ) @xmath325 , ( @xmath3247 ) @xmath326 , si @xmath327 et @xmath328 , ( @xmath3248 ) a ) @xmath329 est un homomorphisme de groupes , \b ) @xmath330 , on note @xmath331 le groupe commutatif engendr par tous ces lments ( pour @xmath322 racines simples et @xmath243 ) , \c ) la formule @xmath332 dfinit un homomorphisme de groupes @xmath333 , not @xmath245 . on note de
la mme manire les lments et leurs images dans @xmath320 .
\1 ) dans le cas classique , steinberg considre lui aussi un quotient de @xmath334 un peu analogue : comparer ( @xmath3247 ) et ( @xmath3248 ) respectivement aux conditions ( b ) et ( c ) de @xcite ou @xcite .
\2 ) on na pas cherch
ici un systme minimal de relations pour @xmath320 .
le lecteur trouvera des rductions faciles par lui - mme et dautres moins videntes dans ( * ? ? ?
* ou 3.8 ) , @xcite ou @xcite .
\3 ) lendomorphisme @xmath335 du groupe @xmath331 dfini par @xmath336 est une involution , car @xmath337 @xmath338 ( car @xmath339 est llvation au carr ) .
en fait pour @xmath340 on a @xmath341 ( @xmath342 avec une notation additive pour hom@xmath343 ) .
en particulier ( @xmath3245 ) et ( kmt5 ) sont semblables .
\4 ) la relation ( @xmath3244 ) : @xmath267 , pour @xmath219 , @xmath238 et @xmath344 est consquence des relations ci - dessus : en effet pour @xmath40 , @xmath228 racines simples , @xmath253 , @xmath345 , on a : @xmath346 ( avec @xmath347 et @xmath348 ) .
cette relation stend au cas @xmath40 non simple daprs ( @xmath3245 ) , ( @xmath3247 ) et ( @xmath3248 ) : si @xmath40 , @xmath228 sont simples et @xmath245 satisfait ( @xmath3244),on a : @xmath349 @xmath350 @xmath351 et @xmath352 .
donc @xmath353 satisfait aussi ( @xmath3244 ) .
\5 ) on peut appliquer ( @xmath3247 ) @xmath354 .
on sait que @xmath355 .
donc @xmath356 @xmath357 et @xmath358 .
[ 1.8 ] \1 ) lhomomorphisme canonique de @xmath235 dans @xmath261 se factorise par @xmath359 , autrement dit les relations ( @xmath3245 ) , ( @xmath3247 ) et ( @xmath3248 ) sont satisfaites dans @xmath261 .
\2 ) pour un anneau @xmath8 , @xmath263 est le quotient du produit libre @xmath360 par les 2 relations suivantes ( pour @xmath40 racine simple , @xmath265 , @xmath238 et @xmath266 ) : ( kmt4 ) @xmath361 , ( @xmath3246 ) @xmath362 si @xmath253 .
\3 ) la relation ( @xmath3246 ) induit un homomorphisme de foncteurs en groupes @xmath363 do nt le noyau @xmath364 est central dans @xmath235 .
le foncteur @xmath261 est le quotient ( comme foncteur en groupes ) du produit semi - direct @xmath365 par @xmath366 antidiagonal . en particulier le quotient ( comme foncteur en groupes ) @xmath367 sinjecte dans @xmath261 .
par dfinition de @xmath368 et ( kmt6 ) , la relation ( @xmath3246 ) est bien satisfaite dans @xmath263 . on en dduit aussitt la relation ( @xmath3248 )
et que limage de @xmath369 est dans @xmath276 .
dautre part les deux involutions @xmath335 et @xmath370 concident sur cette image daprs les calculs de [ 1.7].3 ; ainsi ( @xmath3245 ) est une consquence de ( kmt5 ) .
enfin @xmath371 , do ( @xmath3247 ) . on a ainsi montr 1 )
et que les relations ( kmt4 ) , ( @xmath3246 ) sont vrifies dans @xmath263 .
inversement il faut montrer que ( kmt5 ) , ( kmt6 ) et ( kmt7 ) sont des consquences de ( kmt4 ) , ( @xmath3245 ) , ( @xmath3246 ) , ( @xmath3247 ) et ( @xmath3248 ) .
la relation ( kmt7 ) est un cas particulier de ( @xmath3247 ) , ( kmt6 ) dcoule de ( @xmath3246 ) et de la dfinition de @xmath372 .
pour @xmath40 simple et @xmath266 , on a : @xmath373 ; en effet pour @xmath374 avec @xmath375 et @xmath253 on a : @xmath376 .
ainsi ( kmt5 ) est satisfaite .
\3 ) on a vu en [ 1.6].1 que @xmath146 sinjecte dans @xmath261 , la relation ( @xmath3246 ) ninduit donc aucun quotient dans @xmath146 .
par contre , comme @xmath377 est engendr par les @xmath372 , cette relation se traduit par un homomorphisme @xmath363 ; celui - ci est compatible avec les morphismes @xmath378 et @xmath379 ( pour @xmath323 ) , daprs ( @xmath3248c ) et la relation correspondante dans @xmath146 .
ainsi la relation ( @xmath3244 ) ( [ 1.7].4 ) montre que le noyau @xmath364 est central dans @xmath235 .
la relation ( kmt4 ) montre que @xmath261 est quotient du produit semi - direct indiqu par des relations induites par ( @xmath3246 ) .
la relation ( @xmath3244 ) montre que ce quotient se limite laction antidiagonale de @xmath366 .
[ 1.9 ] \1 ) pour le sgr simplement connexe @xmath160 , on a @xmath380 et @xmath381 , que lon notera aussi @xmath382 ou @xmath383 .
\2 ) si @xmath8 est un corps , lhomomorphisme canonique de @xmath384 dans @xmath385 est un isomorphisme .
\1 ) comme @xmath386 est produit direct des groupes @xmath387 ( pour @xmath40 racine simple ) et sinjecte dans @xmath388 , la relation ( @xmath3246 ) permet didentifier @xmath369 et @xmath386 .
alors ( kmt4 ) se rduit ( @xmath3244 ) do nt on sait quil est vrifi dans @xmath320 ( [ 1.7].4 ) , on a donc bien lidentification propose .
\2 ) on sait dans @xmath388 que @xmath386 et @xmath384 forment un produit semi - direct ( [ 1.6].5 ) le quotient par @xmath389 naffecte donc pas @xmath384 .
si @xmath390 est une extension commutative de sgr , la proposition [ 1.8 ] permet de dfinir un morphisme @xmath391 de foncteurs . en effet @xmath75 et @xmath129 correspondent la mme matrice @xmath17 , donc les mmes @xmath117 , @xmath207 , @xmath321 et @xmath320 .
la compatibilit de @xmath130 avec les racines et coracines permet de dfinir @xmath392 comme le passage au quotient de @xmath393 .
pour tout anneau @xmath8 , les homomorphismes @xmath392 et @xmath153 ont mme noyau ; le groupe @xmath394 est engendr par @xmath395 et le sous - groupe ( distingu ) image de @xmath392 ( [ 1.8].3 ) ; on a @xmath396 .
de plus limage rciproque de @xmath397 ( resp .
@xmath395 ) dans @xmath398 est gale @xmath399 ( resp .
@xmath400 ) .
le noyau ker@xmath401 est central dans @xmath261 ( mais trivial si @xmath75 est un sous - sgr ) .
par construction @xmath402 induit un isomorphisme de @xmath403 sur @xmath404 et de @xmath405 sur @xmath406 , pour tout corps @xmath8 ( [ 1.9].2 ) .
sur un corps @xmath8 , on voit facilement quil est possible didentifier les immeubles combinatoires @xmath407 de @xmath408 et @xmath409 avec leurs facettes et leurs appartements .
les actions de @xmath398 et @xmath394 sont compatibles ( via @xmath392 ) .
on va tudier quelques cas particuliers intressants . [ 1.11 ]
si le sgr @xmath129 est extension semi - directe ( resp .
directe ) du sgr @xmath75 , alors @xmath409 est produit semi - direct ( resp .
direct ) de @xmath408 par un tore @xmath410 o @xmath411 est un supplmentaire de @xmath19 dans @xmath136 et , pour un anneau @xmath8 , laction de @xmath412 sur @xmath398 est donne par les relations @xmath413 et @xmath414pour @xmath415 , @xmath416 , @xmath219 et @xmath238 . on a @xmath417 , donc @xmath418 .
dans @xmath419 la relation ( @xmath3246 ) nimplique que @xmath420 et ( kmt4 ) dcrit laction de @xmath410 sur @xmath359 , do le rsultat . [ 1.12 ] si le sgr @xmath75 est extension centrale torique du sgr @xmath129 , alors @xmath409 est quotient ( comme foncteur en groupes ) de @xmath408 par un sous - tore facteur direct @xmath421 de @xmath146 , central dans @xmath408 .
si lextension centrale est scinde , alors @xmath408 est isomorphe au produit direct de @xmath409 et @xmath421 .
comme @xmath422 est surjective , @xmath423 est facteur direct dans @xmath19 et contenu dans les noyaux ker@xmath228 pour @xmath265 . on pose @xmath424
et la premire assertion est alors une consquence directe de [ 1.8 ] .
si lextension est scinde , @xmath411 a un supplmentaire @xmath425 contenant les coracines et @xmath130 induit un isomorphisme de @xmath426 sur @xmath129 , do le rsultat . [ 1.13 ] si @xmath390 est une extension centrale finie , alors pour un anneau @xmath8 , @xmath427 a pour noyau le groupe fini @xmath428 contenu dans tous les noyaux de racines de @xmath117 .
ce groupe @xmath152 est central dans @xmath263 , le quotient @xmath429 est un sous - groupe distingu de @xmath394 et @xmath430 .
sauf si @xmath8 est un corps algbriquement clos , @xmath394 nest pas forcment gal @xmath429 , car @xmath153 nest pas forcment surjective , voir [ 1.2 ] .
l encore ( @xmath3246 ) dans @xmath409 est consquence de la mme relation dans @xmath408 et ( kmt4 ) montre que @xmath395 normalise limage de @xmath398 .
linjection de @xmath395 dans @xmath394 permet de contrler la situation .
cette algbre `` enveloppante '' est la base de la construction par j. tits de son groupe de kac - moody tudi ci - dessus . on va prouver
ci - dessous un thorme de poincar - birkhoff - witt et introduire des `` exponentielles tordues '' dans le complt de cette algbre .
@xcite , @xcite , @xcite , ( * ? ? ? * viii 12 ) .
\1 ) dans lalgbre enveloppante @xmath431 de lalgbre de lie complexe @xmath65 , j. tits @xcite a introduit la sous@xmath432algbre @xmath433 engendre par les lments suivants : @xmath434pour @xmath204 et @xmath435 , @xmath436pour @xmath204 et @xmath435 , @xmath437 pour @xmath438 et @xmath435 .
la sous - algbre @xmath439 ( resp .
@xmath440 , @xmath441 ) est engendre par les lments du premier type ( resp .
du second type , du troisime type ) .
on oubliera souvent lindice @xmath75 dans les notations prcdentes , dailleurs @xmath439 et @xmath440 ne dpendent que de @xmath17 et non de @xmath75 .
on sait que @xmath433 ( resp .
@xmath439 , @xmath440 , @xmath441 ) est une
@xmath32forme de lalgbre enveloppante correspondante @xmath431 ( resp @xmath442 , @xmath443 , @xmath444 ) et que lon a la dcomposition unique @xmath445 .
\2 ) la graduation de lalgbre de lie @xmath65 par @xmath39 induit une graduation de lalgbre associative @xmath433 par @xmath39 ; le sous - espace de poids @xmath50 est not @xmath446 .
la sous - algbre @xmath439 ( resp .
@xmath440 ) est gradue par @xmath447 ( resp .
@xmath448 ) , la sous - algbre @xmath441 est de poids @xmath4 .
lalgbre @xmath433 est un sous@xmath449module pour la reprsentation adjointe @xmath450 de @xmath431 sur elle - mme . en effet @xmath451 et , pour @xmath438 , @xmath452 agit sur @xmath453 par multiplication par @xmath454 .
la filtration de @xmath431 induit une filtration sur @xmath433 .
le niveau @xmath220 de cette filtration est donc somme directe de @xmath455 et de @xmath456 .
lalgbre de lie @xmath457 est un sous@xmath449module pour laction adjointe , elle a une dcomposition triangulaire @xmath458 ( o @xmath459 et @xmath460 ) et une dcomposition radicielle @xmath461 ( o @xmath462 ) .
\3 ) lalgbre @xmath433 ( ou @xmath439 , @xmath440 , @xmath441 ) a une structure de @xmath32bigbre cocommutative , co - inversible , elle est non commutative ( sauf @xmath441 ) .
sa comultiplication @xmath463 , sa co - unit @xmath464 et sa co - inversion ( ou antiautomorphisme principal ) @xmath465 respectent la graduation et la filtration : elles sont donnes sur les gnrateurs par les formules suivantes ( pour @xmath204 , @xmath438 et @xmath435 ) : @xmath466 , @xmath467 pour @xmath468 et @xmath469 @xmath470 , @xmath471 pour @xmath468 et @xmath472 @xmath473 , @xmath474 pour @xmath468 et @xmath475 .
\4 ) les automorphismes @xmath212 de [ 1.4 ] stendent @xmath431 et stabilisent @xmath433 . on a une action de @xmath216 sur @xmath433 .
ainsi , pour une racine relle @xmath40 et un @xmath476 base ( sur @xmath455 ) de @xmath477 , llment @xmath478 est dans @xmath433 ; cest une base de @xmath479 . on a comme ci - dessus @xmath480 , @xmath481 pour @xmath468 et @xmath482 .
lautomorphisme @xmath483 de @xmath431 stabilise @xmath433 , puisquon a @xmath484 . on note @xmath485 llment suivant @xmath486 ; cet lment de @xmath216 dpend du choix de @xmath476 comme base de @xmath477 , si on change @xmath476 en @xmath487 , on change @xmath485 en @xmath488 .
plus gnralement le foncteur en groupe @xmath261 admet une reprsentation adjointe @xmath489 dans les automorphismes de @xmath433 respectant sa filtration et donc aussi dans les automorphismes de @xmath457 @xcite .
pour @xmath219 et @xmath490 , @xmath239 agit sur @xmath491 comme @xmath492 ; le groupe @xmath146 agit selon la dcomposition de @xmath433 ou @xmath493 selon les poids .
\5 ) lalgbre @xmath494 est gradue par @xmath447 .
on peut la graduer par un _
degr total _
dans @xmath495 do nt chaque facteur regroupe un nombre fini danciens facteurs .
ainsi sa compltion positive ( pour le degr total ) est @xmath496 ; cest une algbre pour le prolongement naturel de la multiplication de @xmath494 .
on peut de mme considrer lalgbre @xmath497 ( en gnral diffrente de @xmath498 ) pour tout anneau @xmath8 .
on peut considrer la compltion positive @xmath499 de @xmath500 selon le degr total ( dans @xmath455 ) ; cest une algbre contenant @xmath501 .
on dfinit de mme @xmath502 qui contient @xmath503 .
lalgbre @xmath502 contient lalgbre de lie @xmath504 .
laction adjointe @xmath450 de @xmath500 sur @xmath500 respecte la graduation .
on peut donc la prolonger en une action de @xmath499 ( ou @xmath502 ) sur elle mme .
lalgbre de lie @xmath505 est un sous@xmath506module . les rsultats prcdents et suivants sont
bien sr encore valables si on remplace @xmath507 par @xmath508 ou par un conjugu de @xmath509 par un lment du groupe de weyl @xmath87 .
on peut en particulier dfinir des compltions ngatives @xmath510 , @xmath511 ou @xmath512 ( selon loppos du degr total ) . [ 2.2 ] lalgbre de lie @xmath493 ( resp .
@xmath513 , @xmath514 ) est le @xmath515module ( resp .
@xmath516module , @xmath517module ) engendr par @xmath518 et les @xmath519 , @xmath520 ( resp .
par les @xmath519 , par les @xmath520 ) pour @xmath204 ( pour laction adjointe de @xmath500 ) .
lalgbre de lie @xmath493 contient la sous - algbre de lie @xmath521 engendre par @xmath522 et les @xmath519 , @xmath520 pour @xmath204 .
comme @xmath493 est une algbre de lie , elle contient @xmath523 ; comme cest un sous@xmath524module de @xmath500 elle contient le @xmath515module @xmath525 engendr par @xmath522 et les @xmath519 , @xmath520 .
si jamais @xmath526 , il existe un nombre premier @xmath9 tel que limage de @xmath525 dans @xmath527 nest pas tout @xmath527 .
mais ceci contredit ( * ? ? ?
le raisonnement pour @xmath528 est analogue .
[ 2.3 ] a ) soit @xmath2 un corps de caractristique @xmath9 .
si , pour tout coefficient de la matrice de kac - moody @xmath17 , on a @xmath529 ou si @xmath530 , alors lalgbre @xmath531 ( engendre par @xmath532 et les @xmath519 , @xmath520 pour @xmath204 ) est gale @xmath533
. de mme lalgbre de lie @xmath534 ( resp .
@xmath535 ) est engendre par les @xmath519 ( resp .
les @xmath520 ) .
\b ) dans le cas simplement lac si @xmath536 pour @xmath537 , on a @xmath538 .
le a ) est clair pour @xmath530 ; pour @xmath539 cest ( * ? ? ?
* p308 ) . on en dduit alors facilement le b )
comme dans la dmonstration prcdente .
on peut aussi utiliser la proposition [ 2.2 ] et ( * ? ? ?
* prop 1 p181 ) pour prouver a ) et b ) .
voir aussi ( * ? ? ?
* prop 1 ) .
[ 2.4 ] soit @xmath540 .
il existe des lments @xmath541}\in\shu$ ] pour @xmath435 , tels que @xmath542}=1 $ ] , @xmath543}=x$ ] et ( si on note @xmath544 ) llment @xmath541}-x^{(n)}$ ] est de filtration @xmath545 dans @xmath431 .
si de plus @xmath546 , @xmath50 , on peut supposer @xmath541}\in\shu_{n\alpha}$ ] .
pour @xmath547 , on peut prendre @xmath541}=\left(\begin{array}{c}h \\ n\end{array}\right)$ ] .
pour @xmath546 , @xmath219 , on peut prendre @xmath541}=x^{(n)}$ ] .
le problme se pose pour @xmath546 , @xmath548 .
des calculs explicites sont faits dans @xcite ; on constate que @xmath541}$ ] nest pas forcment dans @xmath549 .
pour @xmath550 , @xmath551 , on a @xmath552 modulo filtration @xmath545 .
on dduit ainsi lexistence des @xmath553}$ ] de celle des @xmath554}$ ] , @xmath555}$ ] .
la dmonstration se rduit donc au cas o @xmath556 est homogne et lun des gnrateurs du @xmath32module @xmath493 .
le rsultat signal en remarque pour @xmath557 , @xmath558 ou @xmath559 et la proposition [ 2.2 ] montrent quil suffit de dmontrer le lemme suivant ( le cas ngatif est semblable ) .
[ 2.5 ] soient @xmath560 et @xmath561 tels quil existe des lments @xmath541}\in\shu_{n\beta}$ ] satisfaisant aux conditions de la proposition prcdente .
pour @xmath204 , @xmath562 , notons @xmath563 et @xmath564 ( @xmath565 si @xmath566 ) .
alors il existe des lments @xmath567}\in\shu_{n\gamma}$ ] ( pour @xmath435 ) satisfaisant aux conditions de la proposition prcdente . on raisonne par rcurrence sur @xmath20 ; cest vrai pour @xmath568 ou @xmath220 .
si @xmath228 et @xmath77 sont colinaires , on a @xmath569 , donc @xmath570 pour @xmath571 . on peut donc
supposer @xmath228 et @xmath77 indpendants .
on travaille dans lalgbre @xmath572 et pour @xmath573 on note @xmath574 . on a alors
@xmath575 et aussi @xmath576 pour @xmath577 , voir ( * ? ? ?
6 ex.1 p90 et viii 12 lemme 2 p174 ) .
identifions les termes de poids @xmath578 dans les 2 membres de la premire relation ci - dessus , en calculant modulo filtration @xmath545 .
dans le membre de gauche on trouve @xmath579 qui nest diffrent de @xmath580}e_i^{(q)}\in\shu_{n\gamma}$ ] que par le terme de poids @xmath578 de @xmath581 ( pour @xmath582}-x^{(n)}\in\shu_{n\gamma\c}$ ] de filtration @xmath545 ) cest dire par @xmath583 .
daprs la seconde relation ci - dessus ce dernier terme est de filtration @xmath545 .
ainsi le terme de poids @xmath578 du membre de gauche est dans @xmath584 modulo filtration @xmath545 .
dans le membre de droite les termes de poids dans @xmath585 sont ceux de la somme @xmath586 , gale @xmath587 modulo des termes de filtration @xmath545 et de mmes poids .
pour slectionner les termes de poids @xmath578 dans cette dernire somme , on a 2 conditions sur les @xmath588 :
@xmath589 et @xmath590 .
ainsi @xmath591 pour @xmath592 , @xmath593 et @xmath594 ; donc si @xmath595 on a @xmath596 .
ainsi la partie de poids @xmath578 du membre de droite est , modulo filtration @xmath545 , la somme de @xmath597 et de produits @xmath598 avec des @xmath599 .
par rcurrence le @xmath600-ime terme de lun de ces produits est dans @xmath500 modulo filtration @xmath601 , donc chacun de ces produits est dans @xmath584 modulo filtration @xmath545 .
par comparaison des membres de gauche et de droite , on obtient que @xmath597 est dans @xmath584 modulo filtration @xmath545 .
pour @xmath72 , soit @xmath602 une base de @xmath477 sur @xmath455 .
pour @xmath603 racine simple , on choisit @xmath604 et @xmath605 .
par conjugaison par @xmath87 , pour @xmath219 , on peut choisir @xmath606 et @xmath607 de faon que @xmath223=-\alpha^\vee$ ] .
on ordonne la base @xmath608 de @xmath609 de manire quelconque .
pour @xmath610 on choisit des lments @xmath541}$ ] comme en [ 2.4 ] .
pour @xmath611 , on dfinit @xmath612=\prod_{x\in\shb}\,x^{[n_x]}$ ] ( ce produit est pris dans lordre choisi et est en fait fini car presque tous les @xmath613 sont nuls ) .
le poids de @xmath612 $ ] ( et , par dfinition , de @xmath614 ) est la somme pondre @xmath615)=\sum\,n_xpds(x)\in q$ ] des poids des @xmath610 .
ces lments @xmath612 $ ] forment une base de @xmath500 sur @xmath455 .
\1 ) lalgbre gradue @xmath616 dfinie par la filtration de @xmath500 est donc une algbre de polynmes diviss coefficients dans @xmath455 et do nt les indtermines sont les lments de @xmath617 .
\2 ) bien sr ces rsultats sont encore valables sur un anneau @xmath618 quelconque .
si @xmath618 contient @xmath619 , on peut remplacer @xmath541}$ ] par @xmath620 pour tout @xmath610 .
cest une consquence immdiate de [ 2.4 ] daprs ( * ? ? ?
12 n3 prop .
1 ) qui se gnralise sans problme la dimension infinie .
[ 2.7 ] pour @xmath546 , @xmath72 , on peut modifier les @xmath541}$ ] de la proposition [ 2.4 ] de faon satisfaire aux relations supplmentaires suivantes : @xmath621})=\sum_{p+q = n}\,x^{[p]}\otimes x^{[q]}$]et @xmath622})=0 $ ] pour @xmath468 .
pour @xmath546 , @xmath72 , on dit que @xmath623})_{n\in\n}$ ] est une _ suite exponentielle _
associe @xmath556 si elle satisfait aux conditions des propositions [ 2.4 ] ( y compris @xmath541}\in\shu_{n\alpha}$ ] ) et [ 2.7 ] .
plus gnralement , si @xmath624 pour @xmath8 un anneau , il existe des _ suites exponentielles inhomognes _ associes @xmath556 ; on peut prendre @xmath625}=\sum_{p+q = n}\,\lambda^p\mu^q y^{[p]}z^{[q]}$ ] . @xmath626 ( resp .
@xmath616 ) est donc isomorphe comme cogbre gradue ( resp .
bigbre gradue ) une algbre de polynmes diviss , plus prcisment lalgbre symtrique puissances divises @xmath627 .
quitte soustraire @xmath622})$ ] , on peut supposer @xmath622})=0 $ ] , pour @xmath468 . on raisonne par rcurrence sur @xmath20 ; cest clair pour @xmath628 .
supposons le rsultat vrai pour @xmath629 , alors pour @xmath630 et @xmath631 on a @xmath632})=\sum_{p+q = n}\,{[p]}\otimes { [ q]}$ ] ( avec des @xmath633 ) .
la proprit des @xmath620 vis vis de @xmath463 et la relation entre @xmath620 et @xmath541}$ ] montre quil existe des @xmath634 tels que : @xmath621})=\sum_{p+q = n}\,x^{[p]}\otimes x^{[q]}+\sum_{d^\circ(p+r)<n}\,a^n_{p , r}[p]\otimes[r]$ ] .
daprs la proprit de la co - unit @xmath464 , @xmath635 si @xmath636 ou @xmath637 .
par co - commutativit , @xmath638 .
enfin @xmath463 tant compatible la graduation , @xmath635 si le poids de @xmath639 ( de @xmath640.[r]$ ] ) est diffrent de @xmath641 .
daprs la co - associativit et la formule ci - dessus pour @xmath632})$ ] quand @xmath642 on a : @xmath621})\otimes1+\sum_{p+q+r = n}^{r>0}\,x^{[p]}\otimes x^{[q]}\otimes x^{[r]}+\sum_{d^\circ(p+q+r)<n}\,a^n_{p+q , r}[p]\otimes[q]\otimes[r]$ ] @xmath643})+\sum_{p+q+r = n}^{p>0}\,x^{[p]}\otimes x^{[q]}\otimes x^{[r]}+\sum_{d^\circ(p+q+r)<n}\,a^n_{p , q+r}[p]\otimes[q]\otimes[r]$ ] aprs remplacement de @xmath621})$ ] par son expression et simplification , on obtient : @xmath644\otimes[q]\otimes1+\sum_{d^\circ(p+q+r)<n}\,a^n_{p+q , r}[p]\otimes[q]\otimes[r]$ ] @xmath645\otimes[r]+\sum_{d^\circ(p+q+r)<n}\,a^n_{p , q+r}[p]\otimes[q]\otimes[r]$ ] on identifie alors les coefficients de @xmath640\otimes[q]\otimes[r]$ ] dans les 2 membres .
pour @xmath646 , @xmath647 ou @xmath637 , on retrouve des relations connues .
pour @xmath648 , on trouve que @xmath649 .
sachant de plus que @xmath638 , on en dduit que @xmath650 ( pour @xmath651 ) ne dpend que de @xmath639 , on le note donc @xmath652 .
( on laisse au lecteur cet exercice lmentaire , mais ncessitant lexamen de plusieurs cas . )
llment @xmath653}-\sum_{d^\circ(s)<n}^{s\not=0}\,b^n_s[s]$ ] est bien gal @xmath620 modulo filtration @xmath545 et , comme @xmath654 pour @xmath655 , son poids est bien @xmath641 . on a : @xmath656}\otimes x^{[r]}+\sum_{d^\circ(p+r)<n}\,a^n_{p , r}[p]\otimes[r]$ ] @xmath657\otimes1 + 1\otimes[s])-\sum\,b^n_{s}[p]\otimes[r]$ ] o la dernire somme porte sur les @xmath658 tels que @xmath659 , @xmath660 et @xmath651 .
elle est donc gale la seconde somme puisque @xmath661 ou @xmath4 selon que @xmath636 et @xmath662 ou non .
ainsi : @xmath663}\otimes x^{[r]}$ ] on raisonne dans @xmath664 pour une racine @xmath665 ( ou @xmath666 pour @xmath667 ) et un anneau @xmath618 .
si @xmath546 , @xmath665 et @xmath668 , _
lexponentielle tordue _
@xmath669\lambda x= \sum_{n\geq{}0}\,\lambda^nx^{[n]}\in{\widehat\shu}^+_r$ ] nest dfinie que modulo le choix de la suite exponentielle @xmath541}$ ] . ses proprits multiplicatives sont mauvaises : @xmath669(\lambda+\mu)x= \sum_{n\geq{}0}\,(\lambda+\mu)^nx^{[n]}$ ] est en gnral diffrent de @xmath669\lambda x.[exp]\mu x$ ] .
pour @xmath670 il faudrait poser @xmath669\lambda x=([exp]x)^\lambda$ ] ( voir ci - dessous pour @xmath671 ) , mais lextension pour @xmath668 pose problme .
pour tous @xmath551 , @xmath672 , @xmath669(\lambda x).[exp](\mu y)$ ] est un choix possible dexponentielle tordue ( inhomogne ) pour @xmath673 .
par contre @xmath674\lambda x)=1 $ ] et pour le coproduit @xmath675\lambda x=[exp]\lambda x\widehat\otimes[exp]\lambda x$ ] : lexponentielle tordue est un lment de type groupe .
la co - inversion @xmath465 vrifie @xmath676 .
donc @xmath677\lambda x).(\tau[exp]\lambda x)=1=(\tau[exp]\lambda x).([exp]\lambda x)$ ] : lexponentielle tordue est inversible dans
@xmath664 et son inverse vaut @xmath678\lambda x=\sum_{n\geq{}0}\,\lambda^n\tau x^{[n]}$ ] .
bien sr @xmath679}$ ] est lune des suites exponentielles associes @xmath680 , comme la suite @xmath681}$ ] , mais elle nest pas clairement lie ( en gnral ) cette dernire .
pour @xmath618 contenant @xmath619 , ou @xmath219 , on peut considrer les exponentielles classiques @xmath682 qui ont dexcellentes proprits .
\1 ) pour @xmath546 , cherchons sil existe une suite exponentielle @xmath683 autre que @xmath541}$ ] .
daprs la dmonstration de [ 2.7 ] , il suffit de se poser la question quand on change @xmath684}$ ] sans changer @xmath685}$ ] pour @xmath686 . on a alors @xmath687}+y$ ] avec @xmath688 de filtration @xmath689 .
sachant que @xmath690}=\sum_{p+q = n+1}\,x^{[p]}\otimes x^{[q]}$ ] , on a @xmath691 si et seulement si @xmath692 , autrement dit si @xmath693 est primitif dans la cogbre @xmath500 , cest dire si @xmath694 ( * ? ? ? *
ii 1 n 5 cor p13 ) .
ainsi la suite @xmath541}$ ] nest unique qu des lments @xmath695 prs .
autrement dit un autre choix pour @xmath669x$ ] peut scrire @xmath669'x=[exp]x.[exp]x_2.[exp]x_3.\cdots.[exp]x_m.\cdots$ ] .
\2 ) pour @xmath40 racine relle , il y a unicit puisque @xmath696 pour @xmath697 , on a donc toujours @xmath541}=x^{(n)}$ ] et @xmath669x = exp\,x$ ] .
par contre il ny a pas unicit des @xmath541}$ ] pour @xmath698 ou @xmath40 racine imaginaire . on verra en [ 2.12 ] dans des cas affines les bons choix proposs par d. mitzman .
\3 ) pour @xmath546 mais en raisonnant dans @xmath699 , on obtient quil existe @xmath700 tels que @xmath669x = exp(x).exp(y_2).\cdots.exp(y_m).\cdots$ ] . en particulier @xmath677x)^{-1}=\tau[exp]x=\cdots.exp(-y_m).\cdots.exp(-y_2).exp(-x)$ ] .
si les espaces @xmath701 commutent , les exponentielles ci - dessus commutent .
mais lon a naturellement @xmath669\lambda x = exp(\lambda x).exp(\lambda^2y_2).\cdots.exp(\lambda^my_m).\cdots$ ] et non pas @xmath669\lambda x = exp(\lambda x).exp(\lambda y_2).\cdots.exp(\lambda y_m).\cdots$ ] ( qui dailleurs nest en gnral pas dans @xmath699 ) .
il parat donc difficile de faire de @xmath669 $ ] un sous - groupe un paramtre pour @xmath702 .
\4 ) * remarque * : lexpression ci - dessus de @xmath669x$ ] ( pour @xmath546 , @xmath218 ) en terme dexponentielles classiques , montre que , @xmath703 , @xmath541}\in\shu_\q(\oplus_{r\geq{}1}\,{\mathfrak{g}}_{r\alpha\q})\cap\shu_{n\qa}$ ] et @xmath541}.x^{[m]}=\left(\begin{array}{c}n+m \\ n\end{array}\right)x^{[n+m]}$ ] modulo @xmath704 .
ce ntait pas clair auparavant pour @xmath40 imaginaire .
ainsi la base de @xmath500 sur @xmath455 construite en [ 2.6 ] , qui est compatible avec la graduation par les poids dans @xmath39 , est aussi compatible avec la dcomposition de @xmath705 ou @xmath706 en produit tensoriel correspondant la dcomposition en somme directe @xmath707 , condition de regrouper une racine imaginaire et ses multiples positifs .
pour @xmath40 imaginaire , cette base est compatible la filtration par les @xmath708 .
on peut essayer de gnraliser [ 2.8 ] et [ 2.9 ] pour @xmath698 @xmath709 .
on travaille donc dans @xmath710 et , pour simplifier , on se place en rang @xmath220 : @xmath19 est de base @xmath711 et @xmath712 .
on raisonne dans @xmath713 pour un anneau @xmath618 ; ce nest pas le complt de @xmath714 pour la graduation par @xmath39 ; la multiplication de @xmath714 se prolonge car @xmath715 .
ainsi @xmath716 est une algbre commutative avec une comultiplication @xmath717 , mais malheureusement sans co - inversion et dpendant du choix de @xmath711 ( au lieu de @xmath718 ) , voir ci - dessous . on sait que @xmath710 est la bigbre des distributions lorigine du groupe multiplicatif @xmath719 .
la bigbre affine de ce groupe est @xmath720=\z[t , t^{-1}]$ ] et la dualit entre ces 2 bigbres est donne par @xmath721 ( en fait @xmath722 est @xmath723 calcul en llment neutre ) ( * ? ? ?
* ii 4 5.10 ) .
pour @xmath724 , on peut poser @xmath725=\sum_{i\geq{}0}\,(x-1)^i\left(\begin{array}{c}h \\
i\end{array}\right)\in{\widehat\shu}^0_r$ ] ( cest moralement @xmath726 ) . on a bien @xmath725t^n = x^n$ ]
pour @xmath727 ; par contre pour @xmath728 , ceci na un sens que si @xmath729 est nilpotent ( donc @xmath730 ) . on a @xmath731=[x]\otimes[x]\in{\widehat\shu}^0_r{\widehat\otimes}{\widehat\shu}^0_r$ ] : llment @xmath725 $ ] est de type groupe .
le prolongement naturel de @xmath465 devrait donner : @xmath732=\sum_{i\geq{}0}\,(x-1)^i\left(\begin{array}{c}-h \\ i\end{array}\right)=\sum_{p , q}\,(1-x)^{p+q}\left(\begin{array}{c}p+q-1 \\ p\end{array}\right)\left(\begin{array}{c}h \\
q\end{array}\right)$ ] @xmath733 .
mais ce calcul na un sens dans @xmath716 que si @xmath734 est nilpotent , auquel cas @xmath732 $ ] est bien gal @xmath735=\sum_{q\geq{}0}\,(1-x)^qx^{-q}\left(\begin{array}{c}h \\ q\end{array}\right)$ ] , vu le calcul ci - dessus de @xmath736t^{-q}$ ] .
les exponentielles tordues en poids nul semblent donc trop imparfaites .
* p114 - 116 ) on considre des indtermines @xmath737 et @xmath738 .
le polynme @xmath739 est le coefficient de @xmath740 dans le dveloppement de lexpression formelle @xmath741 .
cest un polynme en les @xmath742 coefficients positifs dans @xmath619 . son poids total est @xmath168 si @xmath742 est de poids @xmath600 . on a : @xmath743 , @xmath744 , @xmath745 , @xmath746 , @xmath747 cas particuliers pour certaines spcialisations des indtermines , on a : si @xmath748 pour @xmath749 , alors @xmath750 , si @xmath751 pour @xmath752 , alors @xmath753 , si @xmath754 pour @xmath752 et @xmath755 , alors @xmath756 .
\a ) @xmath757 , si @xmath758 .
\b ) @xmath759 .
\c ) @xmath760 est congru @xmath761 modulo des polynmes de degr total @xmath762 . pour @xmath758 @xmath739 na pas de terme constant .
la premire assertion rsulte de la troisime ligne de la dmonstration de @xcite p.115 puisque @xmath763 .
pour la seconde assertion , on a @xmath764.@xmath765 , do le rsultat .
enfin la dernire assertion est claire .
dans une @xmath766algbre de lie gradue considrons une suite @xmath767 dlments commutant deux deux et de poids @xmath768 . dans lalgbre enveloppante on peut donc calculer @xmath769 qui est de poids @xmath770 .
daprs le lemme la suite des @xmath683 est exponentielle . on peut donc lui associer lexponentielle tordue @xmath771 .
revenons la situation de [ 2.9 ] : @xmath218 et @xmath546 .
supposons que les espaces @xmath772 , @xmath773,@xmath747 , @xmath774,@xmath747 commutent deux deux .
cest vrai dans le cas des algbres de kac - moody affines ( ou semi - simples ) ; mais malheureusement ce sont essentiellement les seuls cas o cela soit vrifi .
pour une suite exponentielle @xmath541}$ ] , il existe des lments @xmath775 pour @xmath749 tels que @xmath669x = exp(x).exp(y_2).\cdots.exp(y_m).\cdots = exp(x+y_2+\cdots+y_m+\cdots)$ ] .
posons @xmath776 , @xmath777,@xmath747 , @xmath778,@xmath747 . on a alors @xmath669x = exp(\sum_{j\geq{}1}\,x_j / j)$ ] .
en identifiant les termes de poids @xmath641 , on obtient @xmath541}=\lambda_n(\underline x)\in\shu$ ] .
grce [ 2.11].a on en dduit facilement par rcurrence que @xmath779 , @xmath780 ; mais les exemples ci - dessous montrent que @xmath781 nest en gnral pas dans @xmath493 .
david mitzman a calcul explicitement une base de lalgbre enveloppante entire @xmath433 dans le cas affine ( et pour un certain sgr @xmath75 ) ( * ? ? ?
* et 4.4.12 ) .
elle sexprime comme la base de [ 2.6 ] ; avec @xmath782}=\lambda_n(h , h,\cdots)=(-1)^n\left(\begin{array}{c}-h \\ n\end{array}\right)$ ] pour @xmath783 et @xmath541}=\lambda_n(x,0,\cdots)=x^{(n)}$ ] pour @xmath556 base de @xmath74 , @xmath219 .
si @xmath548 et @xmath784 le choix de @xmath541}$ ] est @xmath541}=\lambda_n(x , x_2,\cdots , x_m,\cdots)$ ] avec des @xmath785 ; donc @xmath669x = exp(x).\prod_{j=2}^\infty\,exp(\frac{1 } { j}x_j)$ ] , par dfinition des @xmath786 . on
va dcrire ces @xmath787 dans le cas le plus facile , 4.2.5 : @xmath788 est simplement lace ( donc non tordue , de type @xmath789 , @xmath790 ou @xmath791 ) .
voir aussi le corollaire [ 2.3].b .
une algbre de lie @xmath788 de lun de ces types peut scrire comme extension centrale ( avec centre de dimension @xmath220 ) dun produit semi - direct ( avec @xmath3 ) dune algbre de lacets @xmath792 $ ] o @xmath793 est une algbre de lie simple complexe de type @xmath17 , @xmath794 ou @xmath795 .
les racines imaginaires sont les multiples entiers dune racine @xmath796 et , pour @xmath797 , on a @xmath798 o @xmath799 est une sous - algbre de cartan de @xmath793 .
si @xmath800 est une base de coracines dans @xmath799 , lespace @xmath801 admet pour base @xmath802 .
le bon choix ( de mitzman ) pour @xmath803}$ ] est @xmath804 , 3.9.1 et 4.2.3 . on peut remarquer que @xmath803}=\lambda_p(h_{i}\otimes
t^n , h_{i}\otimes t^{2n},\cdots , h_{i}\otimes t^{jn},\cdots)$ ] a pour image @xmath805}\otimes t^{np}=\left(\begin{array}{c}h+p-1 \\ p\end{array}\right)\otimes t^{np}=\left(\begin{array}{c}-h \\ p\end{array}\right)\otimes ( -t^n)^p$ ] dans le quotient @xmath806 $ ] de @xmath807 ( troisime cas particulier de [ 2.11 ] ) .
* lexemple de @xmath808 * : pour le type @xmath809 , on peut prciser encore plus . on a @xmath810\subset end_{\c[t , t^{-1}]}(\c[t , t^{-1}]^2)$ ] .
si @xmath811 , alors , pour @xmath797 , @xmath801 a pour base @xmath812 .
la reprsentation naturelle de @xmath813 sur @xmath814 ^ 2 $ ] induit un homomorphisme @xmath299 de @xmath807 dans @xmath815}(\c[t , t^{-1}]^2)$ ] ou de @xmath816 dans @xmath817 . on a @xmath818})=(-t^n)^p\left(\begin{array}{c}-h \\ p\end{array}\right)=t^{np}\left(\begin{array}{c}h+p-1 \\ p\end{array}\right)\in end_{\z(\!(t)\!)}(\z(\!(t)\!)^2)$ ] . on en dduit
que @xmath819}(\z[t , t^{-1}]^2)$ ] et que , pour @xmath820 dans un anneau @xmath8 , @xmath821\lambda h_n)=diag(u , v)\in end_{k(\!(t)\!)}(k(\!(t)\!)^2)$ ] avec @xmath822 et @xmath823 . en particulier @xmath821\lambda h_n)$ ] est bien inversible et mme dans @xmath824 .
\1 ) pour @xmath218 , @xmath825 est une sous@xmath432algbre de @xmath500 .
pour @xmath219 , @xmath826 admet pour base sur @xmath455 les @xmath827 , @xmath435 .
pour @xmath548 et @xmath435 , on choisit une base @xmath828 de @xmath829 et des suites exponentielles @xmath554}$ ] pour @xmath830 . daprs
[ 2.9].4 ci - dessus , ces @xmath554}$ ] sont dans
@xmath826 et , daprs [ 2.6 ] , les @xmath612=\prod_{x\in\shb_{(\alpha)}}\,x^{[n_x]}$ ] ( pour @xmath831 , @xmath832 , @xmath833 ) forment une base de @xmath826 . on en dduit , en particulier , que @xmath826 est une sous@xmath432bigbre co - inversible de @xmath500 .
si @xmath834 est un ensemble clos , @xmath835 est une algbre de lie ; lalgbre @xmath836 engendre par les @xmath826 pour @xmath837 est une sous - bigbre co - inversible de @xmath500 de base les @xmath612 $ ] pour @xmath838 avec @xmath839 .
lalgbre enveloppante de @xmath840 est @xmath841 et @xmath842 .
les algbres @xmath836 et @xmath843 sont stables par laction adjointe de @xmath836 .
pour @xmath844 ( resp @xmath845 ) on a @xmath846 ( resp .
@xmath847 ) . on a @xmath848 . pour un anneau @xmath8 , on note @xmath849 et @xmath850 .
\2 ) si @xmath851 est clos , lalgbre @xmath836 est gradue par @xmath447 . on peut la graduer par le degr total .
sa compltion positive @xmath852 est une sous@xmath432algbre de @xmath853 .
on peut de mme considrer lalgbre @xmath854 ( en gnral diffrente de @xmath855 ) pour tout anneau @xmath8 .
si @xmath834 est clos , on peut considrer la compltion positive @xmath856 de @xmath836 selon le degr total ( dans @xmath455 ) ; cest une sous@xmath432algbre de @xmath857 . on dfinit
de mme @xmath858 .
lintersection de @xmath858 avec @xmath859 est @xmath860 .
bien sr @xmath858 et @xmath861 sont des sous@xmath862modules de @xmath502 si @xmath109 est un idal de @xmath108 , @xmath863 ( resp @xmath864 ) est un idal de @xmath858 ( resp @xmath865 ) .
\3 ) les rsultats prcdents et suivants sont bien sr encore valables si on remplace @xmath507 par @xmath508 ou par un conjugu de @xmath509 par un lment du groupe de weyl @xmath87 .
on peut en particulier dfinir des compltions ngatives @xmath866 ( selon loppos du degr total ) .
soit @xmath820 un poids dominant de @xmath146 , @xmath867 .
considrons le @xmath868module irrductible @xmath869 de plus haut poids @xmath820 ( * ? ? ?
* et 10 ) .
si @xmath870 est un vecteur non nul de poids @xmath820 , @xmath869 est un quotient du module de verma @xmath871 ( dans le cas non symtrisable , on prend plutt pour @xmath869 le quotient intgrable maximal de @xmath872 comme en ( * ? ? ?
* p28 ) ) .
le sous@xmath449module @xmath873 est une @xmath32forme de @xmath869 ; on peut dfinir @xmath874 , pour tout anneau @xmath8 .
ces modules sont intgrables au sens de @xcite et gradus par @xmath448( translation prs ) avec des espaces de poids de dimension finie .
en particulier on a une reprsentation diagonalisable de @xmath146 dans @xmath875 .
il est clair que , pour tout anneau @xmath8 , @xmath876 est en fait un @xmath877module .
on dfinit de mme des modules intgrables plus bas poids @xmath876 pour @xmath878 ; ce sont des @xmath879modules .
on va tudier ci - dessous deux groupes de kac - moody `` maximaux '' dits _ positif _ et _ ngatif _ qui contiennent le groupe minimal tudi dans la premire partie .
ils se dduisent lun de lautre par lchange de @xmath507 et @xmath508 ; on va donc essentiellement se concentrer sur le groupe positif .
ce groupe sera not @xmath262 , les sous - groupes ou les algbres associes seront en gnral nots avec un exposant @xmath880 ou @xmath881 .
les analogues pour @xmath508 seront nots avec un exposant @xmath882 ou @xmath883 .
soit @xmath851 un ensemble clos de racines .
on considre lalgbre @xmath836 de [ 2.13 ] , qui ne dpend que de @xmath17 et @xmath108 ( et non de @xmath75 ) .
cest une @xmath32bigbre co - inversible , cocommutative , gradue par @xmath447 : @xmath884 ; tous ses espaces de poids sont libres de dimension finie sur @xmath455 .
on considre son dual restreint @xmath885=\bigoplus_{\alpha\in\n\psi}\,\shu(\psi)_\alpha^*$ ] .
cest une @xmath32bigbre commutative et co - inversible ( sa co - inversion est le dual de @xmath465 encore not @xmath465 , qui est un isomorphisme dalgbres ) ; cest dire lalgbre dun schma en groupe affine @xmath886 que lon verra comme un foncteur en groupes : pour un anneau @xmath618 , @xmath887,r)$ ] . on note @xmath888 .
pour @xmath889 on note @xmath890 , pour @xmath891 on note @xmath892 ; @xmath886 et @xmath893 sont des sous - schmas en groupe de @xmath894 .
la dfinition de @xmath895=\z[{\mathfrak{u}}^{ma}_\psi]$ ] dans ( * ? ? ?
* p 19 - 21 ) est plus complique car elle ne se limite pas @xmath851 . dans notre cas le lemme 3 de dit bien que @xmath896 $ ] est le @xmath766dual restreint de @xmath897 et alors @xmath895 $ ] ( dfini page 21 de ) est bien le dual restreint de @xmath836 .
la base de @xmath836 indexe par des @xmath838 explique en [ 2.13].1 fournit par dualit une base @xmath898 de @xmath885 $ ] , indexe par les mmes @xmath614 .
comme @xmath899=\sum_{p+q = n}\,[p]\otimes[q]$ ] , on a @xmath900 .
ainsi @xmath885 $ ] est une algbre de polynmes sur @xmath455 do nt les indtermines @xmath901 sont indexes par les @xmath902 .
on retrouve ainsi un rsultat de olivier mathieu ( * ? ? ? * lemme 2 p41 ) .
inversement si on admet ce rsultat , @xmath885 $ ] admet comme base des monmes , son dual restreint @xmath836 admet une base gradue qui a de bonnes proprits pour le coproduit @xmath463 .
un raisonnement analogue au raisonnement dunicit de [ 2.9 ] permet alors de construire des suites exponentielles .
[ 3.2 ] pour @xmath108 comme ci - dessus et @xmath618 un anneau , @xmath903 sidentifie au sous - groupe multiplicatif de @xmath904 form des produits @xmath905\lambda_xx$ ] , pour des @xmath906 , le produit tant pris dans lordre ( quelconque ) choisi sur @xmath617 .
lcriture dun lment de @xmath903
sous la forme dun tel produit est unique .
\1 ) si @xmath108 est fini et donc form de racines relles , on retrouve le groupe de sommes formelles de @xcite , cest dire lunique schma en groupe lisse connexe unipotent dalgbre de lie @xmath907 , ( * ? ? ?
. 1 p547 ) et aussi [ 3.4 ] ci - dessous .
dans ce cas les exponentielles sont classiques et les produits sont finis .
\2 ) pour @xmath891 , le choix de llment de base @xmath476 de @xmath477 dtermine un isomorphisme @xmath250 du groupe additif @xmath237 sur @xmath283 donn par @xmath908 .
\3 ) dans les produits infinis proposs et pour tout @xmath435 , tous les facteurs sauf un nombre fini sont gaux @xmath220 modulo des termes de degr total @xmath909 .
ces produits infinis sont donc bien dans @xmath910 .
le @xmath911dual de @xmath912 $ ] est @xmath910 .
donc @xmath903 sidentifie lensemble des lments @xmath913 tels que @xmath914 @xmath915 et @xmath916 pour tous @xmath917 $ ] @xmath918 ou encore @xmath919 .
ainsi @xmath903 est lensemble des lments de @xmath910 de terme constant @xmath220 et de type groupe .
parmi ces lments il y a les produits infinis de lnonc .
mais @xmath912 $ ] est une algbre de polynmes dindtermines @xmath901 , @xmath902 .
par dfinition @xmath920\lambda_yy)=\lambda_x$ ] .
ainsi ces produits infinis fournissent une et une seule fois toutes les applications de @xmath921 dans @xmath618 , donc tous les homomorphismes de @xmath912 $ ] dans @xmath618 . [ 3.3 ] soient @xmath922 des sous - ensembles clos de racines .
\a ) @xmath923 est un sous - groupe ferm de @xmath924 et @xmath925 $ ] est une algbre de polynmes dindtermines indexes par @xmath926 .
\b ) si @xmath927 est galement clos , alors on a une dcomposition unique @xmath928 .
\c ) si @xmath109 est un idal de @xmath108 , alors @xmath923 est un sous - groupe distingu de @xmath924 et on a un produit semi - direct si , de plus , @xmath927 est clos .
\d ) si @xmath927 est rduit une racine @xmath40 , alors le quotient @xmath929 est isomorphe au groupe additif @xmath237 pour @xmath40 relle et au groupe unipotent commutatif @xmath930 pour @xmath40 imaginaire .
\e ) si @xmath931 , @xmath932 implique @xmath933 , alors @xmath929 est commutatif
. le groupe @xmath934 est isomorphe au groupe additif @xmath935 .
\f ) si @xmath936 est un autre sous - ensemble clos et si @xmath937 , @xmath938 , @xmath939 implique @xmath940 , alors @xmath941 contient le groupe de commutateurs @xmath942 \a ) @xmath943 est une sous - bigbre co - inversible de @xmath836 , donc @xmath923 est un sous - groupe de @xmath924 .
ce sous - groupe est ferm , dfini par lannulation des indtermines @xmath901 pour @xmath944 .
pour la dernire assertion il suffit dans [ 3.2 ] dordonner @xmath617 de faon que @xmath945 soit en dernier ( droite ) .
lassertion b ) rsulte aussitt de la proposition [ 3.2 ] \c ) plaons nous dabord sur @xmath3 ( ou un anneau contenant @xmath619 )
. on peut alors remplacer dans [ 3.2 ] les exponentielles tordues par les exponentielles habituelles .
des calculs classiques montrent alors que @xmath946 est distingu dans @xmath947 . en fait il est clair que @xmath947
ou @xmath946 est le groupe introduit dans @xcite presque sous le mme nom et on peut donc faire rfrence @xcite .
le normalisateur de @xmath923 dans @xmath924 est ferm dans @xmath924 ( * ? ? ?
* ii 1 3.6b ) .
comme @xmath885 $ ] est une algbre de polynmes et que lon vient de voir que ce normalisateur est gal @xmath886 sur @xmath3 , il en est de mme sur @xmath455 .
\d ) ordonnons @xmath921 de manire que @xmath602 soit en premier .
alors , daprs [ 2.6 ] et [ 3.2 ] , un lment de @xmath903 scrit de manire unique sous la forme @xmath948\lambda_xx)z$ ] avec des @xmath906 et @xmath949 . soit @xmath950 ( vide si @xmath40 est relle ) ; il reste montrer que @xmath951\lambda_xx).(\prod_x\,[exp]\lambda'_xx)=\prod_x\,[exp](\lambda_x+\lambda'_x)x$ ] ( modulo @xmath952 )
. mais on a vu dans la remarque [ 2.9].4 que , pour @xmath953 , @xmath541},y^{[n]}\in\shu^\alpha$ ] ( ces lments commutent donc modulo @xmath954 ) , et que @xmath541}.x^{[m]}=\left(\begin{array}{c}n+m \\ n\end{array}\right)x^{[n+m]}$ ] modulo @xmath954 .
le rsultat est alors clair .
\e ) il est clair que les lments de @xmath955 commutent modulo lidal @xmath956 .
vus [ 3.2 ] et les bonnes proprits de la base des @xmath612 $ ] ( [ 2.9].4 ) , @xmath957 est commutatif isomorphe au produit direct des groupes @xmath958 pour @xmath959 .
\f ) comme en c ) on a le rsultat sur @xmath3 .
le centralisateur de @xmath960 dans @xmath961 est donc gal @xmath961 sur @xmath3 .
mais ce centralisateur est ferm
dans @xmath961 ( * ? ? ?
* ii 1 3.6c ) .
comme @xmath962 $ ] est un anneau de polynmes , ce centralisateur est gal @xmath961 .
soient @xmath108 un sous - ensemble clos de @xmath507 et @xmath20 le plus petit des degrs totaux ( @xmath963 ) des @xmath964 .
choisissons @xmath837 de degr total @xmath20 , alors @xmath965 est un idal de @xmath108 .
ainsi @xmath923 est un sous - groupe distingu de @xmath924 et le quotient @xmath929 est isomorphe au groupe additif @xmath237 ou @xmath930 .
si @xmath40 est relle on a un produit semi - direct .
ce procd fournit une numrotation de @xmath108 par @xmath495 ( ou un intervalle @xmath966 $ ] de @xmath495 ) : @xmath967 .
pour @xmath435 , on note @xmath968 .
daprs le lemme [ 3.3 ] le groupe @xmath969 est distingu dans @xmath924 et @xmath970 est isomorphe @xmath971 .
la proposition [ 3.2 ] montre que @xmath924 est limite projective des @xmath972 .
ainsi @xmath924 est un groupe pro - unipotent ( ou mme unipotent si @xmath108 est fini ) .
pour @xmath973 , notons @xmath974 lensemble des racines de @xmath108 de degr total @xmath975 .
alors les @xmath976 sont distingus dans @xmath924 et @xmath977 est isomorphe au groupe additif @xmath978 .
le groupe @xmath979 est limite projective des @xmath980 .
le _ groupe de borel _ @xmath981 est par dfinition le produit semi - direct @xmath982 pour laction suivante de @xmath146 sur @xmath894 .
pour un anneau @xmath618 , @xmath983 agit sur @xmath984 par des automorphismes de bigbre : sur @xmath985 , laction @xmath986 de @xmath987 est la multiplication par @xmath988 . on en dduit aussitt
laction sur @xmath989 : @xmath990\lambda x=[exp]\alpha(t)\lambda x$ ] si @xmath991 .
il est clair que les groupes @xmath924 de [ 3.3 ] sont stables par laction de @xmath146 .
si @xmath603 est une racine simple on a @xmath992 .
on dfinit comme ci - dessus @xmath993 avec @xmath994 $ ] dans le dual de @xmath995 et isomorphe @xmath237 par @xmath996 , @xmath997 . o. mathieu dfinit un schma en groupe affine _ parabolique minimal _ @xmath998 ( contenant @xmath981 et associ @xmath999 ) et un sous - groupe ferm @xmath1000 ( associ @xmath1001 ) ; il montre que @xmath1002 ( * ? ? ?
* lemme 8 p26 ) .
le schma en groupe @xmath1003 contient les sous - groupes ferms @xmath146 , @xmath251 et @xmath993 , plus prcisment @xmath1004 $ ] est contenu dans le dual de @xmath1005 et sa restriction @xmath1006 ( resp .
@xmath1007 ) est @xmath1008 $ ] ( resp .
il est clair ( ) que @xmath1003 est le groupe rductif de sgr @xmath1010 il agit sur @xmath493 et est isogne au produit de @xmath1011 par un tore dploy .
llment @xmath1012 de @xmath1013 normalise @xmath146 , vrifie @xmath1014 et change @xmath1015 , @xmath1016 ; il agit sur @xmath493 comme @xmath212 ( [ 1.4 ] ) .
enfin la double classe ( grosse cellule ) @xmath1017 de @xmath1018 est un ouvert dense et se dcompose de manire unique sous la forme @xmath1019 .
mathieu se place dans le cas dun sgr @xmath75 libre , colibre , sans cotorsion et de dimension @xmath1020 .
bien sr ces hypothses sont inutiles pour la plupart de ses rsultats ( cest particulirement clair pour la dernire ) . on se placera cependant
dans ce cadre ( depuis lalina prcdent ) jusqu la fin de ce paragraphe o on examinera la gnralisation grce aux rsultats de [ 1.3 ] . on abandonne en gnral dans la suite ( et jusquen 3.17 ) lindice @xmath75 et on abrge parfois @xmath1021 en @xmath1022 , @xmath1018 en @xmath1023 , @xmath1024 en @xmath1025 , etc .
* xviii 2 ) , @xcite , ( * ? ? ?
* i et ii ) on considre des schmas sur @xmath455 , mme si on naura essentiellement
besoin de raisonner que sur des ( sous - anneaux de ) corps , ventuellement de caractristique positive .
soient @xmath1026 et @xmath1027 une dcomposition rduite de @xmath96 .
on considre le schma @xmath1028 et le schma de demazure @xmath1029 .
le schma @xmath1030 est laffinis de @xmath1031 , cest dire @xmath1032)$ ] , il ne dpend que de @xmath96 ( * ? ? ?
* p 40 , l -15 -1 ) . en particulier @xmath1033 .
si @xmath1034 ( pour lordre de bruhat - chevalley ) , il existe une dcomposition rduite @xmath1035 de @xmath1036 extraite de @xmath1037 .
si dans lcriture de @xmath1031 , on remplace par @xmath1022 les @xmath1038 correspondant des @xmath1039 absents de @xmath1035 , on obtient un sous - schma ferm de @xmath1031 clairement isomorphe @xmath1040 .
limmersion ferme @xmath1041 induit un morphisme @xmath1042 entre les affiniss ; cest aussi une immersion ferme , indpendante des choix de @xmath1037 et @xmath1035 ( * ? ? ?
* bas de page 255 ) . on a donc un systme inductif de schmas @xmath1030 et on note @xmath1043 le ind - schma limite inductive de ce systme ; cest le _ groupe de kac - moody positivement maximal _ associ @xmath75 .
chaque morphisme @xmath1044 est une immersion ferme .
on verra essentiellement @xmath1043 comme un foncteur sur la catgorie des anneaux : @xmath1045 .
en fait @xmath1043 est un ind - schma en groupe : pour des dcompositions rduites @xmath1027 et @xmath1046 , il existe un morphisme naturel de @xmath1047 dans @xmath1048 pour un certain @xmath1049 ( * ? ? ? *
p 41 et 45 ) .
ce morphisme se factorise donc par @xmath1050 .
ceci permet de dfinir la multiplication dans @xmath1043 .
la dfinition de linversion est claire .
pour cette structure de groupe le compos @xmath1051 est simplement la multiplication .
\1 ) pour @xmath1052 , on a dfini en [ 2.14 ] un @xmath1053module plus haut poids @xmath875 qui est intgrable , donc un @xmath1054module .
pour un anneau @xmath8 le module @xmath876 est stable par @xmath1055 donc par @xmath1056 .
on obtient ainsi une reprsentation de @xmath1021 dans @xmath875 .
lintgrabilit de @xmath875 @xcite montre en fait que @xmath875 est un @xmath1057module donc aussi un @xmath1058module ( @xmath125 ) .
ce module est `` localement fini '' , plus prcisment pour tout sous@xmath432module @xmath1059 de rang fini de @xmath875 , il existe un sous@xmath432module facteur direct de rang fini @xmath1060 de @xmath875 contenant @xmath1059 , tel que @xmath1061 soit un @xmath1062module pour tout anneau @xmath8 . on en dduit aussitt
un morphisme de @xmath1030 dans @xmath1063 , @xmath1064 . on obtient ainsi une reprsentation de @xmath1043 dans @xmath1063 .
\2 ) soit @xmath1065 le module plus haut poids prcdent : @xmath1066 .
pour un anneau @xmath8 , on dfinit le complt ngatif @xmath1067 de @xmath1068 : @xmath1069 .
il est clair que @xmath1067 est un @xmath1070module gradu ; en particulier cest un @xmath1071module .
\3 ) pour un anneau @xmath8 , laction adjointe @xmath450 de lalgbre @xmath1072 sur elle mme ( [ 2.1].5 ) induit une action , dite adjointe et note @xmath489 , du groupe @xmath1073 sur @xmath1072 .
laction de @xmath276 est dfinie par les poids , do la reprsentation adjointe @xmath489 de @xmath1074 sur @xmath1072 .
cette action stabilise videmment @xmath859 et @xmath1075 .
pour @xmath219 , @xmath238 et @xmath1076 , @xmath1077 .
supposons la matrice @xmath17 de cartan @xmath1078 fini et @xmath87 fini .
soit @xmath1079 le groupe rductif dploy sur @xmath455
de sgr @xmath75 ; daprs lhypothse de mathieu il est semi - simple simplement connexe .
le groupe @xmath1080 est le sous - groupe de borel @xmath1022 de @xmath1079 ( [ 3.2].1 ) et @xmath1081 est un sous - groupe parabolique minimal de @xmath1079 .
pour une dcomposition rduite @xmath1027 dans @xmath87 , le produit dfinit un morphisme @xmath1082 .
si on restreint @xmath299 aux grosses cellules , on obtient un isomorphisme de @xmath1083 sur son image @xmath1084 .
@xmath1085 est llment de plus grande longueur de @xmath87 , cette image est un ouvert dense : la grosse cellule @xmath1086 de @xmath1079 . on a donc la suite
de morphismes : avec @xmath97 et @xmath1087 des immersions ouvertes . de plus limage @xmath1086 de @xmath1087 rencontre toutes les fibres @xmath1088 de @xmath1079 pour @xmath9 premier . [ 3.9 ] dans ces conditions @xmath299 identifie @xmath1089 $ ] @xmath1090 $ ] .
ainsi @xmath1079 sidentifie @xmath1091 lui mme gal @xmath1043 .
si @xmath1092 , on trouve de la mme manire que @xmath1030 est la sous - varit de schubert de @xmath1079 associe @xmath96 ( au moins sur @xmath3 ) .
regardons dabord sur @xmath3 .
comme @xmath1079 est un groupe affine , @xmath299 se factorise par @xmath1093 . on a donc la suite
de morphismes : ce compos est une immersion ouverte dimage dense .
pour les algbres de fonctions , comme @xmath97 est dimage dense @xmath1094 est injectif et @xmath1095 est lidentit par dfinition .
donc @xmath1096 est injectif et @xmath1097 est dominant .
ainsi limage de @xmath1097 contient un ouvert dense @xmath1098 de @xmath1099 .
soit @xmath1100 , @xmath1101 , il existe donc @xmath1102 tels que @xmath1103 , ainsi @xmath1104 .
mais @xmath1105 est injectif , on en dduit que @xmath1106 , @xmath1107 @xmath1108 est injectif .
ainsi @xmath1043 est un sous - groupe de @xmath1079 contenant un ouvert dense : on a @xmath1109 . on a le diagramme : les flches horizontales sont injectives car @xmath1094 et @xmath1110 le sont .
le schma @xmath1111 est produit de groupes additifs et multiplicatifs , la flche verticale de gauche est donc injective .
ainsi toutes les flches sont injectives .
pour montrer que @xmath1090 = \pi^ * \z[{\mathfrak{g}}]$ ] il suffit donc de montrer que @xmath1112 \cap \c[{\mathfrak{g}}]=\z[{\mathfrak{g}}]$ ] .
ceci est clairement vrai si on remplace @xmath455 par @xmath619 .
donc si ce nest pas vrai pour @xmath455 , il existe @xmath1113 et @xmath1114 \cap \q[{\mathfrak{g}}]$ ] tels que @xmath1115 $ ] mais @xmath1116 $ ] .
soit @xmath9 un diviseur premier de @xmath20 ; quitte modifier @xmath130 on peut supposer @xmath1117 .
mais alors @xmath1118 est une fonction non identiquement nulle sur @xmath1088 alors quelle est nulle sur @xmath1119 .
cest absurde car @xmath1119 est un ouvert dense de @xmath1088 .
soit @xmath1120 . on note
@xmath1121 , @xmath1122 , @xmath1123 , @xmath1124 et @xmath1125 .
daprs [ 3.3 ] on a @xmath1126 .
supposons @xmath1127 de type fini et notons @xmath1128 le groupe rductif dploy sur @xmath455 de sgr @xmath1129 .
il est clair que son groupe unipotent maximal positif est @xmath1130 et son sous - groupe de borel positif @xmath1131 .
enfin , pour @xmath1132 , son parabolique minimal associ @xmath600 est @xmath1133 , il vrifie @xmath1134 .
le groupe de weyl ( fini ) @xmath1135 admet un lment de plus grande longueur @xmath1136 .
@xmath1137 est isomorphe au sous - groupe parabolique @xmath1138 de @xmath1043 .
si @xmath1127 nest pas forcment de type fini , la dmonstration ci - dessous prouve que le ind - schma en groupe limite inductive des @xmath1030 pour @xmath1139 est le sous - groupe parabolique @xmath1140 produit semi - direct par @xmath1141 du groupe de kac - moody @xmath1142 . on a @xmath1143 , @xmath1144 et @xmath1145 . on en dduit facilement que @xmath1146 ,
et , comme laffinis dun produit est le produit des affiniss , @xmath1147 . on a dfini @xmath283 et @xmath1148
pour @xmath891 ( en [ 3.2].2 ) ou pour @xmath1149 simple ( en [ 3.5 ] ) .
pour @xmath204 , @xmath1150 est un lment du normalisateur dans @xmath1043 de @xmath1151 ; il induit sur @xmath1151 la rflexion @xmath1152 . de plus @xmath1153 agit sur @xmath493 comme @xmath212 ( [ 1.4 ] et [ 3.7].3 ) .
si @xmath219 et @xmath476 est une base de @xmath477 , il existe @xmath1026 tel que @xmath1154 est une racine simple . on considre
une dcomposition @xmath1155 , @xmath1156 et on a @xmath1157 avec @xmath1158 . on pose alors
@xmath1159 et @xmath1160 , pour @xmath167 dans un anneau @xmath618 .
ces dfinitions de @xmath283 et @xmath250 ne dpendent que de @xmath40 et @xmath476 ( et non des autres choix effectus ) .
pour @xmath891 ( ou @xmath1161 simple ) elles concident avec celles de [ 3.1 ] , [ 3.2 ] ou [ 3.5 ] .
supposons que @xmath1162 , donc @xmath1163 pour @xmath1164 . on doit montrer que @xmath1165 est conjugu de @xmath1166 par @xmath1167 .
mais on sait que les @xmath1168 vrifient les relations de tresse ( ( * ? ?
* ( d ) p551 ) ou calcul classique dans le groupe rductif @xmath1169 quand @xmath1170 est dordre fini ) , que @xmath1171 et on connat les relations de commutation entre les @xmath1153 et les lments du tore .
ainsi , pour toute expression rduite @xmath1172 de @xmath1173 , on peut rduire lexpression ci - dessus de @xmath1167 sous la forme @xmath1174 avec @xmath1175 .
il suffit donc de montrer que si @xmath1176 , il existe une dcomposition rduite @xmath1172 de @xmath96 pour laquelle un @xmath1167 comme ci - dessus vrifie la relation @xmath1177 et @xmath1178 .
daprs @xcite ou , pour un nonc plus prcis ( * ? ? ?
* p85 ) , on est ramen lun des deux cas suivants : @xmath1179 de type fini ou @xmath1180 et @xmath1181 de type fini .
tout se passe alors dans le groupe rductif @xmath1128 et le rsultat est connu .
si @xmath1161 est simple , la concidence des deux dfinitions de @xmath250 et @xmath283 est claire .
si @xmath891 , il existe une suite @xmath1182 telle que @xmath1183 est une racine simple et @xmath1184 @xmath1185 .
pour vrifier la concidence des deux dfinitions de @xmath250 et @xmath283 , on doit vrifier que , si @xmath204 , @xmath891 , @xmath1186 et @xmath1187 , alors @xmath1188 ; cest la dfinition prcise du produit semi - direct @xmath1189 de [ 3.5 ] . on a dfini en [ 1.6 ] le groupe de kac - moody minimal @xmath1190 .
on choisit toujours @xmath75 comme la remarque [ 3.5 ] et on prend les @xmath476 , @xmath1191 de [ 1.4 ] dans la base @xmath617 de [ 2.6 ] .
pour un anneau @xmath8 , les gnrateurs de @xmath1192 sont les @xmath239 , @xmath219 , @xmath238 et @xmath1193 , on les envoie sur les lments de mme nom dans @xmath1194 . on obtient ainsi un homomorphisme de foncteurs en groupes @xmath1195 .
il sagit de vrifier les relations ( kmt3 ) ( kmt7 ) dans @xmath1194 .
la relation ( kmt3 ) espre entre @xmath239 et @xmath1196 provient de la relation constate entre @xmath1197 et @xmath1198 dans aut@xmath1199 .
comme @xmath244 est une paire prnilpotente , il existe @xmath108 finie close dans @xmath117 contenant @xmath40 et @xmath228 .
daprs [ 3.11 ] on peut supposer @xmath1200 .
daprs la remarque [ 3.2].1 cette relation est en fait constate dans @xmath1201 , do le rsultat . les relations ( kmt4 ) ( kmt6 ) nimpliquent que des lments du groupe rductif de rang @xmath220 @xmath1003 , elles sont trivialement satisfaites .
enfin ( kmt7 ) est clair daprs [ 3.11 ] .
\1 ) lhomomorphisme @xmath97 envoie aussi @xmath1202 sur llment de mme nom de @xmath1203 ( pour @xmath40 racine simple ) .
limage @xmath1204 de @xmath1205 est donc engendre par @xmath146 et les @xmath1202 pour @xmath40 racine simple .
par construction @xmath97 est un isomorphisme sur @xmath146 et @xmath1206 si la dcomposition @xmath1207 est rduite ( second alina de [ 3.8 ] ) .
donc @xmath97 est un isomorphisme de @xmath1205 sur @xmath1204 ( que lon note encore @xmath1205 ) .
\2 ) en conjuguant par un lment @xmath20 de @xmath1208 tel que @xmath1209 , on trouve dans @xmath1043 un sous - groupe analogue @xmath894 mais associ @xmath1210 .
ainsi @xmath1043 ne dpend pas du choix de @xmath507 dans sa classe de @xmath1211conjugaison .
par contre lalgbre @xmath1212 dpend a priori de ce choix et si on change @xmath507 en @xmath508 , le groupe obtenu @xmath1213 est diffrent .
\3 ) lhomomorphisme @xmath97 induit un isomorphisme du groupe @xmath283 sur le sous - groupe de mme nom de @xmath1043 : cest clair par dfinition pour @xmath891 , on sy ramne pour @xmath1214 grce [ 3.11 ] .
[ 3.13 ] si @xmath8 est un corps , alors @xmath1215 est injectif .
si @xmath8 est un corps , on notera souvent avec la lettre romaine correspondante lensemble des points sur @xmath8 dun schma .
la proposition nous permet donc didentifier @xmath1216 un sous - groupe de @xmath1217 . daprs ( * ? ? ?
8.4.1 ) @xmath5 est muni dune donne radicielle jumele entire et donc dun systme de tits @xmath1218 avec @xmath1219 .
alors @xmath1220 est un sous - groupe parabolique de @xmath5 , mais par construction il ne peut contenir aucun @xmath1221 pour @xmath40 simple , donc @xmath1222 et ker@xmath1223 est contenu dans @xmath1224 et tous ses conjugus
. daprs ( * ? ? ?
* et 1.2.3.v ) , lintersection de tous ces conjugus est rduite @xmath1225 .
mais par construction @xmath1226 est injectif , do le rsultat .
\1 ) on a vu en [ 3.7].3 que , pour un anneau @xmath8 , @xmath1072 est un @xmath1227module . daprs
[ 2.1].4 la restriction @xmath1228 de cette reprsentation stabilise @xmath1229 ( et @xmath1230 ) et y induit la mme reprsentation que la reprsentation adjointe de @xmath1192 .
\2 ) pour @xmath1052 , on a construit en [ 3.7].1 le module plus haut poids @xmath1231 pour @xmath1043 .
il est clair que la restriction @xmath1079 ( via @xmath97 ) de cette reprsentation est la reprsentation classique plus haut poids de @xmath1079 .
\3 ) en [ 3.7].2 on a construit une reprsentation de @xmath1232 sur le complt ngatif @xmath1067 de @xmath1059 . il est clair que la restriction @xmath1233 ( via @xmath97 ) de cette reprsentation est le complt de la reprsentation de @xmath1079 ci - dessus ( restreinte @xmath1233 ) .
[ 3.15 ] soit @xmath1234 une dcomposition rduite de @xmath1026 et @xmath1235 le morphisme de multiplication .
si @xmath8 est un corps , alors @xmath1236 est surjectif .
il faut revenir un peu sur les dfinitions de [ 3.6 ] .
le groupe @xmath1237 agit sur @xmath1238 par @xmath1239 et @xmath1031 est le quotient schmatique ; on note @xmath299 le morphisme de passage au quotient .
le schma @xmath1030 est laffinis de @xmath1031 , on note @xmath1240 le morphisme naturel . on a donc @xmath1241 . on se place
maintenant sur le corps @xmath8 ( sans forcment le noter dans les noms ) .
soit @xmath1052 un poids
dominant entier rgulier et @xmath869 la reprsentation intgrable de plus haut poids @xmath820 ( ( * ? ? ?
* p246 ) ou [ 3.7 ] ci - dessus ) ; les groupes @xmath1021 et @xmath1023 agissent dessus .
on choisit un vecteur de plus haut poids @xmath1242 ; pour @xmath1026 , @xmath1243 est dfini au signe prs et le sous - espace vectoriel @xmath1244 est de dimension finie .
de plus @xmath1245 si @xmath1246 pour lordre de bruhat - chevalley , mais si @xmath1247 , @xmath1248 a une composante nulle sur @xmath1249 . on note
@xmath1250 ladhrence de @xmath1251 dans @xmath1252 ; cest une sous - varit affine ferme de @xmath1252 . on a @xmath1253 il contient donc @xmath1254 si @xmath1246 ( ( * ? ? ?
* p64 ) formul en caractristique @xmath4 , mais qui se gnralise ) .
ainsi @xmath1255 , contient louvert de @xmath1250 dfini par la non nullit de la coordonne sur @xmath1256 . on a un morphisme @xmath738 de @xmath1238 dans @xmath1250 donn par la `` formule nave '' @xmath1257 ,
il se factorise en un morphisme @xmath1258 de @xmath1031 dans @xmath1250 ( ( * ? ? ?
* p65 , 66 ) en caractristique @xmath4 ) puis , comme @xmath1250 est affine et @xmath1259 une affinisation , en un morphisme @xmath1260 , simplement donn par laction de @xmath1261 sur @xmath1242 [ 3.7].1 . on a donc le diagramme : si @xmath1246 , il existe une dcomposition rduite @xmath1262 de @xmath1263 `` extraite '' de @xmath1037 .
le sous - schma ferm de @xmath1238 form des lments do nt le j - me facteur est @xmath220 si @xmath1039 est omis dans
@xmath1262 est isomorphe @xmath1264 .
on obtient ainsi un diagramme do nt toutes les flches verticales sont injectives : on note @xmath1265 qui contient @xmath1266 .
par rcurrence il suffit de montrer que @xmath1267 . on note
@xmath1268 ouvert de @xmath1238 satur pour le quotient par @xmath1237 .
mais @xmath1269 et cette dernire dcomposition est unique .
donc @xmath1270 est un systme de reprsentants de @xmath1271 et @xmath1237 agit librement sur @xmath1272 .
ainsi @xmath1273 est un ouvert affine de @xmath1031 .
dautre part @xmath1274 avec @xmath1275 pour @xmath1276 . en particulier la coordonne de cet lment sur @xmath1256 est non nulle ; il appartient donc @xmath1277 : on a @xmath1278 ; donc @xmath1279 . le complmentaire de @xmath1272 est runion de sous - schmas ferms obtenus en remplaant dans @xmath1238 lun des facteurs @xmath1023 par @xmath1021 , ce qui , dans limage par @xmath299 revient supprimer ce facteur .
alors le processus de multiplication dcrit dans ( * ? ? ?
* p45 ) montre que limage par @xmath1280 de ce sous - schma ferm est dans un @xmath1281 pour @xmath1247 .
ainsi @xmath1282 et finalement @xmath1283 .
de plus @xmath1284 .
donc @xmath1285 .
soit @xmath1286 un ouvert principal de @xmath1030 contenu dans @xmath1287 .
alors @xmath1288 , ouvert principal donc affine de @xmath1289 . on a donc @xmath1290=k[{\mathfrak{b}}(w)][f^{-1}]$ ] @xmath1291[(f\circ\nu)^{-1}]\subset k[{\mathfrak{e}}^c(\widetilde w)][(f\circ\nu)^{-1}]=k[{\mathfrak{e}}^c(\widetilde w)_{f\circ\nu}]=k[{\mathfrak{e}}(\widetilde w)_{f\circ\nu}]$ ] .
mais @xmath1031 est runion de @xmath1292 ouverts affines ( dmonstration par rcurrence selon lesquisse des pages 54 , 55 de @xcite ) ; donc @xmath1293\subset k[{\mathfrak{e}}(\widetilde w)][(f\circ\nu)^{-1}]$ ] .
ainsi @xmath1290=k[{\mathfrak{e}}(\widetilde w)_{f\circ\nu}]$ ] .
cela prouve que @xmath1259 induit un isomorphisme de @xmath1294 sur @xmath1286 et donc aussi de @xmath1285 sur @xmath1287 . on sait de plus
que @xmath299 induit un isomorphisme du sous - schma ferm @xmath1295 de @xmath1238 sur @xmath1289 .
donc @xmath1296 est surjectif . on note @xmath1298 ( resp .
@xmath1233 ) le sous - foncteur en groupe de @xmath1043 engendr par les sous - groupes @xmath283 pour @xmath891 ( resp .
@xmath1214 ) ; il sidentifie via @xmath97 ( au moins sur un corps ) avec le sous - groupe de mme nom de @xmath1079 . on a @xmath1299 . on rappelle
que @xmath1300 est le systme de gnrateurs de @xmath87 et que @xmath1205 est limage par @xmath97 du groupe de mme nom de @xmath1079 ( remarque [ 3.12].1 ) .
si @xmath8 est un corps , le sextuplet @xmath1301 est une @xmath1297paire raffine ( refined tits system ) . on a donc ( @xcite )
un systme de tits @xmath1302 , une dcomposition de bruhat @xmath1303 , une dcomposition de birkhoff @xmath1304 et @xmath1305 .
vrifions les axiomes de @xcite , voir aussi @xcite .
( rt1 ) : le groupe @xmath6 admet bien @xmath614 , @xmath1306 , @xmath1307 et @xmath1225 comme sous - groupes .
le groupe @xmath1225 normalise les @xmath1308 ( @xmath219 ) donc aussi @xmath1307 et @xmath1306 , il est distingu dans @xmath614 et @xmath1309 est engendr par les lments de @xmath1310 qui sont dordre @xmath1311 .
par construction @xmath6 est runion des @xmath1312 qui sont engendrs par des sous - groupes @xmath1313 ( daprs [ 3.15 ] ) . on a vu que @xmath1313 est engendr par @xmath1314 et @xmath1315 .
donc @xmath6 est engendr par @xmath614 et @xmath1306 .
( rt2 ) pour @xmath1316 , @xmath1317 est dans @xmath5 qui est muni dune donne radicielle jumele @xmath1318 ( * ? ? ?
* prop 8.4.1 ) et daprs le thorme 1.5.4 de on a @xmath1319 .
un rsultat classique dans le groupe @xmath1320 donne donc ( rt2a ) ; tandis que ( rt2b ) est clair .
enfin ( rt2c ) rsulte de [ 3.3 ] ( voir aussi [ 3.5 ] ) .
( rt3 ) soient @xmath1321 , @xmath1322 et @xmath1323 tels que @xmath1324 .
alors @xmath20 et @xmath1325sont dans @xmath5 ainsi donc que @xmath1326 . on dduit
de laxiome ( rt3 ) dans @xmath5 @xcite que @xmath1327 .
[ 3.17 ] si @xmath8 est un corps , @xmath1328 et @xmath1329 .
le sextuplet @xmath1330 est aussi une @xmath1297paire raffine ( [ 1.6].5 et @xcite ) . on a deux
systmes de tits @xmath1331 et @xmath1332 , [ 1.6].5 .
donc @xmath1333 et @xmath1334 .
comme @xmath1335 , on a @xmath1336 .
mais @xmath1337 et @xmath1338 donc @xmath1339 .
[ 3.17b ] si @xmath8 est un corps , linjection @xmath97 induit un isomorphisme simplicial @xmath1340quivariant de limmeuble @xmath1341 associ @xmath1079 sur @xmath8 sur limmeuble @xmath1342 associ au systme de tits @xmath1331 . plus
prcisment ces deux immeubles ont les mmes facettes et les appartements de @xmath1341 sont des appartements de @xmath1342 , mais , en gnral , @xmath1342 a plus dappartements
. les appartements sont isomorphes car associs au mme groupe @xmath614 .
daprs [ 3.17 ] @xmath1343 sinjecte dans @xmath1344 , do une injection @xmath1345 .
mais , par construction , si @xmath40 est simple , cest le mme groupe @xmath1308 qui est transitif sur les chambres ( de @xmath1341 ou @xmath1342 ) @xmath1346adjacentes la chambre fondamentale @xmath1347 ( associe @xmath1224 ) , ce sont donc les mmes chambres .
par rcurrence sur la longueur de galeries , on en dduit que @xmath1348 .
\1 ) on a dcrit ci - dessus le groupe la mathieu @xmath1349 associ un sgr @xmath75 libre , colibre , sans cotorsion et vrifiant de plus une certaine condition de dimension ( [ 3.5 ] ) . ll est clair que cette condition nintervient pas dans les constructions de mathieu , on peut donc sen affranchir . on a dailleurs vu en [ 1.3 ]
que si @xmath75 est libre , colibre , sans cotorsion il est extension semi - directe dun sgr @xmath165 vrifiant de plus cette condition de dimension . on en a dduit en [ 1.11 ]
que @xmath408 est produit semi - direct de @xmath1350 par un tore @xmath1351 .
de mme @xmath1349 est produit semi - direct de @xmath1352 par ce tore @xmath1351 .
ces dcompositions en produit semi - direct proviennent de la mme dcomposition @xmath1353 et sont donc compatibles avec les morphismes de foncteurs @xmath97 de [ 3.12 ] .
\2 ) soit @xmath75 un sgr quelconque . daprs
[ 1.3 ] il existe une extension centrale torique @xmath161 de @xmath75 et une extension semi - directe @xmath1354 de @xmath161 qui est libre , colibre et sans cotorsion .
daprs les corollaires [ 1.11 ] et [ 1.12 ] @xmath408 est quotient de @xmath1355 par un sous - tore central @xmath1356 de @xmath1357 et @xmath1358 est produit semi - direct de @xmath1355 par un sous - tore @xmath1359 de @xmath1360 .
inversement @xmath1355 est le noyau dun homomorphisme @xmath1361 de @xmath1358 sur @xmath1359 . lhomomorphisme @xmath1361 stend au groupe @xmath1362 ( construit ci - dessus ) : on prend @xmath1361 trivial sur @xmath1363 , on connat @xmath1361 sur @xmath1360 et les groupes @xmath1025 de [ 3.5 ] , donc @xmath1361 est dfini sur @xmath1364 et les paraboliques @xmath1023 ; son extension @xmath1362 dcoule alors de la construction de [ 3.6 ] . on dfinit
donc @xmath1365 comme le noyau de @xmath1361 .
enfin @xmath1349 est le quotient ( comme foncteur en groupe ) de @xmath1365 par le tore @xmath1356 . on a ainsi dfini
@xmath1349 et un morphisme de foncteurs @xmath1366 pour tout sgr @xmath75 .
si @xmath75 est libre , colibre et sans cotorsion on a ainsi deux dfinitions de @xmath1349 .
mais alors , daprs [ 1.3 ] , [ 1.11 ] et [ 1.12 ] , @xmath1355 est produit direct de @xmath408 par @xmath1356 et @xmath1358 produit semi - direct de @xmath1355 par @xmath1359 .
il est clair que @xmath1365 et @xmath1362 comme dfinis en 1 ) scrivent de la mme manire comme produit semi - direct .
do lidentification avec les groupes dfinis comme ci - dessus .
\3 ) * fonctorialit .
* la construction de mathieu est clairement fonctorielle en @xmath75 pour les sgr libres , colibres , sans cotorsion et les extensions commutatives de sgr . on a remarqu en [ 1.3].1 que les constructions @xmath1367 et @xmath1368 sont fonctorielles .
une extension commutative @xmath1369 dfinit donc un morphisme de foncteurs en groupes @xmath1370 .
ce morphisme est compatible avec les homomorphismes @xmath1361 sur les tores @xmath1359 correspondants , il induit donc un morphisme @xmath1371 .
ce dernier morphisme change les tores @xmath1356 correspondants , ce qui donne le morphisme @xmath1372 cherch .
le groupe @xmath1349 est donc fonctoriel en @xmath75 pour les extensions commutatives de sgr ( de manire compatible avec la fonctorialit des groupes @xmath261 et les morphismes @xmath1373 ) .
par construction , si @xmath1374 est un tel morphisme , @xmath1375 induit un isomorphisme de @xmath1376 sur @xmath1377 et le morphisme @xmath147 de @xmath146 dans @xmath154 .
si @xmath8 est un corps , @xmath1378 induit lisomorphisme @xmath1379 de @xmath385 sur @xmath1380 [ 1.10 ] et [ 3.12 ] .
dans ( * ? ? ?
* et 1.1.2 ) s. kumar considre un sgr vrifiant les mmes conditions que celles demandes par o. mathieu ( [ 3.5 ] ) .
on reste donc dans ce cadre et bien sr on se place , comme kumar , sur le corps @xmath3 .
pour @xmath1381 clos dans @xmath507 , les groupes @xmath1382 , @xmath1225 et @xmath614 dfinis dans 6.1 sidentifient clairement @xmath1383 , @xmath1225 et @xmath614 dfinis ci - dessus . de mme pour une partie de type fini @xmath1127 de @xmath12 , le groupe parabolique @xmath1384 de kumar sidentifie au groupe @xmath1385 de [ 3.10 ] , en particulier les deux dfinitions des paraboliques minimaux @xmath1313 concident . on a vu que
@xmath1302 est un systme de tits . daprs
[ 5.1.7 ] @xmath6 est le produit amalgam des sous - groupes @xmath614 et @xmath1313 .
mais cest exactement la dfinition de 6.1.16 du groupe de kac - moody la kumar .
donc sur @xmath3 les groupes la kumar et la mathieu concident .
linjection @xmath1386 de [ 3.12 ] permet didentifier @xmath1387 et le groupe minimal la kumar @xcite .
on considre dornavant un sgr fixe @xmath75 qui est libre et le groupe de kac - moody @xmath261 associ sur un corps @xmath2 . on a vu en [ 1.3 ] et [ 1.11 ] que si @xmath75 est un sgr non libre , @xmath408 est un sous - groupe dun groupe @xmath1388 avec @xmath1389 libre .
toute action de @xmath1390 sur une masure induira donc une action de @xmath1391 sur celle - ci .
on peut noter que lessentialis de la masure de @xmath1390 ne dpend pas du choix de @xmath1389 : les appartements sont des espaces affines sous hom@xmath1392 .
cest le choix fait par b. rmy pour ses ralisations coniques dimmeubles jumels . on abrgera frquemment @xmath1391 en @xmath1393 ou mme @xmath5 . on notera
avec la lettre romaine correspondante les points sur @xmath2 dun foncteur en groupe not avec une lettre gothique ; de mme pour les morphismes de foncteurs .
partir de [ 4.2 ] , le corps @xmath2 sera suppos muni dune valuation relle @xmath1394 .
le quintuplet @xmath1395 de [ 1.1 ] satisfait aux conditions de ( * ? ?
* et 1.2.1 ) , on est dans le cas kac - moody ( modrment imaginaire ) .
on peut en particulier dfinir dans @xmath1396 des facettes vectorielles et des cnes de tits positif ou ngatif @xmath1397 @xcite .
ces notions sont galement introduites dans ( * ? ? ?
* 1.3 ) o @xmath1398 est not @xmath1399 . on note
@xmath1400 la chambre vectorielle fondamentale , donc @xmath1401 .
une facette vectorielle est dite sphrique si son fixateur ( dans @xmath87 ) est fini si elle est contenue dans @xmath1402 ( intrieur du cne de tits ) .
daprs [ 1.6 ] le groupe de kac - moody @xmath5 contient un sous - groupe @xmath1225 et des sous - groupes radiciels @xmath1308 pour @xmath219 , chacun isomorphe au groupe additif @xmath1403 par un isomorphisme @xmath1404 . le triplet @xmath1318 est une donne radicielle de type @xmath117 au sens de @xcite .
le groupe @xmath614 dfini dans @xcite concide avec celui de [ 1.6].4 , il est muni dune application surjective @xmath1405 de noyau @xmath1225 .
on reprend , sauf exception signale , les notations de @xcite . on a
en particulier un immeuble @xmath1341 do nt les appartements sont isomorphes @xmath1406 , mais aussi un immeuble @xmath1407 do nt les appartements sont isomorphes @xmath1408 .
ces deux immeubles sont jumels @xcite ; comme dans @xcite on ne garde que les facettes sphriques dans @xmath407 . on a dj vu ces immeubles
de manire combinatoire en [ 1.6].5 , mais on les verra dsormais sous la forme de ces ralisations coniques . en [ 3.17b ] on a vu que @xmath6 ( resp @xmath7 ) agit sur @xmath1341 ( resp .
@xmath1407 ) .
\1 ) on suppose dornavant le corps @xmath2 muni dune valuation relle non triviale @xmath1394 , et ainsi @xmath1410 est un sous - groupe non trivial de @xmath1411 . on note
@xmath1412 lanneau des entiers de @xmath1413 et @xmath1414 son idal maximal . on note
@xmath1409 lespace vectoriel @xmath1396 considr comme espace affine ; on note cependant toujours @xmath4 llment neutre de @xmath1396 .
le groupe @xmath1415 agit sur @xmath1409 par translations : si @xmath1416 , @xmath1417 est llment de @xmath1396 tel que @xmath1418 , @xmath1419 .
cette action est @xmath1211quivariante .
daprs le raisonnement de @xcite on a : \2 ) lemme il existe une action affine @xmath1259 de @xmath614 sur @xmath1409 qui induit la prcdente sur @xmath1225 et telle que , pour @xmath1323 , @xmath1420 est une application affine dapplication linaire associe @xmath1421 .
\3 ) limage de @xmath614 est @xmath1422 .
son noyau est @xmath1423ker@xmath1424 .
\4 ) par construction le point @xmath4 est fix par les lments @xmath1425 et donc ( par conjugaison : ( kmt7 ) ) par tous les @xmath1426 , pour @xmath219 .
daprs ( kmt6 ) pour @xmath1427 , @xmath1428 et @xmath1429 .
donc @xmath1430 est la rflexion @xmath1431 par rapport lhyperplan @xmath1432 , cest dire lapplication affine dapplication linaire associe @xmath1433 et densemble de points fixes @xmath1434 .
\5 ) * dfinitions . *
l_appartement affine tmoin _ @xmath1409
est lespace affine @xmath1409 ( sous lespace vectoriel @xmath1396 muni de @xmath87 , @xmath117 et @xmath73 ) muni de lensemble @xmath1435 des _ murs ( rels ) _ @xmath1436 pour @xmath219 et @xmath1437 et de lensemble @xmath1438 des _ murs imaginaires _ @xmath1439
pour @xmath548 et @xmath1437 .
lespace @xmath1440 , dfini pour @xmath1441 et @xmath1442 est un _ demi - appartement _
si @xmath219 et @xmath1437 . on note @xmath1443 .
ainsi @xmath1409 satisfait aux conditions de @xcite ( une variante prs pour @xmath1438 , 1.6.2 ) ; il est semi - discret si et seulement si la valuation @xmath1394 est discrte .
on choisit @xmath1444 comme point spcial privilgi . on reprend
les notations de @xcite ( sauf pour @xmath1445 , @xmath44 et @xmath1446 , voir ci - dessous ) .
on notera les ressemblances et diffrences avec @xcite .
le cas @xmath1447 est trait dans ( * ? ? ?
* et 3.1 ) . on note @xmath1448 et @xmath1449 .
donc @xmath1450 .
pour un filtre @xmath1451 de parties de @xmath1409 , on considre comme dans @xcite ou @xcite plusieurs variantes denclos : si @xmath1452 , @xmath1453 est le filtre form des sous - ensembles de @xmath1409 contenant un lment de @xmath1451 de la forme @xmath1454 avec pour chaque @xmath40 , @xmath1455 ( en particulier chaque @xmath1456 contient @xmath1451 ) . on note
@xmath1457 , @xmath1458 .
on dfinit aussi ( selon charignon @xcite ) @xmath1459 comme le filtre form des sous - ensembles de @xmath1409 contenant un lment de @xmath1451 de la forme @xmath1460 o @xmath1461 et @xmath1462 . on a @xmath1463 ( enveloppe convexe ferme
de @xmath1451 ) ; @xmath1464 et @xmath1465 concident avec ceux dfinis dans @xcite .
un filtre @xmath1451 est dit _ clos _ si @xmath1466 et , quand on ne prcise pas , lenclos de @xmath1451 dsigne @xmath1464 .
une _ facette - locale _ de @xmath1409 est associe un point @xmath556 de @xmath1409 et une facette vectorielle @xmath1467 dans @xmath1396 ; cest le filtre @xmath1468 .
_ facette _
associe @xmath1469 est le filtre @xmath1470 form des ensembles contenant une intersection de demi - espaces @xmath1456 ou @xmath1471 ( un seul @xmath1455 pour chaque @xmath218 ) , cette intersection devant contenir @xmath1469 un voisinage de @xmath556 dans @xmath1472 .
la facette ferme @xmath1473 est ladhrence de @xmath1470 et aussi @xmath1474 .
les facettes @xmath1469 , @xmath1470 et @xmath1473 sont dites _ sphriques _
si @xmath1467 lest .
sil existe @xmath548 avec @xmath1475 , il se pourrait que @xmath1470 soit lgrement plus gros que la facette dfinie dans @xcite , 1.6.2 .
\6 ) * le groupe de weyl .
* pour @xmath219 , il est clair que le groupe de weyl ( affine ) @xmath1476 , engendr par les rflexions par rapports aux murs , contient les translations de vecteur dans @xmath1477 .
ainsi @xmath1478 ( not @xmath44 dans @xcite ) est un groupe de translations de @xmath1396 , contenu dans @xmath1476 et invariant par @xmath87 . on a @xmath1479 .
le groupe @xmath1480 est not @xmath1445 dans @xcite .
ainsi @xmath1481 est le plus grand sous - groupe de @xmath1482 stabilisant @xmath1435 ( et @xmath1438 ) .
comme @xmath614 est engendr par @xmath1225 et les @xmath1202 , il est clair que @xmath1483 . on a : @xmath1484 .
\7 ) * valuation des groupes radiciels . *
pour @xmath219 et @xmath1485 , on pose @xmath1486 si @xmath1487 avec @xmath1488 .
les @xmath1489 forment une valuation de la donne radicielle @xmath1490 @xcite .
voir aussi @xcite on considre le monode ordonn @xmath1491 de @xcite form dlments @xmath167 , @xmath1492 ( pour @xmath1493 ) et @xmath1494 vrifiant @xmath1495 si @xmath1496 . on note
@xmath1497 lensemble des lments de @xmath1491 qui sont borne infrieure dune partie minore de @xmath1498 . on a @xmath1499 si @xmath1498 est discret et @xmath1500 sinon .
les idaux fractionnaires de @xmath1412 sont indexs par @xmath1497 : @xmath1501 on associe @xmath1502 .
pour @xmath1501 et @xmath219 , @xmath1503 est un sous - groupe de @xmath1308 ; on a @xmath1504 .
soit @xmath1505 un tel filtre .
\1 ) pour @xmath1441 , soit @xmath1506 . pour @xmath218 ( resp .
@xmath117 , @xmath28 ) @xmath1507 ne dpend que de @xmath1508 ( resp .
@xmath1509 ou @xmath1510 , @xmath1511 ) .
si @xmath1505 est un ensemble , @xmath1512 , @xmath1513 , @xmath1514 .
cette fonction @xmath1515 est _ concave _ @xcite : @xmath1516 , @xmath1517 et @xmath1518 . on dit que @xmath1505 est _ presque - ouvert _
si @xmath1519 , @xmath1520 .
on dit que @xmath1505 est _ troit _ si aucun mur ne spare @xmath1505 en deux , autrement dit si @xmath1519 , @xmath1521 ou @xmath1522 ou ( dans le cas discret ) @xmath1523 et sil nexiste pas de @xmath1524 avec @xmath1525 et @xmath1526 .
\2 ) pour @xmath219 , on note @xmath1527 ; @xmath1528 est le groupe engendr par @xmath1529 et @xmath1530 et on note @xmath1531 .
on dfinit @xmath1532 comme le groupe engendr par les @xmath1529 ( pour @xmath219 ) et @xmath1533 ; on verra ( [ 4.12 ] ) que @xmath1534 ( resp .
@xmath1535)nest pas forcment gal au groupe @xmath1536 ( resp .
@xmath1537 ) engendr par les @xmath1529 pour @xmath891 ( resp .
@xmath1214 ) . le groupe @xmath1538 ( @xmath1539 ) est engendr par
tous les @xmath1540 pour @xmath219 .
tous ces groupes sont normaliss par @xmath1541 et on peut donc dfinir @xmath1542 et @xmath1543 .
ces groupes ne dpendent que de lenclos @xmath1508 de @xmath1505 et mme que de @xmath1544 .
si @xmath1545 , on a @xmath1546 , etc .
\3 ) lemme soient @xmath219 et @xmath1505 un filtre comme ci - dessus .
a)@xmath1547 .
\b ) si @xmath1520 , alors @xmath1548 .
si @xmath1549 , alors @xmath1550 .
\c ) @xmath1540 fixe @xmath1505 , @xmath1551 , il existe @xmath1552 fixe point par point par @xmath1420 .
* lemma 3.3 ) \4 ) le groupe @xmath1553 est engendr par les rflexions @xmath1554 pour @xmath219 et @xmath1437 tels que @xmath1555 .
il est contenu dans le fixateur @xmath1556 de @xmath1505 dans @xmath1476 .
il y a bien sr galit si @xmath1505 est rduit un point spcial , mais aussi si @xmath1505 est une facette sphrique et si @xmath1498 est discret . en
effet dans ce dernier cas limage de @xmath1556 dans @xmath87 fixe une facette vectorielle sphrique et on est ramen au cas classique o @xmath1476 est produit semi - direct dun groupe de weyl fini par un groupe discret de translations ; ce cas est trait dans @xcite et ne se gnralise pas si @xmath1557 est dense . daprs [ 4.12].4 ci - dessous il ne se gnralise pas non plus au cas non classique ( non sphrique ) .
\1 ) pour @xmath218 et @xmath1558 , on pose @xmath1559 et @xmath1560 .
de mme @xmath1561 .
par concavit de @xmath1515 , il est clair que @xmath1562 est une sous@xmath1563algbre de lie de @xmath533 .
on peut de mme considrer les compltions positive et ngative : @xmath1564 et @xmath1565 si @xmath1505 est un ensemble , sinon @xmath1566 et @xmath1567 .
ces algbres ne dpendent que de @xmath1508 .
par contre @xmath1509 ou @xmath1544 pourraient tre insuffisants pour les dterminer .
\2 ) lalgbre enveloppante entire @xmath433 est gradue par @xmath1568 . on peut
donc dfinir @xmath1569 avec @xmath1570 .
cest une sous@xmath1563bigbre gradue de @xmath1571 .
le terme de niveau @xmath220 de la filtration induite par celle de @xmath1572 est @xmath1573 .
on peut avoir @xmath1574 . on peut considrer la compltion positive @xmath1575 ( resp
ngative @xmath1576 ) de @xmath1577 pour le degr total ( resp .
son oppos ) ; cest une sous - algbre de @xmath1578 ( resp .
@xmath1579 ) .
\3 ) soit @xmath820 un poids dominant de @xmath146 ( @xmath1052 ) . on a construit en [ 2.14 ]
une forme entire @xmath875 du module intgrable @xmath869 de plus haut poids @xmath820 .
tout poids de @xmath1231 est de la forme @xmath1580 avec @xmath1581 .
pour @xmath1582 et @xmath1583 , on pose @xmath1584 puis @xmath1585 et @xmath1586 ; en particulier @xmath1587 .
par concavit de @xmath1515 , il est clair que @xmath1588 est un sous@xmath1589module gradu de @xmath1590 . on peut
avoir @xmath1591 , mais @xmath1588 ne dpend que de @xmath1592 o @xmath1593 est lensemble des poids de @xmath1059 .
on peut considrer la compltion ngative de @xmath1588 : @xmath1594 si @xmath1505 est un ensemble et @xmath1595 sinon .
il est clair que @xmath1596 est un sous@xmath1589module gradu de @xmath1597 .
bien sr on construit aussi @xmath1588 et sa compltion positive @xmath1598 pour un module intgrable @xmath1059 plus bas poids @xmath1599 .
\1 ) soit @xmath851 un ensemble clos de racines . on lui a associ une @xmath32bigbre @xmath836 ( [ 2.13].2 et [ 3.1 ] ) . on peut donc dfinir @xmath1600 ( avec @xmath1601 ) ; cest une sous@xmath1563bigbre de @xmath1577 et de @xmath1602 ( plus prcisment leur intersection ) .
\2 ) le plus simple pour associer @xmath1603 un sous - groupe de @xmath1604 semble tre de procder comme suit : si @xmath1505 est un ensemble on considre la compltion @xmath1605 ; cest une sous - algbre de @xmath1575 et de @xmath1606 et elle contient @xmath669\lambda x$ ] pour @xmath546 et @xmath1607 .
le monode @xmath1608 est lintersection de @xmath1609 ou @xmath1575 et de @xmath1604 ( qui est un sous - groupe des lments inversibles de @xmath1606 ) . on sait
que linverse dun lment de @xmath1604 est son image par la co - inversion @xmath465 et il est clair que @xmath1609 est stable par @xmath465 ( car @xmath465 est gradue et stabilise @xmath836 ) .
ainsi @xmath1608 est un sous - groupe de @xmath1604 .
daprs [ 3.2 ] il est clair que @xmath1608 est form des produits @xmath905\lambda_x.x$ ] pour @xmath1610 ( il suffit de raisonner par rcurrence sur le degr total ) .
si @xmath1505 est un filtre , le groupe @xmath1608 est la runion des @xmath1611 pour @xmath1612 , donc encore lintersection de @xmath1604 et de @xmath1613 .
\3 ) on sait que @xmath5 et @xmath1604 sinjectent dans @xmath1614 ( [ 3.3]a , [ 3.6 ] et [ 3.13 ] ) . on note donc @xmath1615 . on a aussi @xmath1616 puisque @xmath1617 ( [ 3.17 ] ) .
pour un filtre on a @xmath1618 .
bien sr @xmath1619 contient @xmath1620 ( et @xmath1621 ) .
\4 ) * remarques * : a ) le sous - groupe @xmath1622
de @xmath1415 stabilise les algbres et modules construits en [ 4.4 ] ou ci - dessus .
ainsi @xmath1541 normalise @xmath1623 et @xmath1624 .
\b ) comme @xmath1609 est une sous - algbre de @xmath1575 et de @xmath1606 , il est clair que les groupes @xmath1608 et @xmath1625 stabilisent @xmath1626 et @xmath1627 pour laction adjointe de @xmath1628 sur @xmath1578 , [ 3.7].3 .
\c ) pour @xmath219 et @xmath248 , @xmath1608 est le sous - groupe @xmath1529 de @xmath1629 isomorphe @xmath1630 par @xmath1631 ( dfini en [ 4.2].7 ) .
\d ) si @xmath1545 , on a @xmath1632 et @xmath1633 .
si @xmath1505 est la runion dune famille @xmath1634 de filtres , alors @xmath1635 est le sup ( pris dans @xmath1497 ) des @xmath1636 et @xmath1637 .
ainsi @xmath1638 et @xmath1639 .
\e ) pour @xmath1640 , on a @xmath1641 , @xmath1642 , @xmath1643 .
\f ) @xmath1608 et @xmath1625 ne dpendent
que de @xmath1508 ( cela pourrait tre faux pour @xmath1509 ou @xmath1544 ) .
\5 ) lemme soient @xmath922 des sous - ensembles clos de racines .
\a ) @xmath1644 ( resp .
@xmath1645 ) est un sous - groupe de @xmath1646 ( resp .
@xmath1647 ) .
\b ) si @xmath927 est galement clos , alors on a des dcompositions uniques @xmath1648 et @xmath1649 .
\c ) si @xmath109 est un idal de @xmath108 , alors @xmath1644 ( resp .
@xmath1645 ) est un sous - groupe distingu de @xmath1650 ( resp .
@xmath1647 ) et on a un produit semi - direct si , de plus , @xmath927 est clos .
\d ) si @xmath665 est une racine simple , on a @xmath1651 et @xmath1652 .
les groupes @xmath1653 et @xmath1654 sont normaliss par @xmath1655 .
le a ) est clair .
pour le b ) et le c ) on a des dcompositions analogues dans @xmath1628 ( [ 3.3 ] ) et , par exemple , @xmath1656 , do les rsultats par intersection ( puisque @xmath1657 est gradue ) .
la premire partie de d ) rsulte de c ) .
on sait donc que @xmath1658 et @xmath1659 sont normaliss par @xmath1541 et @xmath1529 .
le mme raisonnement avec @xmath507 remplac par @xmath1660 montre quils sont aussi normaliss par @xmath1530 , do la conclusion .
\6 ) on a vu en [ 3.5 ] que @xmath1661 est normalis par @xmath1662 .
cela signifie en particulier que @xmath1663 .
on peut donc ainsi dfinir dans @xmath1043 un groupe @xmath1664 pour tout @xmath1026 .
\7 ) dans toutes les notations prcdentes un @xmath1665 ( resp .
un @xmath1666 ) peut remplacer @xmath108 quand @xmath844 ( resp .
@xmath1667 ) . on considre aussi le groupe de kac - moody ngativement
maximal @xmath1213 construit comme @xmath1043 en changeant @xmath507 et @xmath508 . plus gnralement on peut changer @xmath9 en @xmath20 et @xmath1668 en @xmath1669 dans ce qui prcde pour obtenir des groupes similaires ( avec des proprits similaires ) dans le cas ngatif .
[ 4.6 ] soit @xmath1505 un filtre de parties de @xmath1409 . il y a trois sous - groupes de @xmath5 associs @xmath1505 et indpendants du choix de @xmath507 dans sa classe de @xmath1211conjugaison .
\1 ) le groupe @xmath1532 ( engendr par tous les @xmath1529 ) est gal @xmath1670 .
\2 ) le groupe @xmath1671 ( engendr par les groupes @xmath1672 et @xmath1532 ) est gal @xmath1673 .
\3 ) symtriquement le groupe @xmath1674 ( engendr par les groupes @xmath1675 et @xmath1532 ) est gal @xmath1676 .
\4 ) on a : @xmath1677 et symtriquement pour @xmath1674 . on refait la preuve de (
3.4 ) en utilisant [ 4.5].5 , [ 4.3].3 et la remarque [ 3.16 ] .
\a ) le groupe @xmath1541 normalise @xmath1532 , @xmath1678 , @xmath1679 , @xmath1680 , @xmath1672 , @xmath1675 , @xmath1671 et @xmath1674 .
le groupe @xmath1542 est contenu dans le fixateur @xmath1681 de @xmath1505 . on note
@xmath1682 , @xmath1683 et @xmath1684 .
ces trois groupes vrifient les mmes dcompositions que 1 ) , 2 ) et 3 ) ci - dessus en remplaant @xmath1538 par @xmath1685 . on verra en [ 5.11].2 que @xmath1686 quand
@xmath1505 est un point spcial ou quand @xmath1505 est une facette sphrique et la valuation est discrte .
\b ) lgalit @xmath1687 est quivalente @xmath1688 ; on verra en [ 5.7].3 quelle nest pas toujours satisfaite .
\c ) on montre galement au cours de la preuve de [ 4.6 ] que @xmath1689 ou @xmath1690 ne dpend pas du choix de @xmath507 dans sa classe de @xmath1211conjugaison .
\d ) dans le cas classique des groupes rductifs , on a @xmath1691 , @xmath1692 et @xmath1693 .
de plus @xmath1694 ( @xmath1695 ) est le mme que le groupe dfini en ( * ? ? ?
* et 6.4.9 ) .
le groupe @xmath1696 est not @xmath1697 par bruhat et tits .
[ 4.7 ] dcomposition diwasawa supposons @xmath1505 troit , alors @xmath1698 .
supposons de plus @xmath1505 presque - ouvert , alors lapplication naturelle de @xmath1699 sur @xmath1700 est bijective voir la preuve de ( * ? ? ?
3.6 ) \1 ) on a aussi @xmath1701 et , de la mme faon avec les groupes maximaux @xmath1702 et @xmath1703 .
\2 ) ainsi quand @xmath1505 est troit , tout sous - groupe @xmath636 de @xmath5 contenant @xmath1532 peut scrire @xmath1704 .
si de plus @xmath1705 avec @xmath1706 et @xmath1707 on a @xmath1708 et , si @xmath1709 , @xmath1710 [ 4.10 ] .
\1 ) supposons que @xmath1505 est un ensemble . daprs
[ 4.5].4b le groupe @xmath1711 stabilise @xmath1575 et @xmath1627 pour laction de @xmath1306 sur @xmath1578 .
le groupe @xmath1712 stabilise @xmath1713 et @xmath1714 pour laction adjointe de @xmath5 et @xmath1712 [ 2.1].4 .
les deux actions concident sur @xmath1715 .
donc @xmath1716 ( [ 4.5].3 ) stabilise @xmath1717 et @xmath1718 pour laction adjointe de @xmath5 .
de mme @xmath1675 stabilise @xmath1577 et @xmath1719 .
ainsi @xmath1577 et @xmath1719 sont stables par @xmath1529 pour tout @xmath219 et donc par @xmath1532 .
finalement on sait que @xmath1577 et @xmath1719 sont stables par les groupes @xmath1532 , @xmath1671 , @xmath1674 et @xmath1541 ( qui normalise les trois autres ) ( [ 4.5].4a et [ 4.6]a ) . ce rsultat est encore vrai si @xmath1505 est un filtre , car alors tous ces objets sont runion filtrante des objets correspondant @xmath1612 .
@xmath1065 un @xmath515module plus haut poids comme en [ 4.4].3 .
supposons dans un premier temps que @xmath1505 est un ensemble .
comme @xmath1588 est un @xmath1720module , le groupe @xmath1721 stabilise @xmath1588 pour laction de @xmath6 [ 3.7 ] .
ce module @xmath1588 est en particulier stable par @xmath1672 et les groupes @xmath1529 pour @xmath891 .
comme @xmath1722 est un @xmath1723module , le groupe @xmath1724 stabilise @xmath1722 , pour la reprsentation de @xmath1725 sur @xmath1726 ( [ 3.7 ] ) qui induit sur @xmath1307 la mme action que @xmath5 ( [ 3.14].3 ) .
comme @xmath5 stabilise @xmath1727 , le groupe @xmath1728 stabilise @xmath1729 .
ainsi le sous - module @xmath1588 de @xmath1727 est stable par laction de @xmath1672 et @xmath1675 .
ce rsultat est encore vrai quand @xmath1505 est un filtre .
le groupe @xmath1532 est engendr par les @xmath1529 pour @xmath219 qui sont dans @xmath1672 ou @xmath1675 .
ainsi @xmath1588 est stable par @xmath1730 , @xmath1731 , @xmath1671 et @xmath1674 .
on sait aussi quil est stable par @xmath1541 .
\3 ) ces rsultats sont encore vrais , mutatis mutandis , pour des modules plus bas poids .
[ 4.9 ] si @xmath1505 est un ensemble , on considre le sous - groupe @xmath1732 de @xmath5 form des lments stabilisant @xmath1577 et @xmath1588 pour tout module @xmath1059 plus haut ou plus bas poids ( comme en [ 4.4].3 ) . on note @xmath1733 .
daprs [ 4.8 ] le groupe @xmath1732 contient @xmath1730 , @xmath1731 , @xmath1671 , @xmath1674 et @xmath1541 .
daprs la remarque [ 4.7].2 il est utile de dterminer @xmath1734 , voir [ 4.11 ] . on note @xmath1735 la runion de @xmath73 et de tous les ensembles de poids des modules @xmath1059 ci - dessus .
le groupe @xmath1732 ne dpend que de @xmath1736 ( et non seulement de @xmath1508 ) .
soit @xmath1323 , on peut crire @xmath1737 avec @xmath1738 dans @xmath1739 ( le groupe engendr par les @xmath1740 qui fixe @xmath4 ) [ 4.2].4 , @xmath1741 et @xmath1416 .
\1 ) pour @xmath1742 un module plus haut ou plus bas poids ou @xmath1743 ou @xmath1744 , @xmath1582 et @xmath1745 , on a @xmath1746 .
considrons galement laction sur @xmath1409 : @xmath1747 .
donc @xmath1748 et ainsi @xmath1749 .
finalement @xmath1750 .
ainsi @xmath1751 ; en particulier le fixateur @xmath1681 de @xmath1505 dans @xmath614 ( pour laction @xmath1259 ) normalise @xmath1732 . on a vu
en [ 4.3].4 que @xmath1752 .
\2 ) on a aussi @xmath1753 , donc @xmath1754 , @xmath1755 et @xmath1681 normalise les groupes @xmath1532 , @xmath1697 , @xmath1538 et @xmath1685 .
\3 ) daprs 1 ) on a @xmath1756 pour @xmath851 .
pour @xmath1757 ( @xmath40 racine simple ) et @xmath1758 , on peut passer aux complts : @xmath1759 .
donc @xmath1760 ; plus prcisment pour @xmath1761 , @xmath561 et @xmath1762 , @xmath1763\lambda x.n^{-1}$ ] est une exponentielle tordue associe @xmath1764 et @xmath1765 et cela passe aux produits infinis .
daprs 2 ) et [ 4.5].6 on a @xmath1766 si @xmath1767 .
ce rsultat stend alors au cas gnral pour @xmath96 .
par intersection avec @xmath5 , on a @xmath1768 .
daprs [ 4.6 ] , on a @xmath1769 et @xmath1770 . en particulier @xmath1681 normalise @xmath1671 et @xmath1674 .
[ 4.11 ] si @xmath1505 est un ensemble , @xmath1771 et @xmath1772 .
de plus @xmath1773 est le stabilisateur ( dans @xmath614 pour laction @xmath1259 sur @xmath1409 ) du @xmath1774enclos @xmath1736 de @xmath1505 , il normalise @xmath1532 , @xmath1672 , @xmath1675,@xmath747 * n.b . * en particulier @xmath1775 . on refait la preuve de @xcite car il y
a quelques changements substantiels .
\a ) soient @xmath1323 et @xmath1776 tels que @xmath1777 et @xmath1778 .
pour @xmath1779 ou
@xmath1719 ou @xmath1577 , @xmath1780 et @xmath1781 , on dfinit @xmath1782 comme la restriction de @xmath1783 @xmath1784 suivie de la projection sur @xmath1785 ( paralllement aux autres espaces de poids ) .
pour tout @xmath1582 , @xmath1786 et @xmath1787 ( en un sens vident ) , donc @xmath1788 .
ainsi @xmath1789 .
il reste dterminer @xmath1773 et @xmath1790 ( ainsi que @xmath1791 ) .
\b ) on a vu en [ 4.8 ] que @xmath1792 .
inversement soit @xmath1793 stabilisant @xmath1577 et donc @xmath1575 .
on peut crire @xmath1794 avec les conventions suivantes : lordre des facteurs @xmath1795 est tel que la hauteur crot de droite gauche et , pour @xmath665 , @xmath1796t_{\alpha1}e_{\alpha1}.\cdots.[exp]t_{\alpha s}e_{\alpha s}$ ] avec @xmath1797 une base de @xmath477 et les @xmath1798 dans @xmath2 ( si @xmath891 on a @xmath1799 et @xmath669=exp$ ] ) .
nous allons montrer que @xmath1800 et pour cela que , @xmath1801 , @xmath1802 . par rcurrence on peut supposer que cest vrai pour @xmath1803 droite de @xmath1795 , alors ces @xmath1803 sont dans @xmath1711 et stabilisent @xmath1575 , on peut donc les supposer gaux @xmath220 .
donc @xmath1804 et ainsi @xmath1805 .
choisissons un lment @xmath1806 ( base de @xmath1807 ) tel que @xmath1808 ; quitte changer @xmath711 en @xmath718 et donc utiliser une autre base de @xmath1809 , on peut supposer @xmath1810 .
par ailleurs @xmath1811t_{\alpha i}e_{\alpha i})=\sum^\infty_{p=0}\,t_{\alpha i}^p ad(e_{\alpha i}^{[p]})$ ] et @xmath1812 a pour terme de poids @xmath40 lexpression @xmath1813 et doit tre dans @xmath1814 .
comme les @xmath1815 pour @xmath1816 et @xmath1817 font partie dune base de @xmath1809 , on doit avoir @xmath1818 pour @xmath1816 et @xmath1819 . pour @xmath1820 et @xmath1821 on obtient le rsultat cherch @xmath1822 \c ) reprenons les notations de [ 4.10 ] : @xmath1746 et @xmath1823 .
donc @xmath1824 si et seulement si , @xmath1825 , @xmath1826 cest dire @xmath1827 ; cest quivalent au fait que @xmath20 stabilise lensemble @xmath1736 .
le groupe @xmath1773 normalise @xmath1532 , @xmath1672 , @xmath1675,@xmath747car ces groupes ne dpendent que de @xmath1736 .
[ 4.12 ] 1 ) soit @xmath556 un point spcial de @xmath1409 et @xmath1828 , alors @xmath1829 est une partie borne de @xmath1830 , car tout lment de @xmath28 est combinaison linaire coefficients dans @xmath1831 dlments de @xmath1735 .
donc @xmath1832 ( car @xmath1833 [ 4.3 ] ) . daprs
[ 4.7 ] et [ 4.11 ] , on a @xmath1834 .
\2 ) si de plus @xmath1835 est lorigine de @xmath1409 , on a @xmath1836 et @xmath1837 pour tout module plus haut ou plus bas poids .
il est clair que ces modules sont stables par @xmath1838 ( [ 3.7 ] et [ 3.14 ] ) .
donc @xmath1839 . on a @xmath1840 par construction et aussi @xmath1841 .
comme @xmath1842 ( [ 3.2 ] et [ 4.5].2 ) , on a @xmath1843 . on se
demande si cette dernire intersection est gale @xmath1844 ( resp .
@xmath1845 ) ( voir [ 1.6].5 ) .
cela prouverait que @xmath1846 ( resp
. peut - tre @xmath1847 ) .
un rsultat analogue a t prouv dans @xcite quand le corps rsiduel de @xmath2 contient @xmath3 .
mais il concerne le groupe minimal @xmath1079 la kumar avec sa structure de ind - groupe ( sur @xmath3 mais tendue aux @xmath1848algbres ) .
il pourrait donc ne pas avoir de rapport avec le rsultat cherch ici , car [ 3.19 ] ne donne une identification que pour les points rationnels sur @xmath3 des groupes
maximaux ou minimaux .
le foncteur en groupes minimal @xmath1079 nest pas forcment le meilleur choix sur tous les anneaux , un foncteur @xmath274 , comme en [ 1.6].1 , peut tre plus adapt , voir 3 ) ci - dessous .
daprs [ 3.17 ] et [ 1.6].1 , [ 1.6].5 ( kmg7 ) , on a toujours @xmath1849 et @xmath1850 mais cela ne permet pas de rpondre aux questions .
\3 ) cas de @xmath808 : on considre le sgr @xmath1851 .
lalgbre de kac - moody correspondante est @xmath1852 $ ] , voir [ 2.12 ] .
si on pose @xmath1853 , on a @xmath1854 , @xmath1855 et les espaces propres sont , pour @xmath1856 , @xmath1857 , @xmath1858 et @xmath1859 .
comme @xmath75 est non libre on considre la masure essentielle et lappartement @xmath1409 correspondant du dbut de cette section : @xmath1860 et @xmath1861 forment donc une base de lespace vectoriel rel @xmath1862 .
\a ) la reprsentation naturelle
@xmath299 de @xmath493 sur @xmath1863}({\z[t , t^{-1}]}^2)$ ] ( [ 2.12 ] ) induit une identification de @xmath1864 et @xmath1865})$ ] qui envoie @xmath1866 sur le groupe des matrices de @xmath1867)$ ] qui sont triangulaires suprieures strictes modulo @xmath1868 .
le foncteur en groupe @xmath274 de [ 1.6].1 peut tre choisi tel que @xmath1869})$ ] pour tout anneau @xmath8 .
on sait que le groupe @xmath1712 est produit libre des sous - groupes @xmath1870)=\left(\begin{matrix}1&k[t ] \cr 0&1\cr \end{matrix}\right)=\langle u_{\alpha_1+n\delta}(k)\mid n\in\n\rangle$ ] et @xmath1871)=\left(\begin{matrix}1&0\cr tk[t]&0\cr \end{matrix}\right)=\langle u_{\alpha_0+n\delta}(k)\mid n\in\n\rangle$ ] , @xcite ; et on a : @xmath1872 o on note @xmath1873 un lment non nul de lidal maximal @xmath1874 de @xmath1412 ( une uniformisante dans le cas discret ) .
cet lment @xmath1875 est dans @xmath1876 ( daprs lexpression de gauche ) et dans @xmath1712 ( expression du milieu ou de droite ) donc dans @xmath1877 .
cependant lexpression de droite montre que @xmath1783 nest pas dans @xmath1878 , puisquil y a unicit des dcompositions dans le produit libre @xmath1712 . on a donc une
inclusion stricte : @xmath1879 .
\b ) la reprsentation naturelle @xmath299 permet didentifier @xmath1880 @xmath1881 ; voir @xcite pour un rsultat voisin .
dans cette identification @xmath1882 ( resp .
@xmath1883 ) devient le groupe des matrices de @xmath1884)$ ] ( resp .
@xmath1885)$ ] ) qui sont triangulaires suprieures strictes modulo @xmath1868 ( lingrdient essentiel de la dmonstration a t expliqu en [ 2.12 ] ) .
on sest demand en 2 ) si @xmath1886})$ ] est gal @xmath1887 ( cest impossible daprs a ) ) ou ventuellement @xmath1888 .
ce dernier groupe est form des matrices de @xmath1889)$ ] qui sont triangulaires suprieures strictes modulo @xmath1868 , il est donc bien gal @xmath1890 . on a donc : @xmath1891 . en particulier @xmath1892 .
mais en fait @xmath1893 et @xmath1894 , car @xmath1876 est le groupe engendr par les matrices lmentaires de @xmath1895})$ ] et on sait que celui - ci est gal @xmath1895})= { \mathfrak{g}}_\#(\sho)$ ] ( * ? ? ?
\c ) on considre @xmath1896 avec @xmath1897 dtermin par @xmath1898 et @xmath1899 .
alors @xmath1900 est topologiquement engendr par @xmath1901)$ ] et @xmath1902)$ ] dans @xmath1884)$ ] . on en dduit assez facilement que @xmath1900 ( resp .
@xmath1903 ) est contenu dans ( en fait gal ) lensemble des matrices @xmath1904)$ ] ( resp . @xmath1889)$ ] ) telles que @xmath1905 , @xmath1906 et @xmath1907 modulo @xmath1908
. de mme @xmath1909 est topologiquement engendr par @xmath1910)$ ] et @xmath1911)$ ] dans @xmath1912)$ ] ; donc @xmath1909 ( resp .
@xmath1913 ) est contenu dans ( en fait gal ) lensemble des matrices @xmath1914)$ ] ( resp . @xmath1889)$ ] ) telles que @xmath1915 modulo @xmath1916 , @xmath1917 modulo @xmath1918 et @xmath1907 modulo @xmath1919 .
ainsi @xmath1903 et @xmath1913 sont contenus dans le groupe @xmath1920 des matrices @xmath1921)$ ] telles que @xmath1915 et @xmath1907 modulo @xmath1919 .
dautre part ( si @xmath1922 ) @xmath1923 est form des matrices diagonales avec @xmath1924 pour @xmath1925 et @xmath1856 . on voit facilement que , si @xmath1926 , la
matrice @xmath1927 est dans @xmath1928 mais pas dans @xmath1929 .
donc @xmath1930 ne satisfait aucune des conditions ( gf@xmath1668 ) de [ 5.3 ] ci - dessous ( si @xmath1926 ) .
\4 ) on considre toujours @xmath808 avec son appartement @xmath1409 despace vectoriel associ @xmath1396 et @xmath1861 , @xmath1931 base du dual @xmath1862 . on
suppose @xmath1932 .
les murs ont alors pour quations @xmath1933 avec @xmath1934 et les coracines vrifient @xmath1935 .
le groupe de weyl ( affine ) associ est @xmath1936 .
le groupe de weyl vectoriel @xmath87 contient la transvection @xmath1937 donne par @xmath1938 .
si lon considre la translation @xmath1939 de vecteur @xmath1940 , llment @xmath1941 de @xmath1476 fixe tous les points de lhyperplan affine dquation @xmath1942 .
cependant le point @xmath693 tel que @xmath1943 et @xmath1944 nest dans aucun mur .
ainsi @xmath1945 alors que @xmath1946 est infini .
ce phnomne disparat si on considre un espace vectoriel @xmath1396 o @xmath1947 et @xmath1940 sont indpendants .
[ 4.13 ] soit @xmath1505 un sous - ensemble de @xmath1409 . on dfinit
le groupe @xmath1948 comme lintersection des groupes @xmath1949 o @xmath1098 est une partie non vide de @xmath1950 , @xmath1951 une extension value et @xmath1949 est le groupe construit de manire analogue @xmath1952 dans @xmath1953 .
( on sait que @xmath1954 sinjecte fonctoriellement dans @xmath1953 [ 1.6 ] . ) si @xmath1505 est un filtre de parties de @xmath1409 , on dfinit @xmath1955 .
si @xmath1505 est un point ou une facette de @xmath1409 , on dira que @xmath1948 est le `` fixateur '' de @xmath1505 voir ci - dessous [ 5.1 ] et [ 5.7].2 .
proprits \1 ) on a une fonctorialit vidente en le corps des diffrents groupes ou algbres dfinis jusquen [ 4.6 ] ( sauf @xmath1732 et @xmath1948 ) . en particulier le sous - groupe @xmath1948
de @xmath1956 contient les groupes @xmath1957 , @xmath1958 , @xmath1697 et @xmath1681 .
daprs [ 4.10 ] , pour @xmath1323 on a @xmath1959 . bien sr si @xmath1545 , on a @xmath1960 .
\2 ) soient @xmath1505 un sous - ensemble de @xmath1409 et @xmath556 un point de @xmath1950 .
il existe une extension value @xmath1951 telle que @xmath556 soit un point spcial de @xmath1961 ( si @xmath1962 ) . daprs
[ 4.12].1 on a @xmath1963 , donc @xmath1964 .
cette relation est encore valable si @xmath1505 est un filtre .
par contre , mme pour un ensemble , on ne sait pas si linclusion @xmath1965 est toujours une galit .
\3 ) si @xmath1505 est un ensemble , @xmath1966 , daprs 1 ) , 2 ) ci - dessus , [ 4.11 ] et lunicit dans les dcompositions de [ 3.16 ] .
donc @xmath1967 , @xmath1968 et ceci est encore valable pour un filtre .
\4 ) si @xmath1505 est troit on a , daprs [ 4.7 ] , @xmath1969 .
ainsi @xmath1970 .
\5 ) on ne va utiliser @xmath1948 que si @xmath1505 est un point ou une facette et , dans ce cas , on ne dfinira que plus tard ( [ 5.11 ] ) le sous - groupe parahorique @xmath1697 ( avec @xmath1971 ) .
les proprits des groupes @xmath1972 sont rsumes dans la proposition suivante .
il nest pas exclus que , pour tout @xmath556 dans @xmath1409 , @xmath1973 @xmath1974 et @xmath1975 [ 4.12].3b . en [ 5.7].3 on verra que ceci
ne peut se gnraliser tout ensemble @xmath1930 la place de @xmath556 .
[ 4.14 ] les groupes @xmath1972 associs aux points de @xmath1409 satisfont aux proprits suivantes : ( p1 ) @xmath1976 ( le fixateur de @xmath556 dans @xmath614 ) .
( p2 ) @xmath1977 .
( p3 ) @xmath1978 avec @xmath1979 et @xmath1980 . on a en fait ( p3 )
@xmath1981 la gnralisation des rsultats de @xcite ( en particulier le lemme [ 4.11 ] ) la caractristique rsiduelle positive a ncessit , sans surprise , le remplacement de lalgbre de lie @xmath1719 par lalgbre `` enveloppante '' @xmath1577 .
les raisonnements de ( * ? ? ?
3.7 ) sont compliqus , entachs dune ou deux erreurs ( do nt je suis responsable ) et utilisent la symtrisabilit .
ctait sans doute trop ambitieux de dfinir ce stade les groupes @xmath1697 au lieu des groupes @xmath1948 . on a pu simplifier et gnraliser en restreignant lambition et en utilisant des extensions de corps ( comme suggr par les travaux de cyril charignon @xcite ou @xcite ) .
dans cette section on ne va utiliser que les proprits ( p1 ) ( p3 ) des groupes @xmath1972 indiques dans la proposition [ 4.14 ] .
la mention qui sera faite des groupes @xmath1948 plus gnraux nest pas indispensable au raisonnement .
[ 5.1 ] la _ masure _ @xmath1982 de @xmath1079 sur @xmath2 est le quotient de lensemble @xmath1983 par la relation : @xmath1984 il existe @xmath1323 tel que @xmath1985 et @xmath1986 .
il est clair que @xmath1987 est une relation dquivalence @xcite .
lapplication @xmath1988 , @xmath1989 est injective daprs ( p1 ) .
elle identifie @xmath1409 son image @xmath1990 , l_appartement _ de @xmath1225 dans @xmath0 .
laction gauche de @xmath5 sur @xmath1983 induit une action de @xmath5 sur @xmath0 .
les appartements de @xmath0 sont les @xmath1991 pour @xmath1875 . laction de @xmath614 sur @xmath1992 se fait via @xmath1259 ; en particulier @xmath1541 fixe ( point par point ) @xmath1992 .
par construction le fixateur de @xmath1993 est @xmath1994 et , pour @xmath1875 on a @xmath1995 .
comme @xmath1972 contient @xmath1996 , il est clair que , @xmath1519 , @xmath1997 , @xmath1998 fixe @xmath1999 .
donc , pour @xmath1442 , le groupe @xmath2000 fixe @xmath2001 .
pour @xmath1505 une partie de @xmath1409 , on note @xmath2003 et pour @xmath1505 un filtre , @xmath2004 .
le groupe @xmath2002 est le fixateur de @xmath1505 ( pour laction de @xmath5 sur @xmath0 qui contient @xmath1409 ) . par construction de @xmath1948 , on a @xmath2005 ( donc @xmath2006 ) et @xmath2007 .
daprs [ 4.5].4d et ( p3 ) on a @xmath2008 et @xmath2009 .
le sous - ensemble @xmath2010 de @xmath5 consistant en les @xmath1875 tels que @xmath2011 est @xmath2012 .
@xmath2013 .
[ 5.3 ] on considre les conditions suivantes : ( gf+ ) @xmath2014 .
( gf@xmath1666 ) @xmath2015 .
( tf ) @xmath2016 . on dit que @xmath1505 a un _ fixateur transitif _
sil satisfait ( tf ) . on dit que @xmath1505 a un _ assez bon fixateur _
sil satisfait ( tf ) et ( gf+ ) ou ( gf@xmath1666 ) . on dit que @xmath1505 a un _ bon fixateur _
sil satisfait ( tf ) , ( gf+ ) et ( gf@xmath1666 ) .
daprs la remarque [ 4.6]c , cette dfinition ne dpend pas du choix de @xmath507 dans sa classe de @xmath1211conjugaison .
la condition ( gf+ ) ou ( gf@xmath1666 ) implique que @xmath2017 ( [ 4.13].1 ) ; le groupe @xmath614 permute les filtres satisfaisant ( gf+ ) , ( gf@xmath1666 ) ou ( tf ) et les fixateurs correspondants ( [ 4.13].1 ) .
daprs ( p3 ) et [ 5.1 ] un point a toujours un bon fixateur .
mais il existe des ensembles ne vrifiant ni ( gf+ ) ni ( gf@xmath1666 ) , [ 4.12].3c .
[ 5.4 ] soit @xmath1505 un filtre de parties de @xmath1409 .
si @xmath1505 a un fixateur transitif , alors @xmath2002 est transitif sur les appartements contenant @xmath1505 .
cest classique et dans ( * ? ? ?
consquence alors @xmath2002 et tous ses sous - groupes normaux ne dpendent pas du choix de lappartement contenant @xmath1505 .
[ 5.5 ] \1 ) supposons @xmath2018 .
si @xmath1505 dans @xmath2019 a un bon ( ou assez bon ) fixateur , alors cest galement vrai pour @xmath1098 et @xmath2020 , @xmath2021 . en particulier
tout appartement contenant @xmath1505 contient aussi son enclos @xmath1508
. inversement si @xmath2022 , le plus petit espace affine contenant @xmath1505 , ( ou si @xmath2023 , donc @xmath2024 ) , @xmath1505 a un assez bon fixateur et @xmath1098 a un bon fixateur , alors @xmath1505 a un bon fixateur .
\2 ) si un filtre @xmath1505 dans @xmath2019 est engendr par une famille @xmath2025 de filtres ( @xmath2026 ) avec des bons ( ou assez bons ) fixateurs , alors @xmath1505 a un bon ( ou assez bon ) fixateur @xmath2027 .
\3 ) supposons que le filtre @xmath1505 dans @xmath2019 est la runion dune suite croissante @xmath2028 ( @xmath2029 ) de filtres avec de bons ( ou assez bons ) fixateurs et que , pour un certain @xmath97 , lespace @xmath2030 a un fixateur fini @xmath2031 dans @xmath2032 , alors @xmath1505 a un bon ( ou assez bon ) fixateur @xmath2033 .
4)soient @xmath1505 et @xmath1098 deux filtres dans @xmath2019 .
supposons que @xmath1098 satisfait ( gf+ ) ( resp .
( gf+ ) et ( tf ) ) et quil existe un nombre fini de chambres vectorielles fermes positives @xmath2034 , @xmath2035 telles que : @xmath2036 .
alors @xmath2037 satisfait ( gf+ ) ( resp .
( gf+ ) et ( tf ) ) et @xmath2038 .
voir ( * ? ? ?
4.3 ) [ 5.6 ] dans 4 ) ci - dessus , les mmes rsultats sont vrais si on change @xmath1665 en @xmath1666 . si @xmath1098 a un bon fixateur , @xmath2039 et @xmath2040 , alors @xmath2037 a un bon fixateur .
si @xmath1505 satisfait ( gf@xmath1666 ) , @xmath1098 satisfait ( gf+ ) , @xmath1505 ou @xmath1098 satisfait ( tf ) , @xmath2036 et @xmath2041 , alors @xmath2037 a un bon fixateur . pour les preuves manquantes ci - dessous , voir ( * ? ? ?
\1 ) si @xmath2042 ou @xmath2043 dans @xmath1409 si @xmath2044 ( cne de tits ) , alors @xmath2045 , @xmath2046 $ ] et @xmath2047 ont de bons fixateurs ( [ 5.6 ] et ( p3 ) ) .
si @xmath2048 , @xmath229x , y]=[x , y]\setminus\{x\}$ ] a un bon fixateur et le germe de segment @xmath2049)$ ] ou le germe dintervalle @xmath2050x , y)=germ_x(]x , y])$ ] est dit _ prordonn _ et a un bon fixateur ( [ 5.5].2 ) .
si @xmath2051 ou @xmath2052 dans @xmath1409 si @xmath2053 ( intrieur du cne de tits ) , la demi - droite @xmath796 dorigine @xmath556 et contenant @xmath693 est dite _ gnrique _ et a un bon fixateur . en effet @xmath796 est runion croissante des segments
@xmath2054 $ ] pour @xmath435 et , si @xmath468 , ce segment a un fixateur fini dans @xmath1476 ( car @xmath2053 ) ; on peut donc appliquer [ 5.5].3 .
de mme la droite contenant @xmath556 et @xmath693 a aussi un bon fixateur .
\2 ) une facette locale @xmath1469 , une facette @xmath1470 ou une facette ferme @xmath1473 a un bon fixateur .
ici et dans la suite @xmath1467 dsigne une facette vectorielle et @xmath2055 une chambre vectorielle de @xmath1396 .
\3 ) un quartier @xmath2056 a un bon fixateur .
on notera que le fixateur du quartier @xmath2057 est @xmath2058 , alors que @xmath2059 , car @xmath2060 et @xmath2061 . on a vu
en [ 4.12].3 que @xmath2062 peut tre plus petit que @xmath2063 . on peut donc avoir @xmath2064 .
\4 ) un germe de quartier @xmath2065 a un bon fixateur ( [ 5.5].2 ) .
le fixateur de @xmath2066 est @xmath2067 car tout lment de @xmath2068 est un produit fini dlments de groupes @xmath1308 pour @xmath2069 .
par contre @xmath1306 nest pas la runion des @xmath2070 pour @xmath2071 , on na sans doute pas daction de @xmath6 sur @xmath0 .
\5 ) lappartement @xmath1409 lui - mme a un bon fixateur @xmath2072 .
par dfinition le stabilisateur de @xmath1409 est donc @xmath2073 .
comme le corps @xmath2 est infini , il rsulte alors de [ 1.6].4 que les appartements de @xmath0 sont en bijection avec les sous - tores dploys maximaux de @xmath1079 .
\6 ) un mur @xmath1439 a un bon fixateur : soit @xmath2074 et @xmath2075 dans une cloison vectorielle de ker@xmath40 , alors @xmath1439 est runion croissante des @xmath2076 do nt le support @xmath1439 a un fixateur fini ( @xmath2077 ) dans @xmath2078 , on conclut grce 1 ) ci - dessus et [ 5.5].3 .
ce rsultat stend au cas du support dune facette sphrique .
le ( bon ) fixateur de @xmath1439 est @xmath2079 .
\7 ) un demi - appartement @xmath2001 est lenclos de deux quartiers do nt des cloisons sont dans @xmath1439 et opposes , daprs 3 ) ci - dessus et [ 5.5].1 , [ 5.5].4 , il a un bon fixateur , qui est gal @xmath2080 .
\8 ) si @xmath2081 si @xmath2082 , alors @xmath2083 a un bon fixateur pour toutes facettes vectorielles @xmath2084 et @xmath2085 de signes quelconques : daprs 2 ) ci - dessus on peut appliquer la remarque [ 5.6 ] aux facettes locales @xmath2086 et @xmath2087 , on conclut grce [ 5.5].1 .
\9 ) lenclos @xmath2088 a un bon fixateur ds que les facettes vectorielles @xmath2084 et @xmath2085 sont de signe opposs ou si lune delles est sphrique . en effet le premier cas est clair par 2 ) ci - dessus et la remarque [ 5.6 ] .
si la facette vectorielle @xmath2084 est sphrique , elle est contenue dans le cne de tits ouvert @xmath2089 ( ou @xmath2090 ) , il existe donc une chambre @xmath2055 et un @xmath2091 tels que , @xmath2092 , @xmath2093 contient @xmath2094 pour un voisinage @xmath2095 de @xmath4 dans @xmath2085 .
alors @xmath2096 et @xmath2097 .
daprs 2 ) ci - dessus et la remarque [ 5.6 ] @xmath2098 a un bon fixateur et on conclut par [ 5.5].1 .
soient @xmath2099 un appartement de @xmath0 et @xmath2100 , la relation @xmath2101 ( resp .
@xmath2102 ) dans @xmath1409 ne dpend pas du choix de @xmath1783 daprs [ 5.7].5 , car @xmath614 stabilise le cne de tits et son intrieur ; on note cette relation @xmath2103 ( resp .
@xmath2104 ) .
soient @xmath2105 , @xmath2106 deux appartements de @xmath0 et @xmath2107 tels que @xmath2108 .
alors @xmath2109 ( car @xmath2045 a un bon fixateur ) . on dfinit donc ainsi une relation @xmath2110 ( et aussi @xmath2111 ) sur @xmath2112 ; on verra en [ 5.14 ] que cest un prordre . il rsulte de [ 5.5].1 que @xmath2113 contient @xmath2047 ( calcul dans @xmath2105
ou @xmath2106 ) et en particulier le segment @xmath2046 $ ] ( qui est le mme dans @xmath2105 et @xmath2106 ) ; on dit quune intersection dappartements est _ convexe pour le prordre _ @xmath2110 .
contrairement aux conventions depuis 4.0 , on va modifier le sgr .
soit @xmath1374 un morphisme de sgr libres qui est une extension centrale finie .
on abrgera @xmath261 en @xmath1079 , @xmath409 en @xmath2114 , @xmath402 en @xmath130 , etc .
alors @xmath19 est un sous - module de @xmath136 de mme rang , donc @xmath2115 et , comme @xmath131 , on a @xmath2116 , @xmath2117 . le morphisme @xmath2118 est dcrit en [ 1.10 ] et [ 1.13 ] , en particulier @xmath2119 , @xmath2120 et le noyau de @xmath130 est dans @xmath1225 et mme dans @xmath1541 puisquil est fini . par construction ( [ 4.2 ] )
les appartements affines @xmath1409 et @xmath2121 sont identiques , les actions @xmath1259 et @xmath2122 de @xmath614 et @xmath2123 sont compatibles et on a les mmes murs puisque pour @xmath219 , @xmath2124 et @xmath2125 . les images @xmath1422 et @xmath2126 sont en gnral diffrentes ( [ 4.2].6 ) , mais on a @xmath2127 et @xmath2128 . pour @xmath2129 , les algbres de lie @xmath2130 et @xmath2131 ont les mmes composantes de poids non nul , [ 2.2 ] ( ce nest pas vrai en poids nul : si @xmath2 est de caractristique positive lapplication de @xmath2132 dans @xmath2133 peut mme ne pas tre injective ) .
ainsi lisomorphisme @xmath130 de @xmath1306 sur @xmath2134 ( [ 3.18].3 ) induit un isomorphisme de @xmath2070 sur @xmath2135 et aussi de @xmath2136 sur @xmath2137 , on identifie ces groupes . on a les rsultats
symtriques pour @xmath2138 et @xmath2139 , etc .
comme les actions de @xmath614 et @xmath2123 sur @xmath1409 sont compatibles et ker@xmath2140 , on a @xmath2141 . pour @xmath2142 , on a donc @xmath2143 et @xmath2144 car ker@xmath2145
. les relations ci - dessus permettent facilement de montrer que lapplication @xmath2146 passe au quotient en une bijection de la masure @xmath0 sur la masure @xmath2147 .
cette bijection est compatible avec @xmath130 et les actions de @xmath5 et @xmath2148 ; elle change les facettes .
comme @xmath2149 et @xmath2150 les appartements de @xmath0 et @xmath2147 sont les mmes .
on identifie ces deux masures .
[ 5.9 ] soient @xmath1374 une extension centrale finie de sgr libres et @xmath1505 un filtre de parties de @xmath1409 .
alors @xmath1505 a un bon ( ou assez bon ) fixateur @xmath2002 pour @xmath2151 si et seulement si il en a un @xmath2152 pour @xmath2153 .
dans ce cas @xmath2154 .
comme les actions de @xmath5 et @xmath2148 sur @xmath0 sont compatibles , on a @xmath2155 .
supposons que @xmath2152 est un ( assez ) bon fixateur ( pour le signe @xmath1665 ) . on a @xmath2156 , donc @xmath2157 ,
do ( gf+ ) pour @xmath1505 et @xmath5 .
soit @xmath1612 , @xmath2158 .
donc ( tf ) est satisfait par le filtre @xmath1505 et le groupe @xmath5 .
supposons que @xmath2002 est un ( assez ) bon fixateur ( pour le signe @xmath1665 ) .
soit @xmath2159 , quitte multiplier gauche @xmath2160 par un lment de @xmath2161 on peut supposer @xmath2162 .
donc il existe @xmath2163 tel que @xmath2164 .
do ( tf ) pour @xmath1505 et @xmath2148 .
pour @xmath2165 , ce calcul montre que @xmath2166 .
mais @xmath1675 , @xmath1672 sont dans @xmath2152 , donc @xmath2167 .
do ( gf+ ) pour @xmath1505 et @xmath2148 , plus la dernire assertion de lnonc .
\1 ) daprs [ 1.3 ] on a une suite dextensions commutatives de sgr @xmath2168 qui sont successivement centrale torique , semi - directe et centrale finie . on note @xmath2169 , @xmath2170 , @xmath2171 , @xmath5 les groupes correspondants et @xmath2172 le compos des morphismes @xmath2173 .
daprs [ 5.8 ] , @xmath2171 et @xmath5 ont la mme masure @xmath0 .
par contre @xmath160 et @xmath2174 ne sont en gnral pas libres , on considre les actions de @xmath2169 et @xmath2170 sur @xmath0 via leurs morphismes dans @xmath5 .
\2 ) on a @xmath2175 ( [ 1.8].3 ) ; on en dduit aussitt que le groupe simplement connexe @xmath2169 permute transitivement les appartements de @xmath0 . le stabilisateur de @xmath1409 dans @xmath2169 est le groupe @xmath2176 ( [ 1.10 ] ) et , daprs la construction de [ 4.2 ] , @xmath2177 .
ainsi tout appartement de @xmath0 est muni dune unique structure dappartement de type @xmath1409 telle que @xmath2169 induise des isomorphismes dappartements ( au sens de @xcite ) .
\3 ) dans ( * ? ? ?
* et 6.2 ) on se place dans le cas @xmath2178 ( libre ) et le groupe @xmath2179 qui y est dfini est gal @xmath2180 .
\1 ) on dit quune paire @xmath2181 forme de deux filtres de parties de @xmath0 est _ conviviale _ sil existe un appartement contenant ces deux filtres et si deux appartements @xmath2182 contenant @xmath2183 sont isomorphes par un isomorphisme fixant lenclos de @xmath2184 et @xmath2185 ( calcul dans @xmath17 ou @xmath2186 ) . on ne considrera
ici que des paires @xmath2187_conviviales _ , telles que les isomorphismes dappartements soient induits par des lments de @xmath5 .
une paire @xmath2181 est donc @xmath2187conviviale ds que @xmath2188 est contenu dans un appartement et y a un assez bon fixateur ( car @xmath2189 et @xmath2190 induisent des isomorphismes dappartements ) .
\2 ) lemme soient
@xmath2183 deux filtres de parties de @xmath1409 .
supposons @xmath2191 , alors pour tous @xmath2192 , @xmath2193 et @xmath2194 sont contenus dans un mme appartement . si @xmath2191
ou si @xmath2184 ou @xmath2185 a un fixateur transitif , alors @xmath2184 et @xmath2185 sont conjugus par @xmath5 si et seulement si ils le sont par @xmath614 .
les consquences de @xmath2191 sont classiques , ( * ? ? ?
* et 3.7 ) .
la dernire assertion rsulte de [ 5.4 ] et [ 5.7].5 .
\3 ) il rsulte donc de la dcomposition diwasawa [ 4.7 ] et de [ 5.7].4 quun filtre troit
dans un appartement
et un germe de quartier sont toujours dans un mme appartement . daprs
[ 5.7 ] et [ 5.5].4 une facette ( ou un germe de segment , un germe dintervalle , une facette locale , une facette ferme ) et un germe de quartier forment une paire @xmath2187conviviale .
\4 ) par convexit ordonne un appartement contenant un point @xmath556 et un germe de quartier @xmath2195 ,
contient le quartier @xmath2056 et son enclos donc son adhrence @xmath2196 . on en dduit quune facette @xmath1470 et un
filtre troit @xmath1451 contenant @xmath556 sont toujours dans un mme appartement .
ainsi , daprs [ 5.7].9 , deux facettes @xmath2197 et @xmath2198 en le mme point forment une paire @xmath2187conviviale ds que @xmath2084 et @xmath2085 sont de signes opposs ou si lune delles est sphrique .
\5 ) soient
@xmath2199 une chambre de @xmath1409 et @xmath1439 lun de ses murs ( avec @xmath2200 ) .
alors @xmath2201 agit transitivement sur les chambres @xmath2202 adjacentes @xmath2203 le long de @xmath1439 ; en particulier nimporte laquelle de ces chambres @xmath2204 est dans un mme appartement que le demi - appartement @xmath2001 , @xcite .
de plus @xmath2204 et @xmath2001 forment une paire conviviale : si @xmath2205 , @xmath2206 a un bon fixateur @xmath2207 ( [ 5.7].7 ) .
\6 ) on a vu en @xcite que deux points de la masure @xmath0 ne sont pas toujours contenus dans un mme appartement .
cest pour cette raison que lon a abandonn le nom dimmeuble pour @xmath0 ; une dfinition abstraite des masures affines nest pas possible dans des termes approchant ceux de @xcite , do la dfinition de @xcite inspire de @xcite .
on reproduit ci - dessous cette dfinition dans une formulation utilisant la notion de paire conviviale .
une paire de points ou une paire de facettes de @xmath0 nest pas toujours conviviale .
de plus dans lexemple [ 4.12].3c de @xmath808 , on a trouv deux points de @xmath1409 do nt le fixateur nest pas assez bon .
par contre il est peut - tre possible que le fixateur de deux points de @xmath1409 soit toujours transitif ( donc quune paire de points dun mme appartement soit toujours conviviale ) .
[ 5.7d ] une _ masure affine _ de type @xmath1409 est un ensemble @xmath2208 muni dun recouvrement par un ensemble @xmath2209 de sous - ensembles appels _ appartements _ tel que : ( ma1 ) tout @xmath2210 est muni dune structure dappartement de type @xmath1409 ( au sens de @xcite est `` isomorphe '' @xmath1409 ) .
( ma2 ) si @xmath1451 est un point , un germe dintervalle prordonn , une demi - droite gnrique ou une chemine solide dun appartement @xmath17 , alors @xmath2211 est une paire conviviale .
( ma3 + 4 ) si @xmath2212 est un germe de chemine vase , si @xmath1451 est une facette ou un germe de chemine solide , alors @xmath2213 est une paire conviviale .
( la notion de chemine , utilise ci - dessus , ne sera dfinie quen [ 5.12].1 . )
la masure affine @xmath2208 est dite _ ordonne _ si elle vrifie laxiome supplmentaire suivant : ( mao ) soient @xmath2214 deux points de @xmath2208 et @xmath2182 deux appartements les contenant ; si @xmath2215 ( dans @xmath17 ) , alors les segments @xmath2046_a$ ] et @xmath2046_{a'}$ ] dfinis par @xmath556 et @xmath693 dans @xmath17 et @xmath2186 sont gaux .
la masure affine @xmath2208 est dite _ paisse _ si toute cloison de @xmath2208 est dans ladhrence dau moins trois chambres .
\1 ) si @xmath2129 a un assez bon fixateur , on a @xmath2216 , puisque @xmath1672 , @xmath1675 , ... ne dpendent pas du sgr extension commutative prs . en particulier le fixateur @xmath2217 de @xmath1505 dans @xmath2169 agit transitivement sur les appartements de @xmath0 contenant @xmath1505 . on note @xmath2218 .
ainsi @xmath2219 .
le groupe @xmath2220 a pour image via @xmath1259 le fixateur @xmath1556 de @xmath1505 dans @xmath1476 . on a aussi @xmath2221 .
il est clair que les facettes @xmath2222 et @xmath2223 ont le mme fixateur dans @xmath1476 , donc @xmath2224 .
\2 ) quand @xmath2225 , on note @xmath2226 ( qui est alors aussi gal @xmath1957 et @xmath1958 ) .
cela se produit si @xmath1505 est rduit un point spcial ou si @xmath1505 est une facette sphrique et si la valuation est discrte ( daprs [ 4.3].4 ) ; cest a priori rare en dehors de ces deux cas daprs @xcite et [ 4.12].4 .
pour gnraliser la dfinition de @xcite , on dfinit un _ sous - groupe parahorique _ comme le groupe @xmath2227 associ une facette sphrique @xmath1451 ou @xmath2228 .
on parle de _ sous - groupe diwahori _
si @xmath1451 ou @xmath2228 est une chambre .
\3 ) le _ type global affine _
dun filtre de @xmath0 qui a un assez bon fixateur ( en particulier une facette ) est son orbite sous @xmath2169 .
dans un appartement @xmath17 de @xmath0 deux filtres ( assez bien fixs ) ont mme type si et seulement si ils sont conjugus par le groupe de weyl @xmath2229 de @xmath17 ( le 1 ) , [ 5.7c].2 et [ 5.10].2 ) .
pour les facettes ( ou les filtres troits assez bien fixs ) la relation `` avoir le mme type global affine '' est la relation engendre par transitivit partir des relations prcdentes dans les appartements .
en effet , tant donnes deux facettes @xmath2183 il existe toujours un quartier @xmath2230 et des appartements @xmath2231 contenant @xmath2230 et respectivement @xmath2183 ( [ 5.7c].3 ) ; de plus @xmath2230 contient forcment un @xmath2232conjugu de @xmath2233 .
le type global affine dune facette concide donc avec le type habituel dans le cas dun immeuble affine discret .
mais en valuation dense et si on suppose @xmath1409 essentiel , le type global affine dune facette locale @xmath1469 est la donne de lorbite @xmath2234 de @xmath556 et de lorbite de @xmath1467 sous @xmath2235 ( cest dire du type ( vectoriel ) de @xmath1467 si @xmath556 est spcial ) .
\1 ) une _ chemine _ dans @xmath1409 est associe une facette @xmath2236 ( sa base ) et une facette vectorielle @xmath1467 ( sa direction ) , cest le filtre @xmath2237 .
la chemine @xmath2238 est dite _ vase _ si @xmath1467 est sphrique , son signe est alors celui de @xmath1467 .
cette chemine est dite _ solide _ ( resp . _ pleine _ ) si la direction de tout sous - espace affine la contenant a un fixateur fini dans @xmath87 ( resp .
est @xmath1396 ) .
une chemine vase est solide .
lenclos dun quartier est une chemine pleine .
si @xmath2239 la chemine @xmath2240 est lenclos de la face de quartier @xmath1472 ; cette face est dite sphrique si @xmath1467 est sphrique .. un raccourci de la chemine @xmath2238 est dfini par un lment @xmath2241 , cest la chemine @xmath2242 .
le _ germe _ de la chemine @xmath2238 est le filtre @xmath2243 form des parties de @xmath1409 contenant un de ses raccourcis .
si @xmath1467 est la facette vectorielle minimale , on a @xmath2244 .
\2 ) la facette vectorielle @xmath1467 est associ un sous - groupe parabolique @xmath2245 de @xmath5 avec une dcomposition de levi @xmath2246 .
le groupe @xmath2247 est engendr par @xmath1225 et les @xmath1308 pour @xmath219 et @xmath2248 , @xcite ou @xcite .
si @xmath2249 pour @xmath1120 ( [ 4.1 ] ) , le groupe @xmath2247 est clairement un quotient du groupe de kac - moody minimal @xmath2250 ( notations de [ 3.10 ] ) .
daprs la remarque [ 3.10 ] et [ 3.13 ] on a en fait galit : @xmath2251 .
ce groupe @xmath2252 induit sur @xmath1409 une structure dappartement @xmath2253 do nt les murs sont les @xmath1439 pour @xmath2254 et @xmath1437 .
lenclos correspondant @xmath2255 de la facette @xmath1451 est une facette ferme de @xmath2253 , on note @xmath2256 son fixateur dans @xmath2251 .
cette construction se ralise aussi si @xmath1467 est une autre facette vectorielle ou si @xmath1451 est rduit un point @xmath2257 ( ou @xmath2222 ) .
le sous - groupe `` parahorique '' de @xmath5 associ @xmath2258 est le produit semi - direct @xmath2259 .
il ne dpend en fait que du germe @xmath2260 , on le note donc aussi @xmath2261 .
les rsultats suivants se dmontrent comme dans @xcite .
\3 ) proposition le groupe @xmath2261 fixe ( point par point ) le germe de chemine @xmath2262 .
\4 ) proposition soient @xmath2263 et @xmath2264 deux germes de chemines de @xmath1409 avec @xmath2263 vas , alors @xmath2265 .
\5 ) proposition une chemine solide
et son germe ont de bons fixateurs .
il en est de mme pour une face de quartier sphrique @xmath1472 et son germe ( linfini ) ; le fixateur ( point par point ) de @xmath2266 est @xmath2267 .
\6 ) proposition soient @xmath2263 un germe de chemine vase et @xmath2264 un germe de chemine solide ou une facette dans @xmath1409 , alors @xmath2268 a un assez bon fixateur ( et mme un bon fixateur si @xmath2264 est vas ) .
\7 ) remarque on obtient donc de nouvelles paires @xmath2187conviviales dans @xmath0 : un germe de chemine vase et un germe de chemine solide ou un germe de chemine vase et une facette .
cela permet en particulier de dfinir des
rtractions de @xmath0 sur un appartement @xmath17 avec pour centre un germe de chemine vase pleine de @xmath17 ( par exemple un germe de quartier ) @xcite .
[ 5.13 ] la masure affine @xmath0 construite en [ 5.1 ] est une masure affine ordonne paisse au sens de @xcite ou [ 5.7d ] .
elle est semi - discrte si la valuation @xmath1394 de @xmath2 est discrte .
laxiome ( ma1 ) rsulte de [ 5.10].2 et ( ma2 ) rsulte de ce que lon a vu en [ 5.7 ] ou [ 5.12].5 : les filtres impliqus dans ( ma2 ) ont de bons fixateurs .
enfin ( ma3 + 4 ) rsulte de [ 5.12].7 .
les dernires assertions se montrent comme dans @xcite . on notera
cependant que lpaisseur dune cloison est le cardinal du corps rsiduel de @xmath2 et donc peut tre finie dans notre cas , plus gnral que celui de @xcite .
[ 5.14 ] 1 ) la masure @xmath0 a donc toutes les proprits dmontres dans @xcite , un certain nombre dentre elles ont dj t prouves ci - dessus . on obtient cependant au moins
deux proprit intressantes supplmentaires : les relations @xmath2110 et @xmath2269 de [ 5.7b ] sont des prordres ( elles sont transitives ) et les rsidus en chaque point de @xmath0 ont une structure dimmeubles jumels .
\2 ) on notera les proprits de laction du groupe @xmath5 : il est transitif sur les appartements et tous les isomorphismes entre appartements do nt lexistence est exige par les axiomes de @xcite ou [ 5.7d ] sont induits par des lments de @xmath5 . on peut dire que laction
de @xmath5 est _ fortement transitive_. dans le cas classique dune action de groupe sur un immeuble affine discret , cette notion de forte transitivit entrane clairement la notion classique ( et lui est sans doute quivalente ) .
[ 5.15 ] les immeubles jumels linfini associs la masure @xmath0 en @xcite sidentifient ( avec leurs appartements et leur action de @xmath5 ) aux immeubles jumels de @xmath5 dfinis par j. tits ( [ 1.6].5 ) .
limmeuble microaffine positif ( resp
. ngatif )
linfini associ la masure @xmath0 en @xcite sidentifie ( avec ses appartements et son action de @xmath5 ) limmeuble microaffine de @xcite ( dans sa ralisation de satake ) associ @xmath5 et la valuation @xmath1394 de @xmath2 ( resp .
lanalogue obtenu en changeant le cne de tits en son oppos ) .
cyril charignon construit directement dans @xcite un objet immobilier contenant @xmath0 , les immeubles microaffines ci - dessus et dautres masures affines associes aux sous - groupes paraboliques non sphriques de @xmath5 .
cest la gnralisation au cas kac - moody de la compactification de satake ( ou compactification polydrique ) des immeubles de bruhat - tits .
cette dernire a t construite abstraitement ( sans laide dun groupe ) dans @xcite . on identifie facilement lappartement canonique ( et son action de @xmath614 ) dun immeuble de @xcite avec celui qui doit lui correspondre , ( * ?
* et 4.4 ) .
il reste donc identifier les fixateurs de points de ces appartements . le fixateur de la facette vectorielle ( sphrique ) @xmath1467 dans limmeuble de [ 1.6].5 est le
sous - groupe parabolique @xmath2245 engendr par @xmath1225 et les @xmath1308 pour @xmath2270 .
il est clair quun lment de chacun de ces groupes transforme une face de quartier de @xmath1409 de direction @xmath1467 en une face parallle ( dans @xmath0 ) .
ainsi @xmath2245 fixe la classe de paralllisme des faces de quartier correspondant @xmath1467 .
un point de lappartement @xmath2271 de limmeuble de @xcite est de la forme @xmath2272 avec @xmath1467 une facette vectorielle sphrique positive de @xmath1396 , @xmath2273 lespace vectoriel engendr et @xmath2274 [ , 4.2 ] . on lui associe le germe ( linfini )
de la face de quartier @xmath2275 pour @xmath2276 .
le fixateur de @xmath2277 est engendr par @xmath2278 , le groupe @xmath2279 et le sous - groupe de @xmath1225 induisant dans @xmath1409 des translations de vecteur dans
@xmath2273 [ , 4.2.3 ] .
les deux premiers groupes engendrent le fixateur point par point du germe de la face de quartier @xmath2275 ( [ 5.12].5 ) et le troisime groupe stabilise videmment ce germe .
ces inclusions entre les fixateurs de points permettent de dfinir des applications surjectives et @xmath2187quivariantes des immeubles de [ 1.6].5 ou @xcite vers les immeubles
correspondants de @xcite .
mais un point @xmath556 et son transform @xmath2280 sont toujours dans un mme appartement et les applications ci - dessus sont injectives sur les appartements .
les inclusions entre fixateurs sont donc des galits et on a bien les identifications dimmeubles annonces .
le groupe @xmath6 nagit pas sur la masure affine @xmath0 .
par contre , daprs [ 3.17b ] il agit sur limmeuble vectoriel @xmath1341 .
le groupe @xmath5 agit sur limmeuble microaffine positif @xmath2281 de @xcite ( dans sa ralisation de satake ) qui est runion disjointe dimmeubles indexs par les facettes sphriques positives .
plus prcisment , une telle facette @xmath1467 on associe limmeuble de bruhat - tits ( non tendu ) @xmath2282 du groupe rductif @xmath2247 sur @xmath2 .
le groupe @xmath2278 agit trivialement sur @xmath2282 et @xmath5 permute les @xmath2282 selon son action sur @xmath1341 : un germe de chemine ou de face de quartier @xmath2283 correspond une facette ou un point @xmath2284 de @xmath2282 .
le fixateur ( point par point ) de cette facette ou ce point @xmath2284 pour laction de @xmath5 sur @xmath2281 est @xmath2285 avec les notations de [ 5.12].2 et @xmath2286 le centre de @xmath2247 .
ce fixateur de @xmath2284 est donc plus grand que @xmath2287 qui est le fixateur ( point par point ) de @xmath2262 ( [ 5.12].3 ) et ( par dfinition ) le _ fixateur strict
_ de @xmath2284 .
le fixateur de la facette sphrique positive @xmath1467 pour laction de @xmath6 sur @xmath1341 est @xmath2288 ( [ 3.10 ] et [ 3.17b ] ) .
on peut donc prolonger laction de @xmath5 sur @xmath2281 en une action de @xmath6 : @xmath2289 agit sur @xmath2282 via @xmath2247 @xmath2290 agit trivialement .
le fixateur dans @xmath6 de @xmath2284 comme ci - dessus est @xmath2291 ; il contient le groupe @xmath2292 @xmath2293 que lon appelle encore fixateur strict de @xmath2284 pour laction de @xmath6 ( mme si ce nest pas justifi par une action sur @xmath0 qui contient le filtre @xmath2262 ) .
pour deux germes de chemines vases positives @xmath2263 et @xmath2264 , la dmonstration dans @xcite de la proposition [ 5.12].4 se gnralise : @xmath2294 .
dans @xcite on peut modifier le quotient en envoyant @xmath2295 sur @xmath2296 par la seconde projection . on obtient une ralisation gomtrique @xmath2297 , de satake au sens fort , de limmeuble microaffine , sur laquelle @xmath5 et @xmath6 agissent .
alors le fixateur strict @xmath2298 ( resp .
@xmath2292 ) est le fixateur de @xmath2299 pour laction de @xmath5 ( resp .
@xmath6 ) .
pour la construction des masures nous avons plong le groupe de kac - moody @xmath261 dans le groupe de kac - moody maximal la mathieu @xmath262 .
plaons nous sur un corps quelconque @xmath8 et notons @xmath2300 , @xmath2301 , etc
. alors @xmath6 apparat comme le complt de @xmath5 pour une certaine filtration .
dautres groupes maximaux ont t dfinis par rmy et ronan @xcite ou carbone et garland @xcite laide dautres filtrations , on va essayer de les comparer en tirant parti des rsultats des sections [ s2 ] et [ s3 ] .
par ailleurs moody , dans un preprint non publi @xcite , a dmontr la simplicit dun groupe analogue @xmath6 en caractristique @xmath4 .
l encore on va utiliser les sections [ s2 ] et [ s3 ] pour montrer dans certains cas la simplicit dun sous - quotient de @xmath6 ( thorme [ 6.19 ] ) .
\1 ) ce groupe @xmath2302 est ladhrence de limage de @xmath5 dans le groupe des automorphismes de son immeuble positif @xmath1341 @xcite .
il contient donc le quotient de @xmath5 par le noyau @xmath2303 de laction de @xmath5 sur cet immeuble .
ce groupe @xmath2304 est en fait le centre de @xmath5 : @xmath2305 ( [ 1.6].5 et ( * ? ? ?
* lemma 1b1 ) ) .
\2 ) pour @xmath2307 , soit @xmath2308 ( resp .
@xmath2309 ) la runion des chambres fermes
de limmeuble @xmath1341 ( resp . de son appartement standard @xmath1399 ) qui sont distance numrique @xmath169 de la chambre fondamentale @xmath2310
. les fixateurs @xmath2311 et @xmath2312 de @xmath2308 et @xmath2309 dans @xmath2313 forment deux filtrations de @xmath1712 qui sont exhaustives ( @xmath2314 ) et spares ( le systme de tits @xmath1218 est satur le fixateur de @xmath1399 dans @xmath5 est @xmath1225 @xcite et @xmath2315 : [ 1.6].5 ) . de plus @xmath2311 est le plus grand sous - groupe distingu de @xmath1712 contenu dans @xmath2312 .
\3 ) on dfinit de la mme manire des fixateurs @xmath2319 et @xmath2320 , pour laction de @xmath1306 sur @xmath1341 ( [ 3.17b ] )
. les filtrations correspondantes de @xmath1306 sont plus diffrentes : lintersection des @xmath2320 contient les groupes @xmath2321 pour @xmath2322 ( le systme de tits @xmath1331 est rarement satur ) et celle des @xmath2319 est souvent triviale ( [ 6.5 ] ) .
\4 ) le groupe
@xmath2323 est le complt de @xmath1712 pour sa filtration par les @xmath2324 .
plus gnralement le groupe @xmath2306 est le complt de @xmath5 pour sa filtration ( non exhaustive ) par les @xmath2324 ; il est muni de la topologie correspondante . en particulier @xmath2323 est ouvert dans @xmath2306 .
le groupe @xmath2302 est le spar - complt de @xmath5 pour la filtration par les fixateurs @xmath2325 .
cest un quotient de @xmath2306 par un sous - groupe @xmath2326 contenant @xmath2327 ( 1 ) ) ; il contient @xmath2323 .
si le corps @xmath8 est fini , @xmath2328 ( * ? ?
* prop.1 ) .
soit @xmath2329 un poids dominant rgulier ( @xmath2330 , @xmath125 ) .
on considre la reprsentation @xmath2331 de @xmath5 ou @xmath6 ( de plus haut poids @xmath2332 ) dans @xmath2333 ( [ 3.7].1 ) .
le groupe @xmath2334 dfini par carbone et garland @xcite est le spar - complt de @xmath5 pour la filtration dfinie par les fixateurs de parties finies de @xmath2335 , par les fixateurs de sous - espaces vectoriels de dimension finie de @xmath2335 .
pour @xmath435 soit @xmath2336 la somme des sous - espaces de @xmath2335 de poids @xmath2337 avec @xmath2338 et @xmath2339 .
ainsi @xmath2334 est le spar - complt de @xmath5 pour la filtration dfinie par les fixateurs @xmath2340 de @xmath2336 .
cette filtration nest pas spare : @xmath2341 ker@xmath2331 .
si @xmath2342 est un vecteur de plus haut poids et si @xmath2343 ker@xmath2331 , il fixe @xmath2342 et @xmath2344 @xmath125 , donc est de la forme @xmath2345 avec @xmath1416 , @xmath1776 et @xmath2346 .
mais @xmath2347 et @xmath2348 , donc @xmath2349 @xmath125 : @xmath2350ker@xmath2351 ( [ 1.6].5 ) .
inversement @xmath2352ker@xmath2351 agit trivialement sur @xmath2335 , donc ker@xmath2353ker@xmath2354 ker@xmath2355 .
on verra plus loin que @xmath2356 ker@xmath2357 ( [ 6.3].4 ) .
si lon veut une filtration spare et donc plonger @xmath5 dans son complt , on modifie la construction en faisant agir @xmath5 sur la somme directe @xmath2358 o @xmath2359 engendrent @xmath28 ; on note alors @xmath2336 la somme des @xmath2336 correspondant aux diffrents facteurs .
le complt de carbone - garland modifi @xmath2360 ainsi obtenu est donc dfini par une filtration contenue dans @xmath1712 , celle des @xmath2361 .
pour comparer @xmath2360 et @xmath2306 on va comparer cette filtration sur @xmath1712 avec celle des @xmath2311 . on note
@xmath2362 le spar complt de @xmath1712 pour la filtration par les @xmath2363 .
\1 ) daprs [ 3.2 ] , [ 3.3 ] et [ 3.4 ] , @xmath1306 est complet pour la filtration par les sous - groupes distingus @xmath2364 o @xmath2365 . on note
@xmath2366 ladhrence de @xmath1712 dans @xmath1306 pour la topologie associe ; cest aussi le complt de @xmath1712 pour la filtration par les @xmath2367 .
\2 ) pour @xmath204 et @xmath2368 , on a @xmath2369 et @xmath2370 , o @xmath1059 est le maximum des valeurs absolues de coefficients non diagonaux de la matrice de kac - moody @xmath17 .
ainsi @xmath2371 et @xmath1153 est un automorphisme du groupe topologique @xmath2372 .
\3 ) ds que @xmath2373 , on a @xmath2374 pour toute suite @xmath2375 . on en dduit que ,
pour @xmath2376 , @xmath2377 est dans le conjugu de @xmath1306 par @xmath2378 et donc @xmath2377 fixe @xmath2379 pour laction de @xmath6 sur @xmath1341 . on a @xmath2380 et mme @xmath2381 , car @xmath2377 est distingu dans @xmath1306 . en particulier @xmath2382 .
\4 ) vue la forme de laction de @xmath2383 sur @xmath2335 , on a @xmath2384 .
par ailleurs , pour @xmath2385 et @xmath2375 , le fixateur de @xmath2386 dans @xmath1306 est dans @xmath2387 , cest dire dans le fixateur ( dans @xmath1306 ) de @xmath2388 .
comme en 2 ) ci - dessus on montre que le poids de @xmath2389 est @xmath2337 avec @xmath2390 si @xmath2391 .
donc , si @xmath2392 avec @xmath2393 , @xmath2394 fixe toutes les chambres de la forme @xmath2388 , cest dire les chambres de @xmath2395 .
si @xmath8 est un corps fini , @xmath1306 et donc @xmath2366 sont compacts .
comme @xmath2362 et @xmath2323 sont spars , les homomorphismes @xmath2400 , @xmath2402 et @xmath2403 sont surjectifs , ferms et ouverts .
ce sont donc des isomorphismes de groupes topologiques si et seulement si ils sont injectifs .
\6 ) * remarques * a ) la filtration @xmath2404 de @xmath1306 permet de dfinir une mtrique invariante gauche sur @xmath6 , pour laquelle @xmath6 est complet et do nt les boules ouvertes sont les @xmath2405 .
daprs la dcomposition de bruhat , le fait que @xmath2406 et la relation de 2 ) impliquant les @xmath1153 , cette mtrique est quivalente celle , invariante droite , do nt les boules ouvertes sont les @xmath2407 .
ainsi @xmath6 est un groupe topologique , dans lequel @xmath1306 est ouvert .
ladhrence @xmath2408 de @xmath5 dans @xmath6 est le spar - complt de @xmath5 pour la filtration induite .
\b ) comme @xmath2408 , @xmath2360 et @xmath2306 sont des spars - complts pour les filtrations dans @xmath1712 de 4 ) ci - dessus , les homomorphismes de 5 ) ci - dessus se prolongent en @xmath2400 : .
par dfinition ker@xmath2409 o @xmath2410 est le noyau de laction de @xmath6 sur limmeuble @xmath1341 .
\c ) le groupe @xmath2327 senvoie trivialement dans @xmath2302 .
daprs les calculs de [ 6.2 ] on a donc un homomorphisme continu @xmath2411 .
celui - ci est surjectif , ferm et ouvert si le corps est fini ( rsultat de u. baumgartner et b. rmy ( * ? ? ?
2.6 ) ) .
comme @xmath2327 centralise @xmath1306 il est contenu dans @xmath2412 .
mais @xmath2417 est gal son normalisateur ( car il fait partie dun systme de tits ) on a donc @xmath2418 et mme @xmath2419 .
soit @xmath2420 .
on crit @xmath2421 avec @xmath1416 et @xmath1793 .
soit @xmath204 , on peut alors crire @xmath2422 avec @xmath2423 et @xmath2424 .
soient @xmath2425 et @xmath2426 , alors @xmath2427 avec @xmath2428 .
un calcul rapide dans @xmath1011 montre alors que @xmath2429 @xmath2430 si et seulement si @xmath2431 et @xmath2432 ; donc @xmath2433 et @xmath2434 .
on dit que lalgbre de lie @xmath2436 est simple au sens du thorme de gabber - kac ( ou _ gk - simple _ ) si tout sous@xmath2437module gradu non trivial de @xmath1230 contenu dans @xmath2438 est rduit @xmath2439 .
de mme on dit que le groupe @xmath6 est _ gk - simple _ si tout sous - groupe distingu de @xmath6 contenu dans @xmath1306 est rduit @xmath2440 , si @xmath2441 ou si @xmath2442 . on espre
que @xmath6 est toujours gk - simple ( voir [ 6.9].1 ou [ 6.8 ] ) .
lassertion correspondante pour @xmath5 est toujours satisfaite ( [ 6.1].1 ) . en caractristique @xmath4 , @xmath1230 est gk - simple dans le cas symtrisable ( et mme conjecturalement dans tous les cas ) .
mais , comme me la indiqu o. mathieu , cela nimplique pas la gk - simplicit en caractristique @xmath9 ; voir ci - dessous lexemple des lacets de @xmath2443 ( [ 6.8 ] ) .
on voit facilement que @xmath1230 est gk - simple si et seulement si @xmath2444 tout @xmath2445 tel que , @xmath2446 , @xmath2447 @xmath2448 est forcment nul .
lensemble de ces @xmath2445 est dtermin par des quations linaires coefficients entiers indpendantes de @xmath8 .
si @xmath2449 est gk - simple , un dterminant dtermin par ces quations est non nul .
ainsi la condition ci - dessus est vrifie pour @xmath2450 si la caractristique @xmath9 de @xmath8 ne divise pas ce dterminant .
malheureusement @xmath2451 est vide ou infini , on pourrait donc avoir @xmath1230 non gk - simple quelque soit la caractristique .
par contre dans le cas affine on na quun nombre fini de dterminants considrer car il y a priodicit des espaces radiciels imaginaires . en effet ,
pour le type affine @xmath2452 , @xmath2453 est ralis comme sous - algbre de lie de @xmath2454 $ ] o @xmath2455 est lalgbre simple de type @xmath2456 ; daprs @xcite ( resp . les calculs
explicites de @xcite ) @xmath2457 ( resp .
@xmath2458 ) est stable par multiplication par @xmath2459 . on calcule dans lalgbre @xmath2466 qui est gradue par les poids dans @xmath447 et le degr total dans @xmath495 ( [ 2.13].2 ) et aussi dans @xmath2467 .
soit @xmath2468 ; sa composante @xmath2469 de degr @xmath20 est dans lalgbre de lie @xmath1230 . soit @xmath2470 de degr @xmath20 tel
que la composante @xmath2471 de @xmath2469 sur @xmath2450 soit non nulle .
comme @xmath1230 est gk - simple , il existe @xmath2461 et @xmath2472 tels que @xmath2473 ( sinon @xmath2474 est de plus bas degr @xmath20 ) .
si @xmath1113 , @xmath2482 et ce groupe est normalis par @xmath2483 ( contenu dans @xmath2484 ) [ 3.3 ] .
soient @xmath2425 et @xmath2485 , alors @xmath2486 vaut @xmath2487 ( voir la dmonstration de [ 2.5 ] ) ; sa composante de poids @xmath2488 est donc gale la somme finie @xmath2489 , o @xmath2490 est la composante de poids @xmath2491 de @xmath2394 dans @xmath2492 .
cette composante de @xmath2493 sexprime comme un polynme en @xmath2332 do nt le terme de degr @xmath2494 est non nul .
comme @xmath8 est infini , il existe un @xmath2425 tel que ce polynme soit non nul et donc @xmath2495 , puisque sa composante de poids @xmath2488 est non nulle . [ 6.7 ] supposons
@xmath1230 gk - simple et @xmath8 infini .
alors @xmath2400 , @xmath2402 et @xmath2403 sont des homomorphismes .
ainsi @xmath2408 , @xmath2360 et @xmath2306 sont des groupes topologiques isomorphes ( et de mme pour @xmath2366 , @xmath2362 et @xmath2323 ) .
soit @xmath2460 avec @xmath758 .
daprs la proposition [ 6.6 ] il existe @xmath686 , @xmath2497 et @xmath2498 tels que @xmath2499 .
mais @xmath2417 est le fixateur de la chambre fondamentale @xmath2310 , donc @xmath2394 ne fixe pas la chambre @xmath2500 qui est distance @xmath2501 de @xmath2310 ( on a une galerie @xmath2502 ) . on a donc montr que @xmath2503
pour @xmath2307 .
avec les inclusions inverses prouves en [ 6.3].4 , on voit que les trois filtrations sont quivalentes et donc @xmath2400 , @xmath2402 et @xmath2403 sont des homomorphismes . [
6.8 ] on considre la matrice de kac - moody @xmath17 de type @xmath2504 avec @xmath2505 .
pour un bon choix du sgr @xmath75 , on obtient des algbres et groupes de lacets : @xmath2506 $ ] , @xmath2507)$ ] , @xmath2508 et @xmath2509 .
si la caractristique @xmath9 de @xmath8 divise @xmath2510 , @xmath1230 contient toutes les matrices scalaires .
celles - ci sont clairement annules par @xmath2511 et sont contenues dans @xmath2438 si le scalaire est dans @xmath2512 $ ] .
nest pas gk - simple ds que @xmath9 divise @xmath2510 ; il est facile de voir que la rciproque est vraie .
notons @xmath2513 et @xmath2514 sa base canonique .
soit @xmath2515 $ ] , on considre les rseaux @xmath2516 , @xmath2517,@xmath747 , @xmath2518 , @xmath2519 et de manire gnrale @xmath2520 pour @xmath550 .
alors @xmath2417 est le stabilisateur dans @xmath6 de tous ces rseaux et il est assez facile de voir que @xmath2521 pour @xmath758 .
galement @xmath2522 est form des matrices
coefficients dans @xmath618 congrues lidentit modulo @xmath2523 et une matrice triangulaire suprieure modulo @xmath2524 . le groupe @xmath1306 ne contient pas de matrice scalaire .
@xmath9 divise @xmath2510 , on peut trouver dans
@xmath2525 une matrice diagonale @xmath2394 do nt tous les coefficients sont gaux modulo @xmath2526 .
si @xmath2527 , alors @xmath2528 a les mmes coefficients que @xmath2394 sauf un ( colonne @xmath600 , ligne @xmath2529 ) qui est dans @xmath2530 , donc @xmath2531 .
la conclusion de la proposition [ 6.6 ] nest pas vrifie .
les sommets de lappartement standard @xmath1399 sont associs aux rseaux de la forme @xmath2532 .
ainsi le fixateur dune partie finie de @xmath1399 est form des @xmath2533 stabilisant un nombre fini de tels sous - modules . on en dduit facilement
que la filtration de @xmath6 par les @xmath2534 est quivalente la filtration par les @xmath2535 .
notons @xmath2536=u^{ma+}\cap(\cap_{i=1}^{m-1}\,(exp\,f_i)b^{ma+}(n)(exp - f_i))$ ] .
on calcule facilement que @xmath2536 $ ] est form des matrices @xmath2537 telles que @xmath2538 et @xmath2539 .
ainsi @xmath2540 et @xmath2541 .
si @xmath2542 est la plus grande puissance de @xmath9 divisant @xmath2510 , on en dduit que @xmath2543 ; donc @xmath2536\subset u^{ma+}_{mn / p^e}$ ] .
comme la filtration par les @xmath2319 est plus fine que celle par les @xmath2536 $ ] , on dduit de linclusion prcdente et de celles prouves en [ 6.3].4 que toutes ces filtrations sont quivalentes .
la conclusion de la proposition [ 6.7 ] est vrifie mme si @xmath1230 nest pas gk - simple . on a aussi @xmath2442 : @xmath6 est gk - simple .
[ 6.9 ] 1 ) on a vu au cours de la dmonstration de la proposition [ 6.7 ] que , si @xmath1230 est gk - simple et @xmath8 infini , les filtrations de @xmath1306 par les @xmath2377 , @xmath2544 ou @xmath2545 sont quivalentes .
en particulier @xmath6 est gk - simple : @xmath2442 ( ( * ? ? ?
8) en caractristique @xmath4 ) .
\2 ) si @xmath6 est gk - simple ou plus gnralement si @xmath2546 , lhomomorphisme @xmath2547 est injectif ( [ 6.3].6.b ) .
si de plus @xmath8 est fini , on en dduit que @xmath2548 et @xmath2549 sont des isomorphismes de groupes topologiques ( [ 6.3].5 ) ; ainsi les groupes topologiques @xmath2408 , @xmath2360 et @xmath2306 sont isomorphes .
inversement pour @xmath8 fini , lisomorphisme de @xmath2408 et @xmath2306 implique que @xmath2546 ( [ 6.1].4 ) ; si la caractristique de @xmath8 est grande , cela implique que @xmath6 est gk - simple ( [ 6.11 ] ) .
\4 ) supposons @xmath1230 gk - simple .
soit @xmath2550 un corps contenant @xmath8 . on considre limmeuble @xmath2551 de @xmath1079 sur @xmath2550 , il contient @xmath1341 . on note
@xmath2552 la boule de @xmath2551 de centre @xmath2310 et rayon @xmath167 et @xmath2553 son fixateur dans @xmath1306 . les groupes @xmath2377 , @xmath2544 et @xmath2553 sont les intersections avec @xmath1306 des groupes
analogues @xmath2377 , @xmath2544 et @xmath2319 dans @xmath2554 .
si @xmath2550 est infini la dmonstration de [ 6.7 ] prouve que @xmath2555 si @xmath2556 .
notons @xmath2557 ( resp .
@xmath2558 ) le complt de @xmath1712 ( resp .
@xmath5 ) pour la filtration par les @xmath2559 .
alors les groupes @xmath2408 , @xmath2360 et @xmath2558 sont des groupes topologiques isomorphes ( et de mme pour @xmath2366 , @xmath2362 et @xmath2557 ) . en particulier @xmath2557 et @xmath2558 ne dpendent pas du choix du corps infini @xmath2550 contenant @xmath8 .
\5 ) supposons @xmath1230 gk - simple .
sil y a injection continue du groupe @xmath2323 associ @xmath8 dans celui associ lextension infinie @xmath2550 ( hypothse raisonnable mais non videmment vrifie ) alors @xmath2400 est un isomorphisme de groupes topologiques et @xmath2560 , @xmath2561 .
il sagit de savoir si @xmath1712 est dense dans @xmath1306 , cest dire si @xmath1306 est topologiquement engendr par les sous - groupes radiciels @xmath2562 pour @xmath2317 . on va rpondre par un contre - exemple et une proposition assez gnrale .
* contre - exemple . * on considre la matrice de kac - moody @xmath2563 avec @xmath2505 et @xmath2564 . on note @xmath2565
les racines simples , @xmath2566 ( resp .
@xmath2567 ) une base de @xmath2568 ( resp . de @xmath2569 ) sur @xmath2570 et @xmath2571 ( resp .
@xmath2572 ) llment non trivial de @xmath2562 ( resp .
@xmath2573 ) . daprs ( * ? ? ?
5.25 ) les plus petites racines de @xmath2574 sont @xmath2575 tandis que @xmath2576 sont des racines imaginaires . de plus @xmath2577 ou @xmath2578 nest pas une racine pour @xmath2579 .
soit @xmath108 lidal de @xmath2580 form des racines positives de la forme @xmath2581 avec @xmath2582 ou @xmath2583 ; il contient toutes les racines relles positives sauf @xmath2584 et @xmath2585 .
notons @xmath2586 , cest un sous - groupe ouvert de @xmath1306 . daprs
[ 3.3 ] @xmath2587 est un groupe commutatif isomorphe @xmath2588 .
si @xmath1712 est dense dans @xmath1306 ce groupe doit tre engendr par les images de @xmath2589 et @xmath2590 .
on fait des calculs dans le quotient de @xmath1075 par lidal @xmath2591 . on utilise que @xmath2383 et @xmath2592 . on note @xmath2593 .
ainsi @xmath2594 , @xmath1075 admet @xmath2595 comme base dun supplmentaire de @xmath2591 et lisomorphisme de @xmath2596 sur @xmath2588 sobtient par la projection vidente .
dans cette algbre quotient @xmath2597 et @xmath2598 .
on peut calculer facilement les 17 mots en a et b de longueur @xmath2599 .
on constate que @xmath2600 et @xmath2601 .
ainsi ces 17 mots constituent toute limage de @xmath1712 dans cette algbre quotient et il y a au maximum 14 mots diffrents .
comme @xmath2587 est de cardinal 16 , il nest pas gal au groupe engendr par @xmath2589 et @xmath2590 ( il contient 8 lments et pas @xmath669e^{(2)}*f$ ] ) .
[ 6.11 ] supposons la caractristique @xmath9 du corps @xmath8 nulle ou strictement plus grande que @xmath1059 ( [ 6.3].2 ) .
alors le groupe @xmath1712 ( resp .
@xmath5 ) est dense dans @xmath1306 ( resp .
@xmath6 ) . daprs [ 2.3 ] lalgbre de lie @xmath2438 est engendre par les @xmath519 pour @xmath204 .
montrons par rcurrence sur @xmath20 que tout lment @xmath2602 est congru un lment de @xmath1712 modulo @xmath2603 , cest clair pour @xmath2464 . daprs
[ 3.3 ] @xmath2604 est un groupe commutatif isomorphe @xmath2605 . daprs
[ 3.2 ] et les bonnes proprits de la base des @xmath612 $ ] ( [ 2.9].4 ) , lisomorphisme @xmath2606 est donn comme suit : on plonge @xmath1306 dans @xmath1075 et llment @xmath2602 est congru @xmath2607 modulo @xmath2608 .
il suffit de montrer le rsultat cherch pour des lments @xmath1783 engendrant @xmath2604 , on va considrer ceux tels que @xmath2609 soit de la forme @xmath2610 $ ] avec @xmath204 et @xmath2611 .
par hypothse il existe @xmath2612 tel que @xmath2613 , donc @xmath711 est congru @xmath2614 modulo @xmath2615 .
considrons @xmath2616 .
on calcule facilement ce produit dans @xmath2617 , il est gal @xmath2618 $ ] .
donc @xmath2619 et @xmath2620 @xmath2621 modulo @xmath2603 . on va maintenant gnraliser le rsultat de simplicit de moody @xcite . on suit son
schma de dmonstration en reproduisant mme les parties inchanges , cause de la disponibilit restreinte de cette rfrence . on note @xmath2624 ( resp .
@xmath2625 ) le sous - groupe de @xmath5 ( resp .
@xmath6 ) engendr par les groupes radiciels @xmath2562 pour @xmath2626 ( resp . et par @xmath1306 ) .
il est normalis par @xmath1225 et les @xmath2627 donc est distingu dans @xmath5 ( resp .
@xmath6 ) . daprs
[ 6.4 ] on a @xmath2628 .
\2 ) le cas a ) rsulte de 1 ) , de la proposition [ 6.11 ] et de ce que @xmath2635 .
pour le cas b ) on peut utiliser le lemme [ 6.15 ] ci - dessous .
par un calcul analogue au prcdent on en dduit que @xmath2377 est parfait modulo @xmath2603 .
[ 6.15 ] on suppose la matrice de kac - moody @xmath17 sans facteur de type affine et le corps @xmath8 infini , alors @xmath2636 il existe @xmath2637 tel que @xmath2638 .
si @xmath8 nest pas extension algbrique dun corps fini on peut supposer @xmath2639 indpendant de @xmath2584 .
on peut supposer @xmath17 indcomposable et mme de type indfini , car le cas de type fini est clair .
daprs @xcite il existe alors @xmath2640 avec @xmath2641 et @xmath2642 @xmath2643 .
soient @xmath2644 et @xmath2645 ; comme @xmath2646 , @xmath2647 .
si @xmath2648 , on a @xmath2649 ; mais les @xmath2650 sont tous de mme signe et non tous nuls , donc @xmath2651 . ainsi @xmath2652
si @xmath556 est dordre @xmath2653 dans @xmath2654 ou , mieux , sil est dordre infini .
[ 6.16 ] on suppose @xmath2625 topologiquement parfait et la matrice de kac - moody @xmath17 indcomposable .
soit @xmath2 un sous - groupe ferm de @xmath6 normalis par @xmath2655 , alors @xmath2656 ou @xmath2657 . le systme de tits @xmath2659 , les groupes @xmath1306 et @xmath2 satisfont aux conditions du thorme 5 de ( * ? ? ? * iv 2 n7 ) sauf peut - tre ( 2 ) ou ( 3 ) ( avec @xmath2660 , @xmath2661 et @xmath2662 ) . on peut donc suivre la dmonstration
de ce thorme jusquau point o ( 2 ) et ( 3 ) sont utiliss .
si @xmath2663 on obtient donc @xmath2664 .
les sous - groupes @xmath2665 sont ouverts donc ferms ; comme @xmath2625est topologiquement parfait , on a @xmath2666 , @xmath2667 ( [ 3.4 ] ) .
soit @xmath2668 , crivons @xmath2669 avec @xmath2670 et @xmath2671 .
la suite @xmath2469 tend vers @xmath220 , donc @xmath2672 tend vers @xmath1783 et comme @xmath2 est ferm @xmath2673 . [ 6.17 ] on suppose @xmath2625 topologiquement parfait et la matrice de kac - moody @xmath17 indcomposable .
soient @xmath2 un sous - groupe de @xmath6 normalis par @xmath2625 et @xmath2674 tels que @xmath2663 et @xmath2652 @xmath2675 .
alors @xmath2676 .
daprs [ 6.16 ] @xmath2677 contient @xmath2625 .
soit @xmath2678 une suite dlments de @xmath2 convergeant vers @xmath1868 .
pour @xmath168 assez grand @xmath2679 . on va montrer que @xmath1868 est conjugu dans @xmath2625 @xmath2680 , donc @xmath2676 .
supposons @xmath2688 et @xmath2689 dj construits . on sait
que @xmath2690 est isomorphe au groupe additif somme des @xmath2450 pour @xmath2691 .
notons @xmath2692 llment correspondant la classe de @xmath2693 . on dfinit
@xmath2694 comme un lment de @xmath2695 do nt la classe modulo @xmath2696 correspond @xmath2697 .
soit @xmath2698 , sa classe modulo @xmath2696 correspond llment @xmath2699 .
donc @xmath2700 .
posons @xmath2708 et supposons @xmath2709 dj construits .
la classe de @xmath2710 dans @xmath2690 correspond un lment @xmath2692 de la somme des @xmath2450 pour @xmath2691 .
on dfinit @xmath2694 comme un lment de @xmath2695 do nt la classe modulo @xmath2696 correspond @xmath2697 .
alors @xmath2711 a la mme image dans @xmath2690 que @xmath2712 ( [ 3.3].f ) qui correspond @xmath2699 .
donc , si on pose @xmath2713 , on a bien @xmath2714 .
[ 6.19 ] on suppose @xmath2625 topologiquement parfait , la matrice de kac - moody @xmath17 indcomposable non de type affine et le corps @xmath8 de caractristique @xmath4 ou de caractristique @xmath9 mais non algbrique sur @xmath10 .
alors pour tout sous - groupe @xmath2 de @xmath6 normalis par @xmath2625 , soit @xmath2 est contenu dans @xmath2715 , soit il contient @xmath2625 . en particulier @xmath2658 est un groupe simple .
si @xmath2663 , les hypothses des lemmes [ 6.17 ] et [ 6.18 ] sont vrifies ( grce au lemme [ 6.15 ] ) .
donc @xmath2703 .
mais @xmath2625 contient les @xmath2716 ( pour @xmath2626 ) et normalise @xmath2 , donc @xmath2 contient aussi les @xmath2562 pour @xmath2626 , cest dire @xmath2657 .
\2 ) la simplicit de @xmath2658 en caractristique @xmath4 est le rsultat essentiel de r. moody dans @xcite .
il considre en fait un groupe dautomorphismes du complt de lalgbre de kac - moody simple associe une matrice de kac - moody @xmath17 ( indcomposable non de type affine ) . il ny a donc pas pour lui dhypothse de gk - simplicit de @xmath1230 ( et @xmath2717 pour son choix de groupe ) .
3 ) pour un corps fini ( de cardinal @xmath2718 ) et une matrice @xmath17 @xmath2719sphrique indcomposable , carbone , ershov et ritter montrent la simplicit de @xmath2720 ( * ? ? ?
ce groupe est ladhrence de limage de @xmath2624 dans aut@xmath2721 , cest dire , par compacit , limage de ladhrence @xmath2722 de @xmath2624 dans @xmath6 .
donc @xmath2723 .
en grande caractristique , [ 6.11 ] montre que @xmath2724 ; on retrouve alors un nonc analogue [ 6.19 ] .
si @xmath8 est fini mais assez grand , on sait aussi que
@xmath2725 est simple ( * ? ? ?
20 ) . on en dduit facilement
le mme rsultat pour @xmath8 extension algbrique dun corps fini .
\4 ) si @xmath17 est indcomposable
de type affine non tordu , @xmath2658 est le groupe des points sur @xmath2726 dun groupe algbrique simple dploy , il est donc simple .
le mme rsultat est sans doute encore vrai si on enlve `` non tordu '' et `` dploy '' .
victor g. kac & dale h. peterson , defining relations of certain infinite dimensional groups , in _ lie cartan et les mathmatiques daujourdhui , lyon ( 1984 ) _ , astrisque n@xmath2729 hors srie ( 1985 ) , 165 - 208 .
guy rousseau , euclidean buildings , in `` gomtries courbure ngative ou nulle , groupes discrets et rigidit , grenoble , 2004 '' , l. bessires , a. parreau & b. rmy , _ sminaires et congrs _ * 18 * ( soc . math .
2008 ) , 77 - 116 .
robert steinberg , gnrateurs , relations et revtements de groupes algbriques , in _ colloque sur la thorie des groupes algbriques , bruxelles ( 1962 ) _ , centre belge rech . math .
, librairie universitaire louvain & gauthier - villars paris ( 1962 ) , 113 - 127 .
jacques tits , twin buildings and groups of kac - moody type , in _ groups combinatorics and geometry ( durham , 1990 ) _ , m. liebeck & j. saxl , london math .
lecture note * 165 * , ( cambridge u. press , cambridge , 1992 ) , 249 - 286 . | pour un groupe de kac - moody dploy ( au sens de j. tits ) sur un corps rellement valu quelconque , on construit une masure affine ordonne sur laquelle ce groupe agit
. cette construction gnralise celle dj effectue par s. gaussent et lauteur quand le corps rsiduel contient le corps des complexes @xcite et celle de f. bruhat et j. tits quand le groupe est rductif . on montre
que cette masure vrifie bien toutes les proprits des masures affines ordonnes comme dfinies dans @xcite . on utilise
le groupe de kac - moody maximal au sens do .
mathieu et on montre quelques rsultats pour celui - ci sur un corps quelconque ; en particulier on prouve , dans certains cas , un rsultat de simplicit pour ce groupe maximal . |
type ii superconductors @xcite allow for a partial penetration of magnetic field into the bulk of the superconducting ( sc ) material when the applied field @xmath4 satisfies @xmath5 . in a seminal work abrikosov @xcite
showed that when the ratio @xmath6 , where @xmath7 is the magnetic field penetration depth and @xmath8 is the coherence length , is greater than @xmath9 the magnetic field penetrates the sc material in the form of flux - lines ( fls ) .
these fls are also called vortices , since they are surrounded by circular currents .
each fl carries a quantized unit of flux @xmath10 called the fluxoid .
the fls have cylindrical shape of radius @xmath11 ( the radius is not sharp since the magnetic field decays exponentially like @xmath12 , where r is the distance from the axis ) and a non - sc core of radius @xmath13 .
due to a repulsive interaction among the fls , they arrange themselves in a triangular lattice referred to as the vortex solid ( vs ) .
this result follows from mean - field theory .
after high - temperature superconductors were discovered in the 1980 s , it became apparent that thermal fluctuations , not included in the mean - field theory @xcite , play an important role at relatively high temperatures and fields , still below @xmath14 and @xmath15 .
these fluctuations can cause the abrikosov lattice to melt into a disordered liquid via a first order transition ( fot ) , which can be roughly estimated using the lindemann criterion known from solid state physics @xcite .
technologically , the melting of the fl lattice is important since the vortex liquid ( vl ) is not actually sc due to the dissipation caused by the fl motion when an electric current passes through the system .
pinning of fls by naturally occurring defects in the form of vacancies , interstitials , twin and grain boundaries etc .
, is effective to impede fl motion in the vs phase , where the fls form a rigid correlated network .
the effectiveness of the pinning manifests itself by leading to high critical currents . in the vl phase pinning of a few vortices
does not inhibit others from moving when a current is applied .
thus for practical purposes the sudden increase in resistivity occurs at the melting transition rather than when @xmath16 for any reasonably non - vanishing currents .
the existence of the melting transition in high-@xmath14 pristine materials has been established through numerous experimental @xcite and numerical @xcite studies . as was mentioned above , disorder in the form of points defects and sometimes more extended defects can and
does occur naturally in laboratory samples .
in addition artificial point defects can be induced by bombarding the sample with electrons originating from particle accelerators .
extended columnar defects in the form of linear damaged tracks piercing through the sample can be induced by heavy ion irradiation .
both naturally occurring and artificially induced defects are situated at random positions in the sample and their effective pinning strength ( i.e. their interaction with fls ) can also vary from defect to defect .
thus defects play the role of quenched disorder .
the adjective `` quenched '' refers to the immobility of these defects during experimental time scales .
introduction of disorder in terms of point defects or columnar pins affects both the properties of the solid and liquid phases and might also shift the location of the melting transition in the h - t plane @xcite . in the case of point pins ,
the vs phase is replaced with a bragg glass phase @xcite , characterized by quasi - long - range order .
the melting transition is predicted to shift towards lower temperatures @xcite . in the case of columnar pins
the vs phase is replaced with a so called pinned bose glass @xcite where fls are trapped by the columnar defects and the whole lattice becomes immobile .
the bose glass phase is similar to the localized phase of a two dimensional repulsive bose gas in the presence of quenched disorder , as will be explained in more detail in the next section .
the effect of both kinds of disorders on the fls melting has been studied experimentally in various high - temperature superconductors .
two common materials that have been extensively investigated are yba@xmath0cu@xmath1o@xmath2 ( ybco ) and bi@xmath0sr@xmath0cacu@xmath0o@xmath3 ( bscco ) , both having critical temperatures ranging between 90 - 120 k at @xmath17 .
the main difference between these materials is their anisotropy parameter @xmath18 , where @xmath19 and @xmath20 denote the effective masses of electrons moving along the @xmath21-axis and the @xmath22-plane respectively .
bscco is much more anisotropic : its anisotropy parameter @xmath23 lies in the range of 50 - 200 compared to the range of 5 - 7 for ybco @xcite .
this fact causes the fls to be much `` softer '' or elastic .
thus in the case of bscco the fls are sometimes described as a collection of loosely connected `` pancakes '' residing in adjacent cu - o planes .
experimental studies on ybco have shown a marked shift in the irreversibility line in the presence of the columnar disorder @xcite .
the irreversibility line in the h - t plane marks the onset of hysteresis effects and is located close to the melting transition on the solid phase side . for bscco , many experimental studies have been conducted @xcite .
the more recent ones have shown @xcite that the melting line is not shifted when the density of columnar defects is relatively low , @xmath24 , but for @xmath25 a shift in the position of the melting transition is observed . here
the matching field @xmath26 is defined as @xmath27 where @xmath28 is the density of the columnar defects and @xmath29 is the flux quantum .
theoretical work on columnar disorder includes bose glass theory @xcite .
radzihovsky @xcite considered the possibility of two kinds of bose glass phases ( strongly or weakly pinned ) depending on whether @xmath30 or @xmath31 .
more recently , columnar as well as point disorder were investigated by goldschmidt @xcite using replica field theory .
he showed that the melting line shifts to lower temperature in the case of point disorder and to higher temperature in the case of columnar disorder . due to the complexity of the problem , especially in the presence of disorder
, simulations have been very useful in studying the fl system .
there have been many simulation studies of the vortex system in the presence of disorder .
however , most simulation work has been confined to the addition of point disorder only . in particular , there is little work done on the effects of columnar disorder on the fl melting .
recent work by wengel and tuber @xcite concentrated on the case of high defect density region @xmath32 .
in contrast , a recent simulation study by sen _ et al . _ @xcite uses a small density of columnar defects , each of infinite strength , but considers an extremely small magnetic field .
similarly , nandgaonkar _ et al .
_ @xcite also investigate the case of a very small magnetic field . in this paper
we consider columnar defects of a finite strength ( @xmath33 ) with a relatively low defect density , @xmath34 , and with realistic magnetic fields as used in the experiments .
we first discuss the method implemented for ybco : following nelson @xcite , we map the system of @xmath35 vortices ( fls ) in a high-@xmath14 superconductor onto @xmath35 interacting bosons in 2-space + 1-time dimensions
. then we do a path integral monte carlo ( pimc ) simulation on this system .
the partition function of @xmath35 vortices can be expressed as : @xmath36 where @xmath37 denotes the 2d position vector of the ith vortex at a height @xmath38 along the @xmath21-axis , @xmath39 is the free energy as a function of temperature @xmath40 , @xmath41 is the length of the sample along the @xmath38-direction .
the london free - energy functional @xmath42 is given by @xmath43 where @xmath44 @xmath45 is the vortex line energy per unit length , the line tension is @xmath46 and @xmath47 . here @xmath48 is the lattice spacing , @xmath49 is the magnetic field along the @xmath38-direction and @xmath23 is the anisotropy parameter .
@xmath50 is the in - plane interaction potential between two fls a distance @xmath51 apart .
@xmath52 denotes the modified bessel function of the first kind .
this expression for the london free - energy is an approximation that neglects the non - local interaction of real vortices and replaces it by in - plane interactions only , which is really justified if the fls do not deviate too much from straight lines along the @xmath38-axis @xcite . with this approximation ,
the system of @xmath35 interacting fls is equivalent to a system of @xmath35 bosons in @xmath53 dimension interacting with a pairwise potential @xmath50 .
the path integral representation of a system of @xmath35 bosons of mass @xmath54 each in two dimensions , interacting through a potential @xmath55 , with @xmath56 being the strength and @xmath7 being the range of the repulsive interaction , is given at finite temperature @xmath57 in terms of the imaginary time action @xmath58 here @xmath59 is the imaginary time and @xmath57 is the temperature of the bose system .
we see that there is a one to one parameter mapping between the boson system and the vortex system @xcite : @xmath60 where @xmath61 is the average density of bosons ( and fls ) and @xmath62 is the area of the sample .
we can write the london free - energy functional in a dimensionless form as follows : @xmath63 where @xmath64 all lengths are measured in units of @xmath65 and energies in units of @xmath66 for bosons and @xmath67 for fls .
we next discretize the integral along the @xmath38-axis by dividing it into @xmath68 segments : @xmath69/kt}\label{xi}\ ] ] where @xmath70 , @xmath54 labels the planes and @xmath71}{kt}=\sum _ { i , m } \frac{\left(\mathbf{r}_{i , m+1}-\mathbf{r}_{i , m}\right)^{2 } } { 2\lambda^{2}\tau } + \sum _ { i < j}\tau k_{0}\left(r_{ij , m}/\tilde{\lambda } \right).\ ] ] we work only in the limit where @xmath72 is large , which amounts to taking @xmath41 very large and in the mapping onto @xmath73 bosons corresponds to the ground state of the bosons at the absolute zero temperature ( @xmath74 )
. we used @xmath75 and discretized the @xmath38-axis into @xmath76 planes .
we used the matrix - squaring method @xcite to calculate the right action so that we can work with a small number of planes along the @xmath38-direction . working with the primitive action
requires the use of a large number of slicing of the @xmath38-direction and is very time consuming @xcite .
the boundary conditions in the @xmath38-direction for a system of bosons are @xmath77 , with all permutations @xmath78 of the indices being allowed .
this is what is meant by the summation over @xmath79 in eq .
( [ xi ] ) . for small @xmath56 ( a small repulsive interaction ) , which corresponds to large @xmath80
, we expect a bose - einstein condensed phase where permutations are important .
this is the superfluid phase . for small @xmath80 ,
repulsion is large and permutations are rare as the bose system is in its classical phase which is a wigner crystal .
these two phases correspond respectively to the fl liquid and solid phases .
the melting line represented by @xmath81 is given approximately by the expression @xmath82 .
near @xmath83 this behavior is more complicated ( see below ) .
we now discuss the method of simulations of bscco : because of the high anisotropy of the bscco system one can not use the simple picture given above . here , instead , we follow a different model which can be cast in a form analogous to ybco .
we take the lawrence - doniach model @xcite as our starting point for bscco .
it leads to the following form of the london free - energy for inter - layer ( il ) josephson coupling @xcite : @xmath84 , \label{quadjos}\ ] ] for @xmath85 , and @xmath86 , \label{linjos}\ ] ] for @xmath87 , where @xmath88 is the inter - layer spacing and @xmath89 is the healing length defined by @xmath90 . for the in - plane ( ip ) coupling we use : @xmath91 in principle one should add to the above interaction an electromagnetic interaction among the pancake vortices @xcite .
the electromagnetic interaction becomes dominant in the limit of infinite anisotropy ( @xmath92 ) .
@xcite argue , ( see also ref . ) that for the value @xmath93 for bscco the josephson coupling still dominates by one order of magnitude over the electromagnetic coupling .
their argument goes as follows : clem @xcite shows that if one has a straight array of pancake vortices along the @xmath38-axis , and one pancake is displaced a distance @xmath94 in the @xmath95-direction than the magnetic energy of the configuration increases by an amount @xmath96 where @xmath97 is euler s constant ( = 0.5772 ... ) . for large @xmath94 ( @xmath98 ) ,
the modified bessel function @xmath99 decays exponentially and thus the energy increases like @xmath100 . for small @xmath94 the bessel function
can be expanded in a power series in @xmath101 @xmath102 and thus the magnetic energy behaves like @xmath103 to leading order in @xmath94 which is the same as the quadratic behavior of the josephson energy in eq .
( [ quadjos ] ) above .
the ratio of the coefficients of the quadratic terms in the magnetic and josephson energies goes roughly like @xmath104 ( where @xmath105 for bscco ) .
thus for anisotropy @xmath93 we get a factor of 0.25 ( a somewhat more precise estimate @xcite gives a ratio of about 0.1 ) .
thus the magnetic interaction is negligible compared to the josephson interaction for @xmath93 . for samples with @xmath106 these interactions are already comparable . for large values of @xmath94
the magnetic interaction increases logarithmically and the josephson interaction increases linearly so the magnetic interaction is always negligible .
the key to the estimate given above is to consider not just two pancake vortices but a whole line with one displaced pancake .
this argument is valid if the deviations of the vortices from straight lines are not too large . as for the in - plane interaction clem showed that a linear array of pancake vortices gives rise to exactly the same magnetic field at a distance @xmath94 away from it as produced by an abrikosov vortex line .
( [ logjos ] ) is consistent with the magnetic interaction of pancake vortices , again when the fls do not deviate too much from straight lines .
equations ( [ quadjos])-([logjos ] ) for bscco can be cast in a form similar to that for the ybco with the following substitutions : @xmath107 with these changes the london free - energy functional would look like@xmath108,\ ] ] for @xmath109 , and @xmath110- \frac{(2r_{g}-(|\mathbf{r}_{i , m}-\mathbf{r}_{i , m+1}|))^{2 } } { 2\lambda ^{2}\tau } , \label{second}\ ] ] for @xmath111 , when now again all lengths are measured in units of @xmath65 . while doing simulations at a fixed magnetic field @xmath49 and temperature @xmath40 , the term @xmath112 will remain constant and would drop out of @xmath113 term in the boltzmann factor .
it however needs be considered during the measurement of the energy .
the second term in eq . ( [ second ] ) can be easily handled at the last stage of the bisection method ( see the next section for a discussion of the bisection method ) .
we can make use of a reduced temperature variable to make some expressions look simpler .
first , using the fact that the temperature dependence of @xmath114 arises mainly through @xmath115 and neglecting the logarithmic corrections , one gets @xcite
@xmath116 and hence@xmath117 defining reduced temperature as : @xmath118 we obtain@xmath119 this shows that the equation for the melting line is approximately@xmath120 note that some authors @xcite use a temperature dependence of @xmath121 in @xmath7 , or even @xcite @xmath122 .
all these choices coincide near @xmath14 .
the choice of temperature dependence of @xmath7 is not expected to have a significant effect on the results .
the technique that we use to simulate our system is called multilevel monte carlo simulation ( mmc ) @xcite .
there are several advantages in using this technique for the simulation of the fls over the usual metropolis monte carlo ( mc ) method . in the discrete model , we work with @xmath35 fls with the @xmath38-axis discretized into @xmath68 planes , thus resulting in @xmath35 beads in each of the @xmath68 planes . in the usual mc method
one would displace a few of these beads in a plane by small random displacements inside a two dimensional box and then would accept or reject the move based on a probability given by the boltzmann factor .
a big disadvantage of using this technique is that it is difficult to move beads appreciably from their original positions over a number of mc steps .
the reason for this is that a bead in a plane belonging to a fl finds itself in a local harmonic potential generated by the kinetic energy term involving this bead and the beads belonging to the same fl on either side of the plane .
this harmonic potential becomes stronger and stronger at lower temperatures and magnetic fields . as a result , in the usual mc simulations beads
keep moving around inside these local harmonic cages and end up sampling only a small part of the phase space .
the other problem with the usual mc method is that there is no natural easy way of implementing fl cutting . if there are two fls twisted around each other and if it is energetically favorable for them to reconnect each other in such a way as to lower their free energy then this step should be permitted in the mc method without regard to the question if this process occurs in reality .
this is so because in the mc simulations phase space is sampled according to the probability distribution and all one needs is to generate configurations weighted by the boltzmann factor , and the path followed in configuration space has nothing to do with any real dynamics .
these two main drawbacks of the usual mc method are easily overcome in the mmc technique .
first , one moves bigger chunks of fls encompassing beads in several planes .
this way one can avoid local harmonic traps .
( this is like taking an aerial route to a destination rather than going through the zigzag maze of roads . )
this is much in the spirit of fourier space monte carlo where one first samples modes with smaller wave numbers and then move onto higher modes .
the method of creating new fl configurations is based on the concept of the conditional probabilities .
it is called the bisection method @xcite because one starts sampling beads by iteratively bisecting the fls . at each stage of the division ,
the beads belonging to that stage are moved with some conditional probability factor @xmath123 .
it is important to make sure that the probabilities are chosen in such a way that detailed balance is satisfied at each stage of the division .
one notes that the @xmath123 s may not be the actual boltzmann factors for the beads to be moved at different levels .
but what is required is that when all @xmath123 s are multiplied together , they cancel in such a way so as to leave the correct boltzmann factor for the whole move .
thus , inherent in this algorithm is the fact that a move would finally be accepted only if it has been accepted at each stage of the bisection method .
the power of this method lies in choosing the appropriate @xmath123 s .
if these conditional probability factors are chosen judiciously , most of the rejections would take place at the initial stages of the bisection process when not too much computational effort has been spent yet .
the cutting and reconnection of fls is implemented naturally in a mmc method : permutation among the 3 or 4 lines chosen to be moved becomes the first among the many hierarchical steps one goes through before a move is finally accepted and the position of the beads updated accordingly .
we typically moved a total of 15 to 20 beads distributed over 5 planes .
permutations were sampled by a random walk algorithm through the space of permutations @xcite(see appendix b ) .
in the case of ybco we worked with a field of @xmath124 g. working in the primitive approximation of the action would require the use of a smaller value of the dimensionless parameter @xmath59 , which would require slicing of the @xmath38-direction into a large number of planes . to avoid this ,
the matrix - squaring method @xcite has been used to get the effective action for bigger values of @xmath59 .
for example nordborg and blatter @xcite use a value of @xmath125 and they work with 100 planes . in this simulation
a value of @xmath126 has been used and the @xmath38-axis has been sliced into 75 planes . choosing a bigger value of @xmath127 by utilizing the matrix - squaring method makes it easier to equilibrate the system as compared to working with the primitive approximation . for bscco
, we did not use the matrix - squaring method because of the complications involved due to the few extra terms in free energy which contribute depending on whether @xmath51 is smaller or bigger than @xmath128 . here
, we used the natural inter - layer spacing @xmath88 to calculate the parameter @xmath127 at different temperatures and then used the mmc technique to efficiently sample the configurational space . as mentioned previously , in the present simulations we included only the josephson coupling .
this approximation works well with ybco but it could be less satisfactory for bscco because of its high anisotropy . for very anisotropic materials
the electromagnetic coupling becomes important @xcite .
as discussed in section 2 , ryu _ et al . _
@xcite estimated that for anisotropy of magnitude @xmath93 the josephson interaction still dominates over the electromagnetic interaction by an order of magnitude , but this will not be the case for @xmath106 which can characterize some samples .
@xcite discuss how to include the electromagnetic interaction in a mc simulation , but they only consider the opposite limit where the josephson coupling is totally neglected . to our knowledge there is no satisfactory treatment of both couplings included on equal footings .
we carried out preliminary simulations which show that the inclusion of the electromagnetic coupling does not shift the position of the transition line much at a field of @xmath129 g , thus supporting our current conclusions .
these results will be reported elsewhere @xcite .
simulations were usually carried out for @xmath130 and @xmath131 fls .
the decay of the structure factors at the transition temperature becomes sharper when one uses a larger number of fls .
however , no appreciable shift in the transition temperature is seen while working with the smaller ( @xmath132 ) or the larger system ( @xmath133 ) .
we did not run our simulations for even larger systems as it becomes computationally very time consuming .
typical simulation times were @xmath134 mc steps .
each mc step involved moving 3 - 4 lines in 5 planes simultaneously .
we usually averaged over 10 - 15 realizations of the columnar disorders , though some results have an average over as many as 20 realizations of the disorder .
columnar disorder is modeled as an array of straight cylindrical wells of typical radius @xmath135=25 - 35 placed randomly throughout the cross - sectional region of the sample and oriented along the @xmath38-direction @xcite .
each columnar defect is of length @xmath136 .
the density of the columnar pins can be varied by changing their number for a given cross sectional area , and the strength is controlled by a positive dimensionless parameter @xmath137 .
if a bead happens to wander inside a columnar well , we include an extra free energy of @xmath138 .
the defect concentration was taken to be a 20 percent ratio of defects to fls which means @xmath139 .
the strength of disorder was set at @xmath140 for bscco .
for ybco , @xmath141 was found to be too large in the sense that the transition became too broad and useful information could not be extracted .
thus , for ybco we kept @xmath142 .
other parameters used for ybco are @xmath143 @xmath144 , @xmath145 , @xmath146 .
parameters for bscco are as follows : @xmath147 , @xmath148 , @xmath149 , @xmath150 .
in this section we describe many different physical quantities that were monitored during the simulation . from the variation of these quantities with the temperature we can extract important conclusions about the different phases of the system . in terms of the reduced temperature , the energy @xmath151 can be simply written as @xcite : @xmath152 where , @xmath153 and @xmath154-\frac{(2r_{g}- ( |\mathbf{r}_{i , m}-\mathbf{r}_{i , m+1}|))^{2}}{2\lambda ^{2}\tau } \big\rangle,\ ] ] for @xmath111 , whereas @xmath155 any discontinuous jump in @xmath156 would indicate a fot . from this discontinuous jump in energy , @xmath113
, we can also calculate the jump in entropy @xmath157 , @xmath158 the translational structure factor @xmath159 is defined as , @xmath160 where @xmath161 stands for the mc average , @xmath162 is a reciprocal lattice vector and is of the form @xmath163 and @xmath164 and @xmath165 are some integers . @xmath166
represent the basis vectors of the reciprocal lattice @xmath167 where @xmath168 , @xmath65 is the nearest neighbor distance and @xmath169 are the unit vectors along the hexagonal unit cell such that @xmath170 in the simulations , @xmath164 and @xmath165 in eq .
( [ struc ] ) are chosen to be 1 and 0 or 0 and 1 .
these choices correspond to the first bragg peak .
we will normally write @xmath171 as simply @xmath172 .
this quantity gives important information about the phase of the system . in an ordered phase where fls sit on a triangular lattice
, @xmath172 is of the order of @xmath35 . in the disordered phase , it saturates to almost zero as the system size increases . in the simulations
, however , there is a problem with measuring @xmath173 , especially in the presence of disorder .
we will return to this point at the end of the section .
we use delaunay triangulation to measure the hexatic order parameter , @xmath174 , which is defined as , @xmath175 where @xmath176 denotes the number of the nearest neighbors of a bead at position @xmath177 , @xmath54 and it is 6 for a perfect hexagonal lattice , @xmath178 stands for the bond angle , that is the angle that vector @xmath179 makes with an arbitrary axis . just like the @xmath172
, this quantity too has a large value in an ordered phase and saturates at a finite value for a system of finite size .
as we allow permutations of fls , we can define a number @xmath180 as that fraction of the total number of fls which belong to loops that are bigger than the size of a `` simple '' loop .
a simple loop is defined as a set of @xmath68 beads connected end to end , @xmath68 being the total number of planes .
loops of size @xmath181 , @xmath182 ... start proliferating at and above the transition temperature and in the corresponding 2d boson system this proliferation is related to the onset of the superfluidity .
the transverse fl fluctuations are measured by @xmath183 which is independent of @xmath184 . at the transition temperature @xmath185
undergoes a large increase for large @xmath38 .
we want to emphasize one important aspect of measuring the translational structure factor .
the usual way of measuring this quantity is by choosing @xmath186 corresponding to the first bragg peak i.e. , @xmath187 in eq .
( [ struc ] ) .
now , it might happen that the configuration of the fls comes close to making an almost perfect hexagonal lattice but its basis vectors are not aligned along the usual major axes of the rhombically shaped cell encompassing the system .
if we use the reciprocal lattice vector @xmath188 to measure the structure factor of such a configuration , we would end up getting a very small value for @xmath173 and might wrongly conclude that the system is in a very disordered state .
this happened many times in our simulations ; we got a very low value of the translational structure factor while the hexatic order was indicating a high degree of orderliness in the fl lattice .
to remedy this situation the translational structure factor is measured at 60 different angles , 1 degree apart , and choose that number which gives the largest possible value of the @xmath189 corresponding to some angle @xmath190 . after implementing this technique
we find that @xmath173 and @xmath174 , which differ so much initially , almost follow each other .
\1 . for ybco , simulations were carried out at a magnetic field of @xmath124 g. at this field we have @xmath191 . by the argument given in ref .
, we conclude that our results should hold qualitatively for any @xmath49 such that @xmath192 .
we checked our simulation results against those by nordborg and blatter @xcite .
first we verified that our code was working fine by comparing our results against those given in ref . .
working in the limit of @xmath193 , we kept @xmath59 fixed as @xmath80 was increased .
we got a sharp transition at @xmath194 .
a jump was also seen in the energy as defined in the previous section and was found to be @xmath195 , again the same as in ref . .
the effect of introducing columnar disorder on the structure factor at the first bragg peak is shown in fig .
[ all1 ] . in fig .
[ all2 ] we depict the jump in the energy for a pure as well as for a disordered system . at various disorder strengths ( @xmath33 s ) for @xmath132 . a clear shift in transition point towards higher @xmath80 s
is seen for @xmath196 for @xmath132 . a jump in the energy
is seen up to @xmath197 . for @xmath198 ,
the rise in the energy is soothed out . ]
we note that at lower strengths of the disorder , the melting line is almost unaffected and the transition takes place at @xmath199 .
however , as we increase the strength of the disorder , the melting transition shifts towards higher values of @xmath80 which means that the melting line shifts towards higher temperatures and/or higher magnetic fields . for @xmath137 up to @xmath200 no change is seen in the melting curve or in the jump in the free energy functional .
however , at @xmath137 equal to @xmath201 the structure factor comes down at around @xmath202 and , the jump in the energy becomes gradual .
this finding is in agreement with several experimental studies where a change in the irreversibility line is seen with the introduction of columnar disorder @xcite .
next , we look at the fl entanglement . @xmath203 denotes the number of fls which do not form a simple loop . by looking at the @xmath180 vs. @xmath114 graph in fig .
[ all3 ] . .
for @xmath204 a sharp jump in the fl entanglement is seen . for @xmath198 ,
the rise of @xmath205 with @xmath80 becomes gradual as compared with the clean system . ]
it is clear that the entanglement is suppressed by as much as almost one order of magnitude just after and in the vicinity of the original transition line at @xmath194 .
this result is expected , as it is well known that point disorder helps in line wandering and entanglement while the columnar disorder has just the opposite effect @xcite since a nearby fl is induced to align along the columnar defect , and thus its transverse fluctuations are reduced .
unfortunately , there have been few studies on columnar disorder in ybco .
@xcite work with vanishingly small magnetic field and infinite pinning strengths , in the vicinity of the lower branch of the melting line .
figure [ 64ybco ] shows the effect of columnar disorder on a system of @xmath131 fls in ybco . for @xmath133 .
smooth lines are provided as a visual aid .
transition for @xmath206 and @xmath207 are seen almost at the same @xmath80 s as with @xmath132 . ]
this graph shows a transition at almost the same @xmath114 as in fig .
this is a confirmation that finite size effects are not important in our simulations . \2 . for the bscco system the @xmath156 vs. @xmath40 graphs for different @xmath33 values
are depicted in fig .
[ 36energy ] ( 36 fls ) and fig .
[ 64energy ] ( 64 fls ) . at @xmath208 g for @xmath132 .
lines were added to help visualize the energy jumps .
a discontinuous change in energy is seen at @xmath209 k ( @xmath210 k ) for @xmath211 respectively . for @xmath212
only a change in slope can be seen at @xmath213k ( @xmath214 k ) instead of a discontinuous jump . ] at @xmath215 g for @xmath133 . a jump in energy
can be seen at @xmath216 k ( @xmath217 k ) for @xmath211 respectively .
again no discontinuity is observed for @xmath212 , but only a change in slope at @xmath218 k ( @xmath219 k ) . ] for the pure system ( @xmath220 ) we see a sharp jump in the energy at exactly the point where the translational as well as hexatic structure factors sharply decline .
this jump in energy is a clear signature of a first order transition . from the energy jump
we can compute the jump in entropy ( @xmath157 ) .
the jump in energy at the transition is around @xmath221 k per vortex per layer , which gives @xmath222 which is small compared to the experimental value of @xmath223 at @xmath224 g. however , these values for @xmath157 , @xmath225 and @xmath49 are in qualitative agreement with the same system studied using a different model @xcite .
introduction of the columnar defects of a finite strength shows some interesting effects .
we put columnar pins at random positions with a concentration fixed at 20 percent of the fls .
we study the bscco system for pins of low strength ( @xmath226 ) as well as for a higher strength ( @xmath141 ) .
columnar disorder of strength up to @xmath226 appear to have little effect on the system .
this can be seen from the energy vs. temperature graph in figs .
[ 36energy ] and [ 64energy ] , where the curve for the pure system and the one in the presence of low disorder strength ( @xmath207 ) fall on top of each other , except in the transition region where the jump in case of the system with columnar pins is more gradual . also , from figs .
[ 36stru ] and [ 64stru ] we see that the @xmath172 is not much affected in the presence of the columnar disorder of strength @xmath207 . the same is true for the hexatic order depicted in fig .
[ 64hexa ] .
this means that for the lower strengths of the disorder , the vortex - vortex interaction is dominant in the range of the temperature where we carried out simulations . as a result
the first order vortex - lattice melting transition ( fot ) line is almost not affected .
g for @xmath132 .
no shift in transition point is seen for @xmath207 . in the presence of columnar pins with @xmath212 , the transition temperature increases from @xmath227 k to @xmath228 k. also , we can see that @xmath172 starts to rise close to @xmath229 k ] the nature of fl melting changes , however , when the strength of the disorder is increased to @xmath212 as seen in figs . [ 36stru ] and [ 64stru ] .
g for @xmath133 . for @xmath207
no significant change in the structure factor is observed . in the presence of columnar pins with @xmath212 , transition temperature increasing from @xmath230 k to @xmath231 k. also , we can see that @xmath172 starts to rise close to @xmath232 k ] we see that at the lower temperatures the order parameter is suppressed .
as we increase the temperature , the order parameter starts to rise and joins the melting curve of the pure system and then falls along it at even higher temperatures .
[ 64hexa ] indicates that the hexagonal structure factor also shows a similar behavior .
g for @xmath133 .
when compared with the previous figure , it can be seen that @xmath174 almost follows @xmath172 . ]
this clearly shows that the fls start to disengage from the pinning centers as they wiggle more . at higher temperatures , columnar pins with the strength chosen here
are not effective in pinning the vortex system .
this becomes clear from the fact that the melting curve for the pure system and the one with columnar disorder join each other at a temperature below the fot .
further interpretation and a possible explanation of this phenomenon will be discussed in more detail in the conclusions section below .
this convergence of the curves for the pure and disordered cases is also borne out in the @xmath156 vs. @xmath40 graphs in fig .
[ 36energy ] and fig .
[ 64energy ] .
initially at low temperatures there is a big difference in energies of the systems with no disorder ( @xmath233 ) and high disorder ( @xmath141 ) .
however as the temperature is increased , the two curves come closer and finally merge together in the liquid phase . the jump in energy in presence of columnar defects
can still be seen in the @xmath156 vs. @xmath40 graph for a disorder strength of @xmath226 .
this tends to suggest that the fot in a bscco system is not affected by disorders kept at as many as 20 percent of the total number of lines used in the simulations as long as we keep the strength of the disorder less or equal to @xmath226 . for @xmath141 ,
the rise of energy with temperature becomes much more gradual and we do not see any discontinuous jump in energy at any temperature . on the other hand an abrupt change in slope of the energy vs temperature graph is observed , suggesting a discontinuity in the specific heat characterizing a second order transition . in fig .
[ 64low ] and fig .
[ 64high ] , the snapshots of fls , projected on a plane , at temperatures less than the transition temperature and at a temperature bigger than the transition temperature are shown .
fls for @xmath234 .
fls have been projected onto one plane . ] and @xmath234 .
fls have been projected onto a single plane .
columnar defects are not drawn to scale .
some fls on the boundary do not seem to make loops .
that is only because virtual images of fls outside the cell are not shown . ] from fig .
[ 64low ] it is easily seen that at low temperatures most columnar defects have captured fls .
also , the fls make simple loops and have cleverly set themselves so as to make a hexagonal lattice and yet occupy as many defects as possible .
the transverse fluctuations of the trapped fls are greatly suppressed at low temperatures . at higher temperature beyond the transition point
, we see that columnar defects are not occupied any more and a lot of fls are entangled
. inspection of snapshots like fig .
[ 64low ] gives support to the assertion of sen _ et al .
_ that the bose glass consists of patches of ordered regions with only short range positional and orientational order .
this phase is different from the bragg - glass in systems with point pins which is characterized by quasi - long - range order .
figure [ 64wander ] shows the mean squared displacements of the fls in the @xmath38-direction at different temperatures for a pure system .
-direction at @xmath215 g for @xmath133 . here
@xmath88 is the distance between adjacent planes .
a large increase in line wandering occurs at @xmath235 k. ] at lower temperatures @xmath236 saturates for large @xmath38 but in the liquid state it grows linearly @xcite .
a large gap in @xmath236 for large values of @xmath38 seen at the transition temperature signals the onset of the entanglement of the fls . in the presence of the columnar disorder of the strength
@xmath141 we see from fig . [ 64wander_col ] that @xmath236 at temperatures less than the transition temperature is suppressed compared to the corresponding @xmath236 in the pure system .
-direction at @xmath215 g for @xmath133 in presence of disorder of strength @xmath212 .
the big jump in @xmath185 has moved to @xmath237 k now . ]
this result is in agreement with the findings in ref . .
also the big gap in @xmath236 occurring at the melting transition has moved towards a higher temperature signaling a shift in the position of the melting transition in the presence of columnar defects .
for lower strengths of disorder ( @xmath238 for ybco and @xmath239 for bscco ) no appreciable shift in the melting line was seen . for these strengths ,
a sharp drop in the translational and hexatic structure factors takes place at the transition .
also , the jump in energy at the transition is not affected much as compared with the case when there is no disorder present .
this suggests that for smaller concentrations of the columnar defects the transition still remains first order .
this result is in agreement with ref . where even with the introduction of the columnar disorder no shift in transition temperatures is seen as long as the concentration of columnar disorder introduced is small .
it is found that for ybco as well as bscco , the melting transition shifts towards higher values of temperature and magnetic field when random disorder is introduced provided its strength exceeds a threshold which is different for ybco and bscco .
the size of the shift , for a given concentration of defects , depends on the strength of the disorder . for ybco ,
a considerable shift in the melting line towards higher temperatures and magnetic fields was seen for @xmath226 and @xmath240 .
this shift was bigger for @xmath240 than for @xmath226 .
similarly , for bscco , a large shift in the melting line towards a higher temperature was found for @xmath141 .
these results are in tune with numerous experimental findings @xcite as well as theoretical prediction @xcite where the irreversibility line is seen to shift towards higher temperature and magnetic fields in the presence of columnar defects .
the qualitative reason for this effect is that due to the interaction with the columnar defects the transverse thermal fluctuations of the fls are reduced and thus the melting transition , as determined from the lindemann criterion , takes place at at a higher temperature . at @xmath141
the jump in energy is not discernible any more . instead , a change of slope corresponding to a specific heat discontinuity is observed .
this means that the transition is probably not of first order in the presence of columnar disorder of higher strengths , but it is rather a continuous ( second order ) transition .
the most dramatic outcome of this study for bscco is that for some values of the applied field and defect strength both the translational and hexatic structure factors start to rise at a certain temperature as the transition is approached from the lower temperature side .
this is an unusual result , that to our knowledge , has not been seen in previous simulations .
this fact can be explained as follows . at low temperatures ,
the free energy is dominated by energy effects rather than entropy considerations , and pinning effects are dominant . as a result
most columnar defects capture a fl , while the rest of the fls adjust themselves in positions such as to minimize the free energy .
however , as the temperature is increased , fls start to decouple from the defect sites , which allow the vortices pinned at interstitial positions to move themselves into a more ordered arrangement , thus increasing the structure factors compared to the situation at lower temperatures . as was mentioned before , due to its high anisotropy , fls in bscco behave more like a collection of two dimensional pancakes rather than rigid rods .
these pancakes are comparatively more difficult to get pinned all at once by a columnar defect . on the other hand ,
fls are much stiffer in ybco and it is easier for a columnar defect to capture a fl all along its length .
it can be shown ( eq .
( 9.49 ) in ref . )
that the depinning temperature for fls , @xmath241 , is inversely proportional to the anisotropy parameter @xmath23 .
thus , we expect fl depinning to occur at comparatively smaller temperatures in bscco than in ybco . since the melting temperatures are not that different between these materials at the fields we consider , the depinning process for bscco occurs further below the melting temperature than in ybco .
this allows the structure factor to increase above the depinning temperature before its ultimate decline at the melting transition .
for ybco , the depinning takes place close to the melting temperature , and it is difficult to detect any rise in the structure factors , especially for small systems , because it is masked by the decrease in order due to the stronger thermal fluctuations . in principle , the rise of the order parameters in the presence of low amount of columnar defects as the melting transition temperature is approached could be observed experimentally , if the appropriate parameters are tuned correctly . in small angle neutron scattering ( sans )
@xcite one can measure the integrated intensity over a bragg peak of wave vector @xmath242 which is proportional to the translational structure factor @xmath243 measured in the simulation .
another commonly used technique is muon spin rotation @xcite which gives information about the width of the magnetic field distribution in the sample .
this is not directly proportional to the structure factor at a given @xmath242 , but is rather given by @xmath244 ^ 2},\ ] ] where @xmath245 is the mean square deviation of vortices from their average position .
it is possible to measure this quantity in the simulations but this was not done it in the present work .
this quantity might not show the unusual rise describe above since it is dominated by @xmath246 which is very likely monotonically increasing with temperature yilding a monotonically decreasing line width .
thus in order to look for the effect observed in this paper we suggest using the sans technique to look at a bscco sample with columnar pins described by a matching field of about 25 g and an applied field of about 125 g. it is not clear to us what is the corresponding @xmath33 parameter describing the pinning strength of the experimental defects . according to blatter _
@xcite @xmath33 lies in the range 0.1 - 1 .
this work is supported by the us department of energy ( doe ) , grant no . de - fg02 - 98er45686 .
consider a rhombically shaped region with side @xmath247 and angle @xmath248 unit vectors are @xmath249 , @xmath250 , with @xmath251 . in practice we took @xmath252 but we leave the discussion here more general .
the green s function @xmath253 which describe the 2d coulomb interaction between one vortex and another including all its images , as is implied by the periodic boundary conditions is given by the solution to london s equation ( see e.g. ref . )
@xmath254 with the parameter @xmath255 setting the scale for the range of the interaction .
the solution is given by @xmath256 with @xmath257 where @xmath258 runs over all reciprocal lattice vectors spanned by@xmath259 for @xmath260 . substituting in eq .
( [ expansion ] ) we obtain @xmath261}{l^{2}\sin ^{2 } \theta /(2\pi \lambda ) ^{2}+n_{1}^{2}-2n_{1}n_{2}\cos \theta + n_{2}^{2}}.\ ] ] we are now going to carry out the summation over @xmath262 analytically .
this can be done by using the formula @xmath263 we have subtracted the constant @xmath177 from @xmath264 to ensure that @xmath265 on the contour of integration when @xmath38 has a large negative imaginary component .
the contour of integration is a square with sides parallel to the real and imaginary axes with the origin in the middle , in the limit that its size goes to infinity . in our case
@xmath266 with appropriate values of @xmath72 , @xmath267 , and @xmath95 .
this function has two simple poles , and the residues can be easily evaluated .
the final answer becomes ( relabeling @xmath268 as @xmath61 ) @xmath269 where @xmath270 this expression is simpler than the one used by nordborg and blatter since it does not have different expressions for even and odd n. it also converges faster in certain regions .
essentially the same method is used for permutation sampling as was used in ref . .
the only difference is that we use permutation space of only the neighboring lines .
this is so because even if we were able to get a permutation step , involving a large number of lines , accepted at the first stage of the algorithm , it would be very likely to get rejected at the following stages .
so we work with only 3 - 5 lines .
sufficiently long segments ( typically 5 planes ) of a number of lines were cut .
care should be taken to make sure that chosen fls are the nearest neighbors in the plane where the reconnection of fls is going to take place .
these points can be implemented easily with the concept of linked lists and pointers @xcite . also , care is to be taken that even though a fl may be far away from some other fl in the rhombically shaped unit cell , it can still permute with it through one of the images of the latter .
these few simple points are very important to implement the whole procedure correctly . just for a check , we tried with the sampling procedure given in ref . .
this gave results in good agreement with the sampling procedure given above .
w. e. lawrence and s. doniach , in _ proceedings of lt 12 , kyoto,1970 _ , edited by e. kanda ( keigaku , tokyo , 1971),p .
361 ; s. doniach , in _ high temperature superconductivity , proceedings _ , edited by k. s. bedell et al .
( addison - wesley , redwood city , 1989 ) , p. 406 . | the effect of columnar pins on the flux - lines melting transition in high - temperature superconductors is studied using path integral monte carlo simulations .
we highlight the similarities and differences in the effects of columnar disorder on the melting transition in yba@xmath0cu@xmath1o@xmath2 ( ybco ) and the highly anisotropic bi@xmath0sr@xmath0cacu@xmath0o@xmath3 ( bscco ) at magnetic fields such that the mean separation between flux - lines is smaller than the penetration length . for pure systems , a first order transition from a flux - line solid to a liquid phase is seen as the temperature is increased . when adding columnar defects to the system , the transition temperature is not affected in both materials as long as the strength of an individual columnar defect ( expressed as a flux - line defect interaction ) is less than a certain threshold for a given density of randomly distributed columnar pins .
this threshold strength is lower for ybco than for bscco . for higher strengths the transition line
is shifted for both materials towards higher temperatures , and the sharp jump in energy , characteristic of a first order transition , gives way to a smoother and gradual rise of the energy , characteristic of a second order transition .
also , when columnar defects are present , the vortex solid phase is replaced by a pinned bose glass phase and this is manifested by a marked decrease in translational order and orientational order as measured by the appropriate structure factors . for bscco , we report an unusual rise of the translational order and the hexatic order just before the melting transition .
no such rise is observed in ybco . |
for any function , @xmath4 , and fixed non - zero indeterminates @xmath5 , we introduce and define the _ generalized @xmath0-factorial function _ , or alternately the _ @xmath0-pochhammer symbol _ , denoted by @xmath6 , as the following products : @xmath7 within this article , we are interested in the combinatorial properties of the coefficients of the powers of @xmath8 in the last product expansions which we consider to be generalized forms of the _ stirling numbers of the first kind _ in this setting .
section [ subsection_intro_gensnumsdefs ] defines generalized stirling numbers of both the first and second kinds and motivates the definitions of auxiliary triangles by special classes of formal power series generating function transformations and their corresponding negative - order variants considered in the references @xcite .
we observe that the definition of provides an effective generalization of many other related factorial function variants considered in the references when @xmath9 .
the special cases of @xmath10 for some integer - valued @xmath11 and @xmath12 lead to the motivations for studying these more general factorial functions in @xcite , and form the expansions of multiple @xmath13-factorial functions , @xmath14 , studied in the triangular coefficient expansions defined by @xcite .
the _ factorial powers _ , or _
generalized factorials of @xmath15 of order @xmath16 and increment @xmath17 _ , denoted by @xmath18 or @xmath19 , studied in @xcite form particular special cases , as do the the forms of the generalized _ roman factorials _ and _ knuth factorials _ for @xmath20 defined in @xcite , and the _ @xmath21-shifted factorial functions _ considered in @xcite .
the results proved within this article , for example , provide new expansions of these special factorial functions in terms of their corresponding _
@xmath2-order @xmath1-harmonic number sequences _ ,
@xmath22 , which generalizes known expansions of stirling numbers by the ordinary _ @xmath2-order harmonic numbers _ , @xmath23 , in @xcite .
still other combinatorial sums and properties satisfied by the symbolic polynomial expansions of these special case factorial functions follow as corollaries of the new results we prove in the next sections .
we first employ the next recurrence relation to define the generalized triangle of stirling numbers of the first kind , which we denote by @xmath24 ( x)_{f(t),n}$ ] , or just by @xmath25 when the context is clear , for natural numbers @xmath26 @xcite .
@xmath27 we also define the corresponding generalized forms of the _ stirling numbers of the second kind _ , denoted by @xmath28 , so that we can consider inversion relations and combinatorial analogs to known identities for the ordinary triangles by the sum @xmath29 from which we can prove the following form of a particularly useful generating function transformation motivated in the references when @xmath30 has a taylor series expansion in integral powers of @xmath16 about zero @xcite @xcite @xcite : @xmath31 . \end{aligned}\ ] ] the negative - order cases of the infinite series transformation in are motivated in @xcite where we define modified forms of the stirling numbers of the second kind by @xmath32 which then implies that the transformed ordinary and exponential zeta - like power series enumerating generalized polylogarithm functions and the @xmath1-harmonic numbers , @xmath33 , are expanded by the following two series variants @xcite : @xmath34 we focus on the combinatorial relations and sums involving the generalized positive - order stirling numbers in the next few sections .
we are motivated by euler s original technique for solving the _ basel problem _ of summing the series , @xmath35 , and later more generally for all even - indexed integer zeta constants , @xmath36 , in closed - form by considering partial products of the sine function @xcite .
in particular , we observe that we have both an infinite product and a corresponding taylor series expansion in @xmath37 for @xmath38 given by @xmath39 then if we combine the form of the coefficients of @xmath40 in the partial product expansions at each finite @xmath41 with the known trigonometric series terms defined such that @xmath42 \sin(z ) = -\frac{1}{3!}$ ] given on each respective side of the last equation , we see inductively that @xmath43 \prod_{1 \leq j \leq n } \left(1 - \frac{z^2}{j^2 \pi^2}\right ) \qquad\longrightarrow\qquad \zeta(2 ) = \frac{\pi^2}{6}. \end{aligned}\ ] ] in our case , we wish to similarly enumerate the @xmath2-order @xmath1-harmonic numbers , @xmath33 , through the generalized product expansions defined in .
we first define a shorthand notation for another form of generalized `` _ _ @xmath1factorials _ _ '' that we will need in expanding the next products as follows : @xmath44 if we let @xmath45 denote the _ primitive @xmath46 root of unity _ for integers @xmath47 , and define the coefficient generating function , @xmath48 , by @xmath49 we can factor the partial products in to generate the @xmath2-order @xmath1-harmonic numbers in the following forms : @xmath50\left((-1)^{p+1 } \prod_{m=0}^{p-1 } \sum_{k=0}^{n+1 } \fcfii{f(t)}{n+1}{k } \zeta_p^{m(k-1 ) } w^k\right ) \\
\notag & = \frac{t^{pn(n+1 ) / 2}}{\left(n!_{f}\right)^p } [ w^{2p}]\left(\sum_{j=0}^{p-1 } \frac{(-1)^{j } w^{j}\ p}{p - j } \fcfii{f(t)}{n+1}{1}^j \widetilde{f}_n(w)^{p - j}\right )
\\ \label{eqn_fkp_partialsum_fcf2_exp_forms_v2 } \sum_{k=1}^{n } \frac{t^{k}}{f(k)^p } & = \frac{t^{n(n+1 ) / 2}}{\left(n!_{f}\right)^p } [ w^{2p}]\left((-1)^{p+1 } \prod_{m=0}^{p-1 } \sum_{k=0}^{n+1 } \fcfii{f\left(t^{1 / p}\right)}{n+1}{k } \zeta_p^{m(k-1 ) } w^k\right ) .
\end{aligned}\ ] ] for a fixed @xmath1 and any indeterminate @xmath51 , let the shorthand notation @xmath52
. then the following expansions illustrate several characteristic forms of these prescribed partial sums for the first several special cases of when @xmath53 : @xmath54 for each fixed integer @xmath55 , the particular partial sums defined by the ordinary generating function , @xmath56 , correspond to a function in @xmath16 that is fixed with respect to the lower indices for the triangular coefficients defined by .
moreover , the resulting coefficient expansions enumerating the @xmath1-harmonic numbers at each @xmath57 are isobaric in the sense that the sum of the indices over the lower index @xmath58 is @xmath59 in each individual term in these finite sums .
the _ elementary symmetric polynomials _ depending on the function @xmath1 implicit to the product - based definitions of the generalized stirling numbers of the first kind expanded through provide new forms of the known @xmath2-order harmonic number , or _
exponential bell polynomial _ , expansions of the ordinary stirling numbers of the first kind enumerated in the references @xcite .
thus , if we first define the weighted sums of the @xmath1-harmonic numbers , denoted @xmath60 , recursively according to an identity for the bell polynomials , @xmath61 , for @xmath62 as @xcite @xmath63 we can expand the generalized coefficient triangles through these weighted sums as @xmath64 this definition of the weighted @xmath1-harmonic sums for the generalized triangles in implies the special case expansions given in the next corollary .
the first few special case expansions of the coefficient identities in are stated for fixed @xmath1 , @xmath51 , and integers @xmath65 in the following forms : @xmath66 these expansions are computed explicitly using the recursive formula in for the first few cases of the lower triangle index @xmath67 .
we will return to the expansions of these coefficients in to formulate new finite sum identities providing functional relations between the @xmath2-order @xmath1-harmonic number sequences in the next section .
the next several properties give interesting expansions of these coefficients recursively over the power @xmath2 that can then be employed to remove , or at least significantly obfuscate , the current direct cancellation problem with these forms phrased by the examples in and in . for any fixed @xmath47 and @xmath65
, we have the following coefficient product identities enumerating the @xmath2-order @xmath1-harmonic numbers , @xmath33 : @xmath68 to begin with , observe the following rephrasing of the partial sums expansions from equations and as @xmath69 \widetilde{f}_n(w)^{p+1-j } \\ & = \frac{(p+1 ) ( -1)^{p } t^{n(n+1 ) / 2}}{t^{\frac{pn(n+1)}{2(p+1 ) } } n!_{f } } \fcfii{f(t^{1 / ( p+1)})}{n+1}{p+2 } \\ & \phantom{= \quad \ } + \sum_{j=0}^{p-1 } \frac{(p+1 ) ( -1)^{j } t^{n(n+1 ) / 2}}{t^{\frac{jn(n+1)}{2(p+1 ) } } ( n!_{f})^{p+1-j } ( p+1-j ) } [ w^{j } ] \left ( \frac{\widetilde{f}_n(w)}{w^2}\right)^{p+1-j}. \end{aligned}\ ] ] the coefficients involved in the partial sum forms for each sequence of @xmath33 are implicitly tied to the form of @xmath70 in the triangle definition of . given this distinction ,
let the generating function @xmath71 be defined equivalently in the more careful definition as @xmath72 .
the powers of the generating function @xmath56 from the previous equations satisfy the coefficient term expansions according to the next equation @xcite . @xmath73
\widetilde{f}_n(w)^{p - j } & : = [ w^{2p - j } ] \widetilde{f}_n(t;\ w)^{p - j } = [ w^{j } ] \left(\frac{\widetilde{f}_n(t;\ w)}{w^2}\right)^{p - j } \\ & \phantom { : } = \sum_{\substack{0 \leq i_1 , \ldots , i_{p - j } \leq j \\ i_1 + \cdots i_{p - j } = j } } \fcfii{f(t)}{n+1}{i_1 + 2 } \cdots \fcfii{f(t)}{n+1}{i_{p - j } + 2 } \end{aligned}\ ] ] then by taking the difference of the harmonic sequence terms over successive indices @xmath47 and at a fixed index of @xmath20 , the stated recurrences for these @xmath2-order sequences result . the generating function series over @xmath16 in the next proposition is related to the forms of the _ euler sums _ considered in @xcite in the context of the generalized zeta function transformations considered in @xcite briefly noted in the introduction .
we suggest the infinite sums over these identities for @xmath20 as a topic for future research exploration in the concluding remarks of section [ section_concl ] .
[ prop_fhnum_fnal_eqn_and_coeff_exps ] for any integers @xmath65 and @xmath57 , we have the following functional relations between the @xmath2-order and @xmath74-order @xmath1-harmonic numbers over @xmath16 and @xmath2 : @xmath75 first , notice that implies that we have the following weighted harmonic number sums for the @xmath2-order @xmath1-harmonic numbers : @xmath76 next , we use twice to expand the differences of the left - hand - side of the previous equation as @xmath77 the second identity is verified similarly by combining the coefficient terms as in the last equations and adding the right - hand - side differences of the @xmath74-order @xmath1-harmonic numbers to the first identity .
there are several immediate for columns of the triangle defined by and that can both be given immediately and that follow from an inductive argument .
the next identities in are given for general lower column index @xmath78 by @xmath79\left(\prod_{j=1}^{n-1 } ( w + f(j)\ t^{-j})\right ) \iverson{n \geq 1 } + \iverson{n = k = 0 } \\
\notag & = \sum_{\substack { 0 < i_1 < \cdots < i_{n - k } < n } } f(i_1 ) \cdots f(i_{n - k } ) \cdot t^{-(i_1 + \cdots + i_{n - k } ) } , \end{aligned}\ ] ] which follows immediately by considering the first products of the form @xmath80 in the context of elementary symmetric polynomials for these specific @xmath81 .
the first several special case columns in terms of the shifted upper index of @xmath82 specified in the expansions of are given in general by the next recurrence relations for all @xmath65 and any @xmath83 .
@xmath84 we begin by observing that by when @xmath85 , we have that @xmath86 which implies the first claim by induction since @xmath87 and @xmath88 . to prove the column - wise recurrence relation given in , we notice again by induction that for any functions @xmath89 and @xmath90 , the sequence , @xmath91 , defined recursively by @xmath92 has a closed - form solution given by @xmath93 thus by the second claim is true .
[ def_cvlpolyanalogs ] for @xmath94 , we suggest the next two variants of the generalized _ stirling convolution polynomials _ , denoted by @xmath95 and @xmath96 , as the right - hand - side coefficient definitions in the following equations : @xmath97 for integers @xmath94 , the analogs to the stirling convolution polynomial sequences defined by each satisfy a respective recurrence relation stated in the next equations .
@xmath98 we give a proof of the second identity since the first recurrence follows almost immediately from this result .
let @xmath94 and consider the expansion of the left - hand - side of according to definition [ def_cvlpolyanalogs ] as follows : @xmath99 for any non - negative integer @xmath8 , when @xmath100 , we see that @xmath101 , which implies the result .
the generating functions for the stirling convolution polynomials , @xmath102 , and the @xmath13-factorial polynomials , @xmath103 , from @xcite each have the comparatively `` _ _ nice _ _ '' closed - form generating functions given by @xmath104 \left(\frac{z e^{z}}{e^{z}-1}\right)^{x } & & \text { for } ( f(n ) , t ) \equiv ( n , 1 ) \\ \notag x \sigma_n^{(\alpha)}(x ) & = \gkpsi{x}{x - n}_{\alpha } \frac{(x - n-1)!}{(x-1 ) ! } = [ z^n ] e^{(1-\alpha)z } \left(\frac{\alpha z e^{\alpha z}}{e^{\alpha z}-1}\right)^{x } & & \text { for } ( f(n ) , t ) \equiv ( \alpha n + 1 - \alpha , 1 ) \\ \notag x \sigma_n^{(\alpha ; \beta)}(x ) & = \gkpsi{x}{x - n}_{(\alpha ; \beta ) } \frac{(x - n-1)!}{(x-1 ) ! } = [ z^n ] e^{\beta z } \left(\frac{\alpha z e^{\alpha z}}{e^{\alpha z}-1}\right)^{x } & & \text { for } ( f(n ) , t ) \equiv ( \alpha
n + \beta , 1 ) .
\end{aligned}\ ] ] the stirling polynomial sequence in is a special case of a more general class of _ convolution polynomial _
sequences defined by knuth in his article @xcite .
these polynomial sequences are defined by a general sequence of coefficients , @xmath105 with @xmath106 , such that the corresponding polynomials , @xmath107 , are enumerated by the power series over the original sequence as @xmath108 polynomial sequences of this form satisfy a number of interesting properties , and in particular , the identity that provides the generating function of the next equation for a variant of the original convolution polynomial sequence over @xmath16 .
this result is useful in expanding many identities for the @xmath109 case as given for the stirling polynomial case in @xcite @xcite
. @xmath110 \mathcal{s}_t(z)^{x}\ ] ] a related generalized class of polynomial sequences is considered in roman s book defining the form of sheffer polynomial sequences .
the polynomial sequences of this particular type , say with sequence terms given by @xmath107 , satisfy the form in the following generating function identity where @xmath111 and @xmath112 are prescribed power series satisfying the initial conditions from the reference @xcite : @xmath113 for example , the form of the generalized , or higher order , bernoulli polynomials ( numbers ) is a parametrized sequence whose generating function yields the form of many other special case sequences , including the stirling polynomial case defined in equation @xcite @xcite .
we expect that the convolution polynomial analogs defined above form a sequence of finite - degree polynomials in @xmath8 , for example , as in the stirling polynomial case when we have that @xmath114 where @xmath115 denotes the special triangle of _ second - order eulerian numbers _ for @xmath26 and where the binomial coefficient terms in the previous equations each have a finite - degree polynomial expansion in @xmath8 @xcite . given the relatively exponential generating functions that enumerate the polynomial sequences of the special case forms in , it seems natural to attempt to extend these relations to the generalized polynomial sequence forms defined by .
however , in this more general context we appear to have a stronger dependence of the form and ordinary generating functions of these polynomial sequences on the underlying function @xmath1 . specifically , for the form of the first sequence in , suppose that the function @xmath30 is arbitrary . based on the first several cases of these polynomials
, it appears that the generating function for the sequence can be expanded as @xmath116 f(z)^{x } \quad \text { where } \quad f(z ) : = \sum_{n=0}^{\infty } g_n(x ) z^n \\ \notag & \phantom{f_n(x ) :}
\implies g_n(x ) = \frac{\sum_{j=0}^{n-1 } f(x)^n \numpoly_n(j;\ x ) x^{n-1-j } ( 1+x)^j f(x+1)^j}{n!\ t^{nx}\ \sum_{j=0}^{2n-1 } \denompoly_n(j;\ x ) x^{2n-1-j } ( 1+x)^j f(x+1)^j}\ \iverson{n \geq 1 } + \iverson{n = 0 } \end{aligned}\ ] ] where the forms @xmath117 and @xmath118 denote polynomial sequences of finite non negative integral degree indexed over the natural numbers @xmath119 .
similarly it has been verified for the first @xmath120 of each @xmath16 and @xmath58 that the following equation holds where the terms @xmath121 involved in the series for @xmath122 are defined through the form of the last equation .
@xmath123 z^k f(z)^n =
\sum_{j=1}^{n - k } \binom{n}{j } [ z^{n - k } ] ( f(z ) - 1)^j + \iverson{n = k}\ ] ] note that the coefficients defined through these forms must also satisfy an implicit relation to the particular values of the polynomial parameter @xmath8 as formed through the last equations .
other different expansions may result for special cases of the function @xmath30 and explicit values of the parameter @xmath15 .
we have defined a generalized class of factorial product functions , @xmath6 , that generalizes the forms of many special factorial functions considered in the references .
the coefficient - wise symbolic polynomial expansions of these @xmath1-factorial function variants define generalized triangles of stirling numbers of the first kind which share many analogs to the combinatorial properties satisfied by the ordinary combinatorial triangles
. surprisingly , many inversion relations and other finite sum properties relating the ordinary stirling number triangles are not apparent by inspection of these corresponding sums in the most general cases . a study of ordinary stirling - number - like sums , inversion relations , and generating function transformations is not contained in the article .
we pose formulating these analogs in the most general coefficient cases as a topic for future combinatorial work with the generalized stirling number triangles defined in section [ subsection_intro_gensnumsdefs ] .
another new avenue to explore with these sums and the generalized @xmath1-zeta series transformations motivated in @xcite is to consider finding new identities and expressions for the euler - like sums suggested by the identity in proposition [ prop_fhnum_fnal_eqn_and_coeff_exps ] .
specifically , if we define a class of so - termed `` _ _ @xmath1-zeta _ _ '' functions , @xmath124 , we seek analogs to these infinite euler sum variants expanded through @xmath125 just as the euler sums are expressed through sums and products of the _ riemann zeta function _ , @xmath126 , in the ordinary cases from @xcite .
for example , it is well known that for real - valued @xmath127 @xmath128 and moreover , summation by parts shows us that for any real @xmath127 and any @xmath129 such that we have a convergent limiting zeta function series we have that @xmath130 which similarly implies that @xmath131 additionally , we seek other analogs to known identities for the infinite euler - like - sum variants over the weighted @xmath1-harmonic number sums of the form @xmath132 when @xmath133 , or more generally for any fixed @xmath129 , and where the right - hand - side series in the previous equation converges , say for @xmath134 . | we introduce a class of @xmath0-factorials , or @xmath0-pochhammer symbols , that includes many , if not most , well - known factorial and multiple factorial function variants as special cases .
we consider the combinatorial properties of the corresponding generalized classes of stirling numbers of the first kind which arise as the coefficients of the symbolic polynomial expansions of these @xmath1-factorial functions .
the combinatorial properties of these more general parameterized stirling number triangles we prove within the article include analogs to known expansions of the ordinary stirling numbers by @xmath2-order harmonic number sequences through the definition of a corresponding class of @xmath2-order @xmath1-harmonic numbers . we state and prove several new properties and functional equations enumerating these generalized @xmath1-harmonic number sequences with suggested corollaries of these results providing applications to generalized infinite euler - like sums expanded by their corresponding classes of infinite @xmath1-zeta and @xmath1-polylogarithm functions .
+ _ keywords : factorial ; multifactorial ; @xmath3-factorial ; pochhammer symbol ; stirling number ; generalized stirling number ; harmonic number ; @xmath1-harmonic number ; stirling polynomial . _ |
the first color - magnitude diagrams ( cmd ) obtained by baade for the dwarf spheroidal ( dsph ) companions of the milky way , and in particular for the draco system ( baade & swope 1961 ) , showed all of the features present in the cmd s of globular clusters .
this , together with the presence of rr lyrae stars ( baade & hubble 1939 ; baade & swope 1961 ) led to the interpretation that dsph galaxies are essentially pure population ii systems . but baade ( 1963 ) noted that there are a number of characteristics in the stellar populations of dsph galaxies that differentiate them from globular clusters , including extreme red horizontal branches and the distinct characteristics of the variable stars . when carbon stars were discovered in dsph galaxies ,
these differences were recognized to be due to the presence of an intermediate - age population ( cannon , niss & norgaard nielsen 1980 ; aaronson , olszewski & hodge 1983 ; mould & aaronson 1983 ) . in the past few years this intermediate - age population has been shown beautifully in the cmds of a number of dsph galaxies ( carina : mould & aaronson 1983 ; mighell 1990 ; smecker - hane , stetson & hesser 1996 ; hurley - keller , mateo & nemec 1998 ; fornax : stetson , hesser & smecker - hane 1998 ; leo i : lee et al .
1993 , l93 hereinafter ; this paper ) .
other dsph show only a dominant old stellar population in their cmds ( ursa minor : olszewski & aaronson 1985 ; martnez - delgado & aparicio 1999 ; draco : carney & seitzer 1986 ; stetson , vandenbergh & mcclure 1985 ; grillmair et al .
1998 ; sextans : mateo et al . 1991 ) .
an old stellar population , traced by a horizontal - branch ( hb ) , has been clearly observed in all the dsph galaxies satellites of the milky way , except leo i , regardless of their subsequent star formation histories ( sfh ) . in this respect , as noted by l93 , leo i is a peculiar galaxy , showing a well populated red - clump ( rc ) but no evident hb .
this suggests that the first substantial amount of star formation may have been somehow delayed in this galaxy compared with the other dsph .
leo i is also singular in that its large galactocentric radial velocity ( 177@xmath53 km @xmath6 , zaritsky et al .
1989 ) suggests that it may not be bound to the milky way , as the other dsph galaxies seem to be ( fich & tremaine 1991 ) .
byrd et al . (
1994 ) suggest that both leo i and the magellanic clouds seem to have left the neighborhood of the andromeda galaxy about 10 gyr ago .
it is interesting that the magellanic clouds also seem to have only a small fraction of old stellar population .
leo i presents an enigmatic system with unique characteristics among local group galaxies . from its morphology and from its similarity to other dsph in terms of its lack of detectable quantities of hi ( knapp , kerr & bowers 1978 , see section [ leoi_prev ] ) it would be considered a dsph galaxy .
but it also lacks a conspicuous old population and it has a much larger fraction of intermediate - age population than its dsph counterparts , and even , a non - negligible population of young ( @xmath7 1 gyr old ) stars . in this paper , we present new _ hst _
f555w ( @xmath1 ) and f814w ( @xmath2 ) observations of leo i. in section [ leoi_prev ] , the previous work on leo i is briefly reviewed . in section [ obs ] , we present the observations and data reduction . in section [ phot ]
we discuss the photometry of the galaxy , reduced independently using both allframe and dophot programs , and calibrated using the ground - based photometry of l93 . in section [ cmd ]
we present the cmd of leo i , and discuss the stellar populations and the metallicity of the galaxy . in section [ discus ]
we summarize the conclusions of this paper . in a companion paper , ( gallart et al .
1998 , paper ii ) we will quantitatively derive the sfh of leo i through the comparison of the observed cmd with a set of synthetic cmds .
leo i ( ddo 74 ) , together with leo ii , was discovered by harrington & wilson ( 1950 ) during the course of the first palomar sky survey .
the distances to these galaxies were estimated to be @xmath8 200 kpc , considerably more distant than the other dsph companions of the milky way . it has been observed in hi by knapp et al .
( 1978 ) using the nrao 91-m telescope , but not detected .
they set a limit for its hi mass of @xmath9 in the central 10(@xmath8 780 pc ) of the galaxy .
recently , bowen et al . (
1997 ) used spectra of three qso / agn to set a limit on the hi column density within 24 kpc in the halo of leo i to be @xmath10 .
they find no evidence of dense flows of gas in or out of leo i , and no evidence for tidally disrupted gas . the large distance to leo i and the proximity on the sky of the bright star regulus have made photometric studies difficult . as a consequence , the first cmds of leo i were obtained much later than for the other nearby dsphs ( fox & pritchet 1987 ; reid & mould 1991 ; demers , irwin & gambu 1994 ; l93 ) . from the earliest observations of the stellar populations of leo i there have been indications of a large quantity of intermediate - age stars .
hodge & wright ( 1978 ) observed an unusually large number of anomalous cepheids , and carbon stars were found by aaronson et al .
( 1983 ) and azzopardi , lequeux & westerlund ( 1985 , 1986 ) .
a prominent rc , indicative in a low z system of an intermediate - age stellar population , is seen both in the @xmath11 $ ] cmd of demers et al .
( 1994 ) and in the @xmath12 $ ] cmd of l93 .
the last cmd is particularly deep , reaching @xmath13 ( @xmath14 ) , and suggests the presence of a large number of intermediate age , main sequence stars .
there is no evidence for a prominent hb in any of the published cmd s .
l93 estimated the distance of leo i to be @xmath15 based on the position of the tip of the red giant branch ( rgb ) ; we will adopt this value in this paper .
they also estimated a metallicity of [ fe / h ] = [email protected] dex from the mean color of the rgb .
previous estimates of the metallicity ( aaronson & mould 1985 ; suntzeff , aaronson & olszewski 1986 ; fox & pritchet 1987 ; reid & mould 1991 ) using a number of different methods range from [ fe / h]=1.0 to 1.9 dex . with the new _ hst _ data presented in this paper , the information on the age structure from the turnoffs will help to further constrain the metallicity .
we present wfpc2 _ hst _ @xmath1 ( f555w ) and @xmath2 ( f814w ) data in one 2.6@xmath16 2.6 field in leo i obtained in march 5 , 1994 .
the wfpc2 has four internal cameras : the planetary camera ( pc ) and three wide field ( wf ) cameras .
they image onto a loral 800@xmath16800 ccd , which gives an scale of 0.046 pixel@xmath17 for the pc camera and 0.10 pixel@xmath17 for the wf cameras . at the time of the observations the camera was still operating at the higher temperature of 77.0 @xmath18c .
figure [ carta ] shows the location of the wfpc2 field superimposed on digitized sky survey image of leo i. the position of the ground - based image of l93 is also shown .
the position was chosen so that the pc field was situated in the central , more crowded part of the galaxy .
three deep exposures in both f555w ( @xmath1 ) and f814w ( @xmath2 ) filters ( 1900 sec . and
each , respectively ) were taken . to ensure that the brightest stars were not saturated , one shallow exposure in each filter ( 350 sec . in f555w and 300 sec in f814w ) was also obtained .
figure [ mosaic ] shows the @xmath1 and @xmath2 deep ( 5700 sec . and 4800 sec .
respectively ) wf chip2 images of leo i. all observations were preprocessed through the standard stsci pipeline , as described by holtzmann et al .
in addition , the treatment of the vignetted edges , bad columns and pixels , and correction of the effects of the geometric distortion produced by the wfpc2 cameras , were performed as described by silbermann et al .
photometry of the stars in leo i was measured independently using the set of daophot ii / allframe programs developed by stetson ( 1987 , 1994 ) , and also with a modified version of dophot ( schechter , mateo & saha 1993 ) .
we compare the results obtained with each of these programs below .
allframe photometry was performed in the 8 individual frames and the photometry list in each band was obtained by averaging the magnitudes of the corresponding individual frames . in summary ,
the process is as follows : a candidate star list was obtained from the median of all the images of each field using three daophot ii / allstar detection passes .
this list was fed to allframe , which was run on all eight individual frames simultaneously .
we have used the psfs obtained from the public domain _ hst _
wfpc2 observations of the globular clusters pal 4 and ngc 2419 ( hill et al .
the stars in the different frames of each band were matched and retained if they were found in at least three frames for each of @xmath1 and @xmath2 .
the magnitude of each star in each band was set to the error - weighted average of the magnitudes for each star in the different frames .
the magnitudes of the brightest stars were measured from the short exposure frames . a last match between the stars retained in each band
was made to obtain the @xmath19 photometry table .
dophot photometry was obtained with a modified version of the code to account for the _ hst _ psf ( saha et al .
dophot reductions were made on average @xmath1 and @xmath2 images combined in a manner similar to that described by saha et al .
( 1994 ) in order to remove the effects of cosmic rays .
photometry of the brightest stars was measured from the @xmath1 and @xmath2 short exposure frames .
the dophot and allframe calibrated photometries ( see section [ transjohn ] ) show a reasonably good agreement . there is a scatter of 2 - 3% for even the brightest stars in both @xmath1 and @xmath2 .
no systematic differences can be seen in the @xmath1 photometry . in the @xmath2 photometry
there is good systematic agreement among the brightest stars , but a small tendency for the dophot magnitudes to become brighter compared to the allframe magnitudes with increasing @xmath2 magnitude .
this latter effect is about 0.02 magnitudes at the level of the rc , and increases to about 0.04 - 0.05 mag by @xmath2 = 26 .
we can not decide from these data which program is ` correct ' .
however , the systematic differences are sufficiently small compared to the random scatter that our final conclusions are identical regardless of which reduction program is used . in the following
we will use the star list obtained with daophot / allframe .
our final photometry table contains a total of 31200 stars found in the four wfpc2 chips after removing stars with excessively large photometric errors compared to other stars of similar brightness .
the retained stars have @xmath20 , chi@xmath21 and @xmath22sharp@xmath23 . for our final photometry in the johnson - cousins system we will rely ultimately in the photometry obtained by l93 .
before this last step though , we transformed the profile fitting photometry using the prescription of holtzmann et al .
( 1995 ) . in this section
, we will describe both steps and discuss the differences between the _ hst_-based photometry and the ground based photometry .
the allframe photometry has been transformed to standard magnitudes in the johnson - cousins system using the prescriptions of holtzmann et al .
( 1995 ) and hill et al .
( 1998 ) as adopted for the _ hst _ @xmath24 key project data .
psf magnitudes have been transformed to instrumental magnitudes at an aperture of radius 0.5 " ( consistent with holtzmann et al .
1995 and hill et al .
1998 ) by deriving the value for the aperture correction for each frame using daogrow ( stetson 1990 ) .
the johnson - cousins magnitudes obtained in this way were compared with ground - based magnitudes for the same field obtained by l93 by matching a number of bright ( @xmath25 ) , well measured stars in the _ hst _ ( @xmath26 , @xmath27 ) and ground - based photometry ( @xmath28 , @xmath29 ) .
the zero - points between both data sets have been determined as the median of the distribution of ( @xmath30 ) and ( @xmath31 ) . in table
[ zeros ] the values for the median of ( @xmath30 ) , ( @xmath31 ) and its dispersion @xmath32 are listed for each chip ( no obvious color terms are observed , as expected , since both photometry sets have been transformed to a standard system taking into account the color terms where needed of the corresponding telescope - instrument system ) .
@xmath33 is the number of stars used to calculate the transformation .
although the value of the median zero point varies from chip to chip , it is in the sense of making the corrected @xmath1 magnitudes brighter by @xmath34 mag than @xmath26 and the corrected @xmath2 magnitudes fainter than @xmath27 by about the same amount
. therefore , the final @xmath35 colors are one tenth of a magnitude bluer in the corrected photometry . lcccc chip & filter & median & @xmath32 & n + chip 1 & f555w & -0.037 & 0.100 & 17 + chip 2 & f555w & -0.110 & 0.103 & 59 + chip 3 & f555w & -0.080 & 0.063 & 43 + chip 4 & f555w & -0.059 & 0.067 & 57 + chip 1 & f814w & 0.035 & 0.075 & 17 + chip 2 & f814w & 0.013 & 0.064 & 53 + chip 3 & f814w & 0.076 & 0.042 & 43 + chip 4 & f814w & 0.080 & 0.041 & 53 + note that the cte effect , which may be important in the case of observations made at the temperature of @xmath36 c , could contribute to the dispersion on the zero - point
nevertheless , if the differences @xmath37 , @xmath38 are plotted for different row intervals , no clear trend is seen , which indicates that the error introduced by the cte effect is not of concern in this case .
the fact that the background of our images is considerable ( about 70 @xmath39 ) can be the reason for the cte effect not being noticeable .
we adopt the l93 calibration because it was based on observations of a large number of standards from graham ( 1981 ) and landolt ( 1983 ) and because there was very good agreement between independent calibrations performed on two different observing runs and between calibrations on four nights of one of the runs . in addition , the holtzmann et al .
( 1995 ) zero points were derived for data taken with the wide field camera ccd s operating at a lower temperature compared to the present data set .
in figure [ 4cmd ] we present four @xmath12 $ ] cmds for leo i based on the four wfpc2 chips .
leo i possesses a rather steep and blue rgb , indicative of a low metallicity .
given this low metallicity , its very well - defined rc , at @xmath40 21.5 , is characteristic of an intermediate - age stellar population . the main sequence ( ms ) , reaching up to within 1 mag in brightness of the rc , unambiguously shows that a considerable number of stars with ages between @xmath8 1 gyr and 5 gyr are present in the galaxy , confirming the suggestion by l93 that the faintest stars in their photometry might be from a relatively young ( @xmath8 3 gyr ) intermediate - age population .
our cmd , extending about 2 magnitudes deeper than the l93 photometry and reaching the position expected for the turnoffs of an old population , shows that a rather broad range in ages is present in leo i. a number of yellow stars , slightly brighter and bluer than the rc , are probably evolved counterparts of the brightest stars in the ms .
finally , the lack of discontinuities in the turnoffs / subgiant region indicate a continuous star formation activity ( with possible changes of the star formation rate intensity ) during the galaxy s lifetime . we describe each of these features in more detail in section [ compaiso ] , and discuss their characteristics by comparing them with theoretical isochrones and taking into account the errors discussed in section [ photerrors ] .
we will quantitatively study the sfh of leo i in paper ii by comparing the distribution of stars in the observed cmd with a set of model cmds computed using the stellar evolutionary theory as well as a realistic simulation of the observational effects in the photometry ( see gallart et al .
1996b , c and aparicio , gallart & bertelli 1997 a , b for different applications of this method to the study of the sfh in several lg dwarf irregular galaxies ) . before proceeding with an interpretation of the features present in the cmd ,
it is important to assess the photometric errors . to investigate the total errors present in the photometry
, artificial star tests have been performed in a similar way as described in aparicio & gallart ( 1994 ) and gallart , aparicio & vlchez ( 1996a ) . for details on the tests
run for the leo i data , see paper ii . in short ,
a large number of artificial stars of known magnitudes and colors were injected into the original frames , and the photometry was redone again following exactly the same procedure used to obtain the photometry for the original frames .
the injected and recovered magnitudes of the artificial stars , together with the information of the artificial stars that have been lost , provides us with the true total errors . in figure [ errors ] ,
artificial stars representing a number of small intervals of magnitude and color have been superimposed as white spots on the observed cmd of leo i. enlarged symbols ( @xmath16 , @xmath41 , @xmath42 ) show the recovered magnitudes for the same artificial stars .
the spread in magnitude and color shows the error interval in each of the selected positions .
this information will help us in the interpretation of the different features present in the cmd ( section [ compaiso ] )
. a more quantitative description of these errors and a discussion of the characteristics of the error distribution will be presented in appendix a of paper ii .
we will adopt , here and in paper ii the distance obtained by l93 @xmath43 from the position of the tip of the rgb .
since the ground based observations of l93 cover a larger area than the _ hst _ observations presented in this paper , and therefore sample the tip of the rgb better , they are more suitable to derive the position of the tip . on another hand ,
since we derive the callibration of our photometry from theirs , we do nt expect any difference in the position of the tip in our data .
the adopted distance provides a good agreement between the position of the different features in the cmd and the corresponding theoretical position ( figures [ leoi_isopa ] and [ leoi_isoya ] ) , and its uncertainty does not affect the ( mostly qualitative ) conclusions of this paper . in figure [ leoi_isopa ] ,
isochrones of 16 gyr ( @xmath44 ) , 3 and 1 gyr ( @xmath45 ) from the padova library ( bertelli et al .
1994 ) have been superimposed upon the global cmd of leo i. in figure [ leoi_isoya ] , isochrones of the same ages and metallicities from the yale library ( demarque et al .
1996 ) are shown . in both cases
( except for the padova 1 gyr old , z=0.001 isochrone ) , only the evolution through the rgb tip has been displayed ( these are the only phases available in the yale isochrones ) . in figure [ leoi_isoclump ] , the hb
agb phase for 16 gyr ( @xmath44 ) and the full isochrones for 1 gyr , 600 and 400 myr ( @xmath45 ) from the padova library are shown .
a comparison of the yale and padova isochrones in figures [ leoi_isopa ] and [ leoi_isoya ] shows some differences between them , particularly regarding the shape of the rgb ( the rgb of the padova isochrones are in general _ steeper _ and redder , at the base of the rgb , and bluer near the tip of the rgb , than the yale isochrones for the same age and z ) and the position of the subgiant branches of age @xmath46gyr ( which is brighter in the padova isochrones ) . in spite of these differences , the general characteristics deduced for
the stellar populations of leo i do not critically depend on the set chosen .
however , based on these comparisons , we can gain some insight into current discrepancies between two sets of evolutionary models widely used , and therefore into the main uncertainties of stellar evolution theory that we will need to take into account when analyzing the observations using synthetic cmds ( paper ii ) . in the following
, we will discuss the main features of the leo i cmd using the isochrones in figures [ leoi_isopa ] to [ leoi_isoz ] .
this will allow us to reach a qualitative understanding of the stellar populations of leo i , as a starting point of the more quantitative approach presented in paper ii .
the broad range in magnitude in the ms turnoff region of leo i cmd is a clear indication of a large range in the age of the stars populating leo i. the fainter envelope of the subgiants coincides well with the position expected for a @xmath8 1015 gyr old population whereas the brightest blue stars on the main sequence ( ms ) may be as young as 1 gyr old , and possibly younger .
figure [ leoi_isoclump ] shows that the blue stars brighter than the 1 gyr isochrone are well matched by the ms turnoffs of stars a few hundred myr old .
one may argue that a number of these stars may be affected by observational errors that , as we see in figure [ errors ] , tend to make stars brighter .
they could also be unresolved binaries comprised of two blue stars .
nevertheless , it is very unlikely that the brightest blue stars are stars @xmath8 1 gyr old affected by one of these situations , since one has to take into account that : a ) a 1 gyr old binary could be only as bright as @xmath47 in the extreme case of two identical stars , and b ) none of the blue artificial stars at @xmath48 ( which are around 1 gyr old ) got shifted the necessary amount to account for the stars at @xmath49 and only about 4% of them have been shifted a maximum of 0.5 mag .
we conclude , therefore , that some star formation has likely been going on in the galaxy from 1 gyr to a few hundreds myr ago .
the presence of the bright yellow stars ( see subsection [ yellow ] below ) , also supports this conclusion . concerning the age of the older population of leo i , the present analysis of the data using isochrones alone does not allow us to be much more precise than the range given above ( 1015 gyr ) ,
although we favour the hypothesis that there may be stars older than 10 gyr in leo i. in the old age range , the isochrones are very close to one another in the cmd and therefore the age resolution is not high . in addition , at the corresponding magnitude , the observational errors are quite large .
nevertheless , the characteristics of the errors as shown in figure [ errors ] make it unlikely that the faintest stars in the turnoff region are put there due to large errors because i ) a significant migration to fainter magnitudes of the stars in the @xmath810 gyr turnoff area is not expected and , ii ) because of the approximate symmetric error distribution , errors affecting intermediate - age stars in their turnoff region are not likely to produce the well defined shape consistent with a 16 gyr isochrone ( see figure [ leoi_isopa ] ) .
finally , the fact that there are not obvious discontinuities in the turnoff / subgiant region suggests that the star formation in leo i has proceeded in a more or less continuous way , with possible changes in intensity but no big time gaps between successive bursts , through the life of the galaxy .
these possible changes will be quantified , using synthetic cmds , in paper ii .
core he - burning stars produce two different features in the cmd , the hb and the rc , depending on age and metallicity .
very old , very low metallicity stars distribute along the hb during the core he - burning stage .
the rc is produced when the core he - burners are not so old , or more metal - rich , or both , although other factors may also play a role ( see lee 1993 ) .
rc area in leo i differs from those of the other dsph galaxies in the following two important ways .
first , the lack of a conspicuous hb may indicate , given the low metallicity of the stars in the galaxy , that leo i has only a small fraction of very old stars .
there are a number of stars at @xmath50 , @xmath51 that could be stars on the hb of an old , metal poor population , but their position is also that of the post turn - off @xmath8 1 gyr old stars ( see figure [ leoi_isoclump ] ) .
the relatively large number of these stars and the discontinuity that can be appreciated between them and the rest of the stars in the herszprung - gap supports the hypothesis that hb stars may make a contribution .
this possible contribution will be quantified in paper ii .
second , the leo i rc is very densely populated and is much more extended in luminosity than the rc of single - age populations , with a width of as much as @xmath52 1 mag .
the intermediate - age lmc populous clusters with a well populated rc ( see e.g. bomans , vallenari & de boer 1995 ) have @xmath53 values about a factor of two smaller .
the rcs of the other dsph galaxies with an intermediate - age population ( fornax : stetson et al .
1998 ; carina : hurley keller , mateo & nemec 1998 ) are also much less extended in luminosity .
the leo i rc is more like that observed in the cmds of the general field of the lmc ( vallenari et al .
1996 ; zaritzky , harris & thompson 1997 ) .
a rc extended in luminosity is indicative of an extended sfh with a large intermediate age component .
the older stars in the core he burning phase lie in the lower part of the observed rc , younger rc stars are brighter ( bertelli et al .
1994 , their figure 12 ; see also caputo , castellani & deglinnocenti 1995 ) .
the brightest rc stars may be @xmath8 1 gyr old stars ( which start the core he - burning phase in non - degenerate conditions ) in their blue loop phase .
the stars scattered above the rc , ( as well as the brightest yellow stars , see subsection [ yellow ] ) , could be a few hundred myr old in the same evolutionary phase ( see figure 1 in aparicio et al .
1996 ; gallart 1998 ) .
the rc morphology depends on the fraction of stars of different ages , and will complement the quantitative information about the sfh from the distribution of sub - giant and ms stars ( paper ii ) .
there are a number of bright , yellow stars in the cmd ( at @xmath54 mag and @xmath55 mag ) .
l93 indicate that a significant fraction of these stars show signs of variability , and two of the stars in their sample were identified by hodge & wright ( 1978 ) to be anomalous cepheids ) estimated for them implies that they should be relatively young stars , or mass transfer binaries . since the young age hypothesis appeared incompatible with the idea of dsph galaxies being basically population ii systems , it was suggested that anomalous cepheids could be products of mass - transfer binary systems .
nevertheless , we know today that most dsph galaxies have a substantial amount of intermediate - age population , consistent with anomalous cepheids being relatively young stars that , according to various authors ( gingold 1976 , 1985 ; hirshfeld 1980 ; bono et al .
1997 ) , after undergoing the he - flash , would evolve towards high enough effective temperatures to cross the instability strip before ascending the agb . ] .
some of them also show signs of variability in our _ hst _ data . in figure [ leoi_isoclump ]
however , it is shown that these stars have the magnitudes and colors expected for blue loop stars of few hundred myr .
this supports our previous conclusion that the brightest stars in the ms have ages similar to these . given their position in the cmd , it is interesting to ask whether some of the variables found by hodge & wright ( 1978 ) in leo i could be classical cepheids instead of anomalous cepheids . from the bertelli et al .
( 1994 ) isochrones , we can obtain the mass and luminosity of a 500 myr blue - loop star , which would be a representative star in this position of the cmd .
such a star would have a mass , @xmath56 , and a luminosity , l@xmath8 350 l@xmath57 . from eq .
8 of chiosi et al .
( 1992 ) we calculate that the period that corresponds to a classical cepheid of this mass and metallicity is 1.2 days , which is compatible with the periods found by hodge & wright ( 1978 ) , that range between 0.8 and 2.4 days .
we suggest that some of these variable stars may be similar to the short period cepheids in the smc ( smith et al .
1992 ) , i.e. classical cepheids in the lower extreme of mass , luminosity and period . if this is confirmed , it would be of considerable interest in terms of understanding the relationship between the different types of cepheid variables
. a new wide field survey for variable stars , more accurate and extended to a fainter magnitude limit ( both to search for cepheids and rr lyrae stars ) would be of particular interest in the case of leo i. the rgb of leo i is relatively blue , characteristic of a system with low metallicity .
assuming that the stars are predominantly old , with a small dispersion in age , l93 obtained a mean metallicity [ fe / h]=2.02 @xmath5 0.10 dex and a metallicity dispersion of @xmath58 } < -1.8 $ ] dex .
this estimate was based on the color and intrinsic dispersion in color of the rgb at @xmath59 using a calibration based on the rgb colors of galactic globular clusters ( da costa & armandroff 1990 ; lee , freedman & madore 1993b ) . for a younger mean age of about 3.5 gyr
, they estimate a slightly higher metallicity of [ fe / h]=1.9 , based on the difference in color between a 15 and a 3.5 gyr old population according to the revised yale isochrones ( green et al .
other photometric measurements give a range in metallicity of [ fe / h]= 1.85 to 1.0 dex ( see l93 and references therein ) . the metallicity derived from moderate resolution spectra of two giant stars by suntzeff ( 1992 , unpublished ) is [ fe / h]@xmath60 dex .
since leo i is clearly a highly composite stellar population with a large spread in age , the contribution to the width of the rgb from such an age range may no longer be negligible compared with the dispersion in metallicity .
therefore , an independent estimate of the age range from the ms turnoffs is relevant in the determination of the range in metallicity . in the following ,
we will discuss possible limits on the metallicity dispersion of leo i through the comparison of the rgb with the isochrones shown in figures [ leoi_isopa ] through [ leoi_isoz ] . as we noted in the introduction of section [ cmd ] , there are some differences between the padova and the yale isochrones , but their positions coincide in the zone about 1 magnitude above the rc .
we will use only this position in the comparisons discussed below .
we will first check whether the whole width of the rgb can be accounted for by the dispersion in age . in subsection
[ ms ] above , we have shown that the ages of the stars in leo i range from 1015 gyr to less than 1 gyr . in figure [ leoi_isoz ]
we have superimposed padova isochrones of z=0.0004 and ages 10 , 1 and 0.5 gyr on the leo i cmd .
this shows that the full width of the rgb above the rc can be accounted for by the dispersion in age alone .
a similar result is obtained for a metallicity slightly lower or higher .
this provides a lower limit for the metallicity range , which could be negligible .
the agb of the 0.5 gyr isochrone appears to be too blue compared with the stars in the corresponding area of the cmd .
however , these agbs are expected to be poorly populated because a ) stars are short lived in this phase and b ) the fraction of stars younger than 1 gyr is small , if any .
second , we will discuss the possible range in z at different ages from a ) the position of the rgb , taking into account the fact that isochrones of the same age are redder when they are more metal rich and isochrones of the same metallicity are redder when they are older and b ) that the extension of the blue - loops depends on metallicity : \a ) for stars of a given age , the lower limit of z is given by the blue edge of the rgb area we are considering : isochrones of any age and z=0.0001 have colors in the rgb above the rc within the observed range .
therefore , by means of the present comparison only , we can not rule out the possibility that there may be stars in the galaxy with a range of ages and z as low as z=0.0001 .
the oldest stars of this metallicity would be at the blue edge of the rgb , and would be redder as they are younger .
the upper limit for the metallicity of stars of a given age is given by the red edge of the rgb : for old stars , the red edge of the observed rgb implies an upper limit of z @xmath7 0.0004 ( see figure [ leoi_isoz ] ) , since more metal rich stars would have colors redder than observed .
for intermediate - age stars up to @xmath8 3 gyr old we infer an upper limit of z=0.001 , and for ages @xmath8 3 - 1 gyr old an upper limit of z=0.004 .
\b ) we can use the position of the bright yellow stars to constrain z : the fact that there are a few stars in blueward extended blue
loops implies that their metallicity is as low as [email protected] or even lower ( figure [ leoi_isoclump ] ) , because higher metallicity stars do nt produce blueward extended blue - loops at the observed magnitude .
this does not exclude the possibility that a fraction of young stars have metallicity up to z=0.004 .
these upper limits are compatible with z slowly increasing with time from z@xmath8 0 to [email protected] , on the scale of the padova isochrones . in summary
, we conclude that the width of the leo i rgb can be accounted for the dispersion of the age of its stellar population and , therefore , the metallicity dispersion could be negligible . alternatively
, considering the variation in color of the isochrones depending on both age and metallicity , we set a maximum range of metallicity of @xmath620.0010.004 : a lower limit of z=0.0001 is valid for any age , and the upper limit varies from z=0.0004 to z=0.004 , increasing with time .
these upper limits are quite broad ; they will be better constrained , and some information on the chemical enrichment law gained , from the analysis of the cmd using synthetic cmds in paper ii .
from the new _ hst _ data and the analysis presented in this paper , we conclude the following about the stellar populations of leo i : \1 ) the broad ms turnoff / subgiant region and the wide range in luminosity of the rc show that star formation in leo i has extended from at least @xmath8 1015 gyr ago to less than 1 gyr ago .
a lack of obvious discontinuities in the ms turnoff / subgiant region suggests that star formation proceeded in a more or less continuous way in the central part of the galaxy , with possible intensity variations over time , but no big time gaps between successive bursts , through the life of the galaxy .
\2 ) a conspicuous hb is not seen in the cmd . given the low metallicity of the galaxy , this reasonably implies that the fraction of stars older than @xmath8 10 gyr is small , and indicates that the beginning of a substantial amount of star formation may have been delayed in leo i in comparison to the other dsph galaxies .
it is unclear from the analysis presented in this paper whether leo i contains any stars as old as the milky way globular clusters .
\3 ) there are a number of bright , yellow stars in the same area of the cmd where anomalous cepheids have been found in leo i. these stars also have the color and magnitude expected for the blue - loops of low metallicity , few hundred myr old stars .
we argue that some of these stars may be classical cepheids in the lower extreme of mass , luminosity and period .
4)the evidence that the stars in leo i have a range in age complicates the determination of limits to the metallicity range based on the width of the rgb . in one extreme ,
if the width of the leo i rgb is atributted to the dispersion of the age of its stellar population alone , the metallicity dispersion could be negligible .
alternatively , considering the variation in color of the isochrones depending on both age and metallicity , we set a maximum range of metallicity of @xmath620.0010.004 : a lower limit of z=0.0001 is valid for any age , and the ( broad ) upper limit varies from z=0.0004 to z=0.004 , increasing with time . in summary ,
leo i has unique characteristics among local group galaxies . due to its morphology and its lack of detectable quantities of hi , it can be classified as a dsph galaxy .
but it appears to have the youngest stellar population among them , both because it is the only dsph lacking a conspicuous old population , and because it seems to have a larger fraction of intermediate - age and young population than other dsph .
the star formation seems to have proceeded until almost the present time , without evidence of intense , distinct bursts of star formation .
important questions about leo i still remain .
an analysis of the data using synthetic cmds will give quantitative information about the strength of the star formation at different epochs .
further observations are needed to characterize the variable - star population in leo i , and in particular , to search for rr lyrae variable stars .
this will address the issue of the existence or not of a very old stellar population in leo i. it would be interesting to check for variations of the star formation across the galaxy and to determine whether the hb is also missing in the outer parts of leo i. answering these questions is important not only to understand the formation and evolution of leo i , but also in relation to general questions about the epoch of galaxy formation and the evolution of galaxies of different morphological types .
the determination of the strength of the star formation in leo i at different epochs is important to assess whether it is possible that during intervals of high star formation activity , leo i would have been as bright as the faint blue galaxies observed at intermediate redshift .
in addition , the duration of such a major event of star formation may be important in explaining the number counts of faint blue galaxies .
we want to thank allan sandage for many very useful discussions and a careful reading of the manuscript .
we thank also nancy b. silbermann , shoko sakai and rebecca bernstein for their help through the various stages of the _ hst _ data reduction .
support for this work was provided by nasa grant go-5350 - 03 - 93a from the space telescope science institute , which is operated by the association of universities for research in astronomy inc . under nasa contract nasa526555 .
c.g . also acknowledges financial support from a small research grant from nasa administered by the aas and a theodore dunham jr .
grant for research in astronomy .
thanks the carnegie observatories for their hospitality .
is supported by the ministry of education and culture of the kingdom of spain , by the university of la laguna and by the iac ( grant pb3/94 ) .
m.g.l is supported by the academic research fund of ministry of education , republic of korea , bsri-97 - 5411 .
the digitized sky surveys were produced at the space telescope science institute under u.s .
government grant nag w-2166 .
the images of these surveys are based on photographic data obtained using the oschin schmidt telescope on palomar mountain and the uk schmidt telescope .
smecker - hane , t. a. , stetson , p. b. , hesser , j. e. & van den bergh , d. a. 1996 in _ from stars to galaxies : the impact of stellar physics on galaxy evolution _
( ed . c. leitherer , u. fritze - van alvensleben & j. huchra ) .
asp conf series , vol 98 , p. 328 . | we present deep @xmath0 f555w ( @xmath1 ) and f814w ( @xmath2 ) observations of a central field in the local group dwarf spheroidal ( dsph ) galaxy leo i. the resulting color - magnitude diagram ( cmd ) reaches @xmath3 and reveals the oldest @xmath4 gyr old turnoffs .
nevertheless , a horizontal branch is not obvious in the cmd . given the low metallicity of the galaxy , this likely indicates that the first substantial star formation in the galaxy may have been somehow delayed in leo i in comparison with the other dsph satellites of the milky way .
the subgiant region is well and uniformly populated from the oldest turnoffs up to the 1 gyr old turnoff , indicating that star formation has proceeded in a continuous way , with possible variations in intensity but no big gaps between successive bursts , over the galaxy s lifetime .
the structure of the red - clump of core he - burning stars is consistent with the large amount of intermediate age population inferred from the main sequence and the subgiant region . in spite of the lack of gas in leo i
, the cmd clearly shows star formation continuing until 1 gyr ago and possibly until a few hundred myrs ago in the central part of the galaxy .
subject headings : galaxies : individual ( leo i ) ; galaxies : evolution ; galaxies : stellar content ; galaxies : photometry ; stars : hertzsprung - russell ( hr - diagram ) . |
experiments at the relativistic heavy ion collider ( rhic ) and the large hadron collider ( lhc ) have both been able to produce and detect an exotic phase of matter , the quark - gluon plasma ( qgp ) , by colliding massive nuclei at relativistic energies .
the energy densities involved in these collisions are so intense that the hadrons participating in the collisions dissolve into their constituent partons , creating a state of matter in which the color charge is no longer confined . despite the short lifetime of this matter ( on the order of @xmath5 fm / c ) , there is strong evidence that it has time to ( locally ) thermalize and can be described as a nearly perfect fluid .
one can compute the expected abundance of different particle species produced from a thermalized qgp and compare it with the ratios of detected particle species at rhic or lhc ; the results are in good agreement with one another @xcite , @xcite .
another place where hydrodynamics has provided particularly good agreement with heavy ion experiments is with the so - called `` collective flow '' @xcite @xcite .
when two nuclei collide the resulting qgp will be anisotropic in space , because of differences in density distributions inside the individual nuclei , even for central collisions . within hydrodynamics ( implying strong interactions ) ,
pressure gradients will convert this spatial anisotropy into an anisotropy in momentum space .
this effect would not be observed if the matter produced were noninteracting .
the full power spectrum resulting from heavy ion collisions is given by the fourier transform of the two particle correlation function @xcite @xmath6 here @xmath7 is the particle count and @xmath8 is the relative momentum angle between two observed particles .
the elliptic flow is controlled by the coefficient @xmath9 . from this power spectrum
, one should be able to extract information about the initial density fluctuations or other physics in the qgp , in analogy with the cosmic microwave background power spectrum .
one way to theoretically investigate this power spectrum is by studying hydrodynamic fluctuations on top of a smooth fluid flow and following how these perturbations propagate through to the particle spectra and on to the @xmath10 .
this approach was taken by staig and shuryak in a series of research papers @xcite @xcite @xcite studying linear perturbations .
one can also study the power spectrum numerically using fully nonlinear hydrodynamic codes . both of these approaches require much analytic or numerical effort to obtain the power spectrum . in order to obviate the need for this extensive effort every time one wants to compare theory and experiment ,
focus has been placed on determining how particular features of the power spectrum are affected by properties of the qgp itself , e.g. the viscosity or freezeout temperature .
particular interest has been placed on the first minimum of the power spectrum @xcite or on the hydrodynamic response of the system @xmath11 where the @xmath12 , defined by @xmath13 are the initial state eccentricities . here
@xmath14 is the initial energy density , and the integrals are over the transverse plane of the collision .
this response is often proposed to depend on hydrodynamic parameters by @xmath15 where @xmath16 is the transverse size of the collision , @xmath17 is the temperature of the collision , @xmath18 is the shear viscosity , and @xmath19 is the entropy density .
this dependence of the response has been investigated by lacey _
one of the questions we investigate in this paper is how robust these conclusions are ; that is , can information about the initial state fluctuations or about properties of the qgp actually be extracted once one has averaged over many events ? full , nonlinear hydrodynamic flow generally requires computer code in order to generate results , because the full equations are too complicated to solve analytically .
however , one can often investigate the approximate behavior of the full solutions by considering small perturbations around a known analytical flow .
if one possesses a given analytical flow , i.e. its energy - momentum tensor @xmath20 , then one can consider energy momentum tensors of the form @xmath21 where @xmath22 is small to the first order .
taking the covariant derivative of this perturbed energy - momentum tensor leads directly to the first order equations for the perturbation . in what follows we will consider perturbations to the energy density , pressure , and four velocity directly , which make up the energy momentum tensor @xmath23 where @xmath24 is the energy density
, @xmath25 is the pressure , @xmath26 is the four - velocity , @xmath27 is the metric , @xmath18 is the shear viscosity , @xmath28 is the bulk viscosity and @xmath29 @xmath30 the simplest flow used to model heavy ion collisions is the so - called bjorken flow , which possesses boost invariance in the longitudinal @xmath31 direction ( the velocity profile is such that a lorentz boost in the @xmath31 direction leaves the flow unchanged ) , rotational invariance in the transverse plane ( the @xmath32 direction ) , and translational invariance in the transverse plane .
what makes this flow so simple is that by the change of coordinates from @xmath33 to @xmath34 defined by @xmath35 the four - velocity profile reduces to @xmath36 with a metric @xmath37 in the past few years another analytical flow has been found that can be used to model heavy ion collisions .
gubser @xcite , @xcite exchanged the translational invariance in the transverse plane associated with bjorken flow for conformal invariance .
this allows the fluid to expand radially at the expense of fixing the trace of the energy - momentum tensor to be zero . having a radially - expanding analytical fluid flow opens up the possibility of investigating more realistic linear perturbations than were permitted by bjorken flow . in what follows
, we shall refer to this new flow as gubser flow .
there exists a coordinate system in which gubser flow is also particularly simple .
first , one can rescale the metric @xmath38 introducing the parameter @xmath39 , which characterizes the transverse scale of the system ( more precisely , @xmath40 does this ) , and the new coordinates @xmath41 and @xmath42 defined by @xmath43 one can complete the transformation .
the rescaled metric in the coordinates @xmath44 is given by @xmath45 and the fluid flow is simply @xmath46 .
note that because the gubser flow is conformal , @xmath47 , @xmath48 , and @xmath14 . in what follows
we first examine rapidity independent perturbations of the gubser flow .
we begin by discussing first order perturbations to the gubser flow in general , and then state the specific setup that we have examined .
we then present our two approaches to the problem .
we state our results for this analysis , and compare them with previous work along these lines . after this semi - analytic approach
, we examine similar perturbations numerically in full nonlinear hydrodynamics .
the power spectra we obtain are analyzed , specifically for the stability of the first minimum and for the dependence of the hydrodynamic response on @xmath2 .
these results are stated and the results of the two approaches are compared in a concluding section .
we begin by outlining the general framework of first order perturbations to the gubser flow .
as derived in @xcite , the background temperature in the rescaled frame is given by @xmath49 where @xmath50 , @xmath51 , @xmath52 is a hypergeometric function , and here and in everything that follows , @xmath53 .
this value was used by gubser @xcite @xcite as the value that best matched experimental results at rhic .
the parameter @xmath54 characterizes the initial temperature of the flow .
the perturbation is implemented by varying the rescaled temperature and four - velocity by a small amount : @xmath55 where @xmath56 is the covariant components of the background four - velocity and @xmath57 as detailed in @xcite we can further split the spacial variables from the timelike one by projecting the scalar @xmath58 onto the basis of spherical harmonics and the two nonzero components of the @xmath59 onto the derivatives of the spherical harmonics using the decomposition @xmath60 with @xmath61 . here , the @xmath62 are simply the coefficients in these expansions ( the ambiguity in their definition can be removed by demanding that @xmath63 for some fixed @xmath64 ) . in our analysis
, we include all @xmath65 .
we will often drop the @xmath66 subscript and simply write @xmath58 and @xmath67 when there is an equation that holds for each @xmath66 . note that we are ignoring the so - called vector modes by making this decomposition , for the most general basis for vectors on @xmath68 depending on the coordinates
@xmath69 will be of the form @xmath70 where @xmath71 is a scalar function and the divergence of @xmath72 on @xmath68 vanishes @xcite .
the equations for @xmath58 and @xmath67 can be written in the form @xmath73 where @xmath74 and the components of the @xmath75 matrix are @xmath76 this matrix equation @xmath77 can be reduced to a second order equation for @xmath58 : @xmath78 all of the information about the perturbed system follows from the solutions of .
we have used to investigate the evolution of an initial gaussian hotspot @xmath79 with the further assumption that there be no initial velocity perturbations .
@xmath80 we have done this in two separate ways ( method a and method b ) , which we will outline in the next two subsections .
the @xmath81 coordinate was varied , keeping @xmath82 fixed for the hotspot .
this fixed a value for @xmath64 by the coordinate relations , .
a standard cooper - frye isothermal freezeout was then used in order to compute the particle spectrum of pions : @xmath83 the distribution function @xmath84 was taken to be a simple boltzmann distribution .
modifications of the freezeout surface were taken into account by keeping terms in the exponent of the boltzmann distribution to the first order in @xmath85 , the modification to the freezeout time .
after obtaining the particle spectrum , we calculated two - particle correlations by multiplying two single particle distributions and averaging over the @xmath32 component of the initial perturbation @xmath86 with @xmath87 .
the two - particle distributions were then fourier decomposed in order to generate a power spectrum ; that is , the coefficients @xmath88 in the expansion @xmath6 the first method that we used to solve was by using a standard numerical ordinary differential equation solver .
the initial conditions were set for each of the @xmath66 equations by decomposing the initial hotspot into spherical harmonics .
these components were then set to be the coefficients @xmath62 in , and the initial condition for each @xmath58 was then simply @xmath89 . the second initial condition was fixed by the demand that there be no initial fluid flow .
this is the approach followed in work by staig and shuryak @xcite .
there is a completely different way to approach the initial value problem described above .
consider a flow in which a perturbation `` turns on '' at an initial time @xmath64 .
@xmath90 here @xmath91 is the total energy - momentum tensor for the system .
this leads to hydrodynamic equations of the form @xmath92 we can thus view an initial value problem for the hydrodynamic equations as a problem of a _ sourced _ fluid flow for @xmath58 . with this application in mind
, we shall now outline how we construct the relevant green s functions for this problem .
observe that the @xmath93s in can be viewed as functions of @xmath94 only .
this means that we can rewrite the equation in terms of the new variable @xmath95 .
if we further expand to first order in @xmath96 , we arrive at @xmath97 = 0 , \label{eq : first order}\end{aligned}\ ] ] now observe that for @xmath98 , we have @xmath99 .
this suggests that perhaps the ansatz @xmath100 with @xmath101 the @xmath1th legendre polynomial and @xmath102 the @xmath1th unknown coefficient , will be useful .
if we substitute into and then apply the integral operator @xmath103 to both sides of the equation , we arrive at an algebraic relation between the unknown coefficients @xmath104 .
unfortunately , in the case @xmath105 these relations involve an infinite number of the @xmath104 . if , however , we limit our ansatz to a finite expansion @xmath106 we arrive at a set of linear relations between the coefficients @xmath107 \cap { \mathbb{n}}\}$ ] . in our analysis , we take @xmath108 , when solving the linear equations , but we only use the first @xmath109 terms when calculating results .
we use a larger @xmath7 when solving the linear equations than when generating results because truncating the series introduces errors , which mainly affect the larger @xmath1 coefficients .
if we regard this set as a column vector @xmath110 and the set of matrix coefficients in the resulting linear equations @xmath111 , then we have reduced our original differential equation to the linear equations @xmath112 for each @xmath66 in , there should be two linearly independent solutions .
we can find them by considering the two `` initial conditions '' @xmath113 in the inviscid case , these correspond to even and odd solutions respectively .
these initial conditions can be implemented as follows . if we split the column vector @xmath110 as @xmath114 and the matrix @xmath111 as @xmath115 where @xmath116 is one of the two initial conditions , @xmath117 are the @xmath118 remaining coefficients , @xmath119 is a @xmath120 matrix and @xmath121 is a square @xmath122 matrix .
we then see that can be rewritten as @xmath123 , which allows one to solve for the unknown coefficients @xmath124 we now address the question of the source .
suppose we have a sourced hydrodynamic system with evolution equation @xmath125 how does this source @xmath126 propagate through to the equation for @xmath58 ?
recall that @xmath58 was defined by @xmath127 we can preform a similar decomposition to this source term as was done with the @xmath128 decomposition ; that is @xmath129 using this decomposition , one finds that the source term on the right side of is @xmath130 where @xmath131 using this source and the method described above , one can construct solutions of sourced hydrodynamic equations in the @xmath132 coordinate system . we have applied this method to the problem outlined above , and our results are described in the following sections .
this method is also interesting in its own right and can be used to analyze two point functions directly , as has been investigated by springer and stephanov @xcite . letting @xmath133 denote the source of the @xmath58 equation , we can write a particular @xmath66 solution in the form @xmath134 with @xmath135 ( x ' ) } \sum_{a } \sum_{b } ( l_{a } r_{b } - l_{b } r_{a } ) p_{a}(x ' ) p_{b}(x ) .
\ ] ] here , @xmath136 is a solution to the homogeneous , unsourced equation , @xmath137 is the heaviside step function , and @xmath138 $ ] is the wronskian of the solutions @xmath139 and @xmath140 that satisfy the left and right boundary conditions of interest .
they have the expansions @xmath141 respectively .
since we have constructed the actual green s functions , we can investigate the true two point functions for a dirac-@xmath58 source using this method .
we have calculated the power spectrum for a first order perturbation using method b for @xmath142 , which corresponds to an initial temperature of @xmath143 mev .
the sample inviscid plot is shown in figure [ fig : inviscid ] .
the width of the gaussian hotspot used in our analysis was @xmath144 , to make it significantly more narrow than the background flow representing the colliding nuclei .
viscous plots looked qualitatively similar and have not been included , for we have followed staig and shuryak @xcite in ignoring viscous corrections to the spectrum .
gev differential power spectrum for @xmath145 .
numerical error is smaller than the symbol size.,scaledwidth=97.0% ] in figure [ fig : inviscid ] we see that the position of the first minimum of the power spectrum is not stable to changes in the initial radial position of the perturbation @xmath81 .
increasing @xmath81 tends to shift the first minimum of the power spectrum to larger @xmath146 values .
we observe the minimum to vary by two or three harmonics , which is broadly consistent with staig and shuryak @xcite . in what follows , however , we note a number of differences between their work and ours ( see , in particular the following two subsections [ subsec : compare ss ] and [ subsec : v_n over epsilon_n ] ) .
it seems reasonable to conclude that the first minimum of the fixed @xmath147 differential power spectrum is not a stable observable , even in linear hydrodynamics .
if one were to average over many events with multiple perturbations , there would not be a well - defined location of this first minimum .
since real hydrodynamics is nonlinear , it seems doubtful that there will be a detectable first minimum to the fixed @xmath147 differential power spectrum in real heavy ion collision data .
let us briefly digress in order to perform an error analysis of our method .
there are two main sources of numerical error : a riemann sum was used in the integration process and the freezeout surface was truncated at the edge in order to not need to access times @xmath148 ( see the comparison subsection below for a discussion of this ) .
the riemann sum error was estimated by performing sums with two different resolutions ; and the error associated with the truncation of the freezeout surface was was estimated by varying where this truncation was made , calculating the changes to the power spectra , and then extrapolating these changes to the edge of the freezeout surface .
the total error was found to be @xmath149 , in the region plotted in figure [ fig : inviscid ] , with the larger errors associated to the minima of the power spectra .
we have also tried to identify potential systematic errors in our study and found one issue that needs to be taken into account .
first , recall that the gubser timelike coordinate @xmath41 is not the same as the physical proper time @xmath150 . from the coordinate transformations
, one sees that the coordinate @xmath41 depends on both of the physical coordinates @xmath150 and @xmath151 .
this means that for a given freezeout surface there is a restricted range of @xmath64 that one must choose from in order to avoid needing information about the flow _ before _ the perturbation is initialized ( when calculating the cooper - frye freezeout integral , for example ) .
figure [ fig : rho naught ] illustrates this effect . 0.45 0.45 on the left , the freezeout surface for @xmath152 mev is plotted . on the right ,
the value of the coordinate @xmath41 is show for the points on the freezeout surface .
we see that if we wish to avoid freezing out for @xmath148 at the edge of the freezeout surface , we mush either choose a @xmath64 sufficiently early , or only integrate out to a certain @xmath151 in the cooper - frye formula ; that is , ignore parts of the freezeout surface .
since we wish to fix the @xmath153 coordinate of our perturbation to be @xmath154 , this leads to the necessary choice of ignoring the very edge of the freezeout surface . in order to keep this systematic error small , we chose to cut the freezeout surface between @xmath155 and @xmath156 in order to include as much of the features of the surface as possible .
this allowed us to choose the @xmath81 value of the perturbation to be anything larger than about @xmath157 ( or , for @xmath143 mev , about @xmath158 fm ) .
this is a rather large value , but bringing it closer to the center would have necessitated ignoring significant portions of the freezeout surface , and in turn would result in a large systematic error of our procedure .
staig and shuryak @xcite have analyzed the same problem that we have outlined above as an initial value problem . in their paper
, they found that the first minimum of the fixed @xmath147 differential power spectrum did not shift more than one harmonic as they varied the @xmath81 position between @xmath159 fm and @xmath160 fm , and the overall shape of the power spectra remained consistent as the coordinate was varied . despite the broadly consistent shape of the differential power spectra between our work and theirs ,
there are some subtle differences in analysis . for this reason ,
we highlight here the few differences between our analysis and that of staig and shuryak .
staig and shuryak presumably evaluated the full cooper - frye integral with a saddle point approximation , where the integrand was evaluated at the maximum of the background exponent . in our analysis
, we found that the integrand in the cooper - frye expression was not sharply peaked for every value of @xmath161 , the momentum angular coordinate , so a full calculation of the particle spectra was necessary .
we evaluated the integral in full , using only a riemann sum in the @xmath151 coordinate ; our numerical errors due to this sum were discussed above .
we also used a smaller value of @xmath54 than staig and shuryak .
they chose to use @xmath162 , corresponding to an initial temperature of @xmath152 mev , in order to allow more time to evolve before freezeout .
this changes the detailed shape of the power spectra , and shifts the peaks and minima somewhat , but it does not change our qualitative conclusion that the first minimum is not stable under variations in the initial coordinate @xmath81 of the hotspot . finally , there is no mention in staig and shuryak s work of the systematic error introduced related to the cutting of the freeze - out surface ( see our discussion above ) .
note , however , that they used values of @xmath81 that necessitate cutting a large portion of the freezeout surface to avoid freezing out for @xmath148 . in this section
we study the linear hydrodynamic response to the perturbations by calculating the ratio @xmath11 where @xmath12 are the initial eccentricities . again , for the initial eccentricities , we have used the normalization @xmath163 where here @xmath14 is the initial energy density , and the integrals are over the transverse plane of the collision .
the hydrodynamic response of the system is frequently characterized by the ratio @xmath11 .
we have calculated this ratio ; in particular we have calculated it for integrated @xmath10 over the @xmath147 range @xmath164 gev .
this was performed with the weighting @xmath165 where the integrals are over @xmath164 gev .
our results for the inviscid case are plotted in figure [ fig : inviscid response ] .
the temperature of the fluid and width of the perturbation are identical to what was used in the corresponding plots above .
we note here that the @xmath12 are ill - defined for the gubser flow , for @xmath166 , the background energy density , exhibits a power law decay . from the background temperature profile , we find the leading order behavior @xmath167 since the background energy density is independent of the azimuthal angle @xmath32 and the perturbation is a gaussian , we have that the numerators of the expression for @xmath12 are finite for every @xmath146 . on the other hand ,
the denominators diverge for large enough @xmath146 , for they are given by @xmath168 if @xmath169 . in order to remedy this ,
we have imposed a cutoff on the volume by only integrating to a finite @xmath170 .
this cutoff , however , artificially causes the response to scale with @xmath170 as @xmath171 for @xmath169 .
this consideration must be held in mind when examining the hydrodynamical response of the gubser flow . to aid with this
, we have shaded the cutoff - dependent region in figure [ fig : inviscid response ] .
using the gubser flow to model the energy density profile of the quark gluon plasma after a heavy ion collision was useful because it provided an analytical flow on which to do linear perturbations .
however , it was not without its drawbacks .
the gubser flow was conformal , whereas the flow in a real heavy ion collision is not , and there were limitations on where the original perturbation could be placed .
the latter difficulty necessitated discarding small pieces of the freezeout surface in the cooper - frye integral .
finally , real viscous hydrodynamics is inherently nonlinear , whereas the gubser analysis above only involved linear perturbations . with these shortcomings in mind ,
we have conducted analysis of the @xmath10 coefficients using numerical hydrodynamics .
the specific code used was the causal viscous hydro code for non - central heavy ion collisions version 0.5.2 by luzum and romatschke @xcite . on top of the central collision background energy density used in the code
, a gaussian of width @xmath172 fm and height @xmath173 was added , again , so that the perturbation would be localized relative to the background energy density .
figure [ fig : hydrocoderesponse ] shows the integrated response over the @xmath147 range @xmath174 gev of a system with @xmath175 as the initial radial position of the perturbation is varied between @xmath176 fm .
the curves representing perturbations with smaller @xmath81 are truncated at smaller @xmath146 .
this is because smaller @xmath81 leads to smaller @xmath10 and @xmath12 , which the code could not accurately calculate with feasible lattice spacings .
we did , however , run codes for multiple lattice spacings in order to get a sense of the errors .
we have only plotted those points that have errors smaller than the symbol size on the plots . from figure
[ fig : hydrocoderesponse ] there does not appear to be a well - defined minimum until at least @xmath177 , except perhaps for the perturbation farthest from the symmetry axis .
after averaging over many events with many initial perturbation locations , one would not expect to find a minimum for @xmath178 , which makes it doubtful that experiments would be able to detect one . with this observation in mind
, we turn to examining the proposed form for the response @xmath179 to see whether numerical hydrodynamics supports this simple dependence on @xmath2 .
figure [ fig : varyetaos ] shows plots of @xmath11 for @xmath180 for perturbations centered at @xmath181 fm from the beamline . as expected , we see that @xmath9 decreases as @xmath2 increases .
we note here , however , that there is not a noticeable change in slope as @xmath2 is varied , already placing strain on the proposed linear dependence . for each value of @xmath2
we have fit @xmath182 to the form @xmath183 the proposed form of the response .
since the coefficients @xmath184 are proposed to depend linearly on @xmath2 , we show in figure [ fig : fit etaos ] a linear fit to these ( @xmath2 , @xmath184 ) values for the @xmath185 fm and the @xmath186 fm cases . for each value of viscosity over entropy density ,
we plot the @xmath184 value for the finest lattice spacing with error bars equal in size to the difference between the @xmath184 values for the finest and second finest lattice spacings ( most of which are too small to be seen on the plot ) .
the @xmath187 form was found to fit the logarithm of the response for each viscosity over entropy ratio well , but we see from figure [ fig : fit etaos ] that as the position of the perturbation is varied , the slope of the best fit line changes dramatically .
it would thus seem from this work that after many events are averaged over one would no longer expect to see this dependence of the response on @xmath146 and @xmath2 .
we have used semi - analytical and numerical methods to investigate the hydrodynamical response of a heavy ion collision system with initial state perturbations .
semi - analytically , using the gubser flow , we found that the first minimum of the response was sensitive to the location of the initial - state perturbation .
moreover , this minimum was located in a region where the initial state eccentricities were ill - defined .
numerically , we found that there did not appear to be a well - defined first minimum in the region @xmath178 . taking these results together ,
we conclude that in the region @xmath178 , which includes the region accessible by experiment , there should not occur a sharp minimum of the response .
furthermore , our investigations do not support the dependence of the measured response on @xmath146 and @xmath2 as @xmath188 for the slope in the linear fits of @xmath182 vs. @xmath2 varies with the location of the initial perturbation .
in addition , for some locations , the linear dependence is not even observed .
we thus find that many of the features thought to contain information about the medium will be washed out when averaging over many event locations , or at the very least they will be modified substantially .
we have also described a method for calculating the green s functions for the @xmath18-independent , linearized gubser flow to first order in @xmath96 .
this would allow one to examine the true two - point functions of this system at central rapidity .
the method could also in principle be extended to remove the central rapidity restriction .
this work was supported by the sloan foundation , award no .
br2012 - 038 , and the doe , award no . de - sc0008132 .
we thank pilar staig and edward shuryak for an illuminating correspondence and j. nagle for fruitful discussions . | the power spectrum of heavy ion collisions is investigated by studying initial state fluctuations on top of a smooth hydrodynamic flow .
in particular , the stability of the location of the first minimum of the power spectrum and the dependence of the hydrodynamic response @xmath0 on @xmath1 and on @xmath2 are discussed . in our study
we develop a new green s function method for the analytic hydrodynamic flow by s. gubser and make use of a fully non - linear hydrodynamics code .
we find that there will be no well - defined first minimum of the response for @xmath3 , due to the fact that all minima in that region are found to be sensitive to the location of the initial perturbations .
also , we find that the often proposed form of the hydrodynamical response , that @xmath4 depend quadratically on @xmath1 and linearly on @xmath2 , should not hold once many events have been averaged over . |
usually in experiments the vortices destroy superfluidity at velocities far below the landau critical velocity .
this is why the superfluid hydrodynamics equations can be expanded in powers of low velocities and one safely uses the first nontrivial terms of this expansion .
nevertheless , there is a number of experiments ( see @xcite ) where the superfluid flow is investigated in small orifices .
it has been shown that in these circumstances the maximum velocity is a decreasing function of the orifice width and may reach the order of the landau critical velocity if the aperture is small enough .
this means that all thermodynamic quantities of the superfluid become nontrivial functions of the not small superfluid velocity ( _ i.e. , _ it depends not only on the usual thermodynamic coordinates such as pressure and temperature ) .
the only assumption one can make ( and we do it ) is that the fluid at rest is isotropic .
this quite general statement of the problem is used in the paper ; we find the complete set of thermodynamic inequalities in this light , _
i.e. , _ the conditions imposed on thermodynamic functions for the superfluid to remain stable .
finally we employ the landau phonon - roton model to calculate the highest velocity compatible with obtained thermodynamic inequalities and show that it can be interpreted as a critical velocity .
this thermodynamic scenario supposedly explains the superfluidity break - up in small orifices .
when deriving general superfluid hydrodynamic equations it is usually supposed @xcite that each infinitesimal volume of the liquid is ( locally ) in equilibrium and this equilibrium is stable . for the state of the liquid to be stable
, it should provide an entropy maximum ( at least local ) for an isolated system . instead of investigating the condition of the entropy maximality ,
it is convenient @xcite to use another , equivalent to the first one , condition , that is the condition of the energy minimality under constant entropy and additive integrals of motion .
thus , to examine if the state is stable or not , one must investigate the second variation of the energy . such analysis will provide sufficient conditions for the energy minimality .
total energy of the superfluid @xmath0 is an integral of the energy density @xmath1 over the entire volume @xmath2 the energy density can be obtained via a galilean transformation @xmath3 here @xmath4 is the superfluid velocity , @xmath5 is the mass density and subscript @xmath6 denotes quantities measured in the frame of reference of the superfluid component ( that is the frame where the superfluid velocity is zero ) .
namely , @xmath7 and @xmath8 are the energy density and the momentum density ( or , equally , the mass flux ) with respect to the superfluid component .
the former is a function of @xmath5 , @xmath8 , and the entropy density @xmath9 .
its differential can be written as @xmath10 where lagrange multipliers @xmath11 , @xmath12 , and @xmath13 are the temperature , the chemical potential , and the so - called relative velocity of normal and superfluid components .
the liquid is isotropic and , consequently , the velocity @xmath13 and the momentum density @xmath8 are parallel to each other , as expressed by @xmath14 this leads to a useful identity for the partial derivatives of @xmath8 with respect to @xmath13 : @xmath15 further transforming , we can rewrite it with the help of in the form @xmath16 where we denoted the total momentum density @xmath17 and the normal velocity @xmath18 . as usual
, stability implies that each `` allowed '' fluctuation increases the total energy of the system @xmath0 . allowed
are the fluctuations leaving conserved quantities unchanged .
this means that the minimality of @xmath0 must be investigated under fixed entropy and all additive integrals of motion : mass , momentum , and superfluid velocity .
while the conservation of mass and momentum is well - known , conservation of the superfluid velocity worths a special comment . really , since the superfluid flow is irrotational , the velocity @xmath19 is a gradient of a scalar : @xmath20 .
the same is true for the time derivative @xmath21 .
this formula expresses the conservation of all three components of the vector @xmath22 consider a macroscopic fluctuation of all the variables @xmath23 , @xmath24 , @xmath25 , and @xmath26 .
they are conserved and this ensures that the first variation of the total energy for a uniform system is identically zero @xmath27 the minimality criterion must be obtained as the condition of the positive definiteness of the second differential quadratic form .
the matrix of this quadratic form is a jacobian matrix @xmath28 : @xmath29 common rule states that it is positive definite if all principal minors @xmath30 in the top - left corner are positive .
we recursively test these minors : * the first positivity condition @xmath31 corresponds to the usual requirement of the heat capacity positivity .
it is shown below that @xmath32 , hence the last inequality eventually becomes @xmath33 * positivity of the next group of minors is easily verified with the following transformation @xmath34 whether the minors @xmath35 are positive is determined by the second multiplier in .
required condition is therefore equivalent to the positive definiteness of the matrix @xmath36 here we used and chosen the direction of the @xmath13 vector as the first coordinate .
this adds to our collection two more inequalities @xmath37 @xmath38 * the same transformation applied to the biggest minors gives : @xmath39 again , the minors @xmath40 correspond to nontrivial principal minors of @xmath41 .
we use the thermodynamic identity to relate the chemical potential @xmath12 and the conventional pressure @xmath42 @xmath43 this gives @xmath44 the following is an explicit representation of @xmath41 sub - matrix corresponding to a four - dimensional space @xmath45 ; as before we let the @xmath46-axis run along @xmath13 direction .
using we obtain @xmath47 + appropriate inequalities are : @xmath48 which is literally a generalized ( to a non - zero inter - component velocity @xmath49 ) positive compressibility requirement , @xmath50 and @xmath51 inequalities , , , , , and are sufficient conditions for the thermodynamic stability .
in a `` stopped - normal - component '' arrangement , the mass flux @xmath52 with respect to the normal component may become more convenient than @xmath8the mass flux relative to the superfluid one .
the obvious relation between them @xmath53 leads to the following reformulation of the inequalities : @xmath54 @xmath55 @xmath56 @xmath57 as a simple application of the derived inequalities , consider them at @xmath58 . from , , , and we get @xmath59 @xmath60 using conventional notation , last inequality reads in the limit @xmath61 @xmath62
here we provide a usage example of the stability criteria for real superfluid @xmath63 . to calculate derivatives involved in the inequalities one take refuge in the microscopic approach .
simple and clear landau phonon - roton model works pretty well in wide temperature and velocity ranges .
we use this model to calculate the contribution of these quasiparticles to the `` modified '' free energy in the frame of reference of the superfluid component : @xmath64 differential of this potential is given by @xmath65 the modified free energy is obtained from the excitation spectrum with a conventional formula @xmath66 we denoted the excitation energy @xmath67 , which is given for two branches by the expressions @xmath68 here and below , subscripts distinguish the quantities related to phonons and rotons , @xmath69 is the sound velocity , @xmath70 is the roton energy gap , @xmath71 is the effective mass , and @xmath72 is the momentum at the roton minimum , @xmath73 , @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 , @xmath79 . ]
. a small dimensionless parameter @xmath80 ensures , _
e.g. , _ that the landau critical velocity is determined by @xmath81 . when integrated , these dispersion laws give the following contributions to the free energy : @xmath82 one can obtain all thermodynamic variables by differentiating this potential .
namely @xmath83 @xmath84 inequality is the first to become invalid .
appropriate validity region is plotted in fig .
the liquid is unstable above the curve . versus temperature @xmath11 at normal pressure .
dashed line corresponds to the equation @xmath85 .
note that the condition @xmath86 holds true over entire stability domain .
the `` stability '' critical velocity @xmath87 coincides with the landau critical velocity @xmath88 at zero temperature and vanishes completely at the critical temperature @xmath89 ( the @xmath90-point ) . in the phonon - roton model
the critical temperature is @xmath91 . ] at zero temperature the critical velocity becomes the landau critical velocity @xmath88 .
it should also be noted that for systems where all quasiparticles can be described hydrodynamically ( in other words , systems lacking roton branch ) inequality at zero temperature includes @xmath92 _ i.e. , _ @xmath93 .
experimentally , the superfluidity break - up in small orifices is believed to have the following nature ( see @xcite ) . until the aperture size is too small
the critical velocity does not depend on the temperature and increases as the size decreases .
this is the very behaviour that is specific to the vortex - related critical velocity .
when the orifice width is narrow enough the vortex - related critical velocity becomes so high , that the break - up scenario and its features change .
the critical velocity does not depend on the aperture any more but decreases when the temperature increases .
this behaviour is commonly associated ( see @xcite ) with the iordanski - langer - fisher mechanism ( see @xcite ) .
nevertheless , this association lacks numerical comparison because no reliable information about the actual orifice shape is available . on the other side
experimentally observed behaviour of the critical velocity can be attributed to the suggested stability criterion .
in other words we provide an alternative explanation of experimental results based on an assumption that in narrow orifices the thermodynamic limit of @xmath87 is reached .
we should also note that our approach to the critical velocity as a stability limit is similar to that used by kramer@xcite .
actually the inequality he employed is not a thermodynamic one .
moreover , generally speaking it is wrong . but numerical results for the critical velocity he obtained using the phonon - roton model do not deviate much from those plotted in fig.[crit ] .
it is a pleasure for us to thank i.a.fomin for useful discussions .
this work was supported in parts by intas grant 01 - 686 , crdf grant rp1 - 2411-mo-02 , rfbr grant 03 - 02 - 16401 , and rf president program . | we investigate general thermodynamic stability conditions for the superfluid .
this analysis is performed in an extended space of thermodynamic variables containing ( along with the usual thermodynamic coordinates such as pressure and temperature ) superfluid velocity and momentum density .
the stability conditions lead to _ thermodynamic inequalities _ which replace the landau superfluidity criterion at finite temperatures . _
p.l.kapitza institute for physical problems _
+ _ russian academy of sciences , _ + _ kosygin str .
, 2 , 119334 , moscow , russia _ + |
this article is motivated by the following result of berestycki , et al . given in @xcite for the laplacian perturbed by a divergence - free drift in dimensions @xmath10 .
let @xmath11 be a bounded @xmath12 regular open set and let @xmath13 be a bounded @xmath7-dimensional vector field such that @xmath14 on @xmath2 in the sense of distributions ( distr . ) , i.e. @xmath15 for @xmath5 , let ( @xmath16 ) be the principal eigen - pair corresponding to the dirichlet problem for the operator @xmath17 .
theorem 0.3 of @xcite asserts that @xmath18 remains bounded as @xmath9 , if and only if the equation @xmath19 has a solution @xmath20 ( called a first integral of @xmath6 ) , such that @xmath21 and @xmath22
. the result can be interpreted intuitively in the following way : functions @xmath20 satisfying are constant along the flow of the vector field @xmath23 ( see section [ sec : ub ] ) , and the existence of ( non - trivial ) first integrals allows for flow lines that are contained in @xmath2 . on the other hand , if no such @xmath20 exist , then the flow leaves @xmath2 with speed proportional to @xmath24 . adding the laplacian @xmath25 to @xmath26 , or equivalently the brownian motion to the flow , results in a stochastic process whose trajectories gradually depart from the integral curves of @xmath6 , but the general picture is similar :
if nontrivial first integrals exist , then the trajectories may remain in @xmath2 with positive probability during a finite time interval , even as @xmath9 . in this case
we are lead to a nontrivial limiting transition mechanism between the flow lines .
the result described in the foregoing enjoys many extensions and has proved quite useful in various applications describing the influence of a fluid flow on a diffusion , see for example @xcite . in the context of a compact , connected riemannian manifold a sufficient and necessary condition for @xmath18 to remain bounded , as @xmath9 , expressed in terms of the eigenspaces of the advection operator @xmath27 , has been given in ( * ? ? ?
* theorem 1 ) .
the purpose of the present paper is to verify that a similar property of the principal eigenvalue holds when the classical laplacian is replaced by the fractional laplacian @xmath4 with @xmath28 .
we consider @xmath29 defined as the set of all the nonzero first integrals in the sobolev space @xmath30 equipped with the norm coming from the dirichlet form @xmath31 of @xmath4 ( see below ) .
the sobolev norm condition on the first integrals reflects smoothing properties of the green function of the fractional laplacian , while is related to the flow defined by @xmath6 .
the main difficulty in our development stems from roughness of general elements of @xmath32 and non - locality of @xmath4 , which prevent us from a direct application of the differential calculus in the way it has been done in @xcite .
instead , we use conditioning suggested by a paper of bogdan and dyda @xcite , approximation techniques for flows given by diperna and lions in @xcite , and the properties of the green function and heat kernel of gradient perturbations of @xmath4 obtained by bogdan , jakubowski in @xcite and chen , et al . in @xcite for @xmath3 and bounded @xmath1-regular open sets @xmath2 . these properties allow to define and study , via the classical krein - rutman theorem and compactness arguments , the principal eigen - pair @xmath33 for @xmath34 and @xmath3 .
our main result can be stated as follows .
[ main - thm ] suppose that @xmath35 is a bounded domain with @xmath1-regular boundary that is simply connected , i.e. @xmath36 - the complement of @xmath37 - is connected .
if @xmath28 , and @xmath38 is of zero divergence , then @xmath39 and the infimum is attained .
here we use the convention that @xmath40 , hence @xmath41 if and only if the zero function is the only first integral .
equality results from the following lower and upper bounds of @xmath18 , @xmath42 @xmath43 the bounds are proved in sections [ sec : lb ] and [ sec : ub ] , correspondingly . in section [ proof - coro ]
we explain that the minimum on the right hand side of is attained , and we finish the proof of the theorem . comparing our approach with the arguments used in the case of local operators , cf .
@xcite , we note that the use of the green function seems more robust whenever we lack sufficient differentiability of functions appearing in variational formulas . recall that in the present case we need to deal with @xmath30 , which limits the applicability of the arguments based on the usual differentiation rules of the classical calculus , e.g. the leibnitz formula or the chain rule .
we consider the use of the green function as one of the major features of our approach .
in addition , the non - locality of the quadratic forms forces a substantial modifications of several other arguments , e.g. those involving conditioning of nonlocal operators and quadratic forms in the proof of the upper bound in section [ sec : ub ] .
finally , we stress the fact that the dirichlet fractional laplacian on a bounded domain @xmath2 _ is not _ a fractional power of the dirichlet laplacian on @xmath2 , e.g. the eigenfunctions of these operators have a different power - type decay at the boundary , see @xcite in this connection . as a preparation for the proof , we recall in section [ sec : p ] the estimates of @xcite for the green function and transition density of @xmath44 for the dirichlet problem on @xmath2 .
these functions are defined using hunt s formula , which in principle requires the drift @xmath45 to be defined on the entire @xmath46 .
we show however , in corollary [ cor010212 ] , that they are determined by the restriction of the drift to the domain @xmath2 . in section [ sec3 ]
we prove that the corresponding green s and transition operators are compact , see lemmas [ lem : gdc1 ] and [ lem : gdc1 kb ] .
this result is used to define the principal eigen - pair of @xmath44 , via the krein - rutman theorem . in theorem [ thm011307 ] of section [ sec4 ]
we prove that the domains of @xmath4 and @xmath44 in @xmath47 coincide . in section
[ sec : pt11 ] we employ the bilinear form of @xmath44 to estimate the principal eigenvalue .
the technical assumption @xmath48 is only needed in sections [ sec : ub ] to characterize the first integrals of @xmath6 by means of the theory of flows developed by diperna and lions in @xcite for sobolev - regular vector fields .
we start with a brief description of the setting and recapitulation of some of the results of @xcite .
further details and references may be found in those papers ( see also @xcite and the references therein ) . in what follows , @xmath49 is the euclidean space of dimension @xmath50 , scalar product @xmath51 , norm @xmath52 and lebesgue measure @xmath53 .
all sets , measures and functions in @xmath49 considered throughout this paper will be borel .
we denote by @xmath54 the ball of center @xmath55 and radius @xmath56 .
we will consider nonempty , bounded open set @xmath35 , whose boundary is of class @xmath1 .
the latter means that @xmath56 exists such that for every @xmath57 there are balls @xmath58 and @xmath59 , which are tangent at @xmath60 ( the _ inner _ and _ outer _ tangent ball , respectively ) .
we will refer to such sets @xmath2 as to @xmath1 domains , without requiring connectivity . for an alternative analytic description and localization of @xmath1 domains we refer to ( * ? ? ?
* lemma 1 ) .
we note that each connected component of @xmath2 contains a ball of radius @xmath61 , and the same is true for @xmath62 .
therefore @xmath2 and @xmath62 have a finite number of components , which will play a role in a later discussion of extensions of the vector field to a neighborhood of @xmath63 .
the distance of a given @xmath55 to @xmath62 will be denoted by @xmath64 _ constants _ mean positive numbers , that do not depend on the considered arguments of the functions being compared .
accordingly , notation @xmath65 means that there is a constant @xmath66 such that @xmath67 for all @xmath68 . as usual , @xmath69 and @xmath70 .
we will employ the function space @xmath47 , consisting of all square integrable real valued functions , with the usual scalar product @xmath71 generally , given @xmath72 , the norms in @xmath73 shall be denoted by @xmath74 .
customarily , @xmath75 denotes the space of smooth functions on @xmath49 with compact support in @xmath2 .
also , @xmath76 denotes the closure of @xmath75 in the norm @xmath77 , where @xmath78 in the last equality we have used plancherel theorem . the fourier transform of @xmath79 is given by @xmath80 the discussion in section [ sec : iso ] and section [ sec : kill ] is valid for any @xmath82 .
let @xmath83 the coefficient is chosen in such a way that @xmath84\nu(y)dy=|\xi|^\alpha\,,\quad \xi\in \rd\,.\ ] ] we define the _ fractional laplacian _ as the @xmath85-closure of the operator @xmath86\nu(y)dy\ , , \quad x\in \rd,\,\phi\in c^\infty_c(\rd).\ ] ] its fourier symbol is given by @xmath87 , cf .
the fractional laplacian is the generator of the semigroup of the isotropic @xmath81-stable lvy process @xmath88 on @xmath49 . here
@xmath89 and @xmath90 are the law and expectation for the process starting at @xmath91 .
these are defined on the borel @xmath92-algebra of the canonical cdlg path space @xmath93 via transition probability densities as follows .
we let @xmath94 be the canonical process , i.e. @xmath95 and define time - homogeneous transition density @xmath96 where @xmath97 according to ( [ eq : trf ] ) and the lvy - khinchine formula , @xmath98 is a probabilistic convolution semigroup of functions with the lvy measure @xmath99 , see e.g. @xcite .
from ( [ eq : dpt ] ) we have @xmath100 it is well - known ( @xcite ) that @xmath101 , hence @xmath102 let @xmath103 be the _ time of the first exit _ of the ( canonical ) process from @xmath2 . for each @xmath104 , the measure @xmath105\ ] ] is absolutely continuous with respect to the lebesgue measure on @xmath2 .
its density @xmath106 is continuous in @xmath107 and satisfies g. hunt s formula ( see @xcite , @xcite ) , @xmath108.\ ] ] in addition , the kernel is symmetric ( see @xcite for discussion and references ) : @xmath109 and defines a strongly continuous semigroup on @xmath47 , @xmath110 we shall denote @xmath111 .
the _ green function _ of @xmath4 for @xmath2 is defined as @xmath112 and the respective _ green operator _ is @xmath113 = \int_{\rd}g_d(x , y)f(y)dy.\end{aligned}\ ] ] we have ( @xcite ) , @xmath114 and @xmath115 see ( * ? ? ?
* lemma 5.3 ) , where the fractional laplacian operator appearing above is defined in the sense of distributions theory .
see formula below for another statement .
thus the generator of @xmath116 is @xmath117 with zero ( dirichlet ) exterior conditions , i.e. with the domain equal to the range of @xmath118 , @xmath119 the following estimate has been proved by kulczycki @xcite and chen and song @xcite ( see also ( * ? ? ?
* theorem 21 ) ) , @xmath120^\alpha}\label{greenestimatestj}\\ & & \approx\ ; |x - y|^{\alpha - d}\left(\frac{\delta_d(x)^{\alpha/2}\delta_d(y)^{\alpha/2}}{|x - y|^\alpha } \land 1\right)\nonumber \ , , \qquad x , y \in d\,.\nonumber\end{aligned}\ ] ] in particular , @xmath121 from it also follows that @xmath122 from ( * ? ? ?
* corollary 3.3 ) we have the following gradient estimate , @xmath123 we define @xmath124 as the subspace of @xmath125 made of those elements for which @xmath126 is given as follows ( cf @xcite ) : @xmath127 ^ 2\nu(y - x)dxdy,\label{011109b}\end{aligned}\ ] ] the above formula can be also used to define a bilinear form @xmath128 on @xmath129 by polarization .
we also have @xmath130 therefore @xmath131 we define @xmath32 as the closure of @xmath75 in the norm @xmath132 . by theorems 4.4.2 , a.2.10 and formula ( 4.3.1 ) of @xcite , we have ( cf . ) ) , @xmath133 if @xmath134 and @xmath135 belongs to the domain of the fractional laplacian , then @xmath136 throughout the remainder of the paper we always assume that @xmath137 , and @xmath138 is a bounded vector field .
for @xmath139 and @xmath140 we let @xmath141 and for each @xmath142 @xmath143 let @xmath144 it follows from ( * ? ? ?
* theorem 2 and example 2 ) that series converges uniformly on compact subsets of @xmath145 . from the results of @xcite
, we know that @xmath146 is a transition probability density function , i.e. it is non - negative and @xmath147 in addition , @xmath148 is continuous on @xmath145 , and @xmath149 where @xmath150 if @xmath151 .
in fact , this holds under much weaker , kato - type condition on @xmath6 , see ( * ? ? ?
* theorems 1 and 2 ) .
we denote by @xmath152 and @xmath153 the law and expectation on @xmath154 for the ( canonical ) markov process starting at @xmath68 and defined by the transition probability density @xmath155 , @xmath152 may also be defined by solving stochastic differential equation @xmath156 .
such equations have been studied in dimension @xmath157 in @xcite under the assumptions of boundedness and continuity of the vector field ; also for @xmath158 .
we refer the reader to ( * ? ? ?
* formula ( 13 ) ) , for a closer description of a connection to and .
we may also define the perturbation series for @xmath159 . in
what follows , objects pertaining to @xmath160 will be marked with the superscript hash ( @xmath161 ) , e.g. @xmath162 if @xmath163 on @xmath2 in the sense of distributions theory , and @xmath164 , then @xmath165
if also @xmath166 and we substitute @xmath167 for @xmath79 in , then we obtain @xmath168 the last equality extends to arbitrary @xmath169 . [ lem : duality ] if @xmath163 on @xmath49 , then we have @xmath170 and @xmath171 for all @xmath139 , @xmath172 , and @xmath173 .
let @xmath174 , @xmath172 . by
( see also ( * ? ? ?
* lemma 4 ) ) , @xmath175 from we conclude that @xmath176 . for a general @xmath177 , by we have @xmath178 where @xmath179 , @xmath180 , @xmath181,\ ] ] and @xmath182,\ ] ] with the convention that @xmath183 , cf . .
using formula @xmath184 times in space , and then integrating in time we see that @xmath185 which yields the identities stated in the lemma . a strengthening of proposition [ lem : duality ]
will be given in corollary [ cor010908 ] below . if @xmath163 on @xmath49 , then and proposition [ lem : duality ]
yield @xmath186 we recall that @xmath2 is a bounded @xmath1 domain in @xmath49 .
hunt s formula may be used to define the transition probability density of the ( first ) perturbed and ( then ) killed process @xcite .
thus , for @xmath139 , @xmath187 we let @xmath188\,.\ ] ] we have @xmath189=\int_d \tp_d(t , x , y)f(y)dy,\quad t>0,\,x\in d,\,f\in l^\infty(d),\ ] ] and @xmath190 where @xmath191 } c_t<+\infty$ ] for any @xmath192 . by ( * ?
* formula ( 40 ) ) , there exist constants @xmath193 such that @xmath194 we define the green function of @xmath2 for @xmath195 : @xmath196 clearly @xmath197 is nonnegative . by blumenthal s 0 - 1 law , @xmath198 for all @xmath139 ( and thus @xmath199 ) if @xmath200 or @xmath201 , see ( [ ptxy_comp ] ) .
it follows from and ( see ( * ? ? ?
* lemma 7 ) ) , that @xmath202 the main result of @xcite asserts that @xmath203 and @xmath204 is continuous for @xmath205 ( for estimates of @xmath206 see @xcite ) .
we consider the integral operators @xmath207 and @xmath208 in light of and , the above operators are bounded on every @xmath73 , @xmath209 .
the analogous operators @xmath210 and @xmath211 , defined for @xmath159 , turn out to be mutually adjoint on @xmath47 , as follows from corollary [ cor010908 ] below . from , and we obtain that @xmath212 is a semigroup of contractions in every @xmath73 , @xmath213 $ ] .
it is strongly continuous for @xmath209 .
let @xmath195 be the @xmath47 generator of the semigroup , with the domain @xmath214 .
we have @xmath215 it has been shown in @xcite that the following crucial recursive formula holds @xmath216 and for all @xmath217 and @xmath218 we have @xmath219 later on we shall also consider the operator @xmath220 corresponding to the vector field @xmath221 , where @xmath5 and we let @xmath222 .
clearly , if @xmath24 is fixed , then there is no loss of generality to focus on @xmath223 .
[ sec4 ] the following pointwise version of is proved in ( * ? ? ?
* lemma 12 ) , @xmath224 with @xmath225 we define @xmath226 after a series of auxiliary estimates , we will prove @xmath227 to be compact on @xmath47 .
[ lem011107 ] there exist @xmath228 such that @xmath229 using ( [ eq : gradestimgreen ] ) we obtain that @xmath230}{\delta_d(x)|x - y|},\ ] ] which we bound from above , thanks to ( [ greenestimatestj ] ) , by @xmath231\delta_d(y)^{\alpha/2}}{[\delta_d(x)\vee |x - y| \vee \delta_d(y)]^{\alpha}}\ , \nonumber\\ & \le c_1\,\delta_d(x)^{\alpha/2 - 1}|x - y|^{\alpha-1-d } \dfrac{\delta_d(y)^{\alpha/2}}{[\delta_d(x)\vee |x - y| \vee \delta_d(y)]^{\alpha-1}}\ , , \end{aligned}\ ] ] and this yields . [ lem051307 ] if @xmath232 , then there is @xmath233 such that @xmath234 we only need to examine points close to @xmath235 .
given such a point we consider the integral over its neighborhood @xmath236 .
the neighborhood can be chosen in such a way that , after a bi - lipschitz change of variables ( see ( * ? ? ?
* formula ( 75 ) ) , ( * ? ? ? * formula ( 11 ) ) ) , we can reduce our consideration to the case when @xmath237 $ ] , @xmath238 $ ] and @xmath239 for @xmath240 .
the respective integral over @xmath236 is then estimated by @xmath241 the result is valid for all bounded open _
lipschitz _ sets ( @xcite ) in all dimensions @xmath242 .
lemma [ lem051307 ] yields the following . [ cordk ] let @xmath243 and @xmath244 .
suppose also that @xmath245 denote @xmath246 then , @xmath247 constants @xmath248 depend only on @xmath249 the above result shall be used to establish that operator @xmath250 ( thus also @xmath227 ) is @xmath251 bounded via schur s test , see ( * ? ? ?
* theorem 5.2 ) . note that we have @xmath252 , for some constant @xmath253 , provided that @xmath254 .
likewise , @xmath255 , provided that @xmath256 .
summarizing , we will require the following conditions : @xmath257 [ lemary ] conditions ( [ c1])-([c3 ] ) hold , if @xmath258 and @xmath259 where @xmath260 .
we have @xmath261 , etc .
[ prop021307 ] operator @xmath227 is compact on @xmath47 .
we note that @xmath262 , where @xmath263 is given by and @xmath253 is some constant .
let @xmath264 be as in lemma [ lemary ] .
by corollary [ cordk ] and the discussion preceding lemma [ lemary ] , there are constants @xmath265 such that @xmath266 using schur s test for boundedness of integral operators on @xmath251 spaces ( see ( * ? ? ? * theorem 5.2 ) ) we obtain @xmath267 for @xmath268 , @xmath269 we let @xmath270 , @xmath271 , and @xmath272 each operator @xmath273 is hilbert - schmidt , hence compact ( ( * ? ? ?
* theorem 4 , p. 247 ) ) , due to the fact that @xmath274 we have @xmath275 , unless @xmath276 .
the latter may only happen if @xmath277 satisfy @xmath278 , with @xmath228 ( cf ) , or equivalently when @xmath279 , and then there are constants @xmath280 such that @xmath281 this estimate for @xmath282 actually holds for all @xmath283 , hence yields @xmath284 . since @xmath227 is a norm limit of compact operators
, it is compact , too . in view of proposition[prop021307 ] we may regard the gradient operator @xmath26 as a small perturbation of @xmath4 when @xmath3 .
in fact @xmath26 is relatively compact with respect to @xmath4 with dirichlet conditions in the sense of ( * ? ? ?
* iv.1.3 ) .
[ thm011307 ] @xmath285 . by virtue of
, we can write @xmath286 by , @xmath118 is injective , therefore so is @xmath287 . in particular
, @xmath157 is not an eigenvalue of @xmath227 . since @xmath227 is compact , by the riesz - schauder theory ( * ? ? ?
* theorem x.5.1 , p. 283 ) , @xmath287 is invertible .
thus , @xmath288 for future reference we remark that @xmath289 is invertible , too .
[ cor011307 ] @xmath290 . by (
* lemma 10 ) , @xmath291 for any bounded function @xmath135 . invoking the argument used in proposition [ prop021307 ] , we conclude that there is a number @xmath292 independent of @xmath135 for which @xmath293 by approximation , and extend to all @xmath294 .
furthermore , if @xmath295 and @xmath296 in @xmath47 , then @xmath297 and @xmath298 on @xmath235 , hence @xmath299 ( see ( * ? ? ?
* theorem 5.37 ) ) .
therefore , @xmath300 .
the following result justifies our notation @xmath301 .
[ prop011009 ] if @xmath302 , then @xmath303 .
let @xmath304 and @xmath305 .
applying @xmath306 to @xmath307 and using , we obtain @xmath308 which , thanks to , concludes the proof .
we shall also observe the following localization principle for our perturbation problem .
[ cor010212 ] operators @xmath306 and @xmath309 do not depend on the values of @xmath6 on @xmath62 . if @xmath310 and @xmath311 are two ( bounded ) vector fields on @xmath49 equal to @xmath6 on @xmath2 , and @xmath312 and @xmath313 are the corresponding green s operators , then , by , @xmath314 which identifies @xmath315 as operators on @xmath47 , hence also as functions defined in section [ sec : deftg ] .
a similar conclusion for @xmath316 follows from the fact that @xmath317 is the generator of the semigroup .
we will identify the adjoint operator of @xmath227 on @xmath47 .
[ lemhs ] @xmath318 .
if @xmath319 , then by corollary [ cor011307 ] and , @xmath320 functions @xmath321 form a dense set in @xmath47 , which ends the proof .
when defining @xmath322 , via , we may encounter the situation when @xmath6 is given only on @xmath2 .
we may extend the field outside @xmath2 by letting e.g. @xmath323 on @xmath62 .
of course such an extension needs not satisfy @xmath324 , even though the condition may hold on @xmath2 .
however , we still have a local analogue of proposition [ lem : duality ] .
recall that the transition probability densities and green function corresponding to the vector field @xmath160 are marked with a hash ( @xmath161 ) .
[ cor010908 ] if @xmath325 on @xmath2 , then for all @xmath139 and @xmath283 , @xmath326 and @xmath327 we shall first prove , or , equivalently , that @xmath328 where @xmath329 is adjoint to @xmath330 on @xmath47 . from applied to @xmath331
we have , @xmath332 applying the operator @xmath118 from the left to both sides of the equality we obtain @xmath333 by lemma [ lemhs ] , @xmath334 taking adjoints of both sides of we also obtain , @xmath335 we have already noted in the proof of theorem [ thm011307 ] that @xmath289 is a linear automorphism of @xmath47
. therefore and give .
furthermore , let @xmath336 denote the adjoint of @xmath195 on @xmath47 .
we have @xmath337 where the last equality follows from ( * ? ? ?
* lemma xii.1.6 ) .
let @xmath338 be the semigroup adjoint to @xmath339 . by ( * ?
* corollary 4.3.7 ) , the generator of @xmath338 is @xmath340 . since @xmath341 , the semigroups are equal .
the corresponding kernels are defined pointwise , therefore they satisfy .
[ lem : gdc1 ] operator @xmath342 is hilbert - schmidt on @xmath47 for each @xmath139 .
let @xmath343 be an orthonormal base in @xmath47 . by and plancherel s identity
, we bound the hilbert - schmidt norm as follows . using and @xmath344 [ lem : gdc1 kb ] @xmath197 is compact on @xmath47 .
let @xmath268 , and @xmath345 .
the integral operator @xmath346 on @xmath47 corresponding to the kernel @xmath347 is compact .
indeed , it has a finite hilbert - schmidt norm : @xmath348 ^ 2dxdy\le n^2|d|^2<\infty.\ ] ] the norm of @xmath349 on @xmath47 may be directly estimated as follows , @xmath350 see , e.g. , theorem 3 , p. 176 of @xcite .
we let @xmath351 .
the functions @xmath352 ( and @xmath353 ) are uniformly integrable on @xmath2 by . since @xmath197 is approximated in the norm topology by compact operators , it is compact . in the special case when @xmath354 , @xmath355 equals @xmath356 , a symmetric contraction semigroup on @xmath47 , whence the green operator @xmath118 is symmetric , compact and positive definite .
the spectral theorem yields the following . [ cor011206 ] @xmath357 is symmetric and compact for every @xmath358 . by and , @xmath197 is also irreducible . krein - rutman theorem ( see @xcite ) implies that there exists a unique nonnegative @xmath359 and a number @xmath360 such that @xmath361 and @xmath362 we shall call @xmath363 the _ principal eigenpair _ corresponding to @xmath195 . from
we have @xmath364 and @xmath365 .
furthermore , @xmath366 the formula yields extra regularity of @xmath367 , as follows .
[ lem011106 ] if holds for some @xmath368 and @xmath369 , then @xmath370 and there is @xmath371 such that latexmath:[\[\label{011106 } starting from , for an arbitrary integer @xmath142 we obtain @xmath373 where @xmath374 and @xmath375 to estimate @xmath376 , we use basic properties of the bessel potentials , which can be found in ( * ? ? ?
* ch.ii.4 ) . recall that for @xmath377 the bessel potential kernel @xmath378 is the unique , extended - continuous , probability density function on @xmath49 , whose fourier transform is @xmath379 thus , @xmath380 for all @xmath381 , see ( 4.7 ) of ibid . if @xmath382 , then by ( * ? ? ?
* ( 4.2 ) ) , @xmath383 is locally comparable with @xmath384 .
by there is a constant @xmath253 such that @xmath385 hence , @xmath386 , which is bounded if @xmath387 , see ( * ? ? ? * ( 4.2 ) ) again .
considering such @xmath184 we conclude that @xmath367 is bounded .
the boundary decay of @xmath367 follows from , and .
the continuity of @xmath367 is a consequence of the continuity of @xmath204 for @xmath388 , and the uniform integrability of the kernel , which stems from .
we say that @xmath20 is a _ first integral _ of @xmath6 if @xmath389 cf . .
we write @xmath390 if @xmath391 and @xmath20 is not equal to @xmath392 a.e .
recall that for @xmath5 , the operator @xmath393 is considered with the dirichlet exterior condition on @xmath2 , i.e. it acts on @xmath394 , see theorem [ thm011307 ] .
the green operator and krein - rutman eigen - pair of @xmath44 shall be denoted by @xmath395 and @xmath33 , respectively .
we also recall that @xmath396 and @xmath397 the proof of shall be obtained by demonstration of lower and upper bounds for @xmath18 ( as @xmath222 ) .
[ prop010906 ] if @xmath398 , then @xmath399 .
_ let @xmath398 . according to proposition [ prop011009 ] , @xmath79 belongs to @xmath400 and @xmath401 .
in addition , @xmath402 . taking the scalar product of both sides of the equality against @xmath79 and using we get the result , because the second term vanishes . according to proposition [ prop010906 ] , @xmath403 suppose @xmath404 , as @xmath405 , but @xmath406 stay bounded . by and corollary [ cor011206 ] ,
the sequence @xmath407 is pre - compact in @xmath47 .
suppose that @xmath20 is a weak limit of @xmath408 in @xmath30 , thus a strong limit in @xmath47 .
we have @xmath409 , and for @xmath410 , @xmath411 dividing both sides by @xmath412 and passing to the limit , we obtain that @xmath413 thus @xmath414 .
fatou s lemma and yield @xmath415 , therefore follows . the proof of uses `` conditioning '' of truncations of @xmath416 by the principal eigenfunction inspired by @xcite and @xcite ( see below ) .
here @xmath20 is a first integral in @xmath32 .
an important part of the procedure is to prove that the truncation of @xmath416 is also a first integral in @xmath32 .
this is true if @xmath417 : suppose that @xmath418 and @xmath419 are bounded and @xmath420 .
let @xmath163 and that @xmath421 a.e .
then , a.e .
we have @xmath422 thus @xmath423 is also a first integral of @xmath6 . however , the a.e .
differentiability of @xmath424 is not guaranteed for @xmath425 .
in fact , for @xmath424 , the condition @xmath426 is understood in the sense of distributions theory , cf , and the calculations in may only serve as a motivation . to build up tools for a rigorous proof of the distributional version of , in section [ sec5.2.1 ] for a sobolev - regular , divergence - free vector field @xmath138 , we consider the incompressible flow @xmath427 of mappings on @xmath46 constructed by diperna and lions in ( * ? ? ?
* theorem iii.1 ) ( the integral curves of @xmath6 ) .
the flow is used in lemma [ lem012707 ] to characterize the first integrals in @xmath428 as those locally invariant under the flow .
then , we are able to conclude that the composition of a first integral in @xmath428 with a lipschitz function is also a first integral in @xmath428 , see corollary [ prop011007 ] , and , finally , use the conditioning of @xmath416 . unless stated otherwise , in this section we consider general @xmath138 such that @xmath429 and @xmath163 a.e . on @xmath46 .
according to ( * ? ? ?
* theorem iii.1 ) , there exists a unique a.e .
defined jointly borelian family of mappings @xmath430 ( flow generated by @xmath6 ) with the following properties : first , for a.e .
@xmath55 , the function @xmath431 is continuous and @xmath432 so that , in particular , @xmath433 second , for all @xmath434 and borel measurable sets @xmath435 , @xmath436 where @xmath437 is the @xmath7-dimensional lebesgue measure , and , third , for all @xmath438 , @xmath439 by and , if @xmath440 are nonnegative , then @xmath441 we let @xmath213 $ ] , @xmath442 , and define @xmath443 note that , @xmath444 , @xmath445 , defines a group of isometries on @xmath446 . if @xmath447 , then @xmath448 for a.e .
@xmath55 and all @xmath434 , because @xmath163 implies @xmath449 .
then for all @xmath434 , @xmath450 dxds.\end{aligned}\ ] ] let @xmath451 , @xmath452 and @xmath453 ( a mollifier ) .
define @xmath454 the function is an approximate solution of the transport equation @xmath455 in the following sense : if @xmath456 @xmath457 $ ] , @xmath458 and @xmath459 , then for all @xmath434 and finite @xmath460 , @xmath461 this follows from ( * ? ? ?
* theorem ii.1 ) .
we shall use @xmath462 to characterize the first integrals @xmath463 as those elements of @xmath32 , which are constant along the flow @xmath464 .
a word of explanation may be helpful .
suppose that @xmath465 , @xmath163 a.e .
, and @xmath20 is an @xmath466-integrable first integral , i.e. @xmath467 , as distributions on space - time @xmath468 , see ( * ? ? ? * ( 13 ) ) . by ( * ? ? ?
* corollary ii.1 ) , there is a unique solution to the transport equation with the initial condition @xmath469 .
the equation is understood in the sense of distributions on space - time , too . by ( * ? ? ?
* theorem iii.1 ) , the solution has the form @xmath470 .
however , since @xmath20 is a first integral , the mapping @xmath471 defines another solution .
thus , by uniqueness , @xmath472 a.e . in our case
this argument needs to be slightly modified since the first integral is defined only on @xmath2 and not on the entire @xmath49 . in particular , the identity @xmath472 is bound to hold only for small times @xmath473 .
[ lem012707 ] let @xmath474 , @xmath475 , and @xmath476
. then @xmath414 if and only for every @xmath477 there is @xmath478 such that @xmath479 suppose that @xmath480 satisfies with @xmath478 and let @xmath481 .
choose @xmath477 so that @xmath482 for all @xmath68 in the support of @xmath483 . using and , for all @xmath484 we obtain , @xmath485 applying ,
we rewrite the rightmost side of to obtain , @xmath486 as a result , @xmath487 letting @xmath488 we see that @xmath489 the limiting passage is justified by boundedness of @xmath6 , integrability of @xmath20 ( cf .
the discussion preceding ) and dominated convergence theorem . since @xmath481 is arbitrary
, we conclude that @xmath390 .
conversely , let us assume for some @xmath476 .
by sobolev embedding theorem ( * ? ?
* theorem v.1 , p. 119 ) , we have that @xmath490 , where @xmath491 .
let @xmath474 , @xmath492 on @xmath2 and @xmath493 on @xmath62 .
note that @xmath494 .
let @xmath495 and @xmath473 be such that @xmath496 by and , both @xmath497 and @xmath498 are supported in @xmath2 . using and we have @xmath499 therefore , @xmath500 since @xmath501 , the remainder @xmath502 satisfies . by hlder inequality , the right hand side of tends to @xmath392 , as @xmath503 .
this proves for @xmath484 . as an immediate consequence of lemma [ lem012707 ] we obtain the following .
[ prop011007 ] if @xmath79 is lipschitz on @xmath504 , @xmath420 and @xmath505 , then either @xmath506 or @xmath507 . by (
* theorem 1.4.2 ( v ) ) , @xmath508 .
we apply @xmath79 to , and use lemma [ lem012707 ] .
in particular , we may consider truncations of @xmath20 at the level @xmath268 , @xmath509 [ cor - x1 ] if @xmath268 and @xmath510 , then @xmath511 .
we shall first prove the upper bound under the assumptions that @xmath6 is bounded and of zero divergence on the whole of @xmath49 , and @xmath512 .
[ prop021007 ] suppose that @xmath513 , @xmath495 , @xmath514 and @xmath20 is bounded .
then , @xmath515 we denote @xmath516 . by (
* theorem 1.4.2 ( ii ) , ( iv ) ) , @xmath517 .
also , @xmath518 . considering , we observe that the right hand side of equals @xmath519 . by proposition [ prop011009 ] ,
the left hand side of is @xmath520 however , the second term on the right hand side vanishes , because @xmath521dz=0,\ ] ] and @xmath522 , for a sufficiently large @xmath523 , is a first integral by virtue of corollary [ cor - x1 ] .
thus follows from .
the following elementary identity holds for functions @xmath524 , @xmath525 ^ 2+u^2(x)\frac{v(y)-v(x)}{v(x)}+u^2(y)\frac{v(x)-v(y)}{v(y)}\\ & & = v(x)v(y)\left[\frac{u(x)}{v(x)}-\frac{u(y)}{v(y)}\right]^2\nonumber.\end{aligned}\ ] ] let @xmath526 , @xmath527 , and @xmath528 . by
, @xmath529 ^ 2\ge \left\{\frac{w^2_n(x)}{\phi_a(x)+\eps}-\frac{w^2_n(y)}{\phi_a(y)+\eps}\right\}\left[\phi_a(x)-\phi_a(y)\right].\ ] ] multiplying both sides by @xmath530 and integrating over @xmath531 , we obtain @xmath532\nu(x - y)dxdy.\ ] ] using , we see that @xmath533 letting @xmath534 , we conclude that @xmath535 . letting @xmath536 and using ( * ? ? ? * part ( iii ) of theorem 1.4.2 ) , we obtain @xmath537 .
this proves .
[ sec5.2.3 ] [ sec6 ] suppose that @xmath538 , with @xmath474 , and @xmath539 on @xmath2 . here , as usual , @xmath2 is a bounded domain with the @xmath1 class boundary . by the discussion in section [ sec : gener ] , @xmath540 has finitely many , say , @xmath541 connected components .
denote them by @xmath542 , and assume that @xmath543 belongs to the compactification of @xmath544 .
we start with the following extension result .
[ prop013110 ] there exist @xmath545 and @xmath546 such that @xmath547 @xmath548 has compact support and is smooth outside of @xmath549 . according to the results of section 4 of @xcite we can find @xmath550 , supported in @xmath551 and such that @xmath552 for @xmath553 , numbers @xmath554 , and points @xmath555 such that @xmath556 here @xmath557 .
we denote by @xmath558 the right hand side of the above equality .
the field @xmath559 is incompressible on @xmath560 .
we choose @xmath545 so that @xmath561 for @xmath562 .
let @xmath563 equal to @xmath157 on @xmath564 and vanish outside of @xmath565 .
the field @xmath566 has the desired properties .
we consider the flow @xmath567 of measurable mappings generated by @xmath548 , which satisfies , and , see theorem iii.2 of @xcite .
condition needs to be modified as follows : there exists @xmath228 , such that @xmath568 let @xmath569 be a mollifier , let @xmath570 and let @xmath571 be the flow generated by @xmath572 .
it has been shown in section 3 of @xcite that @xmath573}\int_{b(0,n)}|\phi(x(t , x))-\phi(x_\eps(t , x))|\wedge 2^ndx=0\label{eq:2.31.10}\ ] ] for all @xmath574 and measurable functions @xmath575 .
we shall need the following modification of lemma [ lem012707 ] .
[ lem012707a ] if @xmath576 then , the conclusion of lemma [ lem012707 ] holds for the flow @xmath567 generated by @xmath548 .
suppose that @xmath577 .
since @xmath578 on @xmath2 , we can repeat the proof of the respective part of lemma [ lem012707 ] .
namely , keeping the notation from that lemma , for every @xmath481 we have @xmath579 such that latexmath:[\[\label{022807a } \int_{d } w(x)u_0(x(-t , x))dx=\int_{d } w(x)u_0(x)dx,\quad @xmath579 can be so adjusted that @xmath581 this part can not be guaranteed directly from the definition of the flow , as the extended field @xmath548 needs not be divergence - free .
equality holds however , when @xmath582 is replaced by @xmath583 for a sufficiently small @xmath495 . indeed , by the liouville theorem
, the jacobian @xmath584 of @xmath585 satisfies @xmath586 since @xmath587 in an open neighborhood of @xmath37 we conclude that @xmath588 on @xmath2 for ( sufficiently small ) @xmath589 . since @xmath20 and @xmath483 are supported in @xmath2 , @xmath590 letting @xmath503 and using we obtain .
the rest of the proof follows that of lemma [ lem012707 ] . having established lemma [ lem012707a ] we proceed with the proofs of corollaries [ prop011007 ] and [ cor - x1 ] and proposition [ prop021007 ] with no alterations .
these results yield .
we consider the principal eigen - pair , say @xmath591 , of @xmath4 for the unit ball in @xmath46 . by rotation invariance of @xmath4 and uniqueness
, @xmath592 is a smooth radial function in the ball .
for @xmath553 we take radially symmetric functions @xmath593 , such that @xmath594 .
let @xmath595 , @xmath596 .
the vector field @xmath557 is of zero divergence and tangent to the spheres @xmath597 for all @xmath56 . indeed , since @xmath598 , @xmath599 as a result @xmath600 for any @xmath601 smooth radially symmetric function @xmath602 .
thus , we conclude that @xmath592 is the principal eigenfunction of @xmath44 for every @xmath24 .
we have @xmath603 , and @xmath604 attains its infimum , @xmath605 , at @xmath606 , see section [ sec : lb ] . in passing we
note that the considered limiting eigenproblems are essentially different for different values of @xmath81 , in accordance with the fact that the `` escape rate '' @xmath607 of the isotropic @xmath81-stable lvy processes from @xmath2 depends on @xmath81 .
we refer the interested reader to @xcite , @xcite , @xcite for more information on the eigenproblem of @xmath4 , see also @xcite . to complete the proof of theorem [ main - thm ] we only need to explain the attainability of infimum appearing on the right hand side of .
this is done in the following .
[ lem1 ] if @xmath608 , then @xmath609 attains its infimum on the set .
if @xmath610 is the infimum , then we can choose functions @xmath611 in the set given in the statement of the lemma , such that @xmath612 and , by choosing a subsequence , that @xmath611 weakly converge to @xmath613 .
since @xmath614 is precompact in @xmath47 , we may further assume that @xmath615 in @xmath47 and a.e .
this implies that @xmath616 , @xmath617 is a first integral and , by and fatou s lemma , that @xmath618 . from the definition of @xmath619
, we conclude that equality actually occurs .
krzysztof bogdan thanks the department of mathematics at stanford university for hospitality during his work on the paper , lenya ryzhik for discussions on incompressible flows and michael frazier for a discussion on schur s test . | we study the principal dirichlet eigenvalue of the operator @xmath0 , on a bounded @xmath1 regular domain @xmath2 . here
@xmath3 , @xmath4 is the fractional laplacian , @xmath5 , and @xmath6 is a bounded @xmath7-dimensional divergence - free vector field in the sobolev space @xmath8 .
we prove that the eigenvalue remains bounded , as @xmath9 , if and only if @xmath6 has non - trivial first integrals in the domain of the quadratic form of @xmath4 for the dirichlet condition . |
casimir forces are a striking manifestation of the zero - point energy of the electromagnetic field in the presence of ` mirrors ' endowed with quite general electromagnetic properties @xcite . in many calculations of the casimir energies and forces
, the presence of the mirrors is modeled by appropriate boundary conditions on the interfaces of the different media , that include macroscopic parameters such as their electric permitivity , magnetic permeability , conductivity , etc .
a first - principles calculation of the casimir energy should consider the microscopic degrees of freedom associated to the mirrors .
this could shed light on some interesting open questions , the role of dissipation on the casimir energy being , perhaps , the most important among them . in a previous paper @xcite
, we considered the casimir effect for scalar and gauge fields interacting with dynamical matter on thin mirrors ( see also ref.@xcite for a concrete model realization ) .
more recently , one of us considered the generalization to the case of finite - width mirrors @xcite .
the interaction between the vacuum scalar field and the mirrors degrees of freedom gives rise , in general , to a nonlocal effective action in terms of which the casimir energy may be calculated @xcite .
moreover , under certain circumstances , it is possible to find a formal expression for the casimir energy in terms of the parameters that define the nonlocal kernel @xcite . in this paper
, we will present an application of the previously developed formalism for the casimir effect with nonlocal boundary interaction terms , to situations where those nonlocal terms may be expanded in a series of local ones . in other words , we will perform a derivative expansion of the nonlocal effective action .
one should expect on physical grounds that , in many relevant cases , such a local description of the mirrors must be reliable .
we show here how one can indeed find such an expansion , and then we shall apply it to derive approximate expressions for the casimir energy .
the structure of this paper is the following .
in section [ sec : scalar ] , we derive the derivative expansion for the nonlocal effective action , which will be written in terms of a set of generalized @xmath0-potentials , i.e. terms proportional to dirac s @xmath0-function and its derivatives .
we will illustrate , in concrete examples , how the coupling between the vacuum field and the microscopic degrees of freedom , together with the boundary conditions that confine the microscopic degrees of freedom inside the mirrors , do determine the different coefficients in the derivative expansion . in section [ sec : deltapot ]
we compute the casimir energy for the resulting generalized @xmath0-potentials .
section [ sec : disc ] contains our final remarks .
let us consider a real scalar field @xmath1 in the presence of two flat mirrors of width @xmath2 centered at @xmath3 .
this scalar field interacts with the microscopic degrees of freedom inside the mirrors , which in the specific examples below will be described by a second scalar field @xmath4 . after integrating the microscopic degrees of freedom of the mirrors ,
the effective action for the scalar field @xmath1 will be of the form @xmath5 where @xmath6 and @xmath7 are concentrated on the positions of each mirror . on general grounds we expect these interaction terms to be nonlocal , i.e. , @xmath8 and a similar expression for @xmath7 . here
@xmath9 denotes the time ( @xmath10 ) as well as the @xmath11 spatial coordinates parallel to the mirror ( which we shall denote by @xmath12 ) .
we have assumed translational invariance in the coordinates @xmath9 , and therefore it is useful to write the effective action in terms of the fourier transform of the field in these coordinates , @xmath13 , with the obvious notation @xmath14 for the argument of this function .
the nonlocal kernel @xmath15 may be expanded as follows @xcite @xmath16 and the functions @xmath17 depend essentially on the nature of the boundary conditions for the microscopic fields ( i.e. , those living inside the mirrors ) while the coefficients , @xmath18 , are obtained by taking into account the ( kinematic and dynamical ) properties of those fields .
( [ eq : nonlocal ] ) results from the assumption that , after integrating the microscopic fields , the most relevant term in the effective action is quadratic in the scalar field ; in other words , we are assuming , as usual , that the media can be described by linear response theory .
the particular case in which the interaction between the thick mirrors and the vacuum field is approximated by a local effective action ( i.e. @xmath19 ) has been considered in ref.@xcite .
as we will see , the nonlocal effects can be evaluated perturbatively by expanding the kernel @xmath15 in powers of the @xmath0-function and its derivatives : @xmath20 where @xmath21 depend on the microscopic fields and their interaction with the vacuum field .
we start our derivation of the expansion with the study of a simple example , namely , the case in which the microscopic field @xmath4 is also a real scalar , endowed with a quadratic action , and linearly coupled to @xmath1 .
as already mentioned , we denote by @xmath2 the width of the mirror , which fills the region @xmath22 .
then , as shown in @xcite , the coefficients @xmath18 adopt the diagonal form : @xmath23 where @xmath24 is the mass of the microscopic field , @xmath25 ( @xmath26 ) denote the eigenvalues of @xmath27 corresponding to the eigenvectors @xmath17 , and @xmath28 is the coupling constant between @xmath1 and @xmath4 .
the precise form of those eigenvalues and eigenvectors depends of course on the boundary conditions for the microscopic field . indeed ,
for the case of dirichlet boundary conditions , we have the eigenfunctions : @xmath29 while in the neumann case , we have instead @xmath30 the eigenvalues are then , in both cases , given by the expression : , where @xmath31 in the dirichlet case , while for neumann boundary conditions : @xmath32 the essential difference is thus the existence or not of a zero mode , which is present only in the neumann case .
one should expect this difference to manifest itself when one tries to perform a local approximation for the nonlocal interaction term , under the assumption that @xmath33 .
indeed , note that , in such a case , the zero mode is multiplied by the @xmath2-independent coefficient @xmath34 , while all the other @xmath35 coefficients are relatively suppressed in such a limit . to make this statement more precise ,
let us perform an expansion of the interaction term , assuming that @xmath2 is small ( this shall be made more clear below , after introducing the other length scale to compare it with ) .
we shall assume , in what follows , neumann boundary conditions for the microscopic field .
writing @xmath6 more explicitly : latexmath:[\[\label{eq : nonlocal1 } s_i^{(0)}(\varphi ) \;=\ ; \frac{1}{2 } \
, g^2 \ , \int \frac{d^d k_\parallel}{(2\pi)^d } \ ; \sum_{m , n=0 } \ , \langle { \widetilde\varphi } |\psi_n^{(\epsilon ) } \rangle \,\frac{1 } { \xi_n^2 + k_\parallel^2 + m^2 } \ , \langle \psi_n^{(\epsilon ) } let us first consider the leading term in a small-@xmath2 expansion , obtained by keeping only the zero mode contribution .
this may be written as follows : @xmath37 where we introduced @xmath38 , which works as an approximant of dirac s @xmath0-function .
then we see that this leading term may be regarded as a local @xmath0-function term , with a momentum - dependent strength , and affected by a coefficient @xmath39 .
it is convenient to introduce the product @xmath40 , since that is the constant that determines the strength of the boundary interaction term : @xmath41 the following terms in the expansion are obtained by expanding the overlaps @xmath42 for small @xmath2 , by using a taylor expansion for the vacuum field . up to the second order in derivatives ,
we find : @xmath43_{x_d=0 } & { \rm if } \ ;
n = 2 k + 1 & ( k = 0,1,\ldots ) \\
\sqrt{2 } \big(\frac{\epsilon}{\pi}\big)^2 \ , \frac{1}{(2k)^2 } \ , ( -1)^k \;\big[\partial^2_d{\widetilde\varphi}(k_\parallel , x_d)\big]_{x_d=0 } & { \rm if } \ ;
n = 2 k & ( k = 1,2,\ldots ) \end{array } \right .
\;.\ ] ] we see that the assumption that one could use to justify the expansion _
a posteriori _ is that the field inside the mirror should not change appreciably inside the mirror , more precisely , the length scale of the spatial variation should be much larger than @xmath2 .
this means that higher powers of @xmath2 will be attached to higher derivatives of the field .
equipped with the expansion ( [ eq : nonlocal3 ] ) , we obtain the corresponding expansion of @xmath6 up to the second order in derivatives : @xmath44 where @xmath45 coincides with eq.([eq : local ] ) , while : @xmath46 } \;.\ ] ] here the prime denotes derivative with respect to @xmath47 .
performing the sum of the series , we obtain @xmath48 ^ 2 } \ , \big\ { \frac{1}{8 } - \frac{1}{4 \epsilon \ , \omega(k_\parallel ) } \ , \tanh\big [ \frac{\epsilon \ , \omega(k_\parallel)}{2}\big]\big\ } \;,\ ] ] where @xmath49
. we may again write this term in a similar fashion to ( [ eq : local ] ) : @xmath50 ^ 2 } \
; \int_{-\infty}^{+\infty } dx_d \,\delta'(x_d ) { \widetilde\varphi}^*(k_\parallel , x_d ) \int_{-\infty}^{+\infty } dx'_d \ , \delta'(x'_d ) { \widetilde\varphi}(k_\parallel , x'_d)\ ] ] with @xmath51 \big\ } \;.\ ] ] we will now summarize the result for the expansion of the nonlocal term , presenting it in a way which shall be useful in the derivation of the casimir energy .
thus , we encode the results as follows : the expansion for the interaction action ( [ eq : nonlocal ] ) may be interpreted as an expansion for the nonlocal potential , so that : @xmath52 where : @xmath53 with @xmath54 ^ 2}$ ] and @xmath55 ^ 2}$ ] .
it is worth noting that the derivative expansion of the effective action for the specific example considered so far does not include terms containing only one derivative of the scalar field ( i.e. terms proportional to @xmath56 , see eq.([eq : vnldelta ] ) ) .
this is due to the fact that the eigenfunctions in eqs.([eq : psid ] ) and ( [ eq : psin ] ) have a definite parity in the interval @xmath57 .
therefore , we expect such terms to show up only if one considered ` non - symmetric ' mirrors in which the boundary conditions that confine the microscopic fields are different on both interfaces @xmath58 , for example @xmath59 . up to here , we considered a microscopic field @xmath4 linearly coupled to @xmath1 .
the expansions presented in eqs.([expansionv ] ) and ( [ expansionv0y2 ] ) remain valid when the microscopic @xmath4-field is nonlinearly coupled to @xmath1 .
to illustrate this fact , we will now consider a different case , i.e. we shall deal with a @xmath60 coupling term for @xmath61 . in terms of the eigenfunctions @xmath62 , the second order term in the expansion of the action @xmath63
is given by : @xmath64 the integrations over normal coordinates ( @xmath47 and @xmath65 ) can be trivially performed by means of the orthogonality conditions of the eigenfunctions @xmath62 , obtaining @xmath66 after integrating over @xmath67 , one sees that the interaction action becomes @xmath68 whence one can read the coefficient @xmath69 in eq.([expansionv0y2 ] ) , @xmath70 where we have used that @xmath71 .
the sum in last equation runs from @xmath72 for neumann boundary conditions , and from @xmath73 for dirichlet ones .
eq.([mu0 ] ) is divergent for large @xmath74 , therefore we introduce a renormalization term , in order to obtain a finite coefficient . for a massless field with dirichlet boundary conditions we obtain : @xmath75,\ ] ] where @xmath76 is a renormalization constant .
we note that , contrary to what happens for the case of a linear coupling , we obtain a contribution whose strength is independent of @xmath2 . in the case of neumann
boundary conditions : @xmath77\;,\ ] ] which in the case of a massless microscopic field becomes : @xmath78 \right\}\;.\ ] ] we note that the previous series can be summed , the exact result being @xmath79 the computation of @xmath80 can be performed along similar lines , although is much more cumbersome and we will not present the details here . to summarize , in this section we have shown that the nonlocal interaction between the mirror and the scalar field admits a derivative expansion of the kind given in eq.([eq : vnldelta ] ) , where the coefficients @xmath81 depend not only on @xmath82 but also on the width @xmath2 of the mirror .
it is possible to adjust the relation of the coupling constants and @xmath2 in such a way that the leading term is finite in the limit @xmath83 , while the nonleading contributions are suppressed by powers of @xmath2 .
potentials proportional to the @xmath0-function and its derivatives have been considered previously by other authors ( see for instance @xcite ) .
we have shown here that these potentials arise naturally , in concrete examples , as the leading terms in a derivative expansion of the nonlocal effective interaction .
in this section , we compute the casimir energy that results from a derivative expansion of the nonlocal effective action , namely , when the interaction term at @xmath84 , after fourier transforming the parallel coordinates , has the local form : @xmath85,\ ] ] where @xmath86 , @xmath87 , @xmath80 are arbitrary real ( in euclidean spacetime ) functions of @xmath82 and @xmath2 .
note that eq.([eq : fexp ] ) proceeds from a nonlocal action in coordinate space : @xmath88,\end{aligned}\ ] ] where now @xmath89 , @xmath90 , @xmath91 depend on @xmath92 .
the interaction term for the remaining mirror , @xmath7 , is obtained by a simple shift . the terms containing derivatives of the @xmath0-function
are expected to be suppressed by powers of @xmath2 , however , for the sake of generality , we will first compute the casimir energy for the above interaction terms exactly , performing the expansion in powers of @xmath2 afterwards .
one of the reasons for this procedure is that the knowledge of the exact casimir energy for ( [ eq : fexp1 ] ) may be useful in other circumstances , not necessarily related to the effective models we are considering here .
introducing the matrix @xmath93 ( a function of @xmath94 ) we can write ( [ eq : fexp1 ] ) as follows : @xmath95 and a similar expression for @xmath7 .
then , we introduce two sets of auxiliary fields @xmath96 , @xmath97 and @xmath98 , @xmath99 , in order to write @xmath100 @xmath101\ ] ] @xmath102 \varphi(x)\big\ } , \label{x}\ ] ] with @xmath103 and @xmath104 an irrelevant constant .
note that the representation ( [ x ] ) makes sense when all the eigenvalues of @xmath105 are greater than zero ( they are real , since the matrix is hermitian ) . using eq.([x ] ) , we may write the generating functional as follows : @xmath106 where we omitted , for the sake of clarity , writing all the arguments . after integrating the field @xmath1
we get @xmath107 \right\ } } \nonumber \\ & \times & \exp{\left\{-\frac{1}{2}\int d^{d+1}x \int d^{d+1}y \left(j^{(0)}(x ) + j^{(a)}(x)\right ) \delta(x , y)\left(j^{(a)}(y ) + j^{(a)}(y)\right)\right\}},\end{aligned}\ ] ] where @xmath108 we first evaluate the term quadratic in the currents more explicitly , @xmath109 \;,\end{aligned}\ ] ] and then , introducing fourier transforms in the parallel coordinates , we can show that @xmath110,\label{qf}\end{aligned}\ ] ] where @xmath111 . to compute @xmath112 we used that @xmath113 and we approximated @xmath114 .
this follows from recalling how we obtained the derivative expansion : an approximant of the @xmath0 was replaced by its limit .
a fourier transformation of the above then yields the expression used for @xmath112 . on the other hand ,
the first derivatives of the free propagator are ill - defined at @xmath115 . in order to obtain eq.([qf ] ) , we have used a symmetric limit regularization so that @xmath116 .
we can now compute the vacuum energy @xmath117 as follows , @xmath118 where @xmath119 where @xmath120 are four @xmath121 block - matrices and @xmath122 is the fourier transform of the matrix @xmath105 defined in eq.([matrixm ] ) .
finally , the subtracted ( i.e. , without self - energies ) energy can be written as @xmath123,\ ] ] where @xmath124 and @xmath125 @xmath126 eq.([eq : energydelta ] ) is the main result in this section ; it gives the exact casimir energy for the generalized @xmath0-potentials , as a function of the coefficients that determine the effective interaction term .
it is interesting to emphasize that the structure of the vacuum energy in eq.([eq : energydelta ] ) is similar to the lifshitz formula for the electromagnetic field in the presence of anisotropic materials , in which it is necessary to introduce @xmath121 reflection matrices to take into account the mixing between te and tm modes @xcite . here
the @xmath121 matrices come from the introduction of two auxiliary fields on each mirror to describe the effective action ( see eq.([x ] ) ) .
the presence of a divergence , when @xmath33 , in @xmath112 does not introduce any divergence in the casimir energy .
indeed , the inverses of the matrices @xmath127 and @xmath128 are finite in that limit @xmath83 . computing explicitly the determinant in eq.([eq : energydelta ] ) , the final expression becomes @xmath129 , \label{energy general}\ ] ] where @xmath130 and @xmath131 depend on @xmath14 and @xmath2 explicitly and also implicitly through the coefficients @xmath81 .
although it is possible to obtain general expressions for @xmath130 and @xmath131 , the result is a rather lengthy expression , which is not very illuminating .
thus we will analyze two interesting particular cases .
let us first consider the calculation of the casimir energy in the framework of the derivative expansion introduced in the previous section .
as already mentioned , on general grounds we expect the terms containing higher derivatives to be suppressed by powers of @xmath2 .
to make this point explicit , we write @xmath132 and @xmath133 , where @xmath134 are of the same order of magnitude in the limit @xmath83 .
we then evaluate the determinat in eq.([eq : energydelta ] ) exactly , and expand the result in powers of @xmath2 .
after a long but nevertheless straightforward calculation , the expansion for the casimir energy adopts the form @xmath135 , \label{expansionenergy}\ ] ] where @xmath136\ , .
\label{expansion2}\end{aligned}\ ] ] the coefficients @xmath134 , depending on @xmath14 , encode all the information about the interaction between the vacuum field and the microscopic degrees of freedom living on the mirrors .
the outcome is dominated by the usual @xmath0-potential result , but with a @xmath14-dependent strength .
it is worth noting that , up to this order , there are no terms proportional to @xmath137 in the casimir energy .
these terms do appear in the fourth order . in order to illustrate the kind of corrections induced by the finite width of the mirrors ,
let us assume that the coefficients @xmath134 are approximately constants .
then the casimir energy for @xmath61 can be written , to first order in @xmath2 as follows : @xmath138+o(\epsilon ^2)\ , , \ ] ] where we introduced the coefficient functions @xmath139 , which can be easily evaluated numerically .
they are plotted in figs.1 and 2 . both of them are monotonic and positive definite functions , interpolating between @xmath72 and a finite value for @xmath140 .
the leading term reproduces the casimir result for perfectly conducting mirrors in this limit , while the second one introduces a correction that falls off faster ( with an extra power of the distance ) . as a function of @xmath141 .
this function reproduces the casimir result for perfect conducting plates in the limit @xmath142 , @xmath143 .
, width=359 ] as a function of @xmath141 .
this function approaches @xmath144 in the limit @xmath142.,width=359 ] as a second example , we now consider the vacuum energy for the generalized @xmath0-potentials , assuming that the coefficients @xmath81 do not depend on @xmath2 .
in this situation we obtain @xmath145 ^ 2 } \right]= \frac{1}{2 } \int \frac{d^dk_\parallel}{(2\pi)^d } \ln \left[1-\frac{\tilde\mu_{eff}^2 \ , e^{-2ak_\parallel}}{[2k_\parallel+\tilde\mu_{eff}]^2 } \right ] , \label{exacte}\ ] ] where @xmath146 .
it is interesting to note that the casimir energy is well defined only when the parameters @xmath147 are such that the coefficient that multiplies @xmath148 is less than 1 , and this is the case if @xmath149 .
this condition have been found before , albeit in a different looking but equivalent form ; indeed , we have seen that , for the auxiliary field representation ( [ x ] ) to make sense , positivity of the eigenvalues of the matrix @xmath105 is a necessary ( and sufficient ) condition . from a physical point of view , this condition can be interpreted as follows : the interaction term at each mirror may be diagonalized , to look like the sum of two decoupled quadratic interaction terms , each one involving a mixture of the field and its normal derivative at the mirror . for the vacuum to be stable , one must have therefore non - negative eigenvalues , since they are the coefficients that affect each decoupled term . vanishing eigenvalues , on the other hand
, are not forbidden physically , rather , one should represent them with just one auxiliary field .
otherwise the redundancy pops up in the form of a zero mode .
it is remarkable that eq.([exacte ] ) corresponds to the casimir energy for a usual @xmath0-potential ( i.e. without derivatives of the @xmath0-function ) with an effective coefficient given by @xmath150 .
however , it is also worth stressing that this equation has been derived assuming a particular regularization for @xmath151 .
while this regularization is well justified in the case of the derivative expansion , a formal calculation which started from the generalized @xmath0-potentials , could give different , regularization dependent results , without any immediate physical reason to chose one from another .
for example , in the framework of dimensional regularization one would obtain @xmath152 , with @xmath153 an arbitrary constant .
different regularizations of this object correspond , physically , to imposing different boundary conditions for the propagator of the vacuum field at the mirror .
if the concrete model for the finite width mirror is unknown , this lack of information manifests itself in the fact that one has many regularizations available , and they give rise to different values of the energy .
however , one knows that , what makes sense physically is not the regularization used ; rather , it is the boundary condition it produces on the propagator .
then , one may regard the boundary condition for the propagator as a renormalization condition which hides the ignorance on the details of the model into a bare coefficient function @xcite .
we have shown , in concrete examples , how the nonlocal induced action which results from the integration of the microscopic fields ( that represent the media composing the mirrors ) can be expanded to produce a local action ; i.e. , one that has point - like support .
this means that it may be written as terms involving the @xmath0-function and its derivatives .
equipped with the general form of that local action , we then derived the casimir energy for the vacuum scalar field .
a conceptually interesting point is that the presence of derivatives of the @xmath0-function , in the effective action , produces a final result for the casimir energy that can be written in the form of the lifshitz formula for the electromagnetic field with @xmath121 reflection matrices .
this analogy is a biproduct of the representation of the effective action in terms of two auxiliary fields when the coefficients from that local action come from a microscopic model , we have shown that the result may be consistently expanded in powers of the width of the mirrors , producing a result which may be interpreted as a dirichlet - like energy plus sub - leading corrections .
it would be interesting to generalize these results to the realistic case of the electromagnetic field coupled to dirac fields describing charges on the mirrors .
besides , we considered also the case when the coefficients are assumed to be independent of the width of the mirrors . in this case , the exact result adopts a quite simple form : the vacuum energy coincides with the one given by the usual @xmath0-potential , with an effective coupling .
that is to say , the effect of the terms proportional to derivatives of the @xmath0-function in the effective action is to renormalize the coupling of the usual @xmath0 potential .
the last result depends , in principle , on the particular regularization used to handle these ( highly singular ) potentials .
however , when one abandons the description in terms of the coefficients for those terms , in favour of another in terms of the boundary conditions on the propagator , the apparent ambiguity disappears @xcite .
c.d.f . thanks conicet , anpcyt and uncuyo for financial support . the work of f.d.m . and f.c.l was supported by uba , conicet and anpcyt .
bib g. plunien , b. mller , and w. greiner , phys . rep . * 134 * , 87 ( 1986 ) ; p. milonni , _ the quantum vacuum _ ( academic press , san diego , 1994 ) ; v. m. mostepanenko and n. n. trunov , _ the casimir effect and its applications _ ( clarendon , london , 1997 ) ; m. bordag , _ the casimir effect 50 years later _ ( world scientific , singapore , 1999 ) ; m. bordag , u. mohideen , and v. m. mostepanenko , phys
. rep . * 353 * , 1 ( 2001 ) ; k. a. milton , _ the casimir effect : physical manifestations of the zero - point energy _
( world scientific , singapore , 2001 ) ; s. reynaud _
et al . _ , c. r. acad .
paris * iv-2 * , 1287 ( 2001 ) ; k. a. milton , j. phys . a : math
. gen . * 37 * , r209 ( 2004 ) ; s.k .
lamoreaux , rep .
. phys . * 68 * , 201 ( 2005 ) ; special issue _ `` focus on casimir forces '' _
, new j. phys .
* 8 * ( 2006 ) . c. d. fosco , f. c. lombardo and f. d. mazzitelli , phys .
b * 669 * , 371 ( 2008 ) [ arxiv:0807.3539 [ hep - th ] ] .
m. bordag , i. v. fialkovsky , d. m. gitman and d. v. vassilevich , arxiv:0907.3242 [ hep - th ] . c. d. fosco and e. losada , phys .
b * 675 * , 252 ( 2009 ) .
r. esquivel , c. villarreal and w. l. mochn , phys .
a * 68 * , 052103 ( 2003 ) ; + r. esquivel - sirvent , c. villarreal , w. l. mochan , a. m. contreras - reyes and v. b. svetovoy , j. phys .
a * 39 * , 6323 ( 2006 ) ; + a. saharian and g. esposito , _ in the proceedings of 11th marcel grossmann meeting on general relativity , berlin , germany , 23 - 29 jul 2006 , pp 2761 - 2763 _ 052103 ( 2003 ) ; + a. saharian and g. esposito , j. phys .
a * 39 * , 5233 ( 2006 ) . c. d. fosco , f. c. lombardo and f. d. mazzitelli , phys .
d * 77 * , 085018 ( 2008 ) [ arxiv:0801.0760 [ hep - th ] ] .
see also i. fialkovsky , y. pismak and v. markov , phys .
d * 79 * , 028701 ( 2009 ) .
m. bordag , m. hennig and d. robaschik , j. phys .
gen_*25 * , 4483 ( 1992 ) .
k. milton , j. phys . a : _ math
. gen _ * 37 * , 6391 ( 2004 ) ; + k. a. milton and j. wagner , j. phys . a * 41 * , 155402 ( 2008 ) [ arxiv:0712.3811 [ hep - th ] ] . f. s. s. rosa , d. a. r. dalvit and p. w. milonni , phys . rev .
a * 78 * , 032117 ( 2008 ) .
a detailed explanation of this point will be provided elsewhere : c. d. fosco , f. c. lombardo and f. d. mazzitelli , _ in preparation_. | we calculate the casimir energy for scalar fields in interaction with finite - width mirrors , described by nonlocal interaction terms .
these terms , which include quantum effects due to the matter fields inside the mirrors , are approximated by means of a local expansion procedure . as a result of this expansion ,
an effective theory for the vacuum field emerges , which can be written in terms of generalized @xmath0-potentials . we compute explicitly the casimir energy for these potentials and show that , for some particular cases , it is possible to reinterpret them as imposing imperfect dirichlet boundary conditions . |
the problem of graceful exit from the pre- to the post - big bang branch , and the recovery of classical dynamics at late times , has remained a major issue for pre - big bang ( pbb ) models @xcite ( for a review , see @xcite ) .
the equations derived from the low energy effective action of string theory , can not provide a smooth transition between the pre - big bang phase and the standard post - big bang phase of decreasing curvature .
there have been many attempts to solve this problem of overcoming the curvature singularity .
if it is assumed that the curvature at the transition time is small enough to use the low energy equations , then a smooth transition can be achieved either by adding a nonlocal dilaton potential to the action @xcite , or by considering an anisotropic universe dominated by some kind of matter with a suitable equation of state @xcite .
if the curvature is very large , higher - order corrections to the low energy effective action have to be added , which can be derived from the loop expansion and from the @xmath0 expansion @xcite .
both of these approaches are based on ad hoc assumptions that have to be imposed by hand for a graceful exit solution .
this is a consequence of our poor knowledge of the non - perturbative regime of string theory .
curvature singularities have been recently studied in the framework of loop quantum cosmology ( lqc ) @xcite , which is a canonical quantization of homogeneous cosmological spacetimes based on loop quantum gravity ( lqg ) .
the gravitational phase variables are the matrix - valued ashtekar connection @xmath1 and the conjugate triad @xmath2 , and the elementary variables are the holonomies of the connection and the fluxes of the triad . in the quantum theory ,
holonomies represent fundamental excitations of quantum geometry which are polymer - like one - dimensional excitations .
holonomies also provide information about the connection which does not have a corresponding operator in lqg / c . for classical frw
cosmology , connection is proportional to @xmath3 and thus holonomies encode information about the expansion ( contraction ) of the universe .
the quantum theory obtained from loop quantization turns out to be different from the wheeler - de witt quantization ( the polymer representation is not equivalent to the usual fock representation ) .
wheeler - de witt quantization does not resolve the cosmological singularity , but in lqc a generic resolution of curvature singularities has been obtained .
the resulting picture is of a universe which bounces when curvature reaches planck scale and is classical at low curvatures .
this picture is based on extensive analytical and numerical investigations for frw flat @xcite , closed @xcite and open @xcite models , bianchi i models @xcite , de sitter @xcite and anti - de sitter models @xcite .
recent investigations for flat models have further revealed the genericity of the bounce for a dense subspace of the physical hilbert space @xcite .
lqc therefore in principle allows us to incorporate non - perturbative effects in pbb models , at least at a phenomenological level . if string theory and lqg both encompass elements of an eventual quantum gravity theory , then it is interesting to explore the phenomenology that results when one applies ideas from one approach to models constructed in the framework of the other .
we focus only on this phenomenology , and do not attempt to address the fundamental issue of the relation between string theory and lqg .
instead our approach is to extract the basic elements of lqc quantization that are relevant to understand singularity resolution , and apply them to pbb models .
we start with the massless dilaton @xmath4 in the einstein frame and cast the problem as a hamiltonian system in connection - triad phase space variables . since there is no external time parameter in quantum gravity ,
subtleties arise in recovering the conventional notion of dynamics and identifying the post- and pre - big bang branches .
these subtleties have been emphasized previously in the quantum cosmology of the pbb scenario @xcite . to resolve them ,
we employ the ideas of relational dynamics used in lqc @xcite and treat the dilaton , which is monotonic , as an internal clock . the change of scale factor ( or other observables ) with respect to the internal clock
can then be found by solving the hamiltonian constraint and computing for example , @xmath5 .
classically , as well in the pbb scenario , in the backward evolution of the post - big bang branch , the scale factor goes to zero as @xmath6 , and it increases with an increase in @xmath4 .
similarly , the forward evolution of the pre - big bang branch results in a decrease in the scale factor as @xmath4 increases , with the scale factor vanishing as @xmath7 .
the pre and post - big bang branches are distinguished by the behavior of the scale factor with respect to the dilaton . in classical general relativity and in pbb scenarios
( without any tree - level corrections ) , the pre- and post - big bang branches are disjoint .
a wheeler - de witt quantum cosmology analysis of the pbb scenario reveals that the pre- and post - big bang phases correspond to different branches of the wavefunction @xcite . at an effective level
, trajectories for the scale factor or the dilaton with respect to proper time can be obtained by recasting the equations via introduction of a parameter @xmath8 , for example : @xmath9 .
the parameter @xmath8 , which plays the role of classical external time , can be thought of as emerging by semi - classical approximations .
we would employ this algorithm in our analysis , using the observation that the underlying loop quantum dynamics can be described by an effective hamiltonian for states which are semi - classical at late times @xcite . as it will turn out , loop quantum geometric effects lead to a non - singular transition between the pre- and post - big bang branches , with respect to the internal clock @xmath4 .
unlike the wheeler - de witt quantization , the pre- and post - big bang phases correspond to the same branch of the wavefunction in lqc @xcite . using the effective hamiltonian ,
this dynamics translates to a non - singular evolution with respect to the time parameter @xmath8 .
since our approach captures the features of the polymer representation in lqg / c , we will refer to it as polymerization " , as in a recent work on the quantization of cosmological spacetimes @xcite .
once the polymerization has been performed in the einstein frame , we transform the dynamical equations to the string frame and study the solutions .
we show that polymerization indeed cures the curvature singularity in both the einstein and string frames .
furthermore , at late times there is a smooth transition to classical dynamics .
thus the graceful exit problem in the pbb scenario is overcome in this approach .
the paper is organized as follows . in sec .
ii we start with the zeroth order effective action in the einstein frame for the dilaton without potential , and use a legendre transformation to cast the problem as a hamiltonian system with connection - triad phase space variables .
we then perform polymerization of the connection and show that generic solutions of the modified dynamical equations in the einstein frame are non - singular . in sec .
iii , we give the modified dynamical equations in the string frame .
interestingly , polymer modifications appear similar to quantum one loop corrections in the string frame . by numerically integrating the dynamical equations we show that singularity resolution is obtained also in the string frame . in sec .
iv we summarize our results and compare the polymer modifications leading to non - singular dynamics with previous attempts in pbb scenarios using one loop quantum corrections in the string frame action .
our starting point is the low - energy effective action derived from string theory in the einstein frame .
we set the antisymmetric @xmath10 field to zero , and assume stabilization of all moduli fields associated with compactification . in the absence of a dilaton potential , the action is ( with units @xmath11 and @xmath12 ) s = - d^4 x .
[ eaction ] variation with respect to the metric and the dilaton gives the equations of motion m_p^2 g _
= _ _ - g _ _ ^ , ^_= 0 .
[ eeq ] for the frw metric these lead to h^2 = ^2 , = - ^2 , + 3 h = 0 , [ frweq ] of which only two equations are independent . in the pbb scenario , there are two disjoint solutions : an expanding post - big bang branch and the contracting pre - big bang branch , separated by a curvature singularity at @xmath13 , as shown in fig . [ e_a_plot ] . in order to implement effects of polymer quantization we cast the problem in the hamiltonian framework with same phase space variables as in lqc .
these are the gravitational degrees of freedom @xmath14 , corresponding to the connection and the triad variables ( the latter kinematically related to metric as @xmath15 ) , and the matter degrees of freedom @xmath16 . in the classical limit , [ dotphi ] ( c , p)=(a , a^2 ) , ( , _ ) = ( , a^3 ) .
we obtain the hamiltonian from the action ( [ eaction ] ) using a legendre transformation : [ classicalh ] = - 3 m_p^2 c^2 p^1/2 + .
the dynamical equations ( [ frweq ] ) can be obtained via hamilton s equations of motion : @xmath17 . in lqc , there is no analog of the connection operator .
information about the connection is instead obtained from the holonomies , whose elements are of the form @xmath18 , where the underlying quantum geometry imposes @xcite [ lambdap ] = . here
@xmath19 is a dimensionless constant , which in lqc is o(1 ) .
it determines the minimum eigenvalue of the area operator in lqg , @xmath20 .
the parameter @xmath21 thus captures the discrete nature of quantum geometry in lqc .
dynamics in lqc is described by a quantum difference equation . however , using geometric methods of quantum mechanics one can find an effective hamiltonian of the form ( [ classicalh ] ) that provides an excellent approximation to the quantum evolution of the states which are semi - classical at late times @xcite .
one of the features of this hamiltonian is the presence of higher order terms in @xmath22 which arise by expressing the hamiltonian in the elements of holonomies .
the effective hamiltonian contains terms of the form @xmath23 , which can be thought of as capturing the non - local character of quantum curvature at leading order in an effective continuum spacetime @xcite . to capture this key feature of lqc
, we will adopt it at an effective level in the present scenario and perform a phenomenological polymerization , [ poly ] c , which leads to _ eff = - 3 m_p^2 p^1/2 + .
[ effham ] for small curvature , @xmath24 , eq .
( [ effham ] ) reduces to eq .
( [ classicalh ] ) .
however significant departures from classical dynamics are expected when @xmath25 is large . as we show below the effective hamiltonian ( [ effham ] )
leads to classical dynamics where expected , and to a significant departure from it when the curvature scalar is large , @xmath26 , producing a non - singular evolution from the pre- to the post - big bang .
let us first consider the hamilton s equation for @xmath27 , = \{p,_eff } = ( c ) ( c ) , [ hameqp ] which leads to [ lqc - nonpert ] h^2 = = ^2(c ) .
this can be cast in the form of a modified friedman equation by using the hamiltonian constraint @xmath28 , and eq .
( [ dotphi ] ) to determine @xmath29 : [ lqcfrw ] h^2 = ( 1 - ) , m_*=m_p .
we obtain the raychaudhuri equation from this by using the klein - gordon equation ( or alternatively from hamilton s equation for @xmath30 ) , [ lqcrai ] = - ^2 ( 1 - ) . .
as is evident from the modified friedman and raychaudhuri equations , corrections originating from polymerization die away rapidly when @xmath31 .
however , when @xmath32 is comparable to @xmath33 , quantum gravity modifications become significant and lead to a bounce .
the klein - gordon equation implies @xmath34 , and substituting into eq .
( [ lqcfrw ] ) we find ( ) ^6=1 + 3t^2 , where @xmath35 is the minimum value of @xmath36 , at @xmath37 .
then it follows that & = & m_p^-1 ( t ) , + h & = & .
[ solutions ] these solutions are regular for all time , as shown in figs .
[ e_phi_plot ] and [ e_h_plot ] .
the curvature stays finite through the bounce ( fig .
[ e_r_plot ] ) .
the scale factor undergoes a bounce from a phase of accelerated contraction as can be seen in fig .
[ e_a_plot ] .
the kinetic energy of the dilaton is also regular and well behaved .
the maximum dilaton kinetic energy and ricci curvature are = m_*^4 , |r_0|=. it is important to note that the polymerization ( [ poly ] ) is truly of non - perturbative character in the sense that @xmath29 is an infinite series in powers of @xmath21 .
however , by writing eq .
( [ lqc - nonpert ] ) in the form ( [ lqcfrw ] ) , this feature is concealed .
then the term in eq .
( [ lqcfrw ] ) that is quadratic in kinetic energy may appear to originate from a truncation of a series expansion in kinetic energy of the dilaton .
no such truncation has been performed , and the appearance of the @xmath38 term is an artifact of our expressing the non - perturbative eq .
( [ lqc - nonpert ] ) in the form of a modified friedman equation ( [ lqcfrw ] ) .
we have seen that lqg - inspired polymerization of the connection variable leads to non - singular evolution between the pre - big bang and the post - big bang in the einstein frame .
a complete understanding requires analyzing the dynamics in the string frame , since a reliable solution of the graceful exit problem must work in both frames @xcite .
the string frame action is s = - d^4 x e^-/m_p .
[ saction ] in an frw universe , this action gives the equations of motion , ^2 - + & = & 0 , [ sf ] + + - & = & 0 , [ sr ] +
+ 3 - & = & 0 .
[ sfrweq ] here a prime denotes a derivative with respect to proper time @xmath39 in the string frame , and @xmath40 .
the transformation between the einstein and string frames is given by & = & e^-()/2m_p d , + a(t)&= & e^-()/2m_p ( ) , ( t ) = ( ) .
[ transform ] using these transformations , the lqc modified equations in the string frame are ^2 - + & = & 0 , [ sfp ] + + 2 - & = & 0 , [ srp ] + + 3 - & = & 0.[sframemodeq1 ]
. solutions of eqs .
( [ sfp])([sframemodeq1 ] ) are illustrated in figs .
[ s_a_plot][s_r_plot ] .
the dynamical evolution in the string frame is such that the hubble parameter ( and hence the curvature ) reaches a maximum and then starts to decrease , thus providing a graceful exit from the singularity problem
. the dilaton monotonically grows with time as in the classical pbb scenario , but its slope decreases with time , and its derivative approaches zero .
this opens up an interesting possibility , as it appears that polymer modifications can possibly alleviate the problem of late - time stabilization of the dilaton @xcite . for a more complete investigation
, one needs to study the dynamics with a non - perturbative potential , which is beyond the scope of the present work .
we have shown that lqc inspired polymerization applied to the pbb scenario leads to a solution of the graceful exit problem , i.e. , a regularization of the singularity that divides the pre- and post - big bang branches in the pbb case .
the polymerization was performed in the einstein frame and the resulting modified dynamical equations were transformed to the string frame .
the graceful exit is achieved both in einstein and string frame without any fine tuning of parameters .
in addition , the dilaton in the string frame approaches a constant value at late times .
thus a phenomenological approach using ideas from lqc within the string - theory based pbb leads to a resolution of a key problem in the standard pbb scenario .
we note that , unlike other attempts to construct graceful exit solutions , which introduce matter that violates the null energy condition to regularize the singularity @xcite , our model regularizes the singularity via gravitational effects , without needing any null energy condition violating matter .
polymerization of the connection induces quantum gravitational effects which are only significant at large curvatures .
as the curvature becomes small they die quickly and lead to classical dynamics at low curvature scales .
thus in comparison with previous attempts , our model does not suffer from the lack of classical solutions at late times and there is a generic graceful exit not only from the curvature singularity but also to the classical branch . an important question which arises
is whether polymerization has anything in common with quantum corrections considered so far in pbb .
the appearance of the @xmath41 factor with the squared kinetic energy term is notable , since this is apparently similar to the term obtained by considering one - loop quantum corrections to the string frame action . with a correction _
q = 2 [ stringcorrlagr ] added to the string frame action @xcite , one obtains a first order correction to the energy - momentum tensor .
this has the form of a perfect fluid with _
q = 3 , p_q = , [ stringeos ] so that the the equations of motion give 3 ^ 2 - 3 + & = & e^/m_p _ p , + 2 + 3 ^ 2 - + - & = & e^/m_p p_q .
[ stringcorreq ] notice that in this case , @xmath42 .
by contrast , using eqs .
( [ sfp])([sframemodeq1 ] ) , we find that for the lqc polymerization , _
= - , p _ = - , w _
[ rqpq ] thus it appears not to be possible to re - cast the lqc - inspired polymerization in terms of a one - loop quantum correction to the low energy string theory action . in our phenomenological approach
we have exported to pbb scenarios one of the key features of quantum geometry in non - perturbative lqc , as understood at the level of an effective continuum spacetime . in
related work on string - based cyclic models @xcite , the results have been encouraging for the resolution of singularities , as in the present analysis .
however , these investigations should be viewed as first steps in the direction of applying insights from non - perturbative quantum geometry approaches to string - based cosmological models .
various questions remain open , including an understanding of the relation of polymerization to an action framework .
it is a priori not clear whether an action with a finite number of terms as understood in conventional treatments can be written to mimic polymerization at an effective hamiltonian level .
another issue is to consider polymerization of the matter degrees of freedom along with the gravitational ones .
furthermore , the polymerization considered in our analysis does not capture effects arising from quantum properties of states , such as dispersions and skewness , which although expected to be small in magnitude may provide useful insights .
we conclude with a remark on the different possible ways to polymerize , alternative to eqs .
( [ lambdap ] ) and ( [ poly ] ) . although it is tempting to consider various other choices , for example treating @xmath21 as a constant or a function of @xmath27 different from eq .
( [ lambdap ] ) , it turns out that such versions of polymerization are ruled out by physical constraints .
firstly , the alternatives do not lead to classical gr at low curvature scales , and secondly , they produce unphysical quantum gravity effects at gauge - dependent scales @xcite .
polymerization based on the correct loop quantization of cosmological spacetimes @xcite does not suffer from these problems .
apart from the functional dependence in @xmath21 , we could also consider polymerizations such as @xmath43 , which may arise from a different quantization scheme .
any such choice ( or a similar trigonometric form ) leads to similar qualitative results , only effecting the value of @xmath19 in eq .
( [ lambdap ] ) .
@xmath44 * acknowledgements : * we thank alejandro corichi , aseem paranjape , luca parisi and kevin vandersloot for helpful discussions , and maurizio gasperini for useful comments .
gdr is supported by infn .
the work of rm is partly supported by stfc .
ps is supported by nsf grant phy-0456913 and the eberly research funds of penn state .
he also thanks the institute of cosmology & gravitation at portsmouth for warm hospitality at the beginning of this work .
m. gasperini and g. veneziano , astropart .
* 1 * , 317 ( 1993 ) [ arxiv : hep - th/9211021 ] . m. gasperini and g. veneziano , phys .
* 373 * , 1 ( 2003 ) [ arxiv : hep - th/0207130 ]
. m. gasperini , j. maharana and g. veneziano , nucl .
b * 472 * , 349 ( 1996 ) [ arxiv : hep - th/9602087 ] .
m. gasperini , j. maharana and g. veneziano , phys .
b * 272 * , 277 ( 1991 ) .
m. giovannini , phys .
d * 57 * , 7223 ( 1998 ) [ arxiv : hep - th/9712122 ] .
m. giovannini , phys .
d * 59 * , 083511 ( 1999 ) [ arxiv : hep - th/9807049 ] .
g. de risi and m. gasperini , phys .
b * 521 * , 335 ( 2001 ) [ arxiv : hep - th/0109137 ] .
m. gasperini , m. maggiore and g. veneziano , nucl .
b * 494 * , 315 ( 1997 ) [ arxiv : hep - th/9611039 ] .
r. h. brandenberger , r. easther and j. maia , jhep * 9808 * , 007 ( 1998 ) [ arxiv : gr - qc/9806111 ] .
d. a. easson and r. h. brandenberger , jhep * 9909 * , 003 ( 1999 ) [ arxiv : hep - th/9905175 ] . s. foffa , m. maggiore and r. sturani , nucl .
b * 552 * , 395 ( 1999 ) [ arxiv : hep - th/9903008 ] . c. cartier , e. j. copeland and r. madden , jhep * 0001 * ,
035 ( 2000 ) [ arxiv : hep - th/9910169 ] .
c. cartier , e. j. copeland and m. gasperini , nucl . phys .
b * 607 * , 406 ( 2001 ) [ arxiv : gr - qc/0101019 ] .
a. ashtekar , arxiv : gr - qc/0702030 .
+ for early works , see , + m. bojowald , living rev . rel .
* 8 * , 11 ( 2005 ) [ arxiv : gr - qc/0601085 ] .
+ for a chronological review , see , + g. date , arxiv:0704.0145v1 [ gr - qc ] .
a. corichi , t. vukasinac and j. a. zapata , arxiv:0704.0007v1 [ gr - qc ] .
+ for a similar polymer quantization , see also , + v. husain and o. winkler , phys .
d * 69 * , 084016 ( 2004 ) [ arxiv : gr - qc/0312094 ] . in general , there can be sub - leading corrections from terms of higher order arising due to quantum properties of the wavefunction .
however , for states which correspond to large classical universes at late times these corrections turn out to be negligible @xcite .
m. bojowald , r. maartens and p. singh , phys .
d * 70 * , 083517 ( 2004 ) [ arxiv : hep - th/0407115 ] ; + p. singh , k. vandersloot and g. v. vereshchagin , phys .
d * 74 * , 043510 ( 2006 ) [ arxiv : gr - qc/0606032 ]
. j. polchinski , `` string theory .
1 : an introduction to the bosonic string '' ( cambridge university press , 1998 ) . | we consider a phenomenological modification of the pre - big bang scenario using ideas from the resolution of curvature singularities in loop quantum cosmology . we show that non - perturbative loop modifications to the dynamics , arising from the underlying polymer representation , can resolve the graceful exit problem .
the curvature and the dilaton energy stay finite at all times , in both the string and einstein frames . in the string frame
, the dilaton tends to a constant value at late times after the bounce . |
it was game theory research , seeking screening strategies to prevent the counterfeiting of websites i.e. phishing attacks .
various ways for websites to counterfeit installed software behaviour were studied . in full screen mode , it was found that , browsers can counterfeit almost anything , including blue screens of death and formatting the hard drive . this is discussed in section ii . from an academic point of view
, full screen counterfeiting eliminates several categories of installed software behaviour , as possible anti - counterfeiting solutions .
one category of installed software behaviour was resistant to counterfeiting .
in section iii i present an explanation of this category from the discipline of cryptography and another explanation from the discipline of game theory .
every solution , in that category , was found to be a user - browser shared secret .
basically mallory can not counterfeit what mallory does not know .
the user - browser shared secret is not known by either bob or mallory .
this means a user - browser shared secret can be used to extend tls to seek further authentication of bob by the computer user . on successful verification of a tls certificate s digital signature , the browser presents a dialogue , fig .
1 . mallory can not counterfeit this dialogue without hacking in , to steal the shared secret .
the computer user now authenticates ( 1 ) her browser dialogue and ( 2 ) bob , by accepting the shared secret and bob s identity credentials which came from the tls certificate .
acceptance of these two proofs of identity is communicated by the user entering her login credentials . in section iv the possibility of bob authenticating alice was studied .
my research found that the absence of trent in this process means that alice must authenticate bob first , during account setup .
this makes alice authenticating bob more important than the other way around . at this point
i returned to researching alice s authentication of bob . our total dependence on alice authenticating bob requires us to make this act as straight forward as possible .
the final solution proposed is a mixed strategy of ( 1 ) a user - browser shared secret to facilitate extending tls , for alice to authenticate bob , ( 2 ) standardisation of the login process via a browser created login window , ( 3 ) using central banks as trent for financial institutions and ( 4 ) utilisation of common knowledge and education to guide alice through the process i.e. to prevent phishing attacks alice must fulfil her role and authenticate bob .
the idea is that a phishing attack is a game of incomplete information . that the user does not even know that a phishing attack is taking place .
it is the successful counterfeiting of the website that does this .
if we can devise a signalling strategy which can not be counterfeited then the computer user will know when a phishing attack is taking place .
they will back away from the phishing website causing the phishing attack to fail .
the idea was to add information , specifically an anti - counterfeiting signalling strategy which would be triggered after the browser has verified the digital signature on bob s tls certificate .
i listed behaviour that installed software is capable of but websites are not capable of .
the idea was your browser is installed software so it has this advantage over websites trying to counterfeit its behaviour .
the following categories were proposed for research : 1 .
drawing outside the browser canvas area .
2 . creation of modal windows .
file manipulation e.g. file creation , copying , renaming etc .
this includes the possibility of formatting the hard disk , though we ca nt use that as evidence either .
access to local data and operating system identifiers e.g. your username , your account login picture or whether or not you have accessed this website before .
microsoft , user account control behaviour . 6 .
existing best practice i.e. inspection of the tls certificate being used by your browser . in my original research
i dismissed or counterfeited every category except number 4 .
every solution in category 4 is actually a secret shared between the computer user and their web browser .
category 4 is discussed below in section iii .
originally i dismissed category 3 believing it to be unworkable .
however references [ 3 ] and sitekey [ 5 ] both use cookies to trigger their solutions .
cookies actually fit category 3 .
this is discussed further in section iv .
a key component of this research was the study of screening strategies .
while it was not my intention mallory , the counterfeiters , never wanted screening strategies to work .
effectively i was researching an arms race where mallory would keep changing her behaviour to prevent my screening strategies from working .
section vi outlines one interpretation of that arms race .
the actual path that i followed was to study the categories listed above .
there is no point in me documenting that research here because it is quite similar to discussions of screening strategies found in [ 1 ] and [ 2 ] .
one phishing attack website that i stumbled upon requested a username and password . even though the genuine website was open access .
this type of phishing is more social engineering than counterfeiting . during my research
the most versatile type of phishing i stumbled upon was full screen counterfeiting .
about ten years ago browsers stopped the javascript function window.open() from creating completely undecorated popup windows .
the function still works , however an address bar is added to the window , as shown in fig .
2 . without this
it would be easy to counterfeit browser controls and inspection of the tls certificate. within the past few years a fullscreen api has been added to html 5 / javascript .
it s still in development so browser specific function names exist like mozrequestfullscreen ( ) and webkitrequestfullscreen ( ) .
once a user performs an action which results in this function call , the browser switches into full screen mode .
the transition to full screen happens whether or not the user has been warned in advance .
yes the browser will show a warning , after the transition .
however each browser has its own warning windows and websites can query your browser and deliver the appropriate counterfeit .
3 shows a firefox warning .
the thing about social engineering is that almost anything goes .
consider a criminal harassing a user with warnings , like that shown in fig .
4 . once the user is accustomed to hitting allow , the criminals can request full screen mode and hope the user falls for it .
after that they can counterfeit the entire computer desktop , blue screens of death etc . for example , fig .
5 shows six bitmap images .
they are deliberately drawn to look fake . when assembled correctly they mimic a desktop .
6 shows a computer desktop just before a full screen counterfeiting attack .
once the user clicks on a button which executes javascript .
the requestfullscreen ( ) function is called , along with code which alters the webpage to show the fake browser , and desktop , controls .
this turns fig . 6 into fig .
note the substitution of the bitmaps from fig . 5 for the genuine browser and desktop .
if the user does not notice that the warning message is different they may click on allow. there are a number of differences between the harassment window and the genuine warning window . along with the message being different
the harassment window can not grey the entire desktop , nor can it be drawn at the top centre of the screen .
if the bitmaps used in fig . 5 were realistic then fig . 6 and fig .
8 would be almost identical . furthermore fig .
7 would only appear odd / unusual because the greyed area extended beyond the perceived canvas area and because the warning window appeared outside of the perceived canvas area .
these are very weak indicators of counterfeiting . from a researchers point of view
many types of installed software behaviour can be counterfeited .
including browser addons , inspection of tls certificates , and microsoft user account control behaviour . as such categories 5 and
6 must be eliminated as suitable anti - counterfeiting solutions .
furthermore we now need to be concerned with counterfeiting of blue screens of death , hackers / criminals blackmailing people with the threat of formatting their hard drives etc .
the purpose here is to demonstrate these mechanisms .
no user testing has been performed .
the academic exercise of demonstrating that this is possible is sufficient to eliminate categories 5 and 6 .
there is anecdotal evidence in [ 7 ] that these tactics will work . in the short term
best practice can be changed to ( 1 ) hit escape to exit full screen , ( 2 ) click on the padlock symbol and ( 3 ) read the tls certificate details window , to ensure you are connected to the correct website .
also we can roll back the full screen api , just like the undecorated window functionality .
i believe the system has been mistakenly identified as a two actor system .
when , in fact , a third actor is present .
the failure of tls to force proper authentication of bob allows mallory to masquerade as bob , to counterfeit his identity i.e. phishing attack .
you might argue that the two actor model is incorrect , that users check for the presence of the padlock symbol .
if they did then phishing attacks would not exist , or the conflict would escalate as outlined in section vi . fig .
9 shows a schematic of the existing two actor model and my three actor model . to distinguish between the actors in the two models
i have given them different names , as shown .
in the existing system alice - browser verifies the digital signature on bob s tls certificate . on success alice - browser and bob proceed to implement tls . in my model hal - browser
verifies the digital signature on bob s tls certificate . on success hal - browser
turns to alice - human and invites her to further authenticate bob .
he does this by displaying a window like that shown in fig .
the problem is : this act is vulnerable to counterfeiting . in this context counterfeiting
is referred to as a phishing attack . shown in fig .
10 is a picture of a turtle which is a shared secret between alice - human and hal - browser .
neither bob nor mallory know this secret .
as such mallory can not counterfeit fig . 10 without hacking into hal - browser to steal the secret .
hacking into thousands of computers to steal these secrets is an entirely different endeavour to tricking people into going to a fake website . once you correctly model the system as a three actor system .
cryptographers know how to appropriately authenticate the three participants .
as such fig .
10 is a relatively obvious step for cryptographers .
dhamija et al also use a user - display shared secret .
they use it to protect a dedicated login window from counterfeiting .
they do not appear to go beyond that and use it to present bob s identity credentials [ 4 ] . with my solution , by entering her login credentials
alice - human is accepting bob s identity credentials and her browser s shared secret .
she is authenticating both bob and her web browser .
hal - browser then proceeds to implement tls .
hence fig .
10 extends tls to ensure proper authentication of bob by alice - human .
alice - human now knows she is looking at a dialogue created by her web browser i.e. it is not a counterfeit , a phishing attack .
she can now examine the identity credentials presented and complete bob s authentication .
i was approaching this as a game theorist seeking screening strategies to prevent counterfeiting . here
follows an outline of the game theory interpretation .
anti - counterfeiting technologies and the screening strategy that accompany them go together like a lock and key pair .
the research involved the study of each category , from section ii , to find screening strategies which would prevent phishing attacks .
the definition of a screening strategy , from [ 2 ] is given since its language is used to frame the discussion that follows . from [ 2 ] : _ a screening strategy is a strategy used by a less informed player to elicit information from a more informed player . _ human interactive proofs ( captcha ) , turning tests and anti - counterfeiting technologies are all specific types of screening strategy . here too authentication , through the confirmation of a shared secret , constitutes a screening strategy .
the less informed player is eliciting the identity of the more informed player .
they are not eliciting the secret because they already know it .
they want to know do you know what the secret is? this is why it s just a point of view that this is cryptography . as a game theorist
i see a screening strategy .
it elicits their identity , as the individual who knows the secret or someone else .
furthermore , the fact that this works while other approaches fail indicates phishing attacks involve the counterfeiting of an identity , not a website .
this is significant because it allows us to prevent any type of counterfeiting .
it recasts counterfeiting as theft of intellectual property , patents , copyright , trademarks , designs etc .
accompanied by identity theft .
the purpose of the identity theft is to undermine law enforcement attempts which would otherwise prevent the intellectual property theft .
this means authentication based solutions can be developed for any type of counterfeiting including manufactured goods like pharmaceutical drugs and currencies .
references [ 3 ] and sitekey [ 5 ] both utilise cookies to trigger mutual authentication . in both cases ,
the cookie is used to authenticate hal - browser to bob .
bob then presents his shared secret , seeking authentication from alice - human .
this behaviour resides outside of the shared secret category of solutions , just discussed .
it s subtle , but this is a new category of behaviour which is resistant to counterfeiting . as a machine hal - browser can discern the difference between identical websites .
he can then treat these websites differently , even if we humans ca nt see the difference .
all solutions in this new category involve the web browser differentiating between websites and then responding in an installed software manner , which is not vulnerable to counterfeiting . here
follow our two examples : 1 .
browsers correctly returning cookies to the websites that created them .
2 . browsers correctly volunteering saved passwords for the websites they belong to . for the solution presented in fig .
10 our browser responds with installed software behaviour which is not vulnerable to counterfeiting .
however that response is not specific to each website .
both cookies and saved passwords require the browser to distinguish between websites and provide a response which is specific to that website . to develop a mechanism for bob to authenticate alice , consider the following thought experiments . on account setup
a bank can create a login page which will trigger the browser save password facility .
now the bank specifies both the username and password
. the password should be very high entropy , impossible to remember by alice - human .
the idea is that the browser will remember it .
since the browser can distinguish between websites it will never enter the data into a phishing website .
create a cookie with a high entropy code stored inside it .
this is equivalent to the password idea , just described i.e. the browser only hands over cookies to the websites that created them .
the cookie could be created with multi factor authentication e.g. a code could be sent to alice - human s mobile phone .
both solutions are equivalent and they are both vulnerable to the same kinds of attack .
i refer to this as cookie as certificate since these solutions use cookies as a poor man s certificate . to identify hal - browser to bob . both [ 3 ] and sitekey utilise cookies as certificates .
the thing is , they do nt have a secret private key and they do nt have a digital signature .
basically they are a number , a password .
all they have going for them is that the browser will only hand them over to the website that created them .
mallory simply attacks the cookie creation process .
while various solutions can be presented they tend to look like an arms race e.g. mallory could try a brute force attack generating lots of cookies .
bob will respond in the usual ways . on closer inspection ,
the asymmetric nature of tls makes it easier for alice - human to authenticate bob , than the other way around .
alice - human can trust trent and proceed accordingly . in bob s interactions with alice - human no third party ( trent ) is vouching for anyone . for bob to use a cookie
, he must create that cookie as part of an account registration process .
whatever process is used , it will always be vulnerable to attack by mallory unless alice - human correctly authenticates bob during that account registration . basically , without trent
, bob can never authenticate alice - human without her authenticating him first .
the cookie creation process will always involve alice - human authenticating bob or it will be vulnerable to mitm .
hence bob can not authenticate alice - human without her authenticating him first .
this makes bob authenticating alice - human less important than the other way around .
this is because tls is asymmetric .
alice - human can enjoy the use of tls certificates .
the user - browser secret combined with the tls certificate is fig .
10 . that s all alice - human needs .
if alice - human will always authenticate bob first then its academic how bob authenticates her .
codes , pictures , mobile phone , email , it s all the same .
he might as well just ask for a username and password .
it s up to her not to type her credentials into a phishing website .
the responsibility always falls on her to properly authenticate bob first .
after that it s academic if he accepts a username and password or some convoluted dance .
effectively this breaks the use of cookies as an anti - counterfeiting mechanism .
yes they are an effective solution .
however they are dependent upon tls and fig .
we re actually going in the wrong direction . instead of worrying about bob authenticating alice - human .
we should focus more on actually getting alice - human to authenticate bob .
the original idea was quite similar to [ 3 ] and sitekey , however it was abansoned , as outlined above .
basically an out of band , multi - factor authentication technique could be used to create a cookie e.g. send a text message to alice - human s phone .
12 would need to be used to create the cookie , just to ensure tls is actually being used .
even then it would still fail i.e. this system is vulnerable to all the attacks that work against sitekey and users would only see fig , 12 once every three , or four , years .
the reason why its a nice solution , why it s inclused here .
is that mallory would not be able to get the login screen up without the cookie .
so even if mallory has successfully harvested correct usernames and passwords , she would still need to create the cookie to get the login screen up .
its possible this solution will still be implemented , however , computer users still need to authenticate bob .
there is no getting away from this .
this is the cookie as username idea .
11 shows fig .
10 with this additional solution implemented . .
hence fig . 12 .
note the shared secret is now a photo of a ruler which lookes like a giraffe . ]
before i dismissed cookies i tried to overcome their shortcomings . fig .
12 is a mock - up of a browser created dialogue for the creation of a cookie .
sitekey , [ 3 ] and the cookie as username idea all require fig .
12 to ensure a mitm attack is not used . since alice - human
will only see this dialogue once every three or four years , this solution will fail . to overcome this problem
we need alice - human to be educated in the art . since it s not practical to force people to take lessons we need another way .
the counterfeiting of currencies is hindered through common knowledge i.e. people use genuine bank notes every day . when they encounter a new note their everyday knowledge will aid them in evaluating that note s authenticity . in the same way fig .
10 will become common knowledge .
effectively krugman s theory of passive learning [ 8 ] .
clearly passive learning and low involvement consumer behaviour are beyond the scope of this paper .
the ideas outlined here need to be combined with user studies like [ 7][9 ] and academic work in the field of education and passive learning . to educate alice - human , in the art
, we can make fig .
10 the second page , in a login wizard . the first page can be used to present messages which will slowly educate the user e.g. fig .
13 and fig . 14 .
effectively this passive learning approach will make the required knowledge , common knowledge .
the idea is to meet users half way i.e. we should adjust our authentication solution to make it more accessible . and we should aid users in its use . where flaws appear the solution can be assessed to ascertain why the system failed .
if the authentication mechanism is flawed we can alter tls , if users are being tricked by some aspect of the system , we can adjust both the educational messages and the login process till they can discern genuine from counterfeit .
dhamija et al found that some users did not know this type of fraud was possible .
some users accepted websites which contained professional looking images and distrusted websites which were plain i.e. few links and few images . some distrusted webpages which did not originate from the bank s main homepage [ 7 ] .
misconceptions like this can be addressed with informational messages on wizard page one .
informing and educating does not mean annoying , so a i know this check box could be added to remove educational messages that the user already knows .
eventually just the login page will appear or questions during setup could also remove it .
the existing system is unhelpful for many reasons .
users must remember to click on the padlock to examine the tls certificate . even then , they do nt really know if www.bobsonlinebanking.com should be used rather than www.bobsonlinebanking.ie .
trent should be an actual trusted third party .
the public probably do nt even know the name of any certificate authorities .
identities presented in tls certificates are de facto worldwide trademarks .
they are these businesses shopfront .
yet we put all of these brand names behind a picture of a padlock . on inspecting fig .
10 alice - human needs to remember three things , her user - browser shared secret , trent and bob . her secret will be easy because it s the same for every website .
once the phishers start attacking browser weaknesses associated with tls , alice - human will need to remember trent .
so why not make it easy for her ?
if central banks sell tls certificates to banks that operate within their regulation domain then those certificates can state that central bank s name as trent .
hence european central bank , federal reserve , bank of england , bank of japan etc . will all appear as trent in their respective countries .
now alice - human has less to remember .
regular ecommerce sites can continue as usual with tls certificates bought on the open market .
also the central banks can outsource the certificate creation process . so the certificate authorities will not be out of pocket
, they just wo nt have their name in lights on fig .
effectively this is an extension of the existing extended validation certificates idea .
the central banks would end up authenticating the identity of these banks .
this is precisely trent s role .
tell users to search for the padlock . with either commercially purchased tls certificates , or certificates they created themselves , the criminals can now display a padlock symbol .
users are now told to click on the padlock and examine the tls certificate details to verify bob s identity .
life is a lot more difficult for the phishers because they now need to trick users into full screen mode .
however , once there , they can counterfeit anything .
this gives them more options for theft / crime .
authentication windows like fig .
10 are introduced and users are directed to only enter their login credentials into these login windows .
hal - browser only displays the authentication window fig .
10 on successful verification of a tls certificate s digital signature .
hence mallory can not persuade hal - browser to display the window without her own tls certificate .
the criminals are likely to attack tls and browser weaknesses associated with tls certificates .
anti - virus , private keys , shared secrets all have a role to play .
alice - human s job is to complete bob s authentication .
a mixed strategy of educating alice , making details of the process common knowledge and altering the authentication process can be used to bring alice - human into the process . at the end of the day ,
only alice - human can authenticate bob .
phishing attacks are not the counterfeiting of a website but the counterfeiting of an identity .
the research found the system to be responsive to strategies which protect the website creator s identity , whereas website content could not be protected .
in fact , the research found that web browsers can counterfeit almost anything .
the asymmetric nature of tls makes alice - human authenticating bob more important than vice versa . without trent ,
bob authenticating alice - human will always require her to authenticate him first , during account setup .
otherwise a mitm attack will occur .
this requirement undermines solutions like sitekey .
the conclusion is that a mixed strategy is called for .
a user - browser shared secret allows tls to be extended so alice - human can complete bob s authentication .
the shared secret protects this act from counterfeiting , from phishing , since mallory can not counterfeit what mallory does not know .
the login process should be standardised into a predictable login window .
this makes details of the interaction common knowledge , guiding alice - human through the process .
furthermore educational messages displayed in a login wizard will educate the public by making this knowledge common knowledge .
this includes the possibility of central banks taking over the role of trent for financial institutions.the mixed strategy means meeting users half way. alice - human must fulfil her role and authenticate bob .
we must make this task as easy as possible .
protecting the authentication process from counterfeiting , through the use of a user - browser shared secret , using central banks for trent , standardising the login window , informing and educating users all bring us closer to alice fulfilling her role .
s. schechter , r. dhamija , a. ozment and i. fischer , the emperor s new security indicators , 2007 ieee symposium on security and privacy ( sp 07 ) , 2007 .
r. dhamija , j. tygar and m. hearst , why phishing works , proceedings of the sigchi conference on human factors in computing systems - chi 06 , 2006 .
p. kumaraguru , s. sheng , a. acquisti , l. f. cranor , and j. hong .
`` lessons from a real world evaluation of anti - phishing training '' e - crime researchers summit , anti - phishing working group , october 2008 . | phishing attacks occur because of a failure of computer users to authenticate bob .
the computer user s role , her job , is to authenticate bob .
nobody else can carry out this task .
i researched the ability of browsers to counterfeit the behaviour of installed software .
the objective was to find a signalling strategy which would protect against counterfeiting i.e. phishing attacks .
the research indicates that a user - browser shared secret can not be counterfeited because mallory can not counterfeit what mallory does not know .
after your browser has verified a tls certificate s digital signature the browser should create a two page login wizard .
the first page should display a random educational message , to inform and educate users about the process .
the second page will show the user ( 1 ) the user - browser shared secret , ( 2 ) the verified identity credentials from the tls certificate and ( 3 ) the input fields for the user to enter her login credentials .
the shared secret prevents counterfeiting , prevents phishing .
computer users can now authenticate bob by examining the tls certificate s identity credentials .
the educational messages will communicate to the user , the issues and pitfalls involved . on accepting bob , as bob , the user can enter her login credentials and login . phishing attacks , game theory , applied cryptography , authentication , security protocols , human factors . |
a possible relation between the central velocity dispersion of the spheroidal component ( @xmath1 ) and the galaxy circular velocity ( @xmath0 ) measured in the flat region of the rotation curve ( rc ) was suggested by whitmore et al .
( 1979 ) . by measuring stellar velocity dispersions and h i line widths for a sample of 19 spiral galaxies they found a a significant decrease in @xmath0@xmath3@xmath1 with increasing bulge - to - disk ratio .
since @xmath1 and @xmath0 probe the potential of the spheroid and dark matter ( dm ) halo , a mean value @xmath0@xmath3@xmath1
@xmath4 implies these components are dynamically separate with the bulge substantially cooler than halo .
gerhard et al .
( 2001 ) derived the @xmath0@xmath2 relation for the sample of giant ellipticals studied by kronawitter et al .
this was explained as an indication of near dynamical homology of these objects which were selected to be nearly round and almost non - rotating elliptical galaxies .
these galaxies form a unique dynamical family which scales with luminosity and effective radius . as a consequence the maximum @xmath0 of the galaxy
is correlated to its @xmath1 . whether the same is true for more flattened and fainter ellipticals is still to be investigated . on the contrary ,
both shape and amplitude of the rc of a spiral galaxy depend on the galaxy luminosity and morphological type ( e.g. , burstein & rubin 1985 ) . as a consequence for spiral galaxies the @xmath0@xmath2relation
is not expected a priori .
nevertheless , ferrarese ( 2002 ) and baes ( 2003 ) found that elliptical and spiral galaxies define a common @xmath0@xmath2 relation .
in particular , it results that for a given @xmath1 the value of @xmath0 is independent of the morphological type . but
@xmath1 and @xmath0 are related to the mass of the supermassive black hole ( hereafter smbh , see ferrarese & ford 2005 for a review ) and dm halo ( e.g. , seljak 2002 ) , respectively . therefore ferrarese ( 2002 )
argued that the @xmath0@xmath2relation is suggestive of a correlation between the mass of smbh and dm halo .
previous works concentrated on high surface brightness ( hsb ) galaxies .
it is interesting to investigate whether the @xmath0@xmath2relation holds also for less dense objects characterized by a less steep potential well .
this is the case of low surface brightness ( lsb ) galaxies , which are disc galaxies with a central face - on surface brightness @xmath5 mag arcsec@xmath6 ( e.g. , impey et al . 1996 ) . in pizzella
( 2005 ) we studied the behavior of lsb galaxies in the @xmath0@xmath2 relation . here
we present our results .
in the past years we started a scientific program aimed at deriving the detailed kinematics of ionized gas and stars in hsb and lsb galaxies in order to study their mass distribution and structural properties ( e.g. , pignatelli et al .
2001 ) .
we measured the velocity curves and velocity dispersion profiles along the major axis for both the ionized - gas and stellar components for a preliminary sample of @xmath7 hsb galaxies [ @xmath8 s0s0/a s in corsini et al .
( 2003 ) ; @xmath9 sa s in bertola et al .
( 1996 ) and corsini et al .
( 1999 ) ; @xmath10 s0sc s in vega beltran et al .
( 2001 ) ; @xmath11 sb
scd s in pizzella et al .
( 2004a ) ] and @xmath12 lsb galaxies ( pizzella et al .
the hsb sample consists of disc galaxies with hubble type ranging from s0 to scd , an inclination @xmath13 and a distance @xmath14 mpc .
the lsb sample consists of disc galaxies with hubble type ranging from sa to irr , an intermediate inclination ( @xmath15 ) , and a distance @xmath16 mpc . in order to complete our sample of disc galaxies we included the 38 sa - scd galaxies previously studied by ferrarese ( 2002 ) and the 12 sb - sc galaxies by baes et al .
( 2003 ) .
finally , we considered a sample of 24 elliptical galaxies with a flat rc and for which both @xmath0 and @xmath1 are available from the literature .
they include 19 objects studied by kronawitter et al .
( 2000 ) who derived @xmath0 by dynamical modeling and 5 objects studied by bertola et al .
( 1993 ) for which @xmath0 is directly derived from the flat portion of their h i rcs .
the addition of these galaxies will allow to test against model - dependent biases in the @xmath0@xmath2 relation .
for all the disc galaxies we obtained the ionized - gas rc .
we rejected galaxies with asymmetric rcs or rcs which were not characterized by an outer flat portion .
the flatness of each rc has been checked by fitting it with a linear law @xmath17 for @xmath18 .
the radial range has been chosen in order to avoid the bulge - dominated region .
the rcs with @xmath19 @xmath20 within 3@xmath21 have been considered not to be flat .
we derived @xmath0 by averaging the outermost values of the flat portion of the rc .
we derived @xmath1 from the stellar kinematics by extrapolating the velocity dispersion radial profile to the centre .
this has been done by fitting the innermost data points with an empirical function ( either an exponential , or a gaussian or a constant ) .
we did not apply any aperture correction to @xmath1 as discussed by pizzella et al .
( 2004a ) . for the elliptical galaxies we obtained @xmath0 either from the dynamical models by gerhard et al .
( 2001 ) or from the flat portion of the h irc .
for these galaxies we relaxed the flatness criterion in favor of their large radial extension which is about 10 times larger than that of optical rcs .
aperture measurements of the stellar velocity dispersion corrected to @xmath22 were adopted to estimate @xmath1 .
in summary , we have @xmath23 disc galaxies [ @xmath24 from our preliminary sample , @xmath10 from ferrarese ( 2002 ) , and @xmath25 from baes et al . ( 2003 ) with flat rcs extending to @xmath26 and @xmath27 elliptical galaxies [ @xmath28 from kronawitter et al .
( 2000 ) with flat rcs extending to about @xmath29 , @xmath30 from bertola et al .
( 1993 ) with flat h i rcs extending to @xmath31 .
a power law has been usually adopted to describe the correlation between @xmath0 and @xmath1 for galaxies with @xmath1@xmath32 @xmath33 .
we find @xmath34 with @xmath0 and @xmath1 expressed in @xmath33 .
the resulting power law is plotted in fig .
1 ( dash - dotted line ) . however , according to a @xmath35 analysis the data are also consistent with a linear law out to velocity dispersions as low as @xmath1@xmath36 @xmath33@xmath37 the resulting straight line is plotted in fig . 1 ( continuous line ) .
the data points corresponding to the 5 elliptical galaxies with @xmath0based on h i data follow the same relation as the remaining disc and elliptical galaxies .
they are mostly located on the upper end of the @xmath0@xmath2 relation derived for disc galaxies , in agreement with the findings of bertola et al .
they studied these elliptical galaxies and showed that their dm content and distribution are similar to those of spiral galaxies . and @xmath1 .
the data points corresponding to hsb galaxies ( _ filled circles _ ) , lsb galaxies ( _ squares _ ) , and elliptical galaxies with @xmath0 obtained from h i data ( _ diamonds _ ) or dynamical models ( _ crosses _ ) are shown .
the _ dash - dotted _ and _ continuous line _ represent the power - law ( eq .
1 ) and linear fit ( eq .
2 ) to hsb and elliptical galaxies .
the _ dotted line _ represents the linear - law fit ( eq .
3 ) to lsb galaxies . for a comparison , the _ dashed line _ corresponds to the power - law fit to spiral galaxies with @xmath38 km s@xmath39 by baes et al .
( 2003).,width=359 ]
the lsb and hsb galaxies do not follow the same @xmath0@xmath2 relation .
in fact , most of the lsb galaxies are characterized by a higher @xmath0 for a given @xmath1 ( or a lower @xmath1 for a given @xmath0 ) with respect to hsb galaxies . by applying to the lsb data
points the same regression analysis which has been adopted for the hsb and elliptical galaxies of the final sample , we find @xmath40 with @xmath0 and @xmath1 expressed in @xmath33 .
the straight line corresponding to this fit , which is different from the one obtained for hsb and elliptical galaxies and happens to be parallel to it , is plotted in fig .
1 ( dotted line ) . to address the significance of this result , which is based only on 8 data points , we compared the distribution of the normalized scatter of the lsb galaxies to that of the hsb and elliptical galaxies .
they are different at a high confidence level ( @xmath41 ) by a kolmogorov - smirnov test .
the fact that these objects fall in a different region of the @xmath0@xmath2 plane confirms that lsb and hsb galaxies constitute two different classes of galaxies .
both demographics of smbhs and study of dm distribution in galactic nuclei benefit from the @xmath0@xmath2 relation .
the recent finding that the mass of smbhs correlates with different properties of the host spheroid supports the idea that formation and accretion of smbhs are closely linked to the formation and evolution of their host galaxy .
a task to be pursued is to obtain a firm description of all these relationships spanning a wide range of smbh masses and address if they hold for all hubble types .
in fact , the current demography of smbhs suffers of important biases , related to the limited sampling over the different basic properties of their host galaxies .
the finding that the @xmath0@xmath2 relation holds for small values of @xmath1 points to the idea that smbhs with masses m@xmath42 m@xmath43 may also exist and follow the m@xmath44-@xmath21 relation .
moreover , it has been suggested that the @xmath0@xmath2 relation is equivalent to one between the masses of smbh and dm halo ( ferrarese 2002 ; baes 2003 ) because @xmath1 and @xmath0 are related to the masses of the central smbh and dm halo , respectively . yet , this claim is to be considered with caution , as the demography of smbhs is still limited , in particular as far as spiral galaxies are concerned .
furthermore , the calculation of the virial mass of the dm halo from the measured @xmath0 depends on the assumptions made for the dm density profile and the resulting rotation curve ( e.g. , see the prescriptions by seljack et al .
2002 ) . a better estimate of the virial velocity of the dm halo @xmath45 can be obtained by constraining the baryonic - to - dark matter fraction with detailed dynamical modeling of the sample galaxies .
the resulting @xmath45@xmath46@xmath1 relation is expected to have a smaller scatter than the @xmath0@xmath2 relation .
if the m@xmath44-@xmath21 relation is to hold , the deviation of lsb galaxies _ with bulge _ from the @xmath0@xmath2 of hsb and elliptical galaxies suggests that for a given dm halo mass the lsb galaxies would host a smbh with a smaller mass compared to hsb galaxies . on the other hand ,
if the fundamental correlation of smbh mass is with the halo @xmath0 , then lsbs should have larger black - hole masses for given bulge @xmath1 .
this should be accounted for in the theoretical and numerical investigations of the processes leading to the formation of lsb galaxies .
the collapse of baryonic matter can induce a further concentration in the dm distribution , and a deepening of the overall gravitational well in the central regions . if this is the case , the finding that at a given dm mass ( as traced by @xmath0 ) the central @xmath1 of lsb galaxies is smaller than in their hsb counterparts , would argue against the relevance of baryon collapse in the radial density profile of dm in lsb galaxies . | we analyzed a sample of high and low surface brightness ( hsb and lsb ) disc galaxies and elliptical galaxies to investigate the correlation between the circular velocity ( @xmath0 ) and the central velocity dispersion ( @xmath1 ) .
we better defined the previous @xmath0@xmath2 correlation for hsb and elliptical galaxies , especially at the lower end of the @xmath1 values .
elliptical galaxies with @xmath0 based on dynamical models or directly derived from the h i rotation curves follow the same relation as the hsb galaxies in the @xmath0@xmath2 plane . on the contrary ,
the lsb galaxies follow a different relation , since most of them show either higher @xmath0 ( or lower @xmath1 ) with respect to the hsb galaxies .
this argues against the relevance of baryon collapse in the radial density profile of the dark matter haloes of lsb galaxies .
moreover , if the @xmath0@xmath2 relation is equivalent to one between the mass of the dark matter halo and that of the supermassive black hole , these results suggest that the lsb galaxies host a supermassive black hole with a smaller mass compared to hsb galaxies of equal dark matter halo . on the other hand ,
if the fundamental correlation of smbh mass is with the halo @xmath0 , then lsbs should have larger black hole masses for given bulge @xmath1 .
galaxies : elliptical and lenticular , cd galaxies : fundamental parameters galaxies : kinematics and dynamics galaxies : spirals |
inferring macroscopic properties of physical systems from their microscopic description is an ongoing work in many disciplines of physics , like condensed matter , ultra cold atoms or quantum chromo dynamics .
the most drastic changes in the macroscopic properties of a physical system occur at phase transitions , which often involve a symmetry breaking process .
the theory of such phase transitions was formulated by landau as a phenomenological model @xcite and later devised from microscopic principles using the renormalization group @xcite .
one can identify phases by knowledge of an order parameter which is zero in the disordered phase and nonzero in the ordered phase .
whereas in many known models the order parameter can be determined by symmetry considerations of the underlying hamiltonian , there are states of matter where such a parameter can only be defined in a complicated non - local way @xcite .
these systems include topological states like topological insulators , quantum spin hall states @xcite or quantum spin liquids @xcite .
therefore , we need to develop new methods to identify parameters capable of describing phase transitions . such methods might be borrowed from machine learning . since the 1990s this field has undergone major changes with the development of more powerful computers and artificial neural networks .
it has been shown that such neural networks can approximate every function under mild assumptions @xcite .
they quickly found applications in image classification , speech recognition , natural language understanding and predicting from high - dimensional data .
furthermore , they began to outperform other algorithms on these tasks @xcite . in the last years physicists started to employ machine learning techniques .
most of the tasks were tackled by supervised learning algorithms or with the help of reinforcement learning @xcite . supervised
learning means one is given labeled training data from which the algorithm learns to assign labels to data points . after successful training
it can then predict the labels of previously unseen data with high accuracy .
in addition to supervised learning , there are unsupervised learning algorithms which can find structure in unlabeled data .
they can also classify data into clusters , which are however unlabelled .
it is already possible to employ unsupervised learning techniques to reproduce monte - carlo - sampled states of the ising model @xcite .
phase transitions were found in an unsupervised manner using principal component analysis @xcite .
we employ more powerful machine learning algorithms and transition to methods that can handle nonlinear data .
a first nonlinear extension is kernel principal component analysis @xcite .
the first versions of autoencoders have been around for decades @xcite and were primarily used for dimensional reduction of data before feeding it to a machine learning algorithm .
they are created from an encoding artificial neural network , which outputs a latent representation of the input data , and a decoding neural network that tries to accurately reconstruct the input data from its latent representation .
very shallow versions of autoencoders can reproduce the results of principal component analysis @xcite . in 2013 ,
variational autoencoders have been developed as one of the most successful unsupervised learning algorithms @xcite .
in contrast to traditional autoencoders , variational autoencoders impose restrictions on the distribution of latent variables .
they have shown promising results in encoding and reconstructing data in the field of computer vision . in this work
we use unsupervised learning to determine phase transitions without any information about the microscopic theory or the order parameter .
we transition from principal component analysis to variational autoencoders , and finally test how the latter handles different physical models .
our algorithms are able to find a low dimensional latent representation of the physical system which coincides with the correct order parameter .
the decoder network reconstructs the encoded configuration from its latent representation .
we find that the reconstruction is more accurate in the ordered phase , which suggests the use of the reconstruction error as a universal identifier for phase transitions .
whereas for physicists this work is a promising way to find order parameters of systems where they are hard to identify , computer scientists and machine learning researchers might find an interpretation of the latent parameters .
the ising model is one of the most - studied and well - understood models in physics .
whereas the one - dimensional ising model does not possess a phase transition , the two - dimensional model does .
the hamiltonian of the ising model on the square lattice with vanishing external magnetic @xmath0 field reads @xmath1 with uniform interaction strength @xmath2 and discrete spins @xmath3 on each site @xmath4 .
the notation @xmath5 indicates a summation over nearest neighbors . a spin configuration @xmath6 is a fixed assignment of a spin to each lattice site
, @xmath7 denotes the set of all possible configurations @xmath8 .
we set the boltzmann constant @xmath9 and the interaction strength @xmath10 for the ferromagnetic case and @xmath11 for the antiferromagnetic case .
a spin configuration @xmath8 can be expressed in matrix form as @xmath12 lars onsager solved the two dimensional ising model in 1944 @xcite .
he showed that the critical temperature is @xmath13 .
for the purpose of this work , we assume a square lattice with length @xmath14 such that @xmath15 , and periodic boundary conditions .
we sample the ising model using a monte - carlo algorithm @xcite at temperatures @xmath16 $ ] to generate @xmath17 samples in the ferromagnetic case and @xmath18 samples in the antiferromagnetic case .
the ising model obeys a discrete @xmath19-symmetry , which is spontaneously broken below @xmath20 .
the magnetization of a spin sample is defined as @xmath21 the partition function @xmath22 allows us to define the corresponding order parameter .
it is the expectation value of the absolute value of the magnetization at fixed temperature @xmath23 similarly , with the help of the matrix @xmath24 , we define the order parameter , as the expectation value of the staggered magnetization .
the latter is calculated from an element - wise product with a matrix form of the spin configurations @xmath25 the mermin - wagner - hohenberg theorem @xcite prohibits continuous phase transitions in @xmath26 dimensions at finite temperature when all interactions are sufficiently short - ranged .
hence , we choose the xy model in three dimensions as a model to probe the ability of a variational autoencoder to classify phases of models with continuous symmetries .
the hamiltonian of the xy model reads @xmath27 with spins on the one - sphere @xmath28 .
employing @xmath10 , the transition temperature of this model is @xmath29 @xcite using a cubic lattice with @xmath30 , such that @xmath31 , we perform monte - carlo simulations to create 10000 independent sample spin configurations in the temperature range of @xmath32 $ ] .
the order parameter is defined analogously to the ising model magnetization , but with the @xmath33-norm of a magnetization consisting of two components .
_ principal component analysis
_ @xcite is an orthogonal linear transformation of the data to an ordered set of variables , sorted by their variance .
the first variable , which has the largest variance , is called the first principal component , and so on . the linear function @xmath34 , which maps a collection of spin samples @xmath35 to its first principal component , is defined as @xmath36 \ , \label{eq : pca}\end{aligned}\ ] ] where @xmath37 is the vector of mean values of each spin averaged over the whole dataset .
further principal components are obtained by subtracting the already calculated principal components and repeating .
_ kernel principal component analysis _
@xcite projects the data into a kernel space in which the principal component analysis is then performed . in this work
the nonlinearity is induced by a radial basis functions kernel . _
traditional neural network - based autoencoders _ consist of two artificial neural networks stacked on top of each other .
the encoder network is responsible for encoding the input data into some latent variables .
the decoder network is used to decode these parameters in order to return an accurate recreation of the input data , shown in .
the parameters of this algorithm are trained by performing gradient descent updates in order to minimize the reconstruction loss ( reconstruction error ) between input data and output data .
_ variational autoencoders _ are a modern version of autoencoders which impose additional constraints on the encoded representations , see latent variables in .
these constraints transform the autoencoder to an algorithm that learns a latent variable model for its input data . whereas the neural networks of traditional autoencoders learn an arbitrary function to encode and decode the input data , variational autoencoders learn the parameters of a probability distribution modeling the data . after learning the probability distribution , one can sample parameters from it and then let the encoder network generate samples closely resembling the training data . to achieve this
, variational autoencoders employ the assumption that one can sample the input data from a unit gaussian distribution of latent parameters .
the weights of the model are trained by simultaneously optimizing two loss functions , a reconstruction loss and the kullback - leibler divergence between the learned latent distribution and a prior unit gaussian . in this work
we use autoencoders and variational autoencoders @xcite with one fully connected hidden layer in the encoder as well as one fully connected hidden layer in the decoder , each consisting of 256 neurons .
the number of latent variables is chosen to match the model from which we sample the input data .
the activation functions of the intermediate layers are rectified linear units .
the activation function of the final layer is a _ sigmoid _ in order to predict probabilities of spin @xmath38 or @xmath39 in the ising model , or _
tanh _ for predicting continuous values of spin components in the xy model .
we do not employ any @xmath40 or dropout regularization .
however , we tune the relative weight of the two loss functions of the variational autoencoder to fit the problem at hand .
the kullback - leibler divergence of the variational autoencoder can be regarded as reguarization of the traditional autoencoder . in our autoencoder
the reconstruction loss is the cross - entropy loss between the input and output probability of discrete spins , as in the ising model .
the reconstruction loss is the mean - squared - error between the input and the output data of continuous spin variables in the xy model . to understand why a variational autoencoder can be a suitable choice for the task of classifying phases , we recall what happens during training .
the weights of the autoencoder learn two things : on the one hand , they learn to encode the similarities of all samples to allow for an efficient reconstruction . on the other hand , they learn a latent distribution of the parameters which encode the most information possible to distinguish between different input samples .
let us translate these considerations to the physics of phase transitions .
if all the training samples are in the unordered phase , the autoencoder learns the common structure of all samples .
the autoencoder fails to learn any random entropy fluctuations , which are averaged out over all data points .
however , in the ordered phase there exists a common order in samples belonging into the same phase .
this common order translates to a nonzero latent parameter , which encodes correlations on each input sample .
it turns out that in our cases this parameter is the order parameter corresponding to the broken symmetry .
it is not necessary to find a perfect linear transformation between the order parameter and the latent parameter as it is the case in .
a one - to - one correspondence is sufficient , such that one is able to define a function that maps these parameters onto each other and captures all discontinuities of the derivatives of the order parameter .
we point out similarities between principal component analysis and autoencoders .
although both methods seem very different , they both share common characteristics .
principal component analysis is a dimensionality reduction method which finds the linear projections of the data that maximizes the variance .
reconstructing the input data from its principal components minimizes the mean squared reconstruction error .
although the principal components do not need to follow a gaussian distribution , principal components have the highest mutual agreement with the data if it emerges from a gaussian prior .
moreover , a single layer autoencoder with linear activation functions closely resembles principal component analysis @xcite .
principal component analysis is much easier to apply and in general uses less parameters than autoencoders .
however , it scales very badly to a large dataset .
autoencoders based on convolutional layers can reduce the number of parameters . in extreme cases
this number can be even less than the parameters of principal component analysis .
furthermore , such autoencoders can promote locality of features in the data .
+ + + + + pixels , from the latent parameter .
the brightness indicates the probability of the spin to be up ( white : @xmath41 , black : @xmath42 ) . the first row is a reconstruction of sample configurations from the ferromagnetic ising model .
the second row corresponds to the antiferromagnetic ising model .
the third row is the prediction from the af latent parameter , where each second spin is multiplied by @xmath43 , to show that the second row indeed predicts an antiferromagnetic state.,title="fig:",scaledwidth=50.0% ] + +
the four different algorithms can be applied to the ising model to determine the role of the first principal components or the latent parameters .
shows a clear correlation between these parameters and the magnetization for all four methods .
however , the traditional autoencoder is inaccurate ; this fact leads us to enhancing traditional autoencoders to variational autoencoders .
the principal component methods show the most accurate results , slightly better than the variational autoencoder .
this is to be expected , since the former are modeled by fewer parameters . in the following results section , we concentrate on the variational autoencoder as the most advanced algorithm for unsupervised learning . to begin with ,
we choose the number of latent parameters in the variational autoencoder to be one . after training for 50 epochs and a saturation of the training loss ,
we visualize the results in . on the left
, we see a close linear correlation between the latent parameter and the magnetization . in the middle
we see a histogram of encoded spin configurations into their latent parameter .
the model learned to classify the configurations into three clusters .
having identified the latent parameter to be a close approximation to the magnetization @xmath44 allows us to interpret the properties of the clusters .
the right and left clusters in the middle image correspond to an average magnetization of @xmath45 , while the middle cluster corresponds to the magnetization @xmath46 . employing a different viewpoint , from we
conclude that the parameter which holds the most information on how to distinguish ising spin samples is the order parameter . on the right panel , the average of the magnetization , the latent parameter and the reconstruction loss are shown as a function of the temperature .
a sudden change in the magnetization at @xmath47 defines the phase transition between paramegnetism and ferromagnetism . even without knowing this order parameter
, we can now use the results of the autoencoder to infer the position of the phase transition . as an approximate order parameter
, the average absolute value of latent parameter also shows a steep change at @xmath20 .
the averaged reconstruction loss also changes drastically at @xmath20 during a phase transition .
while the latent parameter is different for each physical model , the reconstruction loss can be used as a universal parameter to identify phase transitions . to summarize , without any knowledge of the ising model and its order parameter , but sample configurations
, we can find a good estimation for its order parameter and the occurrence of a phase transition .
it is a priori not clear how to determine the number of latent neurons in the creation of the neural network of the autoencoder . due to the lack of theoretical groundwork ,
we find the optimal number by experimenting .
if we expand the number of latent dimensions by one , see , the results of our analysis only change slightly .
the second parameter contains a lot less information compared to the first , since it stays very close to zero .
hence , for the ising model , one parameter is sufficient to store most of the information of the latent representation . while the ferromagnetic ising model serves as an ideal starting ground , in the next step we are interested in models where different sites in the samples contribute in a different manner to the order parameter .
we do this in order to show that our model is even sensitive to structure on the smallest scales . for the magnetization in the ferromagnetic ising model ,
all spins contribute with the same weight .
in contrast , in the antiferromagnetic ising model , neighboring spins contribute with opposite weight to the order parameter .
again the variational autoencoder manages to capture the traditional order parameter .
the staggered magnetization is strongly correlated with the latent parameter , see .
the three clusters in the latent representation make it possible to interpret different phases .
furthermore , we see that all three averaged quantities - the magnetization , the latent parameter and the reconstruction loss - can serve as indicators of a phase transition . demonstrates the reconstruction from the latent parameter . in the first row
we see the reconstruction from samples of the ferromagnetic ising model , the latent parameter encodes the whole spin order in the ordered phase .
reconstructions from the antiferromagnetic ising model are shown in the second and third row .
since the reconstructions clearly show an antiferromagnetic phase , we infer that the autoencoder encodes the spin samples even to the most microscopic level . in the xy model
we examine the capabilities of a variational autoencoder to encode models with continuous symmetries . in models like the ising model , where discrete symmetries are present , the autoencoder only needs to learn a discrete set , which is often finite , of possible representations of the symmetry broken phase .
if a continuous symmetry is broken , there are infinitely many possibilities of how the ordered phase can be realized .
hence , in this section we test the ability of the autoencoder to embed all these different realizations into latent variables .
the variational autoencoder handles this model equally well as the ising model .
we find that two latent parameters model the phase transition best . the latent representation in the middle of shows the distribution of various states around a central cluster .
the radial symmetry in this distribution leads to the assumption that a sensible order parameter is constructed from the @xmath33-norm of the latent parameter vector . in
, one sees the correlation between the magnetization and the absolute value of the latent parameter vector . averaging the samples for the same temperature hints to the facts that the latent parameter and the reconstruction loss can serve as an indicator for the phase transition .
we have shown that it is possible to observe phase transitions using unsupervised learning .
we compared different unsupervised learning algorithms ranging from principal component analysis to variational autoencoders and thereby motivated the need for the upgrade of the traditional autoencoder to a variational autoencoder .
the weights and latent parameters of the variational autoencoder approach are able to store information about microscopic and macroscopic properties of the underlying systems .
the most distinguished latent parameters coincide with the known order parameters .
furthermore , we have established the reconstruction loss as a new universal indicator for phase transitions . we have expanded the toolbox of unsupervised learning algorithms in physics by powerful methods , most notably the variational autoencoder , which can handle nonlinear features in the data and scale very well to huge datasets . using these techniques
, we expect to predict unseen phases or uncover unknown order parameters , e.g. in quantum spin liquids .
we hope to develop deep convolutional autoencoders which have a reduced number of parameters compared to fully connected autoencoders and can also promote locality in feature selection . furthermore , since there exists a connection between deep neural networks and renormalization group @xcite , it may be helpful to employ deep convolutional autoencoders to further expose this connection . _ acknowledgments _ we would like to thank timo milbich , bjrn ommer , michael scherer , manuel scherzer and christof wetterich for useful discussions .
we thank shirin nkongolo for proofreading the manuscript .
s.w . acknowledges support by the heidelberg graduate school of fundamental physics .
stefano curtarolo , dane morgan , kristin persson , john rodgers , and gerbrand ceder .
predicting crystal structures with data mining of quantum calculations . _ physical review letters _ , 910 ( 13 ) ,
doi : 10.1103/physrevlett.91.135503 .
matthias rupp , alexandre tkatchenko , klaus - robert mller , and o. anatole von lilienfeld .
fast and accurate modeling of molecular atomization energies with machine learning .
_ physical review letters _ , 1080 ( 5 ) , jan 2012 .
doi : 10.1103/physrevlett.108.058301 .
zhenwei li , james r. kermode , and alessandro de vita .
molecular dynamics with on - the - fly machine learning of quantum - mechanical forces .
_ physical review letters _
, 1140 ( 9 ) , mar 2015 .
doi : 10.1103/physrevlett.114.096405 .
erin ledell , prabhat , dmitry yu .
zubarev , brian austin , and william a. lester
. classification of nodal pockets in many - electron wave functions via machine learning .
_ journal of mathematical chemistry _ , 500 ( 7):0 20432050 , may 2012 .
doi : 10.1007/s10910 - 012 - 0019 - 5 .
g. pilania , j. e. gubernatis , and t. lookman .
structure classification and melting temperature prediction in octet ab solids via machine learning .
_ physical review b _ , 910 ( 21 ) , jun 2015 .
doi : 10.1103/physrevb.91.214302 .
yousef saad , da gao , thanh ngo , scotty bobbitt , james r. chelikowsky , and wanda andreoni .
data mining for materials : computational experiments withabcompounds . _
physical review b _ , 850 ( 10 ) , mar 2012 .
doi : 10.1103/physrevb.85.104104 .
o. s. ovchinnikov , s. jesse , p. bintacchit , s. trolier - mckinstry , and s. v. kalinin .
disorder identification in hysteresis data : recognition analysis of the random - bondrandom - field ising model .
_ physical review letters _ , 1030 ( 15 ) , oct 2009 .
doi : 10.1103/physrevlett.103.157203 .
louis - franois arsenault , alejandro lopez - bezanilla , o. anatole von lilienfeld , and andrew j. millis .
machine learning for many - body physics : the case of the anderson impurity model . _
physical review b _ , 900 ( 15 ) , oct 2014 .
doi : 10.1103/physrevb.90.155136 .
john c. snyder , matthias rupp , katja hansen , klaus - robert mller , and kieron burke .
finding density functionals with machine learning . _ physical review letters _ , 1080 ( 25 ) , jun 2012 .
doi : 10.1103/physrevlett.108.253002 .
geoffroy hautier , christopher c. fischer , anubhav jain , tim mueller , and gerbrand ceder .
finding nature s missing ternary oxide compounds using machine learning and density functional theory . _
chemistry of materials _ , 220 ( 12):0 37623767 , jun 2010 .
doi : 10.1021/cm100795d .
pierre baldi and kurt hornik .
neural networks and principal component analysis : learning from examples without local minima . _ neural networks _ , 20 ( 1):0 5358 , jan 1989 .
doi : 10.1016/0893 - 6080(89)90014 - 2 .
n. d. mermin and h. wagner .
absence of ferromagnetism or antiferromagnetism in one- or two - dimensional isotropic heisenberg models .
_ physical review letters _ , 170 ( 22):0 11331136 , nov 1966 .
doi : 10.1103/physrevlett.17.1133 .
aloysius p. gottlob and martin hasenbusch .
critical behaviour of the 3d xy - model : a monte carlo study .
_ physica a : statistical mechanics and its applications _
, 2010 ( 4):0 593613 , dec 1993 .
doi : 10.1016/0378 - 4371(93)90131-m . | we employ unsupervised machine learning techniques to learn latent parameters which best describe states of the two - dimensional ising model and the three - dimensional xy model .
these methods range from principal component analysis to artificial neural network based variational autoencoders .
the states are sampled using a monte - carlo simulation above and below the critical temperature .
we find that the predicted latent parameters correspond to the known order parameters .
the latent representation of the states of the models in question are clustered , which makes it possible to identify phases without prior knowledge of their existence or the underlying hamiltonian .
furthermore , we find that the reconstruction loss function can be used as a universal identifier for phase transitions . |
main laws of quantum mechanics forbids the perfect cloning of the quantum states , see corresponding discussion for the pure states in @xcite , @xcite , and for the mixed states in @xcite .
but it is possible to carry out an approximate copying of the quantum states @xcite .
quantum cloning machines ( qcm ) depend on the conditions accepted at its designing .
they can produce identical copies of the initial state ( symmetric qcm ) , nonidentical copies ( non - symmetric qcm ) , the quality of the copying can be either identical for all states ( universal qcm ) or depend on the state ( state - dependente qcm ) .
detailed discussion of the different variants of qcm and theirs possible applications in quantum cryptography and quantum informatics can be found in @xcite , @xcite .
one possible application of the qcm is an eavesdropping of the quantum channel .
the aim of such eavesdropping defines the main properties of the designing qcm .
one can design qcm which copies only part of the quantum state , for instance .
such qcm can be useful if eavesdropper , usually called eve , intends to catch part of the transmitted quantum information only .
some classical analogue of this situation can be classical eavesdropping of the key words in the transmitted classical information . at quantum cloning
we can choose the different parts of the quantum signal in which we are interested .
in this paper we intend to discuss some `` partial '' qcm , which copies one constituent of the two - partite states .
our approach gives the possibility to consider qcm for a mixed states too .
it is well known fact , that any mixed state can be considered as a reduction of a pure state , which is called `` purification '' of the mixed state @xcite .
so , cloning of the mixed state can be considered as a `` partial '' cloning of the `` purification '' of the mixed state .
some difference between the `` partial '' cloning machine and the cloning machine for the mixed states is connected with the corresponding difference of the sets of the initial states , see details below .
note , that the main attention in the present literature was devoted to the cloning of the pure states @xcite , @xcite .
we consider two - partite qubit states , qubits are elements of two - dimensional hilbert space @xmath0 . in order to construct qcm we need in tensor product of three such spaces on the ancilla space : @xmath1 , here different components are marked by indexes .
the first and third qubit components constitute a quantum state which carries information in the quantum channel , and the state of first component is interesting for eve .
the second component is a blank state , where we will copy the first component , the last component is necessary for the realization of the qcm .
let quantum channel carries the quantum state @xmath2 @xmath3 where normalization condition holds , @xmath4 here and below @xmath5 are base vectors in @xmath6 .
we suppose , that eve s goal is a copying of the first component of this state . after tracing one can obtain : @xmath7 @xmath8 so , eve has to realize the cloning to produce the pair of states ( in the first and second components respectively ) closest to @xmath9 we consider here symmetric qcm , so , we suppose , that states in the first and second components have to coincide . then produced state must be symmetric with regard to permutation of the first and second components .
let us introduce the orthonormal basis in the subspace of @xmath10 symmetric regarding this permutation : @xmath11 let s assume , that the second component be in state @xmath12 initially .
description of the qcm is , in essence , the definition of the corresponding unitary operator @xmath13 .
following to @xcite , @xcite , we set @xmath14 @xmath15 where @xmath16 are some vectors , belonging to @xmath17 .
symmetry of qcm is provided by the fact , that right - hand part of this relation contains linear combinations of vectors @xmath18 only .
taking into account ( [ init ] ) , we obtain : @xmath19 @xmath20 @xmath21 generally speaking , the choice of the unitary operator @xmath13 is very broad and corresponding analysis is quite complex even for the lowest dimensions , so usually one admits some additional restrictions .
we suppose as in @xcite , that following conditions ( which guarantee the unitarity of @xmath13 ) are fulfilled : @xmath22 @xmath23 let @xmath24 so as @xmath25 , @xmath26 . in this case qcm
produces the next state from @xmath27 : @xmath28 @xmath29 @xmath30 reducing this state on the first component , we obtain : @xmath31 @xmath32 it is necessary to compare the initial state and state which is produced by the qcm , in other words , we have to choose the measure of the closeness of these states .
there are different measures , specifically , fidelity .
it is defined for the mixed states as @xmath33 ^2 $ ] , this value is not very suitable for the analytical considerations .
we use here more convenient measure : @xmath34 ^2 = w(\zeta , \nu , \psi ) , \ ] ] where @xmath35 + \ ] ] @xmath36 @xmath37 .\ ] ] this value estimates the difference between initial and final states with fixed parameters @xmath38 . for the determination of the qcm parameters we average this value respect to the set of all initial states .
we use here the next parametrization of the initial state @xmath39 : @xmath40 @xmath41 where @xmath42 here the first component has zero phase due to the corresponding freedom of the choice . for the averaging we need in corresponding measure .
supposing that all states @xmath43 are equiprobable , we choose as such a measure @xmath44 @xmath45 simple calculations lead to the conclusion , that @xmath46 takes its minimal value at @xmath47 , @xmath48 .
this value @xmath49 implies , that vectors @xmath50 and @xmath51 are parallel .
the values of the fidelity @xmath33 ^2 $ ] , calculated at @xmath52 , for the states on the `` real '' part of the bloch sphere , @xmath53 @xmath54 , @xmath55 , are plotted in fig.1 .
evidently , that this qcm is state - dependent one , because the quality of the cloning depends on the quantum state .
the construction described above can be used for the cloning of the mixed states .
note , that final state produced by qcm depend on the reduction of the initial state @xmath56 ( [ red ] ) only .
it means , that one can reverse our considerations and take the mixed state @xmath56 as initial one .
then pure state @xmath43 defined by relation ( [ init ] ) belongs to the space of the larger dimension and it is a `` purification '' of the state @xmath56 .
`` purification '' of the given mixed state @xmath56 can be realized by different methods , as it follows from ( [ init ] ) , but this nonuniqueness has not influence in the results .
qcm constructed in accordance with relations ( [ ccm1 ] ) , ( [ ccm2 ] ) produces states ( [ out12 ] ) , ( [ out ] ) , which depend on the parameters of the initial state @xmath56 only .
but the set of the initial states is changing , one has to use another parametrization for this set .
as such parametrization one can take a bloch sphere , see @xcite .
namely , density matrix @xmath56 can be described as @xmath57 where @xmath58 in the spherical coordinates we have : @xmath59 @xmath60 and @xmath61 in order to obtain the parameters of the qcm one has to average the value @xmath62 on the bloch sphere .
we suppose that all states in the bloch sphere are equiprobable , so averaging is reduced to the integral @xmath63 note , that if eve has a priori information about transmitted quantum information she has to choose corresponding weight multiplier .
we are searching in such values @xmath64 which correspond to the minimal value @xmath65 . as a result
we obtain @xmath66 .
the plot of fidelity for these parameters differs in a small way from the preceding one , so we omit it .
we have discussed here a `` partial '' quantum cloning , when only one component of the two - partite pure state is cloning . such cloning can be considered as a variant of the eavesdropping of the quantum channel .
the choice of the parameters of the qcm was realized with help of the some natural criterion .
namely , we seek in parameters corresponding to the minimum of the integral average of the `` distance '' between the initial state and output state .
note , that fidelity of the initial and output states for the most part of the bloch sphere exceeds value 5/6 which corresponds to the universal qcm for the pure states @xcite .
this fact has two reasons .
firstly , described qcm copies only part of the two - partite state .
second , this qcm is state - dependent one , and this non - universality raises the quality of the cloning .
moreover , here was discussed the cloning of the mixed qubit states . in order to consider such device we use the `` purification '' of the mixed state and then apply `` partial '' cloning machine .
let s emphasize , that `` purification '' of the mixed state is not unique , but the output state produced by our qcm does not depend on this nonuniqueness .
parameters of the qcm was sought by minimization of the integral average of the distance between the initial and output states .
note , that this averaging differs from the used above because the sets of the states are different . in this case
the fidelity of the initial and output states on the real part of the bloch sphere exceeds the value 5/6 too .
evidently , that choice of the parameters @xmath67 for the qcm is defined by the strategy of the eavesdropping .
w have discussed here only one possible qcm for the mixed states . evidently , that there are many other variants for the qcm , may be , without restrictions like ( [ ccm1 ] ) , ( [ ccm2 ] ) , ( [ con1 ] ) , asymmetric qcm etc .
w.k.wooters , w.h.zurek_. nature , 299 , 802 ( 1982 ) .
_ d.dieks_. phys .
lett.a , 54 , 1844 ( 1982 ) .
_ h.barnum , c.m.caves , c.a.fuchs , r.jozsa , b.schumacher_. phys.rev.lett . ,
76 , 2818 ( 1996 ) ( arxiv quant - ph/9511010 ) .
_ v.buzek , m.hillery._ phys.rev.a , 54 , 1844 ( 1996 ) .
_ v.scarani , s.ibilsdur , n.gisin._ rev.mod.phys . ,
v.77 , 1225 - 1256 ( 2005 ) .
_ n.j.cerf , j.fiurasek_. optical quantum cloning . in : progress in optics ,
v.49 , ed .
e.wolf , elsevier , 2006 .
_ j. preskill . _ lecture notes for physics : quantum information and computation , 1998 .
_ s.adhikari , a.k.pati , i.chakrabarty , b.s.choudhury._ arxiv : quant - ph/0705.0631 , 2007 . | we discuss the `` partial '' quantum cloning of the pure two - partite states , when the `` part '' of initial state related to the one qubit is copied only .
the same approach gives the possibility to design the quantum copying machine for the mixed qubit states .
* `` partial '' quantum cloning and quantum cloning of the mixed states * * a.ya.kazakov * + laboratory of quantum information , + st .- petersburg state university of aerospace instrumentation , 67 b.morskaya str . , st .- petersburg , 190000 russia pacs : 03.67 -a , 03.67 dd |
studies of interactions between d - branes and closed strings are quite important from various points of view . the interactions play a crucial role in finding non - trivial relationships between open strings and closed strings such as ads / cft @xcite and recently - proposed open - closed dualities @xcite , for example .
studies of the interactions are also important to analyze the dynamics of the systems of multiple d - branes .
however , almost all analyses of the interactions have been carried out by considering static d - branes , and it is a long - standing problem to describe the back reaction of the d - branes .
a d - brane in the worldsheet description is just a boundary of the worldsheet with dirichlet boundary condition , and the d - brane is treated as an infinitely heavy classical source of closed strings ; the position , or the collective coordinate , of the d - brane is fixed at a point in the target space in the dirichlet direction .
there are several attempts to describe the back reaction of d - branes .
some of them are based on the conformal field theory on the worldsheet @xcite . in ref .
@xcite , the disk amplitude for the scattering of closed string states from a d - particle is computed in the bosonic string theory in which the collective coordinate of the d - particle is quantized .
the current conservation condition of the d - particle is obtained by demanding the conformal invariance of the amplitude in that work .
momentum conservation condition between the closed strings and the d - particle is described by using the zero - mode integral in the path integral of the trajectory of the d - particle .
attempts to obtain the momentum conservation condition between closed strings and a d - brane from the viewpoint of conformal invariance can be found in refs .
an annulus amplitude for the scattering of the closed strings from the d - brane is considered and a variant of the fischler - susskind mechanism is proposed there ; the infrared ( ir ) divergence in the open string channel of the annulus amplitude is canceled by adding an appropriate operator to the boundary of the worldsheet .
the momentum conservation condition between the closed strings and the d - particle is explicitly obtained in refs .
@xcite by demanding the conformal invariance of the total amplitude . in ref .
@xcite , the ir divergence is canceled by adding a logarithmic operator @xcite that represents the recoil of the d - particle @xcite .
however , it is also pointed out in ref .
@xcite that the divergence does not exist in the case of d@xmath0-brane with @xmath1 . in ref .
@xcite , the energy conservation condition is also obtained as well as the momentum conservation condition in the case of d - particle , explicitly . however , there is still room to clarify how to define the initial momentum of the d - particle there .
further investigation along the ideas of refs .
@xcite is still important for deeper understanding of recoil of d - branes .
some applications of d - brane recoil to other topics , and related works are found in refs .
@xcite .
a target - space theory that handles second - quantized closed strings with dynamical d - branes may provide us a description of back reaction of d - branes .
some arguments on back reaction and recoil of d - branes along this approach is found in ref .
@xcite . in the present work ,
we propose an alternative method to describe the scattering process between a d - brane and closed strings together with the back reaction of the d - brane in the bosonic string theory .
the impact of the closed strings in this method is represented by a source term of the low - energy effective worldvolume theory of the d - brane , and the initial and the final velocity of the d - brane are described by the time derivative of the expectation values of the massless scalar fields of the worldvolume theory .
we utilize the following approximations to justify our approach : 1 .
field theory limit , namely @xmath2 where @xmath3 is the typical momentum of the open strings on the d - brane .
elastic limit , namely the momenta of the closed strings are small enough and no massive open - string mode is excited on the d - brane .
we also assume that the closed strings do not lose their total momentum in the worldvolume directions of the d - brane , and no internal field on the d - brane gets momentum from the closed strings .
non - relativistic limit , namely the velocity of the d - brane is very small .
tree level approximation in the string theory , namely the string coupling is very small . the above conditions 3 and 4 means that the tension of the d - brane is very large .
this is also consistent with the condition 1 .
a nontrivial problem is how to represent the source term of the worldvolume theory in terms of the quantities of the closed strings .
the basic idea is as follows .
we consider a scattering process between the d - brane and the closed strings that creates @xmath4 massless scalar particles on the worldvolume of the d - brane .
we calculate the probability of the creation of the @xmath4 massless scalar particles in the two different frameworks : one of them is the worldvolume theory of the d - brane with the source term and the other is the perturbative string theory . by comparing the two results , we obtain the relationship between the source term and the momenta of the closed strings , and we obtain the momentum conservation condition between the closed strings and the d - brane up to the relative sign of the momentum of the d - brane .
one of the distinction between the present work and those in refs .
@xcite is that all the diagrams we consider in the string theory are disk diagrams and we need not annulus diagrams .
the organization of this article is as follows .
we consider bosonic strings , and we start by considering a recoiling d - particle for simplicity . in section 2 , we consider the worldvolume theory of the d - particle and review some basic facts necessary for the later discussions .
we calculate the amplitude of the creation of the @xmath4 massless particles explicitly . in section 3
, we consider the scattering process between the d - particle and the closed strings and calculate the amplitude of the process that creates @xmath4 massless open string modes on the d - particle in the framework of the string theory . in section 4
, we compare the results obtained in section 2 with those in section 3 .
we show that the absolute value of the total momentum transfer from the closed strings is exactly equal to the absolute value of the change of the d - particle s momentum , within the above approximations
. we also comment on the distribution of the probability of the @xmath4 massless particle creation .
we see that the expectation value of the total energy of the created particles gives the kinetic energy of the d - particle , correctly . in section 5
, we generalize the results obtained for the d - particle to the case of higher dimensional d - branes .
we compactify the worldvolume of the d - brane to make its mass finite , and we obtain the momentum conservation condition between the closed strings and the d - brane up to the relative sign of the momentum of the d - brane , again .
we provide conclusion , several open problems and discussions in the last section .
we begin with the dirac - born - infelt ( dbi ) action of a d - particle .
it is just an ordinary action for a point particle : @xmath5 where @xmath6 is the tension of the d - particle , and @xmath7 denotes the time derivative of the space coordinate @xmath8 of the d - particle .
we consider a _ non - relativistic _ d - particle in this work .
then @xmath9 and the dbi action is written as @xmath10 where we have dropped the constant term and the @xmath11 contributions .
next , we consider a scattering process between the d - particle and closed strings .
we attempt to include the effects of the closed strings in the above action .
a simple conjecture is that a source term may effectively represent the impact of the closed strings .
therefore , we consider the following effective action @xmath12 and let us examine how this action represents the nature of the recoiling d - particle . here , @xmath13 is an ir cut - off that should be sent to zero finally , and we assume that @xmath14 includes relevant effects of the closed strings . let us calculate the time dependence of the expectation value of @xmath8 at the classical level .
we choose the target space coordinate so that the initial momentum of the d - particle is equal to zero and its final momentum is in the @xmath15 direction .
we assume that the source term is switched on only for the duration @xmath16 .
the expectation value of the field @xmath17 is given by @xmath18 where @xmath19 is the retarded green s function which satisfies @xmath20 and @xmath21 satisfies the equation of motion without the source term , @xmath22 @xmath23 is the initial velocity of the d - particle and @xmath24 , in fact . substituting the explicit form of the retarded green s function , @xmath25 where we should take the limit @xmath26 .
then , @xmath27 where @xmath28 is the source term in the momentum space . ] ] note that @xmath29 for @xmath30 . for @xmath31 ,
@xmath32 the limit @xmath26 should be taken before we take the limit @xmath33 .
note that @xmath34 is real .
@xmath35 should be regarded as the duration of that the closed strings affect the d - particle in the scattering process .
we assume that the effects of the closed strings disappear at the infinitely far future and we define the final velocity of the d - particle @xmath36 in this region ( @xmath31 ) . in the same way
, we assume that the effects of the closed strings disappears at the infinitely past and the initial velocity of the d - particle @xmath37 should be defined in the region of @xmath38 .
then , the change of the velocity of the d - particle is given by @xmath39 , and the problem we should consider is how to rewrite @xmath34 in terms of the quantities of the closed strings .
we will come back to this problem in section [ conservation - section ] .
we calculate the @xmath4 particle creation amplitude with the source term in order to compare it with the corresponding amplitude in the string theory .
we treat the source term as a perturbation .
the action without the source term is @xmath40 we consider only @xmath17 and we define a dimension - less field @xmath41 as @xmath42 .
then the lagrangian and the canonical momentum are @xmath43 we define the creation operator @xmath44 and the annihilation operator @xmath45 as @xmath46 where @xmath47 .
the hamiltonian is given by @xmath48 including the source term , the hamiltonian is @xmath49 where we have defined @xmath50 as @xmath51 we define the @xmath4 particle state as @xmath52 . by using the above notations , the @xmath4 particle creation amplitude @xmath53 is given by @xmath54 at the limit @xmath33 .
@xmath55 is a real number defined in ( [ theta ] ) .
note that @xmath56 gives a poisson distribution where @xmath57 , and the final state is a coherent state .
the detailed calculation to obtain ( [ a - field ] ) is reviewed in appendix [ n - parti - amp ] .
in this section , we calculate the amplitude of the creation of the @xmath4 massless scalar open strings on the d - particle from @xmath58 closed strings , at the tree level .
we choose the spacetime coordinate so that the d - particle recoils in the @xmath15 direction as we did in section [ field ] , and the massless open strings we consider here correspond to the zero mode of the d - particle in the @xmath15 direction . direction .
] we also choose the spacetime coordinate so that the initial velocity of the d - particle is zero .
the amplitude consists of several disk diagrams as shown in fig .
[ fig : diagrams ] , for example .
we calculate a single disk amplitude to begin with , and we consider the full diagrams by using the results on the single disk amplitude .
we calculate a disk amplitude @xmath59 of @xmath58 closed strings and @xmath4 massless scalar open strings on the d - particle .
we consider a unit disk @xmath60 with neumann boundary condition @xmath61 and dirichlet boundary conditions @xmath62 for @xmath63 , where @xmath64 is the complex coordinate on the unit disk .
we explicitly calculate for the case that all the closed strings are closed - string tachyons , for simplicity , however the results should be easily generalized to the cases of arbitrary closed strings .
the open strings can not carry momenta in the dirichlet directions and thus their momenta in the space directions are zero .
the energy of the massless open string is also zero due to the on - shell condition .
therefore , the massless open string labeled by @xmath65 is represented by the vertex operator @xmath66 where @xmath67 is the polarization vector which satisfies @xmath68 and @xmath69 for @xmath70 .
the amplitude is given by @xmath72 where only the terms that are linear with respect to every @xmath73 ( @xmath74 ) should be taken and the other terms should be discarded in ( [ amplitude ] ) .
@xmath75 is a constant , @xmath76 is the coupling constant for the massless open strings and @xmath77 is the closed - string coupling constant .
@xmath78 , @xmath79 and @xmath80 is the ghost fields of holomorphic , anti - holomorphic and the @xmath81-component respectively .
we have fixed the positions of the closed string vertices as @xmath82 and @xmath83 , where @xmath84 is the position of the @xmath85-th vertex operator . is fixed and @xmath86 can still move in the region @xmath87 . ]
the worldsheet coordinates are assigned as follows : @xmath88 s with @xmath89 are for the closed strings and @xmath88 s with @xmath90 are for the open strings .
we have also defined @xmath91 as @xmath92 and @xmath93 as @xmath94 we can easily find that @xmath93 is a constant for the present case . note that @xmath95 due to ( [ vanish ] ) , and the last line of ( [ amplitude ] ) becomes @xmath96 .
we can calculate the contribution of the third line of ( [ amplitude ] ) by using @xmath97 where @xmath98 , @xmath99 and we have substituted @xmath68 . see appendix [ notations ] for the details . therefore the amplitude is given by @xmath100 where we have used the relationship @xmath101 in the last line .
@xmath102 is the disk amplitude of @xmath58 closed - string tachyons without open strings .
note that the momentum conservation condition among the strings appears due to the zero - mode integral .
we do not have the momentum conservation condition among the strings in the dirichlet directions due to lack of the zero - mode integral and thus @xmath103 is allowed . in this case
, we should fix both the position of the closed - string vertex and that of one of the open - string vertices . ) . ]
we will fix the position of the open - string vertex labeled by @xmath105 as @xmath106 , where @xmath107 is a fixed value .
the position of the closed - string vertex will be fixed at the center of the unit disk , again .
the amplitude is given by @xmath108 where we have identified @xmath109 as @xmath110 the factor @xmath111 in the first line of ( [ fact - amp-1 ] ) appears because of the absence of the @xmath107 integral .
the identification ( [ identif ] ) can be justified as follows . in the calculation of @xmath109
, we fix the position of the closed - string vertex at the center of the disk .
although the degrees of freedom of the modular transformation on the disk have not fixed completely , the residual symmetry is just a rotational symmetry the volume of which is @xmath112 .
then let us fix the rotational degree of freedom by hand .
more precisely , let us mark at the point @xmath113 , for example , and fix the value of the @xmath81-coordinate of the marked point by hand .
this corresponds to divide the amplitude by @xmath112 .
now we have fixed the position of the closed string vertex and the @xmath81-coordinate of the marked point on the disk .
thus the faddeev - popov determinant is the same as the case we have more than one vertex ; it is @xmath114 . as a result , we obtain ( [ identif ] ) .
after all , we obtain @xmath115 for all positive integers @xmath58 .
all the closed - string vertices we have used in the calculations are closed - string tachyon vertices , however we can easily find that ( [ fact - amp - pre ] ) holds for arbitrary closed strings if we regard @xmath59 as the disk amplitude of the corresponding @xmath58 closed strings and the @xmath4 massless open strings .
now we can calculate the total amplitude of the creation of the @xmath4 massless scalar open strings by using the results obtained in the previous subsection .
we define the total amplitude as @xmath116 . for simplicity , we demonstrate the calculation of the amplitude in the case of @xmath117 and @xmath118 that is given schematically in fig .
[ fig : diagrams ] .
we can easily calculate the total amplitude by using ( [ fact - amp - pre ] ) as @xmath119 where @xmath120 is the total amplitude for no open - string creation and @xmath121 is the disk amplitude with @xmath4 massless open - string vertices and @xmath122-th closed - string vertex .
the above calculation can be easily extended into the case of @xmath123 , and we obtain @xmath124 in general .
furthermore , we can show that @xmath125 the derivation of ( [ fact - amp ] ) and ( [ string - poisson ] ) is given in appendix [ poisson ] .
in order to rewrite @xmath34 in terms of the quantities of the closed strings , we compare the amplitude obtained in section [ field ] with that in section [ string ] . before comparing them explicitly
, we should notice the following facts : 1 .
+ if we have @xmath4 same particles in the final state , we should divide the cross section by @xmath126 . in other words
, we should multiply @xmath127 to the string amplitude in order to compare it with the amplitude in the worldvolume theory where the factor @xmath127 is already included in the definition of the @xmath4-particle state .
2 . + in the worldvolume theory ,
the massless particle exists in the one - dimensional worldvolume and the calculation in section [ field ] is based on non - relativistic quantum mechanics . on the other hand ,
the strings considered in section [ string ] exist in 26-dimensional spacetime , and the center - of - mass coordinates of the massless open strings are confined in the one - dimensional subspace of it . because of the above difference , we also need the following consideration to obtain the correct normalization .
+ let us consider the probability @xmath128 of that @xmath4 open - string particles of mass @xmath13 whose momenta lie in a small region @xmath129 are created from the closed strings .. on the other hand , we have calculated the string amplitudes for the case of @xmath130 in order to maintain the conformal invariance of the worldsheet .
this discrepancy disappears because we take the limit @xmath33 eventually . ]
@xmath128 is given as @xmath131 we have the delta functions in the first line because the open strings do not have momenta in the dirichlet directions .
therefore , we should also multiply the factor @xmath132 to the amplitude in the sting theory to compare it with the amplitude in the worldvolume theory .
note that the factor @xmath133 in ( [ p(n ) ] ) is necessary to make @xmath128 dimensionless .
3 . + the hilbert space of string
theory in section [ string ] includes the closed - string sector and the massive open - string sector , too . on the other hand ,
the hilbert space of the effective field theory on the d - particle considered in section [ field ] includes only massless open - string states . due to the above difference of the hilbert spaces ,
the absolute normalization of the amplitudes can not be compared directly with each other . with the above things in mind ,
we propose the following identification between the amplitudes in the effective field theory and those in the string theory : @xmath134 the physical meaning of the right - hand side is the following .
the probability of the creation of the one open string through the scattering between the d - particle and the closed strings should be proportional to @xmath135 and that of no creation of the massless open string should be proportional to @xmath136 .
we can cancel the ambiguity of the normalization of the probability which arises due to the problem 3 above , by considering the ratio of the probability .
let us substitute the above results into ( [ amp - compare ] ) . from eqs .
( [ x - dot ] ) , ( [ j - j ] ) , ( [ a - field ] ) and ( [ fact - amp ] ) , we obtain @xmath137 then we obtain the following relationship from ( [ amp - compare ] ) : @xmath138 this is the momentum conservation condition between the d - particle and the closed strings up to the relative sign .
the result ( [ conservation ] ) is independent of the ir cut - off and of the types of the closed strings .
we have considered only @xmath139 in the previous subsection .
let us consider @xmath140 for general @xmath4 and see what happens . in the framework of the effective field theory , @xmath128
is given from ( [ a - field ] ) as @xmath141 and @xmath142 .
the above gives a poisson distribution . in the framework of the string theory ,
@xmath140 is given as @xmath143 from ( [ string - poisson ] ) and the discussions in section [ normalization ] .
thus the identification @xmath144 that is made in ( [ amp - compare ] ) is still valid for @xmath145 . in other words , ( [ st - poisson ] ) indicates that we can also derive the poisson distribution for the massless particles from the string theory .
the above observation strongly suggests that the effective action ( [ effective ] ) with the source term captures the phenomena of the scattering between the d - particle and the closed strings well .
it is interesting to see that the total energy of the massless scalar particles on the worldvolume of the d - particle gives the correct kinetic energy of the d - particle .
the expectation value of the total energy of the scalar particles after the scattering is given by using ( [ st - poisson ] ) as @xmath146 that is exactly the expected value of the kinetic energy of the d - particle after the scattering .
this quantity is also independent of the ir cut - off .
we generalize the results in the previous sections to the cases of higher dimensional d - branes .
we consider a d@xmath0-brane which extends in the @xmath147 directions .
we choose the spacetime coordinate so that the initial momentum of the d - brane is zero and the final momentum of it is in the @xmath15 direction .
we consider an elastic scattering between the d@xmath0-brane and closed strings ; we assume that the momenta of the closed strings parallel to the d@xmath0-brane are conserved within the closed strings , and the total momentum coming from the closed strings is exactly perpendicular to the d@xmath0-brane . in this case
, the d@xmath0-brane does not get any internal energy from the closed strings . in this section ,
we consider the following effective scalar field theory on the d@xmath0-brane : @xmath148 where @xmath13 is an ir cut - off and @xmath149 is the tension of the d@xmath0-brane .
@xmath150 runs from @xmath151 to @xmath152 .
we have chosen the worldvolume coordinate @xmath153 so that it is equal to the zero mode of @xmath154 , where @xmath44 runs from @xmath155 to @xmath0 .
we have @xmath156 gauge field on the worldvolume and the gauge field may contribute to the recoiling process of the d - brane , too .
however , we assume that all the contributions of the gauge field are implicitly included in the source term .
we also assume that the source term is switched on only for the duration @xmath157 .
we compactify the @xmath158 directions on a torus of radius @xmath159 to make the mass of the d@xmath0-brane finite .
let us consider momentum conservation in the @xmath15 direction and we consider only the field @xmath17 .
standard calculations in the field theory lead us to the following results : @xmath160 where @xmath161 is the probability of that the source creates @xmath4 massless scalar particles that correspond to the zero mode of the d@xmath0-brane in the @xmath15 direction , and @xmath36 is the time derivative of the expectation value of @xmath17 at @xmath162 .
we regard @xmath36 as the final velocity of the d - brane in the @xmath15 direction as we did in the case of d - particle .
the integrals over the momenta in the uncompactified theory have been replaced with the summation over the kaluza - klein momenta @xmath163 which are labeled by @xmath164 .
note that the volume of the d@xmath0-brane @xmath165 has appeared in order to get the correct measure of the summation .
we have defined @xmath166 and @xmath167 as @xmath168 and @xmath169 .
@xmath170 is the source term in the momentum space that is given by @xmath171 the probability that the source creates the one massless scalar particle of momentum @xmath172 is given by @xmath173 we are considering the elastic case in which the d@xmath0-brane does not get any internal energy from the closed strings .
then @xmath174 for @xmath175 , because only the massless particle with @xmath176 can be created in this process
. therefore eqs .
( [ lambda - p - q ] ) and ( [ x - dot - h - q ] ) become @xmath177 in the present case . in the case
that the massless scalar open string does not carry the momentum , we can easily see that the calculations in the string theory gives a similar result as that of the d - particle case : @xmath178 where @xmath179 is the total tree - level amplitude of the creation of the @xmath4 massless scalar open strings on the d@xmath0-brane from @xmath58 closed strings .-
brane in the @xmath15 direction . ]
next , we generalize the relationship ( [ amp - compare ] ) to the present case .
let us consider the probability of that one scalar particle of mass @xmath13 whose momenta lies in a small region @xmath180 is created from the closed strings .
is the momentum of the massless particle as that in section [ h - field ] , however it is defined as a 25-dimensional vector in this subsection . ]
the probability written in terms of the string amplitude up to the normalization is @xmath181 where we have considered uncompactified spacetime here .
the delta functions come from the fact that the scalar particle do not have the momentum perpendicular to the d@xmath0-brane . in the present case ,
the spacetime is compactified and @xmath182 appears when we convert the momentum integrals into the summation over the kaluza - klein momenta .
then the probability ( [ prob ] ) should be written as @xmath183 where we have substituted @xmath176 as we did bellow ( [ lambda - p - k ] ) . the same probability in terms of the field theory
is given by @xmath184 , and then the relationship corresponding to ( [ amp - compare ] ) for the present case is @xmath185 from eqs .
( [ a^2-field - h ] ) , ( [ lambda - h ] ) , ( [ x - dot - h ] ) , ( [ a1/a0-h ] ) and ( [ relation - h ] ) , we obtain @xmath186 which is the correct momentum conservation condition in the direction perpendicular to the d@xmath0-brane up to the relative sign .
we can also show that the poisson distribution of the probability of the @xmath4 massless particle creation obtained from the effective field theory is consistent with the calculations of the creation probabilities in the string theory .
the expectation value of the total energy of the created massless particles gives the final kinetic energy of the d@xmath0-brane correctly .
we developed a new method to describe an elastic scattering between a d - brane and closed strings in the non - relativistic region .
the momentum conservation condition between the d - brane and the closed strings was obtained up to the relative sign of the d - brane s momentum .
an interesting extension of the present work may be investigation of non - elastic scatterings in which the closed strings lose their total momentum in the worldvolume directions . in this case , the massless modes on the d - brane get non - zero momentum and non - zero energy from the closed strings . how to obtain the energy conservation condition between the d - brane and the closed strings , that have not been discussed in this paper ,
is also an important problem which we should solve in the future .
if we choose a particular spacetime coordinate so that @xmath187 , energy conservation holds through the scattering process .
this suggests that the lorentz covariant generalization of the present work may solve the problem of the energy conservation .
generalization to superstring theory is interesting , too . in the present work ,
the momentum conservation condition between the d - brane and the closed strings was _ derived _ up to the relative sign and a natural question is how to fix the relative sign .
however , we can take a different standpoint in the construction of more generalized formalism to solve the above open problems ; the momentum conservation condition with correct relative sign can be utilized as a prerequisite .
for example , the exact momentum conservation condition for a d - particle , @xmath188 fixes the relative phase between the amplitudes in the field theory and those in the string theory as @xmath189 in our notation . we can also obtain a similar condition between the amplitudes for a d@xmath0-brane .
the new relationship between the amplitudes like ( [ abs - amp - comp ] ) can be a starting point of solving the open problems .
we hope that the present work can be a step to the deeper understanding of the interactions between d - branes and closed strings . *
acknowledgments * the author would like to thank john f. wheater for the collaboration at the initial stage of this project .
the author also thanks the members of the group of theoretical particle physics and cosmology at niels bohr institute and nordita for valuable discussions and comments .
the author wish to thank the organizers of rtn workshop `` the quantum structure of spacetime and the geometric nature of fundamental interactions '' held at copenhagen on 15 - 20 september 2003 where the present work has been initiated .
this work was supported in part by nishina memorial foundation .
we review the derivation of ( [ a - field ] ) in this appendix . in the interaction representation ,
@xmath190 the time evolution of the state in the schrdinger representation is @xmath191 the @xmath4 particle state is defined as @xmath52 and @xmath192 where @xmath193 the following formulae are useful : @xmath194 therefore @xmath195 where @xmath196 is a real number .
thus @xmath4 particle creation amplitude @xmath53 is given as @xmath197 at the limit @xmath33 .
we use the following worldsheet action in this paper : @xmath198 the green s function on the unit disk is given as @xmath199 if we have neumann boundary condition in the @xmath200 direction , and @xmath201 in the case of dirichlet boundary condition instead .
we assume that we have dirichlet boundary condition in the @xmath15 direction here . if the point @xmath202 is on the boundary of the unit disk , @xmath203 where @xmath204 . if the point @xmath205 is also on the boundary , @xmath206 eq .
( [ vanish - int ] ) can be shown by using ( [ cor-3 ] ) and @xmath207^{2\pi}_{0 } \nonumber \\ & = & 0 . \label{vanish}\end{aligned}\ ] ]
we derive ( [ fact - amp ] ) and ( [ string - poisson ] ) in this appendix . to begin with , we define @xmath208 as the disk amplitude with @xmath209-th , @xmath210-th , @xmath211 , and @xmath212-th closed - string vertices and the @xmath4 massless scalar open - string vertices . for example , @xmath213 by using ( [ fact - amp - pre ] ) .
@xmath214 is then written by the summation of all the possible connected and disconnected diagrams : @xmath215 next , let us calculate @xmath216 by using a modification of ( [ cal - a - n-0 ] ) . in order to obtain @xmath216 , we should modify , for example , @xmath217 in the right - hand side of ( [ cal - a - n-0 ] ) as @xmath218 and the extra factor @xmath219 appears in front of the original term . in the same way
, we can show that all the terms in the right - hand side of ( [ cal - a - n-0 ] ) get the same extra factor through the modification and we can conclude that @xmath220 @xmath221 is also calculated by a similar modification of @xmath216 .
for example , @xmath216 contains @xmath222 and this term should be modified as @xmath223 and the extra factor @xmath219 appears in front of the original term , again .
we can also show that all the terms in @xmath216 get the same extra factor through the modification and we can conclude that @xmath224 the same procedure holds for the calculation of @xmath116 from @xmath225 and therefore we can conclude that @xmath226 a. sen , `` open - closed duality at tree level '' , _ phys . rev .
* 91 * ( 2003 ) 181601 , hep - th/0306137 ; a. sen , `` open - closed duality : lessons from matrix model '' , _ mod .
* a19 * ( 2004 ) 841 , hep - th/0308068 ; a. sen , `` rolling tachyon boundary state , conserved charges and two dimensional string theory '' , hep - th/0402157 . | we propose a new method to describe a recoiling d - brane that is elastically scattered by closed strings in the non - relativistic region .
we utilize the low - energy effective field theory on the worldvolume of the d - brane , and the velocity of the d - brane is described by the time derivative of the expectation values of the massless scalar fields on the worldvolume .
the effects of the closed strings are represented by a source term for the massless fields in this method .
the momentum conservation condition between the closed strings and the d - brane is derived up to the relative sign of the momentum of the d - brane . plus
0.2pt minus 0.1pt addtoresetequationsection |
consider the following stochastic differential equation @xmath0 here @xmath1 is an incompressible periodic vector field , @xmath2 is a 2-dimensional brownian motion , and @xmath3 ( molecular diffusivity ) is a small parameter .
we further assume that the stream function @xmath4 , such that @xmath5 is itself periodic in both variables , that is the integral of @xmath6 over the periodicity cell is zero . for simplicity of notation assume that the period of @xmath7 in each of the variables is equal to one .
it is well known ( see for example @xcite ) , that with @xmath8 fixed , the solution of ( [ e1 ] ) scales like a diffusion process with constant diffusion matrix when time goes to infinity .
more precisely , there exists the limit , called the effective diffusivity , @xmath9 where @xmath10 and @xmath11 are the coordinates and @xmath12 is the initial distribution of the process @xmath13 , which we can take to be an arbitrary measure with compact support .
the measure on @xmath14 , { \mbox { \bbc r}}^2)$ ] , induced by the process @xmath15 , converges weakly , when @xmath16 , to the measure induced by the diffusion process with constant matrix @xmath17 .
we are interested in the behavior of the effective diffusivity when the molecular diffusion @xmath18 tends to zero .
assume that all the critical points of @xmath7 are non degenerate .
we distinguish two qualitatively different cases , depending on the structure of the stream lines of the flow given by @xmath1 . in the first case
, there is a level set of @xmath7 , which contains some of the saddle points , and which forms a lattice in @xmath19 , thus dividing the plane into bounded sets , invariant under the flow .
a standard example of a cellular flow , which has been studied in several of the papers cited below , is the flow with the stream function @xmath20 . in this particular example
the separatrices ( the level sets of @xmath7 containing saddle points ) form a rectangular lattice . in the second case
, there is more than one unbounded level set of @xmath7 containing critical points , and thus there are ` open channels ' in the flow , and some of the solutions of the equation @xmath21 go off to infinity .
an example of a flow with open channels is the flow with the stream function @xmath22 .
indeed , the horizontal axis @xmath23 is an unbounded stream line of the flow .
[ pic1 ] since @xmath1 is periodic , we may consider @xmath21 as the flow on the torus .
the torus is then a union of the sepatatrices and a finite number of open domains , bounded by the separatrices , and invariant under the flow . in @xcite fannjiang and
papanicolaou considered cellular flows for which the separatrices form a rectangular lattice on @xmath24 and the stream function satisfies certain symmetry conditions .
they showed that in this case @xmath25 that is the effective diffusivity is enhanced by a factor of order @xmath26 compared to case of the diffusion process @xmath27 without the advection term .
moreover , they found the constant matrix @xmath28 explicitly .
their proof is based on a variational principle applied to a symmetric operator associated to the generator of the process @xmath29 . in @xcite heinze
provided certain upper and lower estimates on the effective diffusivity in the case of cellular flows , for which the separatrices form a rectangular lattice on @xmath24 .
there are earlier physical papers ( @xcite , @xcite , @xcite ) , arguing that the asymptotics in ( [ e5 ] ) is true for particular flows .
our first result is the rigorous proof of this statement for general cellular flows .
[ t1 ] assume that an infinitely smooth periodic stream function @xmath30 defines a cellular flow , and that its critical points are nondegenerate .
then the asymptotics of the effective diffusivity for the process ( [ e1 ] ) is given by ( [ e5 ] ) .
let @xmath31 be the noncompact connected level set of @xmath7 .
this level set contains some of the saddle points of @xmath7 and forms a lattice in @xmath19 . without loss of generality
we may assume that @xmath32 when @xmath33 .
the corresponding level set on the torus will be denoted by @xmath34 . the process @xmath35 consists of the ` fast ' part , which is the periodic motion along the streamlines , and the ` slow ' diffusion across them .
the motion is almost periodic away from the separatrices .
however , once the trajectory is in a sufficiently small neighborhood of the level set @xmath31 , it is likely to continue along it , and may go from cell to cell in a time much shorter than it would take the ` slow ' diffusion to cover the same distance . the rough outline of the proof of theorem [ t1 ] is the following .
we introduce a markov chain , which can be viewed as a discrete time version of the process @xmath35 .
the state space for the markov chain is @xmath36 .
note , that due to the periodicity of @xmath7 , the process @xmath35 can be viewed as a process on the torus . in order to define the transition probabilities , we introduce stopping times for the process @xmath35 .
the stopping time @xmath37 is the first time when @xmath35 hits @xmath36 , and @xmath38 is defined as the first time after @xmath39 when the process @xmath35 returns to @xmath36 , after having traveled ` past ' a saddle point .
the transition times of the markov chain are random .
we show that the study of the asymptotics of the effective diffusivity can be reduced to the study of the asymptotics of transition probabilities and of the expectations of the transition times for the markov chain .
the limit of the transition probabilities as @xmath40 is determined by the behavior of the process @xmath35 in an arbitrarily small neighborhood of @xmath36 .
the asymptotics of the expectations of the transition times , on the contrary , is determined by the event that the trajectory of @xmath35 wanders away from the level set @xmath36 . in order to study the transition times
we use the results of freidlin and wentzell @xcite .
for a given stream function @xmath7 they introduce a graph and a mapping @xmath41 from the plane into the graph , such that each connected level curve of @xmath7 gets mapped into a point on the graph , with the level sets containing critical points mapped into vertices .
then they demonstrate that the process @xmath42 on the graph converges to a limiting markov process .
the asymptotics of the expectations of the transition times for our markov chain is related to the limiting process on the graph .
now consider the flows with ` open channels ' . assuming that the channels are directed along the @xmath43 axis
, we prove that @xmath44 that is the diffusion across the channels is not qualitatively enhanced , compared with the process @xmath45 .
the effective diffusivity in the direction of the flow is enhanced by a factor of order @xmath46 .
[ t2 ] assume that an infinitely smooth periodic stream function @xmath30 defines a flow with open channels , which are directed along the @xmath43 axis , and that its critical points are nondegenerate .
then the asymptotics of the effective diffusivity for the process ( [ e1 ] ) is given by ( [ oc ] ) .
* remark * since the matrix of effective diffusivity is symmetric and positive definite , from theorem [ t2 ] it follows that the off - diagonal terms @xmath47 are bounded uniformly in @xmath18 .
however we do not make a statement here on their asymptotic behavior .
+ the proof of theorem [ t2 ] is much simpler than that of theorem [ t1 ] , and is an easy application of the results of freidlin and wentzell @xcite .
the paper is organized as follows . in section [ se2 ]
we describe the construction of the discrete time markov chain associated with a cellular flow , and relate the question of effective diffusivity for the process @xmath35 to the study of transition probabilities and transition times for the markov chain . in sections
[ se3 ] and [ se4 ] we study of transition probabilities and transition times respectively . in section [ se5 ]
we prove theorem [ t2 ] . in section [ se6 ]
we prove several technical lemmas used in the previous sections .
consider the set @xmath48 on the torus , with some @xmath49 , and let @xmath50 be the connected component of this set , which contains @xmath36 .
thus @xmath50 is a thin tube around @xmath36 , whose width , however , is much larger than a typical fluctuation of the process @xmath51 in fixed time ( see the picture below ) .
let @xmath52 be the connected components of @xmath53 , and let @xmath54 be the saddle points , which belong to @xmath36 .
while the numbers of the connected components and of the saddles are the same for topological reasons , their equality is not used in the proofs , and the numbering of @xmath55 s is not related in any manner to that of @xmath56 s . if there are points , which are carried to @xmath57 by the flow @xmath58 , and to @xmath56 by the flow @xmath59 , then the set of such points ( a subset of @xmath36 ) is denoted by @xmath60 .
we assume ( for the sake of simplicity of notation only ) that @xmath61 is empty , that is the separatrices do not form `` loops '' . in a neighborhood of each curve @xmath62 we may consider a smooth change of coordinates @xmath63 , where @xmath64 is defined by the conditions : @xmath65 on @xmath62 , and @xmath66 ( this way @xmath64 is defined up to multiplication by @xmath67 and up to an additive constant ) . the same change of coordinates can be considered in @xmath68 . in this case
@xmath69 , we define @xmath70 let @xmath71 .
define the stopping times @xmath72 .
then @xmath73 are defined inductively as follows .
assume that @xmath74 , and @xmath75 .
then @xmath76 thus , @xmath77 is the first time after @xmath39 that the process either exits @xmath50 , or goes past a saddle point different from @xmath56 .
define @xmath78 .
let @xmath79 . since almost every trajectory of @xmath35 does not contain any of the points @xmath56
, @xmath80 is a markov chain with the state space @xmath81 .
the stopping times @xmath38 are the consecutive times when the process @xmath35 hits the separatrix @xmath36 after exiting @xmath50 or after having passed past a saddle point . note , that the case , when a point @xmath82 travels , due to diffusion , against the flow @xmath1 , and returns to @xmath83 , does not count as having passed past a saddle point .
[ pic2 ] it is not difficult to see that @xmath80 satisfies the doeblin condition , with the unique ergodic set . therefore there exists a unique invariant measure @xmath84 on @xmath81 ( @xcite ) .
note that @xmath85 also forms a markov chain with the state space @xmath86 , which satisfies doeblin condition for each @xmath18 , and has a unique ergodic set .
let @xmath87 be the stochastic transition function for this chain ( it only depends on the first component of the original point , as is reflected in the notation ) .
then @xmath88 is the invariant measure . since
, due to the presence of diffusion , the distributions of the transition times have exponentially decreasing tails , we can apply the law of large numbers to the function @xmath89 on @xmath90 to obtain @xmath91 @xmath92 in the arguments , which led to ( [ t ] ) , we considered @xmath13 as a process on the torus . in order to keep track of the displacements of @xmath93 , as the process on @xmath19
, we introduce another markov chain , on the extended phase space @xmath94 .
let now @xmath95 be the saddle points of @xmath7 on the plane .
then any @xmath96 can be uniquely identified with a pair @xmath97 , where @xmath98 and @xmath99 , [ a^2_i ] ) \in { \mbox { \bbc z}}^2 $ ] .
( @xmath100 $ ] and @xmath101 $ ] are the integer parts of the first and second coordinates of @xmath56 ) .
thus , we have the mapping @xmath102 .
let @xmath103 and @xmath104 be the components of this mapping .
we define the markov chain @xmath105 as follows @xmath106 note that the second component @xmath107 almost surely takes the values in some finite subset @xmath108 of @xmath109 .
thus @xmath105 is a markov chain on @xmath110 , where @xmath111 .
it is not difficult to see that @xmath105 satisfies doeblin condition with the unique ergodic set . applying the law of large numbers to the vector valued function @xmath112 , defined on @xmath113
, we obtain that there exists @xmath114 , such that @xmath115 since @xmath116 in distribution , and due to ( [ t ] ) , we conclude that @xmath117 . applying the central limit theorem to the same function @xmath118
, we obtain that there exists a matrix @xmath119 , such that @xmath120 due to ( [ t ] ) , we have @xmath121 in sections [ se3 ] and [ se4 ] we shall obtain the asymptotics of the stochastic transition functions for the chain @xmath122 , and of the functions @xmath123 . assuming that this is accomplished
, the next lemma will allow us to obtain the asymptotics of the effective diffusivity , using ( [ dd ] ) .
first we introduce the notations and formulate the assumptions needed for the lemma .
let @xmath124 be a locally compact separable metric space .
let @xmath125 be the set of bounded continuous functions on @xmath124 .
let @xmath126 be a family of stochastic transition functions on @xmath124 .
assume that \(a ) the family of measures @xmath127 is tight for any compact set @xmath128 .
\(b ) @xmath129 is weakly feller , that is @xmath130 if @xmath131 .
\(c ) for any @xmath131 and any compact @xmath132 ,
@xmath133 \(d ) there exist unique invariant measures @xmath84 .
there exist @xmath134 , such that @xmath135 ( that is @xmath136 are uniformly exponentially mixing ) .
let @xmath137 be such that @xmath138 for all @xmath18 .
let @xmath122 be the stationary markov chain , with the stochastic transition function @xmath139 .
since @xmath140 is exponentially mixing , the central limit theorem can be applied to @xmath141 , and thus @xmath142 converges weakly as @xmath143 to a mean - zero gaussian distribution .
we denote the covariance matrix of the limiting distribution by @xmath144 .
[ l1 ] suppose that assumptions ( a)-(d ) hold .
then + ( a ) @xmath145 + ( b ) if @xmath146 are uniformly bounded , @xmath147 , and @xmath148 uniformly on any compact , then @xmath149 + ( c ) if @xmath137 is such that @xmath150 for all @xmath18 , then @xmath151 * proof : * from ( a ) it follows that for each @xmath152 the family of measures @xmath153 is tight for any compact set @xmath128 .
let us assume that for a certain @xmath152 , for any @xmath154 , uniformly on any compact set @xmath128 we have @xmath155 note that this is true for @xmath156 by ( c ) . combining ( [ st ] ) and the fact that @xmath153 is tight we obtain that for any compact set @xmath128 and for any @xmath157 there is a compact set @xmath158 such that @xmath159 for sufficiently small @xmath18 .
next we justify ( [ st ] ) for @xmath160 instead of @xmath152 .
@xmath161 @xmath162 where @xmath163 .
the first term on the right hand side tends to zero by ( [ st ] ) , while the second term tends to zero by ( [ del ] ) as @xmath164 is bounded and tends to zero uniformly on any compact .
we therefore have established ( [ st ] ) for all @xmath152 .
let us prove part ( a ) of the lemma .
fix an arbitrary @xmath165 .
then for @xmath131 we have @xmath166 and the limit is uniform in @xmath18 by ( d ) .
the weak convergence of @xmath167 to @xmath168 now follows from ( [ st ] ) . to prove part ( b ) we write @xmath169
the difference of the first two terms on the right hand side tends to zero as @xmath170 weakly .
the difference of the last two terms tends to zero since for any @xmath171 there is a compact set @xmath128 for which @xmath172 for sufficiently small @xmath18 ( since @xmath170 weakly ) , and @xmath173 on any compact set . in order to prove part ( c ) of the lemma it is sufficient to consider @xmath174 ( scalar valued ) . in this case
@xmath175~.\ ] ] due to uniform mixing ( d ) @xmath176 where @xmath177 does not depend on @xmath18 .
in order to prove that @xmath178 we therefore only need to establish that @xmath179 \rightarrow 0\ ] ] for any fixed @xmath152 . the left hand side of ( [ nox ] ) can be written as @xmath180 let @xmath181 .
then @xmath182 uniformly on any compact by ( [ st ] ) , and the conclusion follows by part ( b ) .
in this section we shall identify the limit of the transition probabilities for the chains @xmath105 on @xmath183 and verify the conditions ( a)-(d ) for these chains .
recall the @xmath184 coordinates which we may consider inside each cell @xmath185 near its boundary , that is in @xmath186 .
let @xmath187 in order to find the limit of the transition probabilities we shall demonstrate that in a small neighborhood of @xmath31 after a random change of time the process @xmath29 is well approximated by the process @xmath188 with the generator @xmath189 in @xmath190 coordinates .
let @xmath82 be a point on @xmath31 .
we introduce the stochastic transition function @xmath191 as follows : let @xmath192 ( if @xmath193 and @xmath194 do not belong to the boundary of the same cell , then @xmath195 . consider the @xmath190 coordinates in @xmath196 , so that @xmath197 , and @xmath64 increases in the direction of the flow . since @xmath198 is a closed contour , points with coordinates @xmath190 and @xmath199 belong to @xmath200 $ ] , and consider the process @xmath188 with the generator @xmath201 in @xmath190 coordinates , which starts at the origin ( corresponding to the point @xmath193 ) .
let @xmath202 be the time of the first exit from the following domain : @xmath203 . then define @xmath204 \}.\ ] ] the summation over @xmath205
is needed to account for the fact that @xmath193 and @xmath194 may both belong to the same edge @xmath62 , in which case they both belong to the boundaries of two cells , and we need to consider two sets of @xmath190 coordinates . the function @xmath129 is a stochastic transition function on @xmath31 , and it can be considered as a stochastic transition function on @xmath110 .
it clearly satisfies conditions ( a ) and ( b ) preceding lemma [ l1 ] . for @xmath206 , and for the stopping times
@xmath207 defined in the previous section , let @xmath208 be the transition function for the chain @xmath93 considered on @xmath209 : @xmath210 \}~.\ ] ] note that this definition is similar to that of @xmath211 . here
, however , we do not use the @xmath190 coordinates since with small probability the process @xmath13 may travel outside of the domains @xmath185 for which @xmath212 before time @xmath213 ( due to the presence of the small diffusion term @xmath13 may go past the saddle point @xmath56 , thus traveling to one of the neighboring domains before time @xmath213 ) . [ l2 ] for any closed interval @xmath214 , and any bounded continuous function @xmath215 on @xmath31 , @xmath216 notice that lemma [ l2 ] implies the condition ( c ) for the chain @xmath105 . before we start the proof of lemma [ l2 ] we state and prove the following preliminary lemma .
[ x ] let @xmath217 and @xmath218 be the following two diffusion processes on @xmath219 with infinitely smooth coefficients : @xmath220 @xmath221 with @xmath222 .
suppose that for a certain constant @xmath223 the following bound on the coefficients holds : latexmath:[\[|\nabla v^i| , |\nabla a^{ij}| , |v^i_1| , |v^i_2| , |a^{ij}_1| , @xmath11 stand for the vector ( matrix ) entries of the coefficients .
let @xmath12 be the initial distribution for the processes @xmath225 and @xmath226 .
then for some constant @xmath227 and for any @xmath228 we have @xmath229 * proof : * let us assume that @xmath230 in order to avoid vector and matrix indices . by ito
s formula , for any stopping time @xmath231 , @xmath232)ds + \ ] ] @xmath233)^2 ds~.\ ] ] from the estimates on the coefficients and their derivatives it follows that the expression in ( [ eql1 ] ) can be estimated as follows @xmath234 in particular , for @xmath235 we have @xmath236 let @xmath237
. then @xmath238 . by gronwall
s lemma applied to r(t ) we have @xmath239 define the stopping time @xmath240 .
then , by ( [ eql2 ] ) @xmath241 @xmath242 which yields the lemma .
+ + * proof of lemma [ l2 ] : * since the kernel @xmath211 is smooth in both variables and satisfies condition ( a ) preceding lemma [ l1 ] , to prove the uniform weak convergence stated in lemma [ l2 ] it is sufficient to demonstrate that for an arbitrary closed interval @xmath243 and an arbitrary @xmath244 there is @xmath245 such that @xmath246 suppose that @xmath247 .
( if @xmath248 and @xmath249 do nt belong to the boundary of the same cell , then @xmath250 is equal to zero . ) for the sake of simplicity of notation let us assume that @xmath214 and @xmath243 , that is @xmath248 and @xmath249 belong to the adjacent edges of @xmath31 . without loss of generality
we can assume that @xmath251 in @xmath196 for sufficiently small @xmath18 .
we can consider the process @xmath29 in @xmath190 coordinates in the following domain ( see the picture below ) @xmath252 [ pic3 ] as above we consider the process @xmath188 in @xmath190 coordinates in the domain @xmath253 note that @xmath254 is estimated from below by the probability that @xmath29 leaves @xmath255 through any of the copies of @xmath249 ( which corresponds to @xmath29 making a finite number of rotations inside @xmath196 , and then leaving @xmath185 through the segment @xmath249 ) .
let @xmath256 be the copies of @xmath249 in @xmath190 coordinates ( @xmath257 can be obtained from @xmath258 by a shift by @xmath259 along the @xmath64 axis ) .
for an initial point @xmath260 let @xmath261 be the probability that @xmath29 leaves the domain @xmath255 through @xmath258 .
as stated above , for @xmath262 we have @xmath263 similarly , for @xmath264 let @xmath265 be the probability that @xmath188 leaves the domain @xmath266 through @xmath258 .
if @xmath262 and @xmath249 belong to different edges of @xmath198 ( as we have assumed ) then by ( [ p0 ] ) we have @xmath267 from the definition of @xmath268 it is easily seen that the sum converges uniformly in @xmath193 for @xmath262 . in order to prove ( [ est1 ] ) it is sufficient to demonstrate that for each @xmath269 and each @xmath270 there is @xmath245 such that @xmath271 indeed , for every positive @xmath272 we can select @xmath273 such that @xmath274 . by ( [ est2 ] ) we can ensure that if we take @xmath275 to be sufficiently small then @xmath276 which implies ( [ est1 ] ) due to ( [ ineq1 ] ) . for the sake of simplicity of notation we shall only prove ( [ est2 ] ) for @xmath277 .
the generator of the process @xmath29 is @xmath278 which in @xmath279 coordinates becomes @xmath280 amounts to a random time change for the process @xmath29 , which does not affect any of the transition probabilities .
we shall denote the time - changed process with the generator @xmath281 while @xmath282 where @xmath283 and @xmath284 are one dimensional wiener processes in @xmath285 and @xmath64 variables respectively .
we can not apply lemma [ x ] to ( [ f1 ] ) and ( [ f2 ] ) directly , as some of the coefficients on the right hand side of ( [ f1 ] ) may be unbounded near the saddle point @xmath286 . to circumvent this problem
we shall take a sequence of steps ( justified below ) , which will single out a small neighborhood of the saddle point , where yet another coordinate system will be considered .
let us take @xmath287 small enough , so that there exist @xmath288 , such that for any @xmath289 $ ] the probability of the event that @xmath188 passes through the interval @xmath290 before leaving the domain @xmath291 through @xmath292 differs from @xmath293 by less than @xmath294 for any @xmath262 . due to the smoothness of the transition kernel of the process @xmath188 , for some @xmath295 the interval @xmath128 can be replaced by any set @xmath296 as long as @xmath296 is contained in @xmath128 and has lebesgue measure at least @xmath297 .
step 2 . if necessary
make @xmath298 from step 1 smaller , so that @xmath299 whenever @xmath300 and @xmath301 \times [ \theta(a_j ) - \delta ' , \theta(a_j ) + \delta']$ ] . step 3 .
take @xmath302 and let @xmath303 , and @xmath304 .
take @xmath305 sufficiently small so that whenever @xmath306 the process @xmath29 starting at @xmath307 passes through the set @xmath308 before leaving @xmath309 with probability at least @xmath310 for small enough @xmath18 .
let us split the set @xmath311 into intervals @xmath312 of length @xmath298 ( we can assume that each of the intervals included the endpoints ) . construct on @xmath158 the intervals @xmath313 of length @xmath314 , such that @xmath315 .
let @xmath316 be the first time when a process either exits @xmath309(@xmath317 ) or reaches @xmath158 .
let us take @xmath275 so small that @xmath318 step 5 .
let us take @xmath275 so small that @xmath319 + assuming that the steps 1 - 5 are valid let us prove ( [ est2 ] ) . by the markov property @xmath320 by steps 3 , 5 , and 2 , the second factor on the right hand side can be estimated as follows : @xmath321 while by step 4 @xmath322 combining the above inequalities and using step 1 we obtain @xmath323 @xmath324 which implies ( [ est2 ] ) .
it remains to justify the construction in steps 1 - 5 .
the validity of steps 1 and 2 follows from the fact that the transition kernel of the process @xmath188 is smooth . to justify steps
4 and 5 it is sufficient to consider both processes @xmath13 and @xmath188 in a compliment to a neighborhood of the saddle point , where lemma [ x ] applies . in order to justify step 3
we note that by morse lemma in a neighborhood @xmath325 of the saddle point @xmath57 there is a smooth change of variables , such that in the new variables the stream function is @xmath326 , and the interior of @xmath185 corresponds to the first quadrant @xmath327 . in the new variables
the generator of the process @xmath13 , after a random change of time , becomes @xmath328 , where @xmath329 is a differential operator with first and second order terms , with bounded coefficients , and @xmath330 .
we shall consider the operator @xmath331 in the domain @xmath332 .
make a further change of variables in @xmath333 : @xmath334 in the new variables , after dividing all the coefficients of the operator by @xmath335 , which amounts to a random change of time for the process , the operator can be written as @xmath336 where @xmath337 is a differential operator with first and second order terms .
all the coefficients of @xmath338 can be made arbitrarily small in @xmath333 by selecting a sufficiently small neighborhood @xmath325 of the point @xmath57 , and then taking @xmath18 to be sufficiently small .
the construction in step 3 now follows from lemma [ x ] by comparing the process whose generator is the operator ( [ star ] ) with the deterministic process with generator @xmath339 . +
* remark * to verify condition ( d ) ( uniform mixing ) preceding lemma [ l1 ] for the chain @xmath105 it is sufficient to show ( see @xcite , page 197 ) that there is an integer @xmath340 , an interval @xmath341 , and a constant @xmath342 , such that @xmath343 where @xmath344 is the lebesgue measure on @xmath248 .
the proof of estimate ( [ est101 ] ) is absolutely similar to that of lemma [ l2 ] .
in this section we shall study the asymptotics of the integral @xmath345 , which enters in the expression ( [ dd ] ) for the effective diffusivity .
we shall demonstrate the following : @xmath346 @xmath347 where @xmath348 is a positive function and @xmath248 is an arbitrary closed interval @xmath214 . from parts ( a ) and ( b ) of lemma [ l1 ]
it then follows that @xmath349 where @xmath350 is the invariant measure on @xmath351 for the kernel @xmath211 , defined in section [ se3 ] .
the proof of formulas ( [ fr1 ] ) and ( [ fr2 ] ) will rely on a sequence of lemmas stated below .
we shall study separately the probability of the event that the process @xmath352 starting form @xmath353 reaches @xmath354 before time @xmath355 , and the expectation of the time it takes for the process starting from @xmath356 to reach @xmath81 .
consider the process @xmath357 together with the process @xmath188 , whose generator in @xmath190 coordinates is @xmath358 in the domain @xmath359 .
we follow the process @xmath188 till it exits @xmath360 .
let @xmath361 be the corresponding transition kernel .
thus @xmath361 coincides with a gaussian distribution on @xmath362 , and has two point masses at @xmath363 .
similarly let @xmath364 be the transition kernel for the process @xmath357 , which starts at @xmath82 and is stopped at the time @xmath365 .
we have the following : [ a ] for any continuous function @xmath366 , such that @xmath367 , there exists @xmath342 , such that @xmath368 furthermore , for any closed interval @xmath214 , @xmath369 the proof of lemma [ a ] is completely similar to that of lemma [ l2 ] .
we introduce the following notation : @xmath370 is the first time the process @xmath357 leaves @xmath371 ; similarly , @xmath372 and @xmath373 are the first instances when @xmath357 leaves @xmath185 and @xmath374 respectively . in order to estimate the probability that the process @xmath357 starting from @xmath82 reaches @xmath356 before time @xmath213 we shall need the following @xcite [ b ] there exists a constant @xmath342 , such that @xmath375 this lemma is the same as lemma 4.7 of @xcite ( it must be observed that the proof of lemma 4.7 of @xcite goes through for any @xmath376 ) . in the event that @xmath377 , after the stopping time @xmath365 the process @xmath357 may exit @xmath378 either through @xmath354 or through @xmath81 .
the next lemma estimates the probability that the process exits the domain through @xmath356 .
[ c ] there exists @xmath342 , such that for any @xmath379 @xmath380 * proof : * let @xmath381 be the generator of the process @xmath357 in the domain @xmath382 . then the probability in ( [ lce ] ) is equal to the solution @xmath383 of the equation @xmath384 in @xmath378 with the boundary conditions @xmath385 , @xmath386 . let @xmath387 .
then @xmath388 is the solution of the equation @xmath389 with the boundary conditions @xmath390 . by lemma [ b ] , since @xmath391 is bounded uniformly in @xmath18 , the solution @xmath388 is estimated as follows : @xmath392 this implies the statement of the lemma . using the markov property of the process @xmath357 with respect to the stopping time @xmath365
, we obtain that there is @xmath342 , such that for any @xmath393 we have the following : @xmath394 where @xmath395 is the domain which contains @xmath396 ( one of the domains @xmath185 ) , and the second inequality is due to lemma [ c ] and ( [ pt1 ] ) .
furthermore , due to ( [ pt2 ] ) we can evaluate the asymptotics of the event @xmath397 as follows : @xmath398 the next lemma allows us to estimate the expectation of the time it takes for the process starting at @xmath356 to return to @xmath81 .
[ d ] for each of the domains @xmath185 there exists a constant @xmath399 , such that @xmath400 note that lemma [ d ] , together with ( [ e1e ] ) and ( [ e2 ] ) implies ( [ fr1 ] ) and ( [ fr2 ] ) since the expectation of the time it takes for the process to reach @xmath356 can be estimated by lemma [ b ] .
it remains to prove lemma [ d ] .
we introduce notations and state several technical lemmas needed for the proof of lemma [ d ] . since @xmath32 on @xmath401 , we may assume without loss of generality that @xmath402 inside @xmath185 in a small neighborhood of @xmath185 .
then there is a region @xmath403 , whose boundary consists of @xmath198 and a level curve @xmath404 , and , by selecting a sufficiently small @xmath405 , we can ensure that each level set of @xmath7 in @xmath406 is connected and there are no critical points of @xmath7 in the closure of @xmath406 other than on @xmath198 .
for @xmath407 let us define the following functions : @xmath408 in each case the integration is over the level set @xmath409 .
let @xmath410 be a small constant , to be specified later .
consider the function @xmath411 , which solves the equation @xmath412 with boundary conditions @xmath413 . while it not used here explicitly , we note the fact that the operator in the left hand side of ( [ eqf ] ) after dividing it by the function @xmath414 becomes the generator of the limiting diffusion process on the edge of the graph corresponding to the domain @xmath185 ( cf @xcite and section [ se5 ] of this article ) . we need the following lemma , which will be proved in section [ se6 ] .
[ f ] there is a function @xmath415 , which satisfies @xmath416 , such that @xmath417 for all @xmath418 .
further , there is a constant @xmath342 such that @xmath419 and @xmath420 .
let us select constants @xmath421 and @xmath422 such that @xmath423 . define the subsets @xmath424 and @xmath425 of @xmath406 as follows : @xmath426 let @xmath427 be the first time the process @xmath352 leaves @xmath424 , similarly @xmath428 is the first time the process leaves @xmath425 .
let @xmath429 be the deterministic process @xmath430 and let @xmath431 be the time it takes the process @xmath429 starting at @xmath193 to make one rotation along the level set , @xmath432 .
the next lemma shows that for times of order @xmath431 the process @xmath357 is in a certain sense close to the deterministic process @xmath429 .
the lemma is proved in section [ se6 ] .
[ close ] for any @xmath157 there is @xmath433 such that @xmath434 there exist @xmath435 and @xmath433 such that @xmath436 one of the main ingredients of the proof of lemma [ d ] is the following lemma , which is a particular case of the main result ( theorem 2.3 ) of @xcite .
@xcite [ f ] there is a constant @xmath399 , such that on each level set @xmath437 we have @xmath438 the limit is uniform on each level set , and @xmath439 satisfies @xmath440 .
lemma [ f ] is different from lemma [ d ] in that the initial point in ( [ ef ] ) belongs to a fixed level set of @xmath7 , while in ( [ ed ] ) the initial point is asymptotically close to @xmath401 as @xmath40 .
+ + * proof of lemma [ d ] : * we shall demonstrate that there exists a function @xmath415 , such that @xmath441 , for which @xmath442 uniformly in @xmath443 .
let us show that ( [ tt1 ] ) is sufficient to prove the lemma .
as in the proof of lemma [ c ] , from ( [ tt1 ] ) it follows that @xmath444 uniformly in @xmath443 . for @xmath445 by the markov property @xmath446 the first term on the right side of ( [ equ ] ) is estimated from above by @xmath447 due to ( [ tt1 ] ) .
the second term has the following asymptotics due to ( [ in01 ] ) and lemma [ f ] @xmath448 for sufficiently small @xmath18 , where @xmath449 satisfies @xmath450 . the last term on the right side of ( [ equ ] )
is estimated from above by @xmath451 due to lemmas [ b ] and [ c ] , and the repeated use of ( [ equ ] ) .
therefore @xmath452 for sufficiently small @xmath18 , and @xmath453 which satisfies @xmath454 .
since @xmath455 can be selected arbitrarily small we obtain @xmath456 .
it remains to prove ( [ tt1 ] ) .
let @xmath457 .
we shall prove that for some @xmath458 for all sufficiently small values of @xmath455 @xmath459 for all @xmath460 . from ( [ m2 ] ) it follows that @xmath461 due to the estimate on the derivative of @xmath215 from lemma [ f ] for @xmath443 the right side of ( [ es02 ] ) is estimated from above by @xmath462 , which implies ( [ tt1 ] ) .
now we need to prove ( [ m2 ] ) .
applying ito s formula to @xmath463 we obtain @xmath464 while from ( [ eqf ] ) @xmath465 thus , what we want to show is that there is a constant @xmath158 such that for all @xmath460 @xmath466ds | + \ ] ] @xmath467 ds| + \ ] ] @xmath468ds |\leq k_1t(x)~.\ ] ] since @xmath469 and @xmath470 are bounded , and @xmath158 can be taken arbitrarily large , it is sufficient to estimate only those of the terms in ( [ bg ] ) which contain the second derivative of @xmath215 . by lemma [ close ]
we have @xmath471 , and the second derivative of @xmath215 can be estimated by lemma [ f ] .
therefore for the last term containing @xmath472 we have @xmath473 the estimate @xmath466ds | \leq ct(x)\ ] ] follows from lemma [ close ] and the estimates on @xmath472 and @xmath474 of lemma [ f ] .
this completes the proof of lemma [ d ] .
+ + * proof of theorem [ t1 ] : * the effective diffusivity @xmath17 is related to the variance @xmath119 of the limit of the functional of the markov chain , and to the integral of the expectation of the transition times via formula ( [ dd ] ) .
as shown in lemma [ l2 ] and in the remark following it , lemma [ l1 ] applies , and therefore there exists the limit @xmath475 .
the asymptotics of the integral @xmath476 is given by ( [ ai nt ] ) .
this completes the proof of the theorem .
in this section it will be convenient to consider the process @xmath477 , which is the same as @xmath13 , but only accelerated by the factor @xmath478 , that is @xmath479 .
this process satisfies the equation @xmath480 note , that as a process on the torus , @xmath477 is uniformly ( in @xmath18 ) exponentially mixing .
following @xcite we consider the finite graph @xmath481 which corresponds to the structure of the level sets of @xmath7 on the torus . the graph @xmath481 is constructed as follows : we identify all the points which belong to each connected component of each level set of @xmath7 . this way each of the domains
@xmath185 , bounded by the separatrices , gets mapped into an edge of the graph , while the separatrices themselves get mapped into the vertices .
let @xmath482 label the edge of the graph and let @xmath483 be the coordinate on the edge .
then the process @xmath484 can be considered as a process on the graph .
it is proved in @xcite ( theorem 2.2 ) that the process @xmath484 converges to a certain markov process on the graph with continuous trajectories , which is exponentially mixing .
we state the result here in less generality than in @xcite , but this is sufficient for our purposes .
@xcite there is a markov ( diffusion ) process @xmath485 on the graph @xmath481 , which is exponentially mixing , and has continuous trajectories , such that for any @xmath486 , the process @xmath484 converges to @xmath485 weakly in @xmath14,g)$ ] .
we are now in the position to prove theorem [ t2 ] . + * proof of theorem [ t2 ] * : the displacement of the process @xmath477 in the direction @xmath43 ( the direction of the channels ) is given by @xmath487 where @xmath488 is the @xmath43 component of the velocity field .
therefore , @xmath489 @xmath490 where @xmath491 is distributed according to the invariant ( lebesgue ) measure @xmath12 on @xmath492 . for a function @xmath493 ,
let @xmath494 , @xmath495 be the function defined on the graph , other than on the vertices , which is equal to the average of @xmath215 over the corresponding connected component of the level set of @xmath7 @xmath496 where @xmath429 is the solution of the deterministic equation @xmath497 , the initial point @xmath193 belongs to the level set , and @xmath431 is the time of one revolution around the level set .
it is easily seen that for any initial point @xmath193 which does not belong to any of the separatrices of @xmath7 we have @xmath498 therefore , @xmath499 @xmath500 where @xmath501 is the measure on @xmath481 , which is invariant for the process @xmath485 .
the integrals @xmath502 and @xmath503 can be made arbitrarily small by selecting sufficiently large @xmath504 due to uniform mixing of the processes @xmath505 and @xmath485 .
therefore , @xmath506 which shows that the asymptotics for @xmath507 is as stated in the theorem .
now let us consider the asympotics for @xmath508 .
note that @xmath509 , thus the arguments leading to ( [ ts ] ) do not provide the asymptotics of @xmath508 .
let @xmath510 be those of the separatrices of @xmath7 on the torus which , when unfolded onto the plane , are non - compact .
let us select a point @xmath56 on each of @xmath511 .
let us introduce the sequence of stopping times @xmath512 , @xmath340 , which are the consecutive times when @xmath505 makes the transition to a different level set @xmath511 .
thus @xmath513 is a markov chain on the set @xmath514 .
we can also consider the markov chain @xmath515 on the extended phase space @xmath516 .
the third component @xmath517 is defined as follows : if @xmath513 is considered on the plane , then @xmath518 where @xmath519 is the @xmath520 coordinate of the point corresponding to the separatrix containing the point @xmath513 .
similarly we can introduce the stopping times @xmath521 for the process @xmath485 on the graph , which are the consecutive times when @xmath485 visits different vertices @xmath522 of @xmath481 , corresponding to the unbounded separatrices of @xmath7 . together with the markov chain @xmath523
we can consider the chain @xmath524 on @xmath525 , where @xmath526 is defined the same way as @xmath527 .
let @xmath528 be the invariant measure for the chain @xmath529 , and let @xmath530 be the invariant measure for the chain @xmath531 .
let @xmath215 be the function defined on the state space of the chain @xmath529 , which is equal to the third component : @xmath532 .
the function @xmath533 is defined the same way on @xmath534 . by the central limit theorem applied to the chain @xmath535 ,
there is a number @xmath119 such that @xmath536 similarly , @xmath537 the effective diffusivity in the @xmath520 direction is then different form @xmath538 by the factor @xmath539 , @xmath540 from theorem 2.2 in @xcite and the uniform mixing of the markov chains @xmath535 and @xmath531 it easily follows that @xmath541 and @xmath542 this completes the proof of theorem [ t2 ] .
* proof of lemma [ f ] * : for any function @xmath543 we have @xmath544 where the integrals are over the level set @xmath545 .
in particular @xmath546 , and therefore equation ( [ eqf ] ) can be written as @xmath547 from the definition of the coefficients @xmath548 , and @xmath549 it easily follows that @xmath550 further , with the help of morse lemma and ( [ diff ] ) it is easily seen that @xmath551 let @xmath552 be the point where @xmath411 achieves its maximum , thus @xmath553 . from ( [ b1 ] )
it follows that @xmath554 thus , the estimate on the first derivative of @xmath215 stated in the lemma follows from ( [ abq ] ) .
rewrite ( [ eqf ] ) as @xmath555 from ( [ abq ] ) in now follows that @xmath556 for some @xmath342 .
differentiating both sides of ( [ eqf ] ) we obtain @xmath557 the estimate on @xmath558 now follows from the estimates on the first two derivatives and from ( [ abq ] ) and ( [ bq ] ) .
this completes the proof of the lemma .
+ + * proof of lemma [ close ] * : the proof is based on the use of lemma [ x ] .
we can not however apply lemma [ x ] to the pair of processes @xmath13 and @xmath429 directly , since the rotation time @xmath431 grows logarithmically in @xmath18 when @xmath460 .
let @xmath559 .
consider a process @xmath560 , which solves the equation @xmath561 on each of the segments @xmath562 $ ] , with a finite number of jump discontinuities @xmath563 , @xmath564 .
then for any positive @xmath565 there are positive @xmath566 and @xmath298 such that under the conditions @xmath567 we have @xmath568 note that it is sufficient to establish the following : for any pair of points @xmath569 such that @xmath570 and @xmath571 we have @xmath572 where @xmath573 and @xmath574 are the solutions for the deterministic flow ( [ flow ] ) .
let us take @xmath575 the time it takes for the trajectory of ( [ flow ] ) to make one rotation along the level set @xmath576 is equal to @xmath577 , where @xmath549 was defined in ( [ cff ] ) and is a smooth function for sufficiently small positive @xmath7 , which satisfies @xmath578 the number of full rotations of the trajectory starting from @xmath579 is equal to @xmath580 $ ] and the time it takes to make these rotations is equal to @xmath580t(a_0)$ ] .
it takes @xmath580t(b_0)$ ] to make the same number of rotations for the trajectory starting at @xmath581 .
due to ( [ asym ] ) the difference is estimated as follows ; @xmath582 t(a_0 ) - [ \frac{t}{t(a_0 ) } ] t(b_0)| \leq { \rm const } |\ln \varepsilon
| \varepsilon^{- \alpha_2 } ||a_0 -b_0||~.\ ] ] here we used the facts that @xmath583 and that @xmath584 .
now consider the images of @xmath579 and @xmath581 under the flow ( [ flow ] ) for time @xmath585 .
using the reduction of the flow to a linear system in a neighborhood of each of the saddle points ( hartman - grobman theorem @xcite ) , it is easy to show that @xmath586 combining this with ( [ ec1 ] ) and with the fact that the speed of motion in ( [ flow ] ) is bounded , we obtain ( [ one1 ] ) with @xmath298 and @xmath566 defined in ( [ gr ] ) . this in turn implies ( [ form1a ] ) as noted above .
note that @xmath587 and that by making @xmath566 smaller ( if necessary ) we can satisfy @xmath588 , where @xmath272 is the same as in ( [ tsx ] ) .
observe that for some @xmath589 we have @xmath590 select the points @xmath591 in such a way that @xmath592 for @xmath593 , where @xmath128 is the constant from lemma [ x ] ( applied to the pair of processes @xmath13 and @xmath429 ) . by ( [ length ] )
there is the estimate @xmath594 .
let @xmath595 be the piecewise continuous process , which is defined by the conditions : @xmath596 and @xmath597 on @xmath598 , @xmath599 . by lemma [ x ] @xmath600 due to continuity of @xmath13 formula ( [ y1 ] )
provides an estimate on the sum of the jumps of the process @xmath595 . from ( [ form1a ] ) it now follows that @xmath601 this , together with ( [ y1 ] ) implies ( [ tsn ] ) for any @xmath602 .
since @xmath603 is constant and @xmath604 is piecewise constant , we have @xmath605 this , together with ( [ y1 ] ) and the condition that @xmath606 implies ( [ tsx ] ) for any @xmath607 .
this completes the proof of the lemma .
+ + * acknowledgements * : i am grateful to prof
. s. molchanov for introducing me to this problem and for many useful discussions .
s. childress .
_ alpha - effect in flux ropes and sheets _ , phys .
earth planet internat . ,
20 ( 1979 ) , pp 172 - 180 . j. l. doob .
_ stochastic processes .
_ j. wiley and sons , 1953 .
a. fannjiang and g. papanicolaou .
_ convection enhanced diffusion for periodic flows_. siam j. appl .
vol 54 , no 2 , pp 333 - 408 , 1994 . m. i. freidlin .
_ dirichlet problem for equations with periodic coefficients_. probability theory and appl .
9 , pp 133 - 139 , 1964 .
m. i. freidlin and a. d. wentzell .
_ random perturbations of hamiltonian systems_. memoirs of the american math . society .
vol 109 , no 523 , may 1994
. s. heinze .
_ diffusion - advection in cellular flows with large peclet numbers _ , max - plank - institute , preprint no 63 , 2001
. l. perko , _ differential equations and dynamical systems _ , springer - verlag , 1991 .
b. shraiman , _ diffusive transport in a rayleigh - benard convection cell _ , phys .
a , 36 ( 1987 ) , p 261 .
a. m. soward _ fast dynamo action in steady flow_. j. fluid mech . , 180 ( 1987 )
, pp 267 - 295 . | we consider the motion of a particle in a periodic two dimensional flow perturbed by small ( molecular ) diffusion .
the flow is generated by a divergence free zero mean vector field .
the long time behavior corresponds to the behavior of the homogenized process - that is diffusion process with the constant diffusion matrix ( effective diffusivity ) .
we obtain the asymptotics of the effective diffusivity when the molecular diffusion tends to zero . |
w ursae majoris ( w uma ) variables are eclipsing overcontact binaries with orbital periods ranging from 0.2 1.0 day .
these systems consist of main sequence stars with spectral a - k type sharing a common convective envelope due to filled roche lobes . in some cases
these binaries host o or b spectral type component surrounded with a common radiative envelope , whose true physical understanding is still lacking . in low mass overcontact binaries , it has been argued that the secondary component is oversized with respect to its expected zams radius and at an advanced evolutionary stage ( stepien 2006a ) .
the role of common envelope is to distribute the energy uniformly over the surface of the stars ( lucy 1968 ) , having similar brightness with a few percent difference exhibiting chromospheric activity ( vilhu & walter 1987 ) .
the overcontact binaries are important astrophysical sources as they help to understand the underlying mechanism of the merging process ( eg .
v 1309 sco ; tylenda et al .
2011 ) , stellar dynamo process ( eg . qian et al .
2005 ) , contributing in understanding the galactic structure because of their high number density ( 1/500 ms stars ; rucinski 2002 ) , binary evolution theories ( eg . yakut & eggelton 2005 ) and also serve as distance estimators ( rucinski & duerbeck 1997 ) .
all the overcontact binaries are classified in three broad categories , a - type , w - type ( binnendijk 1970 ) and b - type ( csizmadia & klagyivik 2004 ) . in the a - type
, the less massive component eclipses the massive one causing the primary minimum and an opposite scenario is observed in case of the w - type . in general , a - types often have low mass ratio ( q @xmath3 0.3 ) , relatively long orbital periods ( p @xmath4 0.3 days ) , whereas w - types have mass ratios , q @xmath4 0.3 and short orbital periods ( p @xmath3 0.3 days ) .
it * had been suggested earlier that a - types are in an * advanced evolutionary stage compared to the w - types ( eg .
hilditch 1989 ) but later was overruled as a - types have more mass and angular momentum ( gazeas & niarchos 2006 ) . but a number of overcontact binaries known to harbor a third component causing the sinking of angular momentum and hence the discrepancy in the evolutionary status could be resolved by constraining their age .
the high temperature difference @xmath4 1000 k between the components in overcontact binaries forms the basis for b - type classification ( csizmadia & klagyivik 2004 ) and * systems in this class * are also known as poor thermal overcontact systems ( rucinski & duerbeck 1997 ) . many of the close binaries of several types and overcontact binaries light curves exhibit asymmetry in the brightness of maximum light , known as * oconnell * effect ( * oconnell * 1951 ; milone 1969 ; davidge & milone 1984 ) and is often associated with a dark spot on the primary component .
the strong evidence for the presence of the spot comes from the study of h@xmath0 line in overcontact binary systems .
the first detailed study was performed by barden ( 1985 ) on four w uma systems showing that the h@xmath0 line is a strong signature of the magnetic - associated activity in these systems .
the study of h@xmath0 line is also important as the magnetic field plays a key role in the evolution of overcontact binaries via the magnetic braking process ( stepien 1995 ) .
the presence of spots are related to the chromospheric activity causing the filling of the h@xmath0 line and varying the equivalent width along with the orbital period ( kaszas et al .
moreover due to this activity overcontact binary systems are also good x - ray emitters ( mcgale et al .
1996 ; stepien 2001 ; chen et al . 2006 ) and the related x - ray emission is connected to the stellar dynamo activity arising from the synchronous fast rotating convective common envelope ( gondoin 2004 )
. an illustrative study of vw cep ( kaszas et al .
1998 ) , ae phe and yy eri ( maceroni et al . 1994 ; vilhu and maceroni 2007 ) clearly suggest that the activity is related to primary / massive component ( as it has deep convective zones ) which is in agreement with the theoretical studies ( rucinski 1992 , 1994 ) .
deep ( f @xmath5 50 % ) low mass ratio ( q @xmath6 0.25 ) overcontact binaries ( dlmr ) are considered to be important sources and are possible progenitors for fk com - type and blue stragglers ( qian et al .
* although * a different naming / classification * was * adopted , most of them are a - type overcontact binaries .
they have a period domain ranging from [email protected] ( j13031 - 0101.9 ) to [email protected] ( kn per ) .
qian et al .
( 2006 ) found that few of the systems undergo secular period decrease .
the coupled action of angular momentum loss ( aml ) and thermal relaxation oscillation ( tro ; lucy 1976 ; flannery 1976 ; robertson & eggleton 1977 ) in the overcontact binary , leads to increase in the lifetime of overcontact phase . at this stage
the binary can meet hut s criteria i.e. j@xmath8 @xmath4 1/3 j@xmath9 ( hut 1980 ) or can encounter dynamical instability ( rasio & shaprio 1995 ) which results in merging of the components .
such mergers are rare but v1309 sco can be considered as a prototype for such events .
tylenda et al . (
2011 ) concluded that v1309 sco was a cool overcontact binary system and instabilities caused the secular period decrease of about 24.5 min over a duration of six years .
based on the formation models of the cool overcontact binaries ( stepien 2004 , 2006a , b , 2009 ) , stepien ( 2012 ) concluded that the loss of mass and angular momentum through magnetic winds played a crucial role in the merging process .
the variability of asas j082243 + 1927.0 ( v1 ) at @xmath0@xmath10=08@xmath11 22@xmath12 [email protected] and @xmath14@xmath10=+19@xmath15 26@xmath16 58@xmath17 was discovered in asas ( asas j082243 + 1927.0 ; pojamanski 2002 ) .
later light curves of this system were reported by gettel et al .
( 2006 ) with star i d 10100383 , pepper et al .
( 2008 ; i d kp400793 ) and terrell et al .
( 2012 ) . due to the small orbital period 0.@xmath728 and shape of the light curve ,
it was classified as an ew type variable .
the colours of the variable are b - v = 0.57 ( terrell et al . 2012 ) , j - h = 0.31 and h - k = 0.33 ( gettel et al . 2006 ) . due to the large scatter in the earlier reported light curves ,
the nature of the eclipses and presence of any oconnell effect were not clearly visible .
v1 exhibits total eclipses and hence mass ratio can be accurately estimated for such systems ( terrell & wilson 2005 ) .
moreover , study of the h@xmath0 line among overcontact binaries has been rare when compared to photometry . keeping this objective in view , we performed v band ccd photometry and spectroscopy concentrating on the variation of h@xmath0 line at various phases for this variable .
the v band photometric observations were made on january 19 , 20 and 22 , 2013 with the 2.0-m telescope of the iucaa - girawali observatory ( igo ) .
iucaa faint object spectrograph and camera ( ifosc ) was used which employes a 2k @xmath18 2k , thinned , back - illuminated ccd with a pixel size of 13.5@xmath19 m , gain 1.5e@xmath20 / adu and read out noise 4e@xmath20 .
the ccd provides a 10.5@xmath21 @xmath18 10.5@xmath21 image with a plate scale of 0.3 arcsec pixel@xmath22 .
the observations were performed in the air mass range 1.01 1.80 .
the images were acquired in bessell s v filter with an integration time of 10 s. the extinction corrections were not * applied * as the comparison is close to the variable and of similar brightness ( see figure 1 ) and transformation to the standard system was not done .
the figure 1 shows the location of variable ( v1 ) , comparison ( tyc 1386 - 1630 - 1 ) and check ( tyc 1386 - 121 - 1 ) stars . the v magnitudes and colour indexes of variable , comparison and check stars are shown in table 1 .
these stars were selected as they are nearby and of similar brightness to v1 . the magnitude difference between variable and comparison stars
along with check and comparison stars observed on 19 january 2013 are shown in figure 2 .
the magnitude difference between check and comparison stars was found to be constant @xmath1 0.02 @xmath23 0.002 .
spectroscopic observations were performed on december 56 , 2013 , with the 2.3-m vainu bappu telescope ( vbt ) of the vainu bappu observatory ( vbo ) equipped with the optomechanics research ( omr ) spectrograph along with a detector of 1k @xmath18 1k ccd . the obtained spectra cover a range of 3000 centered around h@xmath0 line with a 600 lines / mm dispersion resulting in a resolution of 2.6 /pixel .
the exposure time was 3540 min for both the variable v1 and a spectrophotometric standard ( bd+08 2015 ) .
the _ fene _ arc lamp was observed for wavelength calibrations .
we used different packages made available in _ iraf _ to reduce the data . in both photometric and spectroscopic reduction , bias and flat - field correction
were made and later _ apphot _ and _ onedspec _ packages were used to measure the magnitudes and to obtain the spectra . later
the spectra were normalized for further studies .
we determined the period of the variable using _
period04 _ package ( lenz & breger 2005 ) and times of minima using the kwee & van woerden s method ( 1956 ) .
the ephemeris determined for the variable is min@xmath24 = 2456312.2997(89 ) + [email protected](2)e .
the obtained period is similar to the one reported in previous studies .
the v band light curve covering all the phases with 418 data points is shown in figure 3 .
visual inspection of the light curve clearly shows the total eclipse at secondary minimum along with an oconnell effect notable at phase 0.25 .
based on the colour indices of the variable , i.e. , b - v=0.57 and j - h=0.31 , the primary component temperature was fixed at 5960 k. the photometric solutions were obtained using the latest version of wilson - devinney ( w - d ) code v2013 with an option of non - linear limb darkening via a square root law ( wilson & devinney 1971 ; wilson 1979 ; wilson 1990 ; van hamme & wilson 2007 ; wilson 2008 ; wilson & van hamme 2013 ) .
the new code has 60 parameters among which 50 are * adjustable*. in this version , a high precision star spot algorithm is incorporated which allows * one * to develop time varying spots by adjusting times of onset , size , and disappearance which affects the light curve * ; * the new code also permits the spot to drift , independent of star rotation ( wilson 2012 ) .
the distance of binaries can be estimated via direct distance estimation ( dde ) process using a control integer ifcgs=1 in the code , but to adjust this parameter , the code requires well calibrated photometry in at least two passbands ( wilson et al .
we adopted the following * model and * method ( eg .
ravi , sriram & vivekananda rao 2012 ; shanti , sriram & vivekananda rao 2013 ) .
a convective outer envelope was assumed for both components ; gravity darkening co - efficients @xmath25 = 0.32 ( lucy 1967 ) and bolometric albedos @xmath26 = 0.5 ( rucinski 1969 ) were taken as fixed parameters . the value of the limb darkening coefficients of components @xmath27 were fixed at 0.64 ( van hamme 1993 ) for v band .
the adjustable parameters were the following : temperature of secondary component @xmath28 , orbital inclination ( _ i _ ) , the dimensionless potentials of the primary component @xmath29= @xmath30 and the bandpass luminosity of the primary star @xmath31 .
the overcontact configuration i.e. mode 3 option of w - d method was used to determine the parameters after overruling the mode 2 option .
+ as no spectroscopic mass ratio was determined for this variable , the grid search method was adopted . the total eclipse at secondary
minimum suggests that it is a low mass ratio system ( 0.1 @xmath3 q @xmath3 0.2 ) ( terrell & wilson 2005 ; wilson 2006 ) .
we searched the parameter space of q in the range of 0.02 @xmath3 q @xmath3 10.0 along with other adjustable parameters . during the computations , the solutions gradually converged from detached ( mode 2 ) to overcontact ( mode 3 ) configuration .
the resulting sum of the weighted square deviations , @xmath32 over the selected range of mass ratio q were noted and a minima at q = 0.121 was observed .
later , it was made as an adjustable parameter along with others in the differential correction routine . during the computations , high residuals
were observed at phase 0.25 and hence a solution with a dark spot over primary / secondary component and group of spots were tested to model the observed oconnell effect .
first computation was * done * with an * application of a dark spot * over * the * primary , which resulted in lowering the residuals ( dark 1 in table 2 ; fig .
3 , top panel ) . a relatively dark and large spot over * the * secondary , resulted in a good fit ( dark 2 in table 2 ) .
we also attempted a solution by varying the spot longitude over * the * primary ( ifsmv1=1 ) and drift of the spot was made independent * of * the star rotation ( eg .
f1@xmath33=0.80 ) but the overall solution remained almost the same except that the longitude and latitude of the spot varied slightly for the fixed size of the spot .
if f1@xmath33 is varied below 0.8 then the solution was found to be diverging .
we also found that * the * secondary temperature is higher than the primary s although an occultation is observed at secondary minimum ( table 2 ) .
based on the solutions , we found that the best value of mass ratio to be q @xmath1 0.106 . to check the consistency of the solution ,
the temperatures and inclination were varied by 510% and no significant difference in the respective solutions were found .
the table 2 shows the photometric elements obtained from the best solutions and the corresponding theoretical fits are shown in figure 3 . the fill - out factor or degree of overcontact parameter ,
f = @xmath34 was derived for all the solutions and were found to be in the range of @xmath1 6172 % .
the obtained values of q and f reveals that the variable is a low mass ratio overcontact binary with a high degree of geometrical contact .
+ the h@xmath0 line was observed at phases : 0.15 , 0.33 , 0.51 and 0.81 for the variable and its mean equivalent width ( ew ) was found to be 1.60 @xmath23 0.13 .
figure 4 shows the variation of the h@xmath0 profiles with respect to the spectrophotometric standard .
it is evident that at phase 0.51 , the variable h@xmath0 line is possibly filled - in with respect to the standard star at a significance of about 2.7@xmath35 whereas at other phases , * the * filled - in feature is relatively weak .
the * fill - in * , especially at phase 0.5 is possibly caused by chromospheric emission .
this is also supported by the observed oconnell effect exhibited in the light curve due to * the * presence of a dark spot on the primary or secondary . as continuous spectroscopic observations are lacking , determining the differences in the profiles arising from individual component and obtaining the h@xmath0 ew for each component were not possible .
however at phase 0.5 , when the secondary is being eclipsed , we argue that the ew@xmath36@xmath37 = 1.09 of h@xmath0 line is most probably arising from the primary component .
the intrinsic equivalent width of the line at phase 0.5 was calculated using the relation + ew@xmath36@xmath37=a @xmath18 ew@xmath36 and ew@xmath38@xmath37=a @xmath18 ew@xmath38 , where a = 1 + q@xmath39 ( @xmath40)@xmath41 ( webbink 2003 ) , ew@xmath36 , ew@xmath38 are measured equivalent widths and ew@xmath36@xmath37 , ew@xmath36@xmath37 are intrinsic equivalent widths . a small sample of periods and ews for overcontact binaries viz .
ae phe , yy eri ( vilhu and maceroni 2007 ) , vw cep ( kaszas et al . 1998 ) suggests a correlation between the two parameters ( fig .
4 ) . since the conclusion of presence of the h@xmath0 filled - in
was limited to visual inspection and hence a continues spectroscopic monitoring of the variable is important to unveil the nature of the h@xmath0 line .
the v band light curve shows a total eclipse at secondary minimum and exhibits * the oconnell * effect .
these features were not clear in the previous published light curves .
however the variable oconnell effect is not unusual in overcontact binaries .
based on the derived solutions , perhaps the variable can be classified as a w - type w uma system , * as the * secondary temperature was found to be relatively higher ( table 2 ) .
however the shape of the light curve , derived mass ratio q @xmath10.106 and f @xmath1 72% indicate that the variable is a low mass ratio overcontact binary with a high degree of geometrical contact , a unique characteristic of a - type w uma systems .
the best fit solutions indicate the presence of a dark spot on primary or secondary ( table 2 ) .
the presence of spots do not affect the mass ratio of a system and hence an a - type classification is justifiable .
in general , the occultation at secondary minimum suggests that the primary is hotter than the secondary component but our solutions indicate that the secondary is slightly hotter .
this phenomenon is probably caused due to the migration of the spot on the stellar surface , responsible for causing variable oconnell effect and could * account * for reversed minima , making an a - type to a w - type and vice - versa .
qian & yang ( 2005 ) reported a - type to w - type and then w - type to a - type features in the light curves of the overcontact binary fg hya .
overall , there are three such low mass ratio overcontact binaries which resembles these features viz .
fg hya ( zola et al . 2010 ; qian & yang 2005 ) , v802 aql ( samec et al .
2004 ; yang et al . 2008 ) and v902 sgr ( samec & corbin 2002 ) .
if the dark spot / spots are responsible for the ambiguity of v1 s classification ( a - type or w - type ) , a radial velocity spectroscopy is necessary to unveil the true nature of the source . the combination of low mass ratio ( q @xmath6 0.25 ) and high filling factor in overcontact binaries ( f @xmath5 50 % ) are considered to be the progenitors for fk com - type and blue straggler stars .
the gradual decrease of mass ratio makes these systems evolve as single fast rotating stars due to the inducement of darwin s instability when the orbital angular momentum is more than three times of spin angular momentum , i.e. , j@xmath9 @xmath5 3 j@xmath42 ( hut 1980 ) .
the merging is also possible when the degree of overcontact is high which helps in causing the dynamical instability ( rasio & shapiro 1995 ) .
table 3 lists the high filling factor ( f @xmath4 50% ) and low mass ratio ( q @xmath3 0.25 ) overcontact binary systems . for these systems , we studied the correlation between the mass ratio and period ( fig .
interestingly , we found systematic trends along the mass ratio with a cut - off .
we fitted a broken power law model @xmath43 , where p1 is break ( a value of log q ) , p2 and p3 are power - law indexes before and after the break and p4 is the normalization .
we found that the break is at log q = -1.071 @xmath23 0.04 and the error bar indicates the 90 % confidence limit .
the line in figure 5 shows the best fit and obtained a cut - off mass ratio q = 0.085 . the mass transfer from primary to secondary results in a short period high mass ratio systems ( i.e. trend below q@xmath2 @xmath6 0.085 ) . however above q@xmath2 @xmath4 0.085 , as a system has to evolve from high to low am state , the simultaneous processes of mass transfer from primary to secondary and mass loss of secondary component and partial transferred mass from lagrangian point l2 can drive the system towards short period low mass system configuration .
moreover as period decreases , the critical roche lobes shall be shrinking , making f to increase and finally evolve as a single rapid rotator supported from the sample ( table 3 ) i.e. high f values are associated with low mass ratio systems . theoretically a critical mass ratio has been predicted because till now no overcontact binaries with low q @xmath3 0.07 have been discovered .
rasio & shaprio ( 1995 ) constrained a minimum mass ratio of 0.09 and while considering the rotation of the secondary , the critical mass ratio , q@xmath2 lowers to 0.076 ( li & zhang 2006 ) . assuming convective and radiative envelope in both secondary and primary
, the q@xmath2 was found to be around 0.0940.109 ( arbutina 2007 ) and taking into account differential rotation of the primary , the q@xmath2 converges at 0.0700.074 .
the mass ratio of the variable v1 is found to be @xmath10.106 , close to the theoretically predicted values .
we have also plotted the period - colour relation for 110 overcontact binary systems from the data of terrell et al .
the line in figure 6 shows the short period blue envelope ( spbe ) theoretical line i.e. b - v@xmath44 = 0.04 @xmath18p@xmath45 ( rucinski 1997 ) and the variable v1 is found to be above it .
any overcontact binary above spbe line has to further evolve towards longer period and cooler temperature ( towards lower right in figure 6 ) .
overall , the variable v1 is an important overcontact binary source as it has the low mass ratio and high fill - out factor ( see table 3 ) and these kinds of sources are considered to be at a stage of merger , potentially forming a fk - com / blue straggler type star .
moreover the solutions suggest that the a dark spot is affecting the occultation at secondary minimum and hence v1 is an interesting source for future observations which may unveil the variable oconnell effect .
the h@xmath0 6563 line emission in overcontact binaries is sparsely studied .
this line along with ca h and k , fe ii and the mg ii resonance line in the uv are the primary indicators of chromospheric activity in solar type stars.the strong absorption of the h@xmath0 line forms the basis of a zero point for the measure of chromospheric activity , however few stars are found to be associated with the weak level of activity and shows weak absorption .
the fill - in effect is generally explained as the consequence of chromospheric activity which decreases the equivalent width of the h@xmath0 absorption line .
this scenario was first observed in overcontact binaries v566 oph , ah vir and w uma by barden ( 1985 ) .
the h@xmath0 variation in vw cep was studied by barden ( 1985 ) , frasca et al .
( 1996 ) and kaszas et al .
( 1998 ) and concluded that activity is associated with the primary rather than secondary component in overcontact binaries . from fig .
4 it is clear that at phase 0.5 the h@xmath0 is more filled - up suggesting a high level of chromospheric activity associated with the primary .
the conclusion was based on the visual inspection , hence a continues spectroscopic monitoring is needed to explain the nature of the h@xmath0 profile .
because h@xmath0 line ew varies due to chromospheric activity , we studied the correlation between period and ew for a few overcontact binaries ( fig .
it is clear that as the orbital period decreases ( i.e fast rotation ) , the ew of the h@xmath0 line is found to be decreasing ( which is measure of increasing activity ) , leading towards the saturation at lower periods .
overcontact binaries are potential sources to estimate the distance of binary systems .
rucinski & duerbeck ( 1997 ) introduced * a * calibration scheme to determine the distance of overcontact binaries and the best calibration is given by m@xmath46= -4.44 log p + 3.02 ( b - v)@xmath47 + 0.12 , where m@xmath46 can be predicted with an uncertainty of up to @xmath23 0.25 mag . using this relation and m@xmath46 = 11.61
, the distance was estimated to be @xmath1279 pc .
gettel et al .
( 2006 ) also estimated the distance to the variable to be @xmath1306 pc based on the relation log d = 0.2 v@xmath48 - 0.18 log ( p ) -1.60 ( j - h ) + 0.56 .
considering the error of @xmath230.25 mag in the value of m@xmath46 , the difference in estimated distances is found to be within the error bars . based on
well studied overcontact binaries , gazeas ( 2009 ) obtained three dimensional correlations as stated below : + log @xmath49 = 0.725 log p - 0.076 log q + 0.365
+ log @xmath50 = 0.725 log p + 0.924 log q + 0.365 + log @xmath51 = 0.930 log p - 0.141 log q + 0.434 + log @xmath52 = 0.930 log p + 0.287 log q + 0.434 + log @xmath53 = 2.531 log p - 0.512 log q + 1.102 + log @xmath54 = 2.531 log p + 0.352 log q + 1.102 + using these correlations ( gazeas 2009 ) , the absolute parameters obtained are @xmath55= 1.10 @xmath23 0.08 m@xmath56 , @xmath57= 0.17 @xmath23 0.05 m@xmath56 , @xmath58= 1.15 @xmath23 0.04 r@xmath56 , @xmath59= 0.42 @xmath23 0.02 r@xmath56 , @xmath31= 1.66 @xmath23 0.24 l@xmath56 and @xmath60= 0.22 @xmath23 0.02 l@xmath56 . in conclusion ,
we report the photometric solution and variation of the h@xmath0 line of a overcontact binary asas j082243 + 1927.0 .
the high value of f and low q suggests that it is a low mass ratio overcontact binary with high degree of contact , making it as one of the potential sources for merger studies .
we found a cut - off mass ratio in mass ratio period plane at q@xmath2 = 0.085 , closely matching with various theoretically predicted ones .
the h@xmath0 line varied at different phases showing the relatively fill - in effect at secondary minimum and is arising arguably from the primary component .
a small sample of h@xmath0 ew and period of overcontact binaries shows a correlation indicating a saturation towards chromospheric emission .
a spectroscopic radial velocity study is essential to resolve the ambiguity of the variable s classification ( a - type or w - type ) , to confirm the photometric mass ratio and to constrain the explicit nature of filled - in h@xmath0 line .
moreover a long term photometric study of this variable is necessary to know its secular period variation and explore the variable oconnell effect .
we acknowledge the director of iucaa , pune and the director of iia , bengaluru for allocating the time for the observations at igo and vbo .
we also thank dr .
vijay mohan , iucaa for helping us with the observations .
we thank the referee for the valuable comments and suggestions which improved the quality of the paper . k.s
acknowledge the support from ugc - bsr research start - up grant , government of india . ccccccc name & v & b - v + + v1 ( variable)&11.62 @xmath230.13&0.57@xmath23 0.01 + tyc 1386 - 1630 - 1 ( comparison ) & 11.58 @xmath230.13&0.48 @xmath230.08 + tyc
1386 - 121 - 1 ( check ) & 11.99 @xmath23 0.21 & 0.54 @xmath230.08 + ccccccc parameters & & unspotted & dark 1 & dark 2 + + a@xmath61 = a@xmath62 & & 0.50 & 0.50 & 0.50 + g@xmath61 = g@xmath62 & & 0.32 & 0.32 & 0.32 + @xmath63(k ) & & 5960 & 5960 & 5960 + @xmath28 ( k ) & & 6080 @xmath2334 & 6078 @xmath23 30 & 6038@xmath2328 + q & & 0.110 @xmath230.004 & [email protected]&[email protected] + i@xmath64 & & 75.60 @xmath231.04 & [email protected]&[email protected] + @xmath65 & & 1.9218 @xmath230.0039 & [email protected]&[email protected] + fill - out factor ( f % ) & & 73.82 @xmath231.32 & [email protected]&[email protected] + @xmath66 & pole & 0.5278 @xmath230.0036 & [email protected] & [email protected] + & side & 0.5865 @xmath230.0055 & 0.6128 @xmath230.0019 & [email protected] + & back & 0.6068 @xmath230.0057 & 0.6349 @xmath230.0020 & [email protected] + @xmath67 & pole & 0.1985 @xmath230.0154 & [email protected] & 0.1879 @xmath230.0090 + & side & 0.2062 @xmath230.0182 & [email protected] & 0.2059 @xmath230.0109 + & back & 0.2365 @xmath230.0348 & [email protected] & 0.2089 @xmath230.0219 + @xmath68 & & [email protected]&[email protected]&[email protected] + spot colatitude ( @xmath69 ) & & & [email protected] & 79.45 @xmath231.37 + spot longitude ( @xmath69)&&&[email protected]&[email protected] + spot radius(@xmath69)&&&[email protected]&30.93 @xmath231.69 + t@xmath70/t@xmath71 & & & 0.74 @xmath230.04 & 0.64 @xmath230.07 + @xmath72w(o - c)@xmath73 & & 0.0088 & 0.0024&0.0022 + acerbi , f. , barani , c. , & martignoni , m. 2014 , new astronomy , 31 , 1 acerbi , f. , barani , c. , & martignoni , m. 2011 , raa , 11 , 843 arbutina b. , 2007 , mnras , 377 , 1635 arbutina b. , 2009 , mnras , 394 , 501 barden , s. c. 1985 , , 295 , 162 binnendijk , l. 1970 , vistas astron .
, 12 , 217 broens , e. 2013 , , 430 , 3070 chen , w. p. sanchawala , k. & chiu , m. c. 2006 , , 131 , 190 csak , b. , kiss , l. l. , vinko , j. , & alfaro , e. j. 2000 , a&a , 356 , 603 csizmadia , s. , & klagyivik , p. 2004 , a&a , 426 , 1001 davidge & milone 1984 , apjs , 55 , 571 essam , a. , saad , s. m. , nouh , m. i. , dumitrescu , a. , elkhateeb , m. m. & haroon , a. 2010 , new astronomy , 15 , 227 flannery , b. p. 1976 ,
apj , 205 , 217 frasca , a. , sanfilippo , d. & catalano , s. 1996 , 313 , 532 gazeas , k. d. , & niarchos , p. g. 2006 , , 370 , 29 gazeas , k. d. 200 , communications in asteroseismology , 159 , 129 gettel , s. j. , geske , m. t. , & mckay , t. a. 2006 , , 131 , 621 goderya , s. n. , leung , k. c. , & schmidt , e. g. 1997a , ap&ss , 246 , 291 goderya , s. n. , leung , k. c. , & schmidt , e. g. 1997b , ap&ss , 254 , 295 gondoin , p. 2004 , a&a , 415 , 1113 hilditch , r. w. 1989 , ssrv , 50 , 289 hut , p. 1980 , a&a , 92 , 167 kaszas , g. , vinko , j. , szatmary , k. , hegedus , t. , gal , j. , kiss , l.l . & borkovits , t. 1998 , a&a , 331 , 231 kilmartin , p. m. , bradstreet , d. h. , & koch , r. h. 1987 , aj , 319 , 334 kwee , k. k. , & van woerden , h. 1956 , ban , 12 , 327 lapasset , e. 1980 , aj , 85 , 1098 labadorf , c. , samec , r. g. , figg , e. r. , fronaulkner , d. r. , & van hamme , w. 2009 , baas , 41 , 212 lenz , p. , & breger , m. 2005 , commun .
144 , 41 li , l. , & zhang , f. 2006 , , 369 , 2001 li , l .- f . ,
liu , q .- y . ,
zhang , f .- h . , & han , z .- w . 2001 , aj , 121 , 1091 lucy , l. b. 1967 , z. astrophys . , 65 , 89 lucy , l. b. 1968 , 151 , 1123 lucy , l.b . 1976 , , 205 , 208 maceroni , c. , vilhu , o . , vantveer , f. , & vanhamme , w .
1994 , a&a , 288 , 529 mcgale , p. a. , pye , j. p. , & hodgkin , s. t. 1996 , mnras , 280 , 627 maciejewski , g , czart , k. , niedzielski , a. , & karska , a 2003 , ibvs , 5400 , 1 milone , e. f. 1969 , communication of the konkoly observatory , 65 , 457 nelson , r. h. , terrel , d. , & milone , e. f. 2014 , new astronomy reviews , 59 , 1 oconnell 1951 , pub .
riverview college obs . , 2 , 85 pepper , j. , stanek , k. z. , pogge , r. w. , latham , d. w. , depoy , d. l. , siverd , r. , poindexter , s. , & sivakoff , g. r. 2008 , , 135 , 907 pojamanski , g. 2002 , acta astron . , 52 , 397 pribulla , t. , chochol , d. , rovithis - livaniou , h. , & rovithis , p. 1999
, a&a , 345 , 137 pribulla , t. & vanko , m. 2002 , coska , 32 , 79 pribulla , t. , rucinski , s. m. , blake , r. m. , et al .
2009 , aj , 137 , 3655 qian , s. b. & liu , q. 2000 , a&a , 355 , 171 qian , s. b. 2001a , mnras , 328 , 635 qian , s. b. 2001b , mnras , 328 , 914 qian , s. b. & yang , y. g. 2004 , aj , 128 , 2430 qian , s. b. & yang , y. , zhu , l. , he , j. & yuan , j. 2006 , ap&ss , 304 , 25 qian , s. b. , liu , l. , soonthornthum , b. , zhu , l. y. & he , j. j. 2007 , aj , 134 , 1475 qian , s. b. , he , j .- j .
, soonthornthum , b. , et al .
2008 , aj , 136 , 1940 qian , s. b. , yang , y .- g . , soonthornthum , b. , et al .
2005 , aj , 130 , 224 qian , s. b. & yang , y. 2005 , mnras , 356 , 765 qian , s. b. , liu , l. , zhu , l. y. , he , j. j. , yang , y. g. & bernasconi , l. 2011 , aj , 141 , 151 rasio , f. a. , & shapiro , s. l. 1995 , apj , 438 , 887 ravi , k. y. , sriram , k. & vivekanada rao , p. 2012 ,
basi , 40 , 51 robertson , j. a. , & eggleton , p.
, , 179 , 359 rucinski , s. m. 1969 , acta astron . , 19 , 245 rucinski , s.m . , 1992 , aj 103 , 960 rucinski , s.m . , 1994 , aj 107 , 738 rucinski , s.m . , 1997 , aj , 113 , 407 rucinski , s. m. , & duerbeck , h. w. 1997 , pasp , 109 , 1340 rucinski , s. m. , & lu , w. 1999 , aj , 118 , 2451 rucinski , s. m. 2002 , pasp , 114 , 1124 samec , r. g. , & corbin , s. 2002 , the observatory , 122 , 22 samec , r. g. , martin , m. , & faulkner , d. r. , 2004 .
stars , 5527 samec , r. g. , labadorf , c. m. , hawkins , n. c. , faulkner , d. r. & van hamme , w. 2011 , aj , 142 , 117 senavci , h. v. , nelson , r. h. , ozavci , i. , selam , s. o. , & albayrak , b. 2008 , new astronomy , 13 , 468 shanti priya , d. sriram , k. & vivekananda rao , p. 2013 , raa , 13 , 465 stepien , k. 1995 , , 274 , 1019 stepien , k. , schmitt , j. h. m. m. , & voges , w. 2001 , a&a , 370 , 157 stepien , k. 2004 , in stars as suns : activity , evolution and planets , ed .
a. k. dupree , & a. o. benz , asp , iau symp .
, 219 , 967 stepien , k. 2006a , acta astron . ,
56 , 199 stepien , k. 2006b , acta astron . , 56
, 347 stepien , k. 2009 , mnras , 397 , 857 stepien , k. 2006 , act .
, 56 , 199 stepien , k. 2012 , a&a , 531 , a18 szalai , t. , kiss , l. l. meszaros , sz .
vinko , j. & csizmadia , sz .
2007 , a&a , 465 , 943 terrell , d. , gross , j. , & conney , w. r. jr .
2012 , , 143 , 99 terrell , d. , & wilson , r. e. 2005 , ap&ss , 296 , 221 tylenda , r. , hajduk , m. , kaminski , t. et al .
2011 , a&a , 528 , a114 ulas , b. , kalomeni , b. , keskin , v. , kose , o. , & yakut , k. 2012 , new astronomy , 17 , 46 van hamme , w. 1993 , , 106 , 209 van hamme , w. & wilson , r. e. 2007 , apj , 661 , 1129 vilhu , o. , & walter , f. m. 1987 , , 321 , 958 vilhu , o. , & maceroni , c. 2007 , iuas , 240 , 719 webbink r.f .
2003 , in turcotte s. , keller s. , keller s.c .
, carallo r.m .
( eds . ) , asp conf.ser.vol.293 , 3d stellar evolution contact binaries , astron.soc.pac . ,
san francisco , p.76 .
wilson , r. e. , & devinney , e. j. 1971 , apj , 166 , 605 wilson , r. e. 1979 , apj , 234 , 1054 wilson , r. e. 1990 , apj , 356 , 613 wilson , r. e. , & van hamme , w. 2013 , computing binary star observables wilson , r. e. 2006 , aspc , 349 , 71 wilson , r. e. 2008 , apj , 672 , 575 wilson , r. e. , & van hamme , w. & terrel , d. 2010 , apj , 723 , 1469 wilson , r. e. 2012 , aj , 144 , 73 yakut , k. , kalomeni , b. , & ibanoglu , c. 2004 , a&a , 417 , 725 yakut , k. & eggleton , p. p. 2005
, , 629 , 1055 yang , y .- g . , qian , s .- b . , & zhu , l .- y . 2005 , aj , 130 , 2252 yang , y .- g
qian , s .- b .
, & zhu , l .- y .. liu , l. , & nakajima , k. 2008 , pasj , 60 , 803 yang , y .- g . , qian , s .- b . , zhu , l .- y . , & he , j .- j . 2009 , aj , 138 , 540 yang , y .- g . ,
qian , s .- b . & soonthornthum , b. 2012 , aj , 143 , 122 yang , y .- g .
, qian , s .- b . , zhang l. y. , dai h. f. , soonthornthum , b. 2013 aj , 146 , 35 zhu , l .- y . ,
qian , s .- b . , soonthornthum , b. , & yang , y .- g .
2005 , aj , 129 , 2806 zola , s. , gazeas , k. , kreiner , j. m. , ogloza , w. , siwak , m. , koziel - wierzbowska , d , & winiarski m. 2010 , mnras , | both high precision ccd photometric and h@xmath0 line studies are presented for an overcontact binary asas j082243 + 1927.0 . the light curve exhibits a total eclipse at secondary minima along with an * oconnell * effect . the light curve was modeled using the wilson devinney code and the best solution provides the mass ratio q @xmath1 0.106 and fill - out factor f @xmath1 72% .
these parameters indicate that the system is a low mass ratio overcontact binary with a high degree of geometrical contact .
the h@xmath0 line equivalent width varied at different phases and it is found that the line is possibly filled - in at secondary minima . from a small sample of overcontact binaries , we found a correlation between the orbital period and h@xmath0 line equivalent width of the primary component .
based on a sample of high filling factor and low mass ratio contact binaries , a mass ratio cut - off is observed at q@xmath2 = 0.085 in mass - ratio period plane . it was observed that below q@xmath2 @xmath3 0.085 , period decreases with an increase in q and above it , period increases as the mass ratio increases .
interestingly , the observed mass ratio cut - off value lies close to the critical mass ratio range as predicted in the literature .
the observational evidence of the cut - off of the mass ratio and its variation with orbital period * are * discussed in terms of mass transfer and angular momentum loss . based on the results
, we suggest that , asas j082243 + 1927.0 is at the verge of merger , eventually forming a fast rotating star .
[ firstpage ] binaries : close binaries ; binaries eclipsing ; stars : individual ( asas j082243 + 1927.0 ) |
the nature of dark matter remains one of the outstanding questions of modern astrophysics .
the success of the cold dark matter cosmological model ( albeit with `` dark energy '' now required : @xmath3cdm ) argues strongly for a major component of the dark matter being in the form of an elementary particle .
however , the inventory of baryons which we can observe locally falls far short of the total inferred from observations of the cosmic microwave background fluctuations @xcite , leaving open the possibility that there may be a significant baryonic component of dark matter .
furthermore , although @xmath3cdm is very successful in describing the growth of structure in the universe on large scales , we still lack a direct detection of any of the candidate dark matter particles . lacking this decisive piece of observational evidence , some authors have proposed models which include a large component of baryonic dark matter . in particular
there have been many papers dealing with the possibility that cold , self - gravitating molecular clouds constitute a major component of the dark matter @xcite . a variety of different forms , including isolated , clustered , and fractal ,
have been considered for the clouds , but all proposals involve dense gas of high column - density , in contrast to the diffuse gas in the interstellar medium which is easily detected in emission and/or absorption .
one of the fundamental predictions of a model featuring dense gas clouds is the gamma - ray emission resulting from cosmic - ray interactions within the clouds @xcite .
because of the potentially large total mass of gas involved , this process may yield a diffuse flux in the galactic plane comparable to the flux from known sources for photon energies around 1 gev @xcite .
considering the high quality data on diffuse emission acquired by the egret detector aboard the compton gamma ray observatory @xcite , it is worth considering this source of gamma - ray emission in detail as it is possible to use these data to constrain the dark matter models ( see @xcite ; @xcite ) .
most previous investigations of this problem have neglected the self - shielding and cascade phenomena which can be important at high column densities @xcite , and have employed emissivities appropriate to the low - density limit .
these effects alter the emergent gamma - ray spectrum , and we note that this could be relevant to the observed excess galactic flux above 1 gev @xcite .
we have noted elsewhere @xcite that massive ( @xmath4 ) aggregates of dense gas clouds could potentially account for many of the unidentified discrete sources detected by egret @xcite .
here we present detailed calculations of the gamma - ray spectra arising from cosmic - ray interactions with dense gas clouds .
we have used a monte carlo simulation code , geant4 , developed for simulating interaction events in detectors used in high - energy particle physics . not surprisingly , we find that the predicted spectra differ substantially between high and low column - density clouds , and we discuss the interpretation of our results in the context of the observed galactic gamma - ray emission .
our calculations are undertaken for cold , dense molecular gas in clouds of radius @xmath5 cm , similar to those proposed by @xcite to explain the extreme scattering events @xcite during which compact extragalactic radio sources are magnified and demagnified as a plasma `` lens '' moves across the line of sight ( see @xcite for a criticism of this model ) .
however , the results of our calculations depend primarily on the column - density of the individual clouds , @xmath6 , under consideration , and their fractional contribution to the galaxy s dark matter halo , and our results can be taken as representative of other models which are characterised by similar values of these quantities .
previous calculations of gamma - ray spectra from cosmic - ray irradiation assumed single interactions of protons with the interstellar medium ( @xcite and references therein ) . in order to investigate cosmic - ray interactions with dense gas , where cascade processes and particle transport are important
, we have used a monte carlo code , geant4 , to derive gamma - ray production spectra .
this code is a general purpose monte carlo code of particle interactions and is widely used for simulation of high - energy particle detectors in accelerator experiments .
cross - sections and interactions of various hadronic processes , i.e. , fission , capture , and elastic scattering , as well as inelastic final state production , are parametrized and extrapolated in high and low particle energy limits , respectively . the @xmath7 production in this code , which is important because of the @xmath8 decay that dominates
the emissivity of the gas at high energies , has been tested against accelerator data @xcite .
initially we experienced one slight difficulty in applying geant4 to our physical circumstance : the low - energy hadron interaction code , called gheisha , did not conserve energy very accurately ( geant4 bug reports no . 171 and 389 ) .
a `` patch '' was available for gheisha ( geant ver .
4.4.1 ) , but this patch appeared to introduce further problems of its own in the energy deposition distribution ( geant4 bug report no .
these difficulties have been overcome by the geant team , and we are not aware of any such problems in the latest release ( geant ver .
4.5.1 ) .
our calculations assume a spherical cloud of molecular hydrogen of uniform density and temperature ( 10 k ) . the radius of the sphere was assumed to be @xmath9 au . protons and electrons
are injected randomly at a surface point of the cloud and particles subsequently emanating from this surface are counted as products .
the adopted spectra of cosmic - ray protons and electrons were taken from @xcite ( here we use the `` median '' flux ; note that the units on his equation ( 3 ) should read @xmath10s@xmath11sr@xmath11gev@xmath11 ) , and @xcite , respectively . in figs .
[ fig : f1 ] and [ fig : f2 ] we plot these spectra together with observational data .
( note that the adopted spectra are those for cosmic rays in the galaxy , whereas the measured points are subject to the modulating influence of the magnetic field of the solar wind ) .
the simulated range of kinetic energy of cosmic rays is from 10 mev to 10 tev .
we divided this energy range into four and superposed the resulting spectra with appropriate weight factors in order to increase the simulation statistics at higher energies , considering the rapidly falling spectrum of cosmic rays .
the density of molecular hydrogen , @xmath12 , was varied from @xmath13 to @xmath14 g @xmath15 in factors of 10 .
this corresponds to the column density , @xmath16 , of @xmath17 g @xmath10 , respectively , where @xmath18 is the incident angle of a cosmic ray into a cloud and @xmath19 for random injection .
figure [ fig : f3 ] shows the resulting gamma - ray spectra obtained in our geant4 simulation for proton injection into a cloud of column density @xmath20 .
( the quantity plotted is @xmath21 , with @xmath22 being the emissivity as given in 3 . )
the dashed , dotted and dot - dashed lines show the spectral components classified by the parent processes producing gamma - rays , i.e. , @xmath7 decay , bremsstrahlung and positron - electron annihilation , respectively .
the latter two components are necessarily omitted in calculations which assume single interactions only ( i.e. the thin material limit ) .
the error bars are calculated from monte carlo statistics .
although the @xmath7 decay component shows a broad peak at @xmath23 mev ( note that the quantity plotted is the emissivity multiplied by @xmath24 ) and dominates above about 200 mev , the electron bremsstrahlung component broadens the emissivity peak .
the bremsstrahlung and annihilation contributions are negligible in the limit of small column - density ; their fractional contribution to the emission is greatest for clouds which are just thick enough ( @xmath25 ) to attenuate the bulk of the incident proton power .
this is as expected considering that the interaction mean free path of gev protons is @xmath26 g @xmath10 and is in accordance with the result of @xcite who treated a similar problem by solving one - dimensional transport equations .
the resulting gamma - ray emissivities for clouds of various column densities are shown in figure [ fig : f4 ] ( proton injection ) and figure [ fig : f5 ] ( electron injection ) . here
the emissivities are defined for irradiation by cosmic - rays of all species ( see 2.3 , equation ( 3 ) ) ; to take account of the contribution of heavier nuclei than helium , the emissivity due to proton irradiation ( figure 4 ) has been multiplied by a nuclear enhancement factor @xcite of 1.52 @xcite . note that for high densities the monte carlo statistics are rather poor , since the yield itself is low .
figure [ fig : f4 ] includes a comparison of our calculated gamma - ray production functions with that of @xcite ( corresponding to the `` thin material '' limit ) .
the results are consistent with those of @xcite for column densities less than about @xmath27 g @xmath10 , except in the energy range @xmath28 mev where the effect of the maximum energy assumed in the monte carlo simulation is evident .
we note the very low values of the emissivity at energies @xmath29 mev , for column densities @xmath30 .
a slightly steeper spectrum in the @xmath31@xmath32 mev region comes from our omission of the contribution of heavy nuclei , which were taken into account in @xcite
. a somewhat surprising feature of these curves is that the power - law index above 1 gev is almost the same as the input cosmic - ray proton flux for column densities less than about 1000 g @xmath10 ( for higher column densities the statistics of the simulations are not good enough to decide whether this result still holds ) .
this is already indicated by @xcite , but is contrary to the expectation of @xcite who suggested a spectral change above 1 gev at the point where self - shielding becomes important . for some purposes the energy - integrated emissivities are of more interest than their differential counterparts , so we present these quantities in figures [ fig : f6 ] and [ fig : f7 ] , for cosmic ray protons and electrons , respectively . at low column densities , where cascades and self - shielding are unimportant
, there is very little variation of the integrated emissivity with cloud column density , and the thin material limit can be adopted for columns less than @xmath33 for protons ( @xmath34 for electrons ) . above this point , however , the emissivity drops rapidly with increasing cloud column - density . in order to gauge the sensitivity of these calculations to the assumed cosmic - ray spectra ,
we have computed our results for two different incident cosmic - ray proton spectra , and three different incident cosmic - ray electron spectra , as shown in figures 6 and 7 , respectively .
the differences are seen to be small in comparison with the variation as a function of cloud column density , but at a fixed column - density the systematic uncertainties associated with the input cosmic - ray spectra are nevertheless significant .
@llll@ + + @xmath6 & @xmath35 & @xmath36&@xmath37 + + @xmath38 & @xmath39 & @xmath40 & @xmath41 + @xmath42 & @xmath43 & @xmath44 & @xmath45 + @xmath46 & @xmath47 & @xmath48 & @xmath49 + @xmath50 & @xmath51 & @xmath52 & @xmath53 + @xmath54 & @xmath55 & @xmath56 & @xmath57 + @xmath58 & @xmath59 & @xmath60 & @xmath61 + @xmath62 & @xmath63 & @xmath64 & @xmath65 + @xmath66 & @xmath67 & @xmath68 & @xmath69 + @lllll@ + + @xmath70&10&100&1000&10000 + + @xmath6 & @xmath71 & @xmath71&@xmath71 & @xmath71 + @xmath38&@xmath72&@xmath73&@xmath74&@xmath75 + @xmath42&@xmath76&@xmath77&@xmath78&@xmath79 + @xmath46&@xmath80&@xmath81&@xmath82&@xmath83 + @xmath50&@xmath84&@xmath85&@xmath86&@xmath87 + @xmath54&@xmath88&@xmath89&@xmath90&@xmath91 + @xmath58&@xmath92&@xmath93&@xmath94&@xmath95 + @xmath62&@xmath96&@xmath97&@xmath98&@xmath99 + @xmath66&@xmath100&@xmath101&@xmath102&@xmath103 + using the gamma - ray production spectra obtained in the previous section , we have calculated the diffuse gamma - ray emission from the galaxy as follows .
the predicted gamma - ray spectrum for each case is @xmath104 where @xmath105 is the spectrum returned by the simulation in units of photons / mev / primary , for an individual cloud , appropriate to the incident cosmic - ray spectrum .
the quantity @xmath106 is the intensity of cosmic rays at a distance @xmath107 along the line of sight , in units of primaries @xmath10 s@xmath11 sr@xmath11 , and @xmath108 is the mean density in gas clouds of column density @xmath6 .
the galactic variation of the spectrum @xmath109 is not well constrained by existing data , and consequently we adopt the simplifying assumption that the shape of the cosmic - ray spectra ( both electrons and protons ) is the same everywhere in the galaxy , with variations only in the normalisation . with this assumption
it is convenient to recast the calculation as @xmath110 ( @xmath111 ) , where the emissivity is @xmath112 ( @xmath113 ) with @xmath114 the cosmic - ray mean intensity in the solar neighbourhood and @xmath115 is the weighted column density ( @xmath116 ) of the cloud population along the line - of - sight under consideration .
this formulation is convenient because the emissivity , @xmath22 , describes the properties of the gas clouds themselves and is independent of the galactic variations in mean dark matter density and cosmic - ray density ; conversely the quantity @xmath117 characterises these properties of the galaxy , and is independent of the properties of the gas clouds themselves .
the emissivity shown in figures 4 and 5 is the quantity @xmath21 , whereas figures 6 and 7 show @xmath118 . for the inner galactic disk , where we are interested in @xmath119
, we need to average over the whole solid angle , @xmath120 , under consideration : @xmath121 . in order to calculate
@xmath117 we need to adopt models for both the galactic cosmic - ray distribution and the galactic distribution of the clouds . the quantity @xmath108 , the density in cold , dense gas clouds , is only weakly constrained by direct observation , because the hypothetical clouds constitute a form of _ dark _ matter .
we therefore proceed by adopting a conventional dark matter density distribution for the galaxy , namely a cored isothermal sphere , as our model cloud density distribution , with a fiducial normalisation which is equivalent to the assumption that all of the dark matter is in the form of dense gas clouds .
this corresponds to the model @xmath122 in terms of cylindrical coordinates @xmath123 , with @xmath124 .
we have adopted a core radius of @xmath125 kpc based on the preferred model of walker ( 1999 ) .
( this choice corresponds to walker s preferred value of cloud column density @xmath126 . )
walker s model exhibits a core radius which is a function of cloud column density , but we have fixed the core radius at 6.2 kpc for all of our computations .
this choice permits more straightforward consideration of the observational constraints because @xmath117 is independent of @xmath6 in this case .
it then remains to specify the cosmic - ray energy - density as a function of position in the galaxy .
@xcite ( hereafter wlg92 ) constructed numerical models of cosmic - ray propagation in the galaxy ; they did not give any analytic forms for their model cosmic - ray distributions , but an appropriate analytic approximation can be deduced from the results which they obtained .
they found that the cosmic - ray radial distribution reflects , in large part , the radial dependence of cosmic - ray sources , with a modest smoothing effect introduced by diffusion .
we have therefore adopted wlg92 s preferred model ( their model 3 ) for the radial distribution of sources as our model for the radial distribution of cosmic rays . the various spectra of cosmic - ray isotope ratios considered by wlg92 favour models in which the diffusion boundaries are in the range @xmath127 kpc above and below the plane of the galaxy .
we adopt the midpoint of this range .
wlg92 do not give a simple functional form for the vertical variation of cosmic - ray density within this zone , so we have simply assumed an exponential model : @xmath128 . we know that in wlg92 s models the cosmic - ray density is fixed at zero at the diffusion boundaries , and consequently the scale height of the exponential should be approximately half the distance to the diffusion boundary , i.e. @xmath129 kpc .
these considerations lead us to the model cosmic - ray mean intensity distribution @xmath130 : @xmath131,\ ] ] in terms of cylindrical coordinates @xmath132 . here
@xmath133 kpc is the radius of the solar circle , while @xmath134 kpc , @xmath135 kpc and @xmath136 .
this distribution has the character of a disk with a central hole . these models for @xmath108 and
@xmath137 allow us to compute the quantity @xmath117 , as per equation 4 , and the resulting variation over the sky is plotted in figure 8 . for reference we give the values of @xmath138 evaluated at the cardinal points , as follows : @xmath139 , @xmath140 , @xmath141 , and @xmath142 . in order to compare with the egret results of @xcite
, we have also evaluated the average of @xmath117 over the inner galactic disk : @xmath143 .
several physical processes contribute to the observed diffuse gamma - ray intensity . in order of decreasing fractional contribution
these are thought to be pion production and bremsstrahlung from cosmic - ray interactions with _ diffuse _ galactic gas , inverse compton emission from cosmic - ray electrons interacting with ambient photons , and an isotropic background , which is presumably extragalactic and due to many faint , discrete sources .
the sum of these contributions offers a good model for the diffuse emission which is actually observed in the 100 mev1 gev band @xcite ; however , the @xmath144 gev emission is poorly modeled @xcite with the prediction @xcite amounting to only @xmath145 of the observed intensity .
some authors @xcite have used the low - energy ( @xmath146 gev ) data to argue that the agreement between model and data allows no room for any significant unmodeled emission and that these data therefore place tight constraints on any baryonic contribution to the dark matter halo of the galaxy .
this line of argument is clearly suspect because of the failure of the same model to account for the high - energy ( @xmath144 gev ) data .
however , even if we ignore the high - energy data the calculations presented in 2.3 demonstrate that the gamma - ray constraints on high column density gas clouds in the dark halo are much weaker than those on low column density gas because the emissivity of the latter is much greater ; consequently , we revisit the published constraints in the following section ( 3.2 ) , returning to the question of the high - energy data in 3.3 .
@xcite argued that the known ( diffuse ) gas accounts for essentially all of the gamma - ray intensity in the @xmath147 mev band observed by the cos b satellite @xcite , and he suggested that any contribution from baryonic clouds should amount to no more than @xmath148 ( @xmath147 mev ) at high galactic latitudes . toward the galactic poles we found ( 3 ) @xmath149 , implying that the average emissivity of the material in the dark halo should be @xmath150 for @xmath151 . from table 1 ( or
figures 6 and 7 ) we see that this requirement is not met for the models which are in the thin material limit but that any model with @xmath152 is acceptable . in other words
the constraint imposed by @xcite permits the galaxy s dark halo to be entirely baryonic , provided the individual components have a column - density @xmath152 .
@xcite pointed out that obervations at low galactic latitudes can provide tighter constraints on any baryonic dark halo , because of the higher intensity expected when looking through the cosmic - ray disk edge - on .
this point is manifest in the large values of @xmath153 , which we computed in 2.3 ( see also figure 8) .
@xcite argued that a suitable constraint on any baryonic component of the dark halo is that it should not contribute more than the uniform background intensity component observed in any given field .
the strongest constraint then comes from observations of the ophiuchus region @xcite , for which the requirement imposed by @xcite is an intensity contribution from the dark matter halo of @xmath154 in the band 300 mev @xmath155 500 $ mev . for this line of sight
our calculation yields @xmath156 , and in the thin material limit , for which we find @xmath157 ( 300 mev @xmath155 500 $ mev ; see table 1 ) , this corresponds to a predicted intensity of @xmath158 in the 300500 mev band , implying that at most 8% of the dark halo can be resident in low column density clouds .
kpc and a spherical halo ) , a difference which is accounted for by their choice of a smaller core radius for the dark halo .
@xcite chose a core radius of 3.5 kpc , which yields @xmath159 , whereas we have employed a core radius of 6.2 kpc ( see 2.3 ) .
] however , our calculations extend to clouds of higher column - densities , and we find that for @xmath160 the limit relaxes to the point where all of the dark halo is permitted to be in the form of high column density clouds .
the constraints discussed in the previous subsection make reference only to the low - energy ( @xmath146 gev ) gamma - ray data .
as noted earlier in this section , the observed intensity of the inner galactic disk at high photon energies ( @xmath161 gev ) is substantially greater than expected @xcite .
much effort has been expended on explaining this discrepancy , with most of the attention given to models in which the galactic cosmic - ray electron and/or proton spectra differ from their locally measured values @xcite . however
, none of these models offers a satisfactory explanation of the observed mean gamma - ray spectrum of the galactic disk , and consequently a successful match to the low energy data should not be taken to mean that the emission model is correct . in turn
this suggests that the constraints formulated by @xcite and @xcite , on the basis of the low - energy data , may be too restrictive .
the question then arises as to what constraints the gamma - ray data do in fact place on unmodeled emission , given the current state of understanding of the observed emission .
the galactic diffuse emission model used in @xcite is based on @xcite , and contains the following contributions : @xmath162 where @xmath163 is the gamma - ray intensity contributed by cosmic - ray interactions with diffuse atomic hydrogen , and similarly for the ionised and molecular components of the interstellar medium ( hii and @xmath164 , respectively ) .
the emissivity for these components @xcite is , of course , computed in the thin material limit , and each component therefore has the same spectral shape , differing only in normalisation . here
@xmath165 is the gamma - ray flux by inverse compton emission ( cosmic - ray electrons up scattering low - energy photons ) , and @xmath166 is the isotropic extragalactic background flux .
although the atomic and ionised hydrogen components can be observed directly via their line emission , and are thus well constrained , this is not true for the molecular component .
the molecular hydrogen column is assumed to be proportional to the co emission line strength , as measured by the columbia co survey @xcite , for example , but the constant of proportionality ( usually denoted by @xmath167 ) is unknown and one is forced to determine its value by fitting to the gamma - ray data .
although the uncertainty in the best - fit determination of @xmath168 ( averaged over the whole sky ) is small , the systematic uncertainties are acknowledged to be much larger , `` at least 10%15% '' @xcite . in turn , this estimate of the uncertainty is small in comparison with the differences among the various values of @xmath167 which have been reported in the literature ( see the review by @xcite ) and the likely range of variation in @xmath167 within the galaxy @xcite .
although these are substantial uncertainties , molecular hydrogen contributes only a fraction of the total observed gamma - ray intensity roughly 20% of the local surface density of gas in the galactic disk is thought to be in molecular form @xcite so the implied fractional uncertainty in the total galactic emission is perhaps as small as 3% .
moreover , uncertainty in @xmath167 affects only the normalisation and not the spectral shape of the predicted emission from diffuse molecular hydrogen , so the observed high - energy excess can not be explained in this way even if @xmath167 could assume an arbitrarily large value .
the other main contributions to uncertainty in the diffuse model are associated with ( 1 ) unmodeled spatial variations in the cosmic - ray spectral energy density and ( 2 ) unmodeled spatial variations in the low - energy photon spectral energy density ; the former affects all of the galactic contributions to @xmath169 , whereas the latter affects only the inverse compton component .
these uncertainties affect both the normalisation and the spectral shape of the predicted gamma - ray emission ; however , the uncertainties are difficult to quantify . for our purposes it
is not actually necessary to quantify the uncertainties on the model input parameters ; it suffices to use the discrepancy between model and data as a measure of the uncertainty in our understanding of the observed emission . in turn
this measure determines the constraints which we can apply to any putative unmodeled emission , such as the contribution from dense gas which we are concerned with here . at photon energies
@xmath170 the fractional discrepancy is roughly 60% @xcite , in the sense that the observed emission is 1.6 times larger than the model , and we henceforth adopt @xmath171% of the total observed intensity as our estimate of the unmodeled emission .
although this estimate is derived from data at high energies , the effects of the various contributing processes are all very widely spread , and _ the estimate therefore applies independent of photon energy . _ the constraints appropriate to high / low galactic latitudes can now be re - evaluated . at high galactic latitudes
the observed intensity is @xmath172 for @xmath173 @xcite , implying that any unmodeled emission should be @xmath174 in this band .
this result is actually slightly stricter than the criterion used by @xcite and thus leads us to tighten our high - latitude constraints , relative to those quoted in 3.2 : the observed high - latitude gamma - ray intensity constrains the amount of low column - density gas to @xmath175% of the total density of the galactic dark halo , with this fraction rising to 100% for gas clouds of column density @xmath176 . at low galactic latitudes we can make use of the mean intensity of the inner galactic disk , which has been accurately determined by @xcite .
for example at 1 gev the mean intensity ( @xmath177 , @xmath178 ) is @xmath179 , and our calculation of @xmath180 for this region yields ( 2.3 ) @xmath181 , implying that the emissivity of the galactic dark halo material must be , on average , @xmath182 . by comparison ,
the actual emissivity of low column - density gas is computed to be ( 2.3 , table 2 ) @xmath183 , implying that @xmath184% of the galaxy s dark halo may be comprised of low column density gas .
for higher column densities the emissivity falls , and table 2 shows that for @xmath185 the emissivity is only @xmath186 . the gamma - ray data on the inner galactic disk thus indicate all of the galaxy s dark halo to be made of dense clouds of column - density @xmath152 .
the constraint we have just given is based on the mean spectrum of the inner galactic disk , in contrast to those given by @xcite who employed limits based on the angular structure of the observed gamma - ray intensity .
specifically , @xcite required that the putative contribution of emission from a baryonic component of the galaxy s dark halo be less than that of the isotropic component of the observed intensity ; this procedure seems to us to be less reliable than the procedure we have employed , for two reasons .
first , even if the dark halo were spherically symmetric the emission from any baryonic component would not be , both because our point of observation is quite distant from the centre of the galaxy and because the resulting gamma - ray emission is strongly dependent on the galactic cosmic - ray distribution , which in turn is strongly concentrated in the disk of the galaxy . indeed the cosmic - ray distribution appears to correlate with the distribution of interstellar matter @xcite , thus complicating the interpretation of the observed gamma - ray intensity distribution . in particular this coupling leads to gamma - ray emission from a baryonic dark halo being correlated with the diffuse gas column density , even if the dense gas is uncorrelated with the diffuse gas .
second , on any given line of sight , such as the ophiuchus field considered by @xcite , a highly structured dark matter halo might exhibit , by chance , a low dark matter column density .
the chances of this are good if , as in the case of @xcite , the field is specifically chosen to have a low `` background '' intensity .
the gamma - ray spectra arising from cosmic - ray interactions with gas clouds of various column - densities have been calculated using a monte carlo event simulator , geant4 .
our calculations reproduce the analytic result in the low column - density limit , where only single particle interactions need to be considered , but exhibit significant differences for clouds of column - density @xmath187 where the emissivity declines substantially for photon energies @xmath188 .
the low emissivity of dense gas means that the baryonic content of the galaxy s dark halo is not so tightly constrained by the gamma - ray data as had previously been thought .
for @xmath176 we find that the existing gamma - ray data , taken in isolation , do not exclude purely baryonic models for the galactic dark halo .
aharonian , f. a. , & atoyan , a.m. 2000 , 362 , 937 asakimori , k. et al .
1998 , , 502 , 278 barwick , s. w. et al .
1998 , , 498 , 779 berezhko , e.g. , & vlk , h.j .
2000 , , 540 , 923 bertsch , d.l .
1993 , , 416 , 587 bloemen h. 1989 , , 27 , 469 boezio , m. et al .
1999 , , 518 , 457 boezio , m. et al .
2000 , , 532 , 653 bsching , i. , pohl , m. , & schlickeiser , r. 2001 , 377 , 1056 casadei , d. , & bindi , v. astro - ph/0302307 cavallo , g. , & gould , r. j. 1971 , nuovo cimento , 2b , 77 combes , f. & pfenniger , d. 1997 , , 327 , 453 dame , t. m. et al .
1987 , , 322 , 706 dannheim , d. `` test of @xmath7-production in geant4 '' , cern summer student report , 15 september 1999 ( unpublished : available at cern website ) de paolis , f. et al .
1995 , , 74 , 14 dermer , c.d .
1986 , , 157 , 223 fiedler , r. et al .
1994 , , 430 , 581 fukugita m. 2004 , in iau symp . 220 , in press ( astro - ph/0312517 ) gaisser , t. k. & shaefer , r. k. 1992 , , 394 , 174 gerhard , o. & silk , j. 1996 , , 472 , 34 gilmore , g. 1994 , , 92 , 539 golden , r. l. et al .
1984 , , 287 , 622 golden , r. l. et al .
1994 , , 436 , 769 hartman et al .
1999 , , 123 , 79 henriksen r. n. , & widrow l. m. 1995 , , 441 , 70 hunter , s.d . , digel , s.w . ,
de geus , e.j . , & kanbach g. 1994 apj 436 , 216 hunter , s.d .
1997 , , 481 , 205 ivanenko , i. p. et al .
1993 , in proc . of the 23rd international cosmic ray conference , calgary , 2 , 17 kalberla , p.m.w . ,
shchekinov , yu.a .
& dettmar , r .- j .
1999 , 350 , l9 kawamura , y. et al .
1989 , , 40 , 729 kniffen d.a .
1996 , suppl .
120 , 615 kobayashi , t. et al . 1999 , proc .
. cosmic - ray conf .
( salt lake city ) , 3 , 61 mckee , c.f .
2001 asp conf ser 243 , 155 mori , masaki 1997 , , 478 , 225 moskalenko , i.v .
& strong , a.w .
1998 , 493 , 694 pfenniger , d. , combes , f. , & martinet , l. 1994 , , 285 , 79 ryan , m. j. et al . 1972 , , 28 , 985 salati , p. , chardonnet , p. , luo , x. , silk , j. , & taillet , r. 1996 , 313 , 1 sanuki , t. et al . , 2000 , , 545 , 1135 sciama , d.w . 2000a , , 312 , 33 sciama , d.w .
2000b , , 319 , 1001 seo , e. s. et al .
1991 , , 378 , 763 shikaze , y. 2001 , ph.d .
thesis , university of tokyo skibo , j.g . and ramaty , r. 1993 , suppl . , 97 , 145 smith , l. h. et al . 1973 , , 180 , 987 stephens , s.a .
& badhwar , g.d .
1981 , , 76 , 213 strong , a.w . ,
moskalenko , i.v . & reimer , o. 2000 , , 537 , 763 tang , k .- k .
, , 1984 , 278 , 881 torii , s. et al . , , 2001 , 559 , 973 umebayashi , t. & nakano , t. , , 1981 , 33 , 617 walker , m. & wardle , m. , 1998 , , 498 , l125 walker , m. 1999 , , 308 , 551 walker , m. , mori , m. & ohishi , m. 2003 , , 589 , 810 wardle , m. & walker , m. 1999 , , 527 , l109 webber , w. r. et al .
1987 , in proc . of the 20th international cosmic ray conference , moscow , 1 , 325 webber , w.r . ,
lee , m. & gupta , m. 1992 390 , 96 ( wlg92 ) wouterloot , j.g.a . ,
brand , j. , burton , w.b . , & kwee , k.k .
1990 230 , 21 | gamma - ray spectra from cosmic - ray proton and electron interactions with dense gas clouds have been calculated using a monte carlo event simulation code , geant4 .
such clouds are postulated as a possible form of baryonic dark matter in the universe .
the simulation fully tracks the cascade and transport processes which are important in a dense medium , and the resulting gamma - ray spectra are computed as a function of cloud column - density .
these calculations are used for predicting the galactic diffuse gamma - ray spectrum which may be contributed by baryonic dark matter ; the results are compared with data from the egret instrument , and used to constrain the fraction of galactic dark matter which may be in the form of dense gas clouds . in agreement with previous authors ,
we find useful constraints on the fraction of galactic dark matter which may be in the form of low column - density clouds ( @xmath0 ) .
however , this fraction rises steeply in the region @xmath1 , and for @xmath2 we find that baryonic dark matter models are virtually unconstrained by the existing gamma - ray data . |
colloidal suspensions present the possibility to develop novel materials via self - assembly . of particular interest are colloidal crystals , whose optical properties can generate iridescent colours , and provide a means by which photonic crystals may be produced@xcite , while further applications range from lasers @xcite to display devices @xcite , with recent advances demonstrating tunable colours through control of lattice spacing with an external field @xcite .
further to the practical importance of colloidal crystals , their well - defined thermodynamic temperature allows colloidal dispersions to be viewed as mesoscopic ` model atoms ' @xcite .
recently , the ability to tune the colloid - colloid interactions has led to the observation of a wide variety of structures @xcite .
of particular interest here , to first order ac electric fields can induce dipolar interactions between the colloidal particles , leading to anisotropic interparticle potentials and exotic crystal structures , some of which are not observed in atomic and molecular systems @xcite , while external control of the colloid - colloid interactions allows direct observation of phase transitions @xcite .
furthermore , direct microscopic observation at the single - particle level allows an unprecedented level of detail to be accessed @xcite , opening the possibility of tackling long - standing problems in condensed matter , such as freezing @xcite .
the introduction of a rotating ac field opens up even more possibilities . in this case , the dipolar interactions lead to an attraction in the plane of rotation and to repulsions above and below .
studies with a rotating magnetic field on granular matter indeed produced disc like patterns consistent with expectations @xcite . unlike granular matter , since colloidal dispersions exhibit brownian motion , thermodynamic equilibrium structures ( ie crystals ) , may be obtained @xcite . in previous work snoswell
_ @xcite showed that lattice spacing within quasi-2d colloidal crystals could be controlled _ in - situ _
, by means of coplanar rotating electric field .
the interparticle dipolar interactions in the plane of the electric field may be treated to first order as a circularly symmetric attraction , due to the time averaging effect of a rapidly rotating field ( 1000hz ) on relatively large particles on the micron lengthscale , where the diffusive timescale is of the order of seconds @xcite . in considering the interactions between the particles ,
the asymmetry between the colloids ( 10nm-1@xmath0 m ) and smaller molecular and ionic species must be addressed .
a number of coarse - graining schemes have been developed where the smaller components are formally integrated out @xcite .
this generates a one - component picture , where only the effective colloid - colloid interactions are considered , and the complexity of the description is vastly reduced .
the equilibrium behaviour of the colloids in the original multi - component system may then be faithfully reproduced by appeal to liquid state theory @xcite and computer simulation @xcite .
central to the success of this one - component approach is the use of a suitable effective colloid - colloid interaction @xmath1 . in this study
, we use a simple numerical treatement in which we can predict the lattice spacing in the quasi 2d crystal from the electric field strength .
we consider a model system of charged colloids , in a rotating electric field @xcite . by exploiting the knowledge both of the electrostatic repulsions and dipolar attractions , we present a direct , quantitative comparison of a tunable interaction and a material property of the crystalline ` rafts ' formed .
we combine experimental measurements of the crystal lattice constant @xmath2 as a function of field strength @xmath3 with monte - carlo simulations according to a screened coulomb repulsion plus dipolar attraction where the only fitting parameter is the debye screening length . in the simulations , we use pairwise interactions , in other words
we assume that at the higher densities at which the crystalline rafts are formed , the system is still accurately described by interactions calculated for two particles in isolation .
we note that deviations from this assumption of pairwise additivity have been measured both in the case of strongly charged colloids @xcite and in the case of _ repulsive _ dipolar interactions @xcite .
we further compare simulation results with the minimum of the effective potential , which we take as a measure of the lattice constant of the crystalline rafts , which we also determine from experimental data .
this paper is organised into six sections . in section
[ theory ] we present expressions for the effective interactions between the colloids , summing the attractions and repulsions to provide an effective one - component description of the system .
section [ experimental ] describes our experimental metholodogy .
section [ simulation ] outlines the monte - carlo simulation technique employed .
the comparison of simulation and experimental results is presented in section [ results ] and in section [ tunability ] we extrapolate our findings to maximise the tunability of the crystal lattice constant , which may be useful for applications .
we conclude our findings in section [ conclusions ] .
in the following we will consider a system consisting of two particles in a surrounding medium .
we shall assume that these particles are charged , leading to a repulsive interaction , and that the rotating ac electric field induces a dipole moment in the two particles and thus induces an attractive interaction . to describe this system , we start from the derjaguin , landau , verwey and overbeek ( dlvo ) approach @xcite , which consists of attractive van der waals interactions at short range , and long - ranged repulsive electrostatic interactions .
the van der waals interactions are very short - ranged , and are neglected , as electrostatic repulsions inhibit the close approach at which van der waals interactions become important .
we shall therefore assume that the only relevant attractions result from the long - ranged dipolar interactions induced by the rotating electric field . in the linear poisson - boltzmann regime
, the electrostatic repulsions may be expressed as a hard core yukawa , or screened coulomb interaction @xcite , @xmath4 where @xmath5 where @xmath6 is boltzmann s constant , @xmath7 is temperature and @xmath8 is the colloid diameter .
the potential at contact , @xmath9 is given by @xmath10 where @xmath11 is the colloid charge , @xmath12 is the bjerrum length and @xmath13 is the inverse debye screening length where @xmath14 is the total number density of monovalent ions .
now the regime of linear poisson - boltzmann theory in which equation ( [ eqyuk ] ) holds corresponds to relatively weak charging .
although this is not the case here , a potential of the yukawa form is recovered at larger separations , if a smaller , renormalized charge is considered @xcite .
we tabulate measurements of the @xmath15-potential of dilute suspensions in table [ tablekappasigma ] .
these values suggest that we expect a renormalised charge , in the conditions under which the colloids form the crystalline rafts , rather high colloid concentration , hence high counter ion concentration @xcite . therefore , noting that the debye length is much smaller than the colloid radius , we follow bocquet _ et .
_ @xcite and take the following expression for the renormalised charge @xmath16 : @xmath17 which we substitute for @xmath11 in equation ( [ eqepsilonyuk ] ) .
this expression gives good agreement with measurements of the effective colloid charge for particles with a comparable @xmath15-potential @xcite .
we recall that many - body effects can lead to a density - dependence in the effective colloid - colloid interactions @xcite .
however , we shall neglect these effects in the present work .
the attractive potential between the colloids resulting from the rotating ac electric field is treated to first order as a dipolar interaction : @xmath18 where @xmath19 is the induced dipole moment of each particle , @xmath20 is the dielectric constant and @xmath21 is the permittivity of free space .
the dipole moment may be calculated from the strength of the electric field @xmath22 where @xmath23 is the clausius - mosotti factor and takes values between 1 and -1/2 depending upon the origin of the dipole moment .
we note that in this case of an alternating field , the root - mean - square of the time dependent field is taken .
substituting @xmath24 leads us to a potential at contact @xmath25 which has a cubic dependence on the colloid diameter , in the case of all other contributions being unchanged . in considering the value of the clausius - mosotti factor ,
two regimes are relevant , corresponding to applied electric fields of high and low frequency , in which the dipolar interactions result from dielectric permittivity differences between the particles and the solvent and differences in conductivity respectively @xcite .
the crossover frequency @xmath26 between these regimes is given by @xmath27 where @xmath28 are conductivities and the indices @xmath19 and @xmath29 refer to the colloidal particles and the medium respectively .
following @xcite the conductivity of the particles @xmath30 is taken to be the sum of the bulk conductivity @xmath31 and the conductivity on the surface @xmath32 .
@xmath33 in principle equation ( [ eqconductivity ] ) holds for particles larger than 1 @xmath0 m . in the case of much smaller particles
, one should take into account that the contribution of the diffuse layer to the conductivity of the particles . noting that the smallest particles used in the experiments are 757 nm in diameter and that the conductivity of the de - ionised water is relatively low , neglect the double layer contribution and
thus use equation ( [ eqconductivity ] ) .
since the largest particles studied are 2070 nm in diameter and the frequency is 1 khz and the bulk ionic strength is @xmath34 mmol as determined from conductivity measurements , we work in the low frequency regime where the contributions from the conductivities dominate @xcite .
the clausius - mosotti factor therefore takes the following form : @xmath35 due to the surface conductivity , @xmath30 is much larger than the conductivity of the medium @xmath36 .
the value of @xmath37 is therefore close to one .
we now combine the contributions from the electrostatic repulsions and the dipolar attractions , yielding the expression .
@xmath38 @xmath39 has the form indicated in fig .
we see a minimum in the potential , which one might expect to provide a first approximation to the lattice constant in the 2d colloidal crystal . in sections [ results ] and [ tunability ] we shall compare this minimum to our simulation results .
colloidal crystals underwent self - assembly as a result of an applied electric field .
we used anionic , sulphate stabilised polystyrene latex particles , either synthesised using a standard technique , surfactant free emulsion polymerisation @xcite in the case of @xmath40757 and 945 nm , or particles produced by the same method but purchased from microparticles gmbh for @xmath401390 and 2070 nm .
particle electrophoretic mobility was measured using a brookhaven zetaplus light scattering instrument .
particle sizes were determined by scanning electron microscopy , either in - house using a jeol jsm-6330f , in the case of @xmath40757 and 945 nm or by microparticles gmbh for @xmath401390 and 2070 nm are are listed in table [ tablekappasigma ] .
experiments were performed with dilute aqueous suspensions ( 0.5 - 1.5 wt% ) .
schematics of the experimental set up are shown in fig .
[ figschematic ] all glass surfaces were chemically washed with 0.1 m koh and washed with copious quantities of milliq water .
particles were deionised by direct contact with ion exchange beads before being made up to the desired electrolyte concentration with kcl .
very low salt conditions are required as even moderate ( mmol ) salt concentrations lead to electrohydrodynamic pattern formation due to ion flow @xcite .
for further details of the experimental set up and procedure the reader is referred to snoswell _ et .
_ @xcite .
we consider our experiment as a 2d system .
however , occasionally we noticed some overlap of the crystalline rafts .
we do not include these results in our analysis . in order to treat the system in 2d
, one might expect the gravitational length @xmath41 , where @xmath29 is the bouyant mass of the particle and @xmath42 is the acceleration due to gravity , to be much less than some characteristic length such as the particle diameter . due to the fairly small density mismatch between polystyrene and water , in fact even for the largest particles @xmath43 nm , @xmath44 .
however , the dipolar interactions between the particles are attractive in - plane , but strongly repulsive in the vertical direction .
this promotes the formation of ` rafts ' and ` sheets ' , and these large assemblies of many particles have very small gravitational lengths .
thus we argue that the system behaves in a quasi 2d manner .
the lattice parameter @xmath2 was taken as the average of typically ten crystalline ` rafts ' , measured across the ` raft ' , of around ten lattice spacings . the response to changing
the electric field strength was determined to be less than 100 ms , we waited 5s after changing the field strength before acquiring data .
we use standard monte - carlo ( mc ) simulations in the nvt ensemble @xcite .
the particles interact via equation ( [ equ ] ) in two dimensions . to mimic the experiments , we initialise the system in a random configuration at a relatively low concentration , corresponding to an area fraction @xmath45 .
we confirmed that different area fractions gave indistinguishable results , however smaller area fractions often led to longer equilibration times .
the attractive interactions cause the particles to approach one another , and form crystallites which then coalesce to form a crystalline ` rafts ' which contain of order @xmath46 particles , in a qualitatively similar way to the experimental system .
each simulation was equilibrated for typically 30000 mc moves per particle , followed by 3000 production moves per particle .
we recall that a potential of the form @xmath47 converges in 2d .
the lattice constant for this 2d hexagonal crystal was taken as @xmath48 where @xmath49 is the pair correlation function , and @xmath50 the minimum between the first and second peaks of @xmath49 .
we found that when we used @xmath51 , that typically one crystalline domain was formed , around the centre of the simulation box .
we therefore do not use periodic boundary conditions , and consider the lattice spacing of the single crystal formed in the simulation .
one parameter is the number of particles typically present in each crystalline domain .
this is estimated to be around @xmath52@xmath53 from experimental data ( see fig .
[ figpix ] ) .
this value is governed by the overall concentration of the colloidal suspension .
we have considered the effects of varying @xmath54 as shown in fig .
[ figfinitesize ] .
the results of a considerable range of @xmath54 resulted in a variation of a around one percent in @xmath2 for the @xmath55 nm system for a field @xmath3=20 @xmath56 .
each simulation was repeated four times except @xmath57 .
we found a slight trend to tighter binding for larger @xmath54 , however , as shown in fig .
[ figfinitesize ] , this effect is rather small .
we henceforth use one simulation per state point unless otherwise stated . the slight scatter in the simulation data
is perhaps indicative the existence of different magic numbers for these crystalline rafts @xcite .
these may be thought of as 2d hexagonal clusters , whose ` magic numbers ' , ie low energy states , are expected to include @xmath58 , @xmath59 , @xmath60 ... close to a ` magic number ' the binding may be expected to be relatively tighter .
a more detailed exploration of this phenomenon lies beyond the scope of this work .
overall , the variation in @xmath2 as a function of @xmath54 is smaller than the experimental scatter , nonetheless we take both @xmath61 and @xmath62 values for @xmath2 when comparing with experimental data .
the slightly larger value for @xmath63 in fig .
[ figfinitesize ] is attributed to the small size of the crystalline ` raft ' , such that the surface particles which are more widely spaced make a greater contribution .
the domain size is of order 100 particles , thus , for larger @xmath54 , defects and grain boundaries may lead to a smaller contribution to the measurement of @xmath2 .
this also applies to experimental data . for fitting ,
all parameters are known , except the debye length @xmath64 which we take as a free parameter for each particle size and salt concentration , although we note that the effective colloid charge is itself a function of the debye length , equation ( [ eqzeff ] ) .
it has previously been shown that comparison of simulation and experimental data can yield a reasonable _ local _ measure of the debye length @xcite .
furthermore , in the region of the sample in which the electric field is applied , the colloid volume fraction is much higher , and due to the colloidal counter - ions the ionic strength may increase , leading to a reduction in the debye length with respect to the bulk .
however , it is the debye length in this region which is relevant for the effective colloid - colloid interactions . the bulk debye length that may be determined , for example from conductivity measurements ,
may therefore be taken only as an upper bound .
system parameters for different particle sizes are tabulated in table [ tablekappasigma ] , and the debye length is plotted in fig . [ figkappasigma](a ) and the contact potential of the yukawa interaction [ equation ( [ eqepsilonyuk ] ) ] in fig . [
figkappasigma](b ) are plotted as a function of particle diameter .
clearly , our mc simulations may be expected to provide a reasonable treatment of the crystalline rafts in ( quasi ) equilibrium , rather than to describe the formation process .
we therefore restrict our analysis to a simple characterisation of the crystal rafts , and primarily consider the lattice spacing .
we furthermore note both here and in previous work @xcite that considerable variation in shapes of crystalline rafts was observed , but that this had little impact upon the measurement of the lattice spacing @xmath2 . likewise , at the level of this work , we neglect possible local variations in @xmath2 due to the proximity of an interface . in any case
we note that for display applications , all lattice spacings contribute to the diffraction .
a few words on equilibration are in order .
this applies both to the experimental system , and to the simulations .
neither system is strictly in equilibrium , in that case we might expect a rather regularly - shaped raft such as a hexagon .
however , the insensitivity of the lattice parameter either as a function of @xmath54 ( fig .
[ figfinitesize ] ) , and the very close agreement between statistically independent simulation runs lead us to conclude that our approach is sufficient to compare lattice parameters between experiment and simulation .
the effect of changing the electric field strength is readily demonstrated in fig . [ figpix ] , which shows optical microscope images of the region of the sample in which the field is applied .
two field strengths are shown , 29 kvm@xmath65 ( a ) and 80 kvm@xmath65 ( b ) for @xmath55 nm and we see a correspondingly tighter lattice in the case of the higher field strength .
similar behaviour was observed in @xcite .
figure [ figpix](c ) illustrates simulated coordinates for @xmath66@xmath67 @xmath68 , @xmath69 , @xmath70@xmath71 @xmath68 , which corresponds to @xmath3=29.0 vm@xmath65for the @xmath55 nm system ( see table [ tablekappasigma ] ) . a more quantitative comparison between experiment and simulation is shown in fig .
[ figg ] , which plots the radial distribution function @xmath49 .
we see that the structure appears to be crystalline , with reasonably long - ranged correlations , although , as is clear from fig .
[ figpix ] , the system is too small to test for truly long - ranged order .
we thus note that it is difficult to distinguish the true crystal from a hexatic phase in this system .
however , our motivation is to model the experimental system , with @xmath72 and thus we argue that identification of the hexatic phase lies beyond the scope of this work .
having illustrated the general behaviour of the system , we now turn our attention to the fitting of the experimental data , with the model described in section ii .
the experimental data was fitted to the theory by taking the debye - length as the fitting parameter .
the main results are shown in figs .
[ figsmallexpmc ] and [ figbigexpmc ] for @xmath73 and 945 nm , and @xmath74 and 2070 nm systems respectively .
these concern the lattice parameter @xmath2 as a function of applied field strength @xmath3 . in general ,
the simulation is able to capture the behaviour of the system in a reasonably quantitative manner .
furthermore , simply plotting the minimum in the potential @xmath75 given by equation ( [ equ ] ) provides a first approximation to the lattice constant @xmath2 .
figure @xmath76(a ) shows the results for @xmath73 nm particles . due to the small particle size , and the inverse cubic dependence of the strength of the dipolar attractions on the particle diameter , as characterised by @xmath77 [ equation ( [ eqepsilondip ] ) ]
the attraction for a given field strength is comparitively small , so we find relatively larger lattice parameters for this system . at low field strengths ( @xmath78 @xmath56 )
we see an increase of the lattice constant for both the experimental data and the simulation , compared to the minimum in the potential .
apparently , as the crystal approaches melting , fluctuations become more pronounced , leading to an increase in @xmath2 , which is not captured in equation ( [ equ ] ) , or perhaps a molten surface layer increases the apparent value of @xmath2 .
we leave this intriguing question for further investigations .
meanwhile , at high field strengths , we find some deviation between simulation , experiment and equation ( [ equ ] ) . apparently , due to the long ranged nature of the interaction , second nearest neighbours experience an attraction of sufficient strength that the crystal is compressed by its own cohesion , leading to a smaller lattice constant .
in the case of the @xmath73 nm particles , we find that for small field strengths , ( @xmath79 @xmath56 ) , the system does not form crystallites , rather it remains as a colloidal liquid .
we determined this both by experimental observation and with simulations . in the latter case , we identified crystallisation with a splitting in the second peak of the pair correlation function @xmath49 .
the kink around @xmath80 @xmath56 in fig .
[ figsmallexpmc ] is likely related an artifact of measuring multiple rafts , so defects and grain boundaries contribute to the value of @xmath2 . at weak field strengths the rafts
tend to be smaller , thus increasing this effect . at larger particle sizes , ( fig .
[ figbigexpmc ] ) , we again see reasonable agreement between experiment and simulation , albeit with some deviation between simulation and equation ( [ equ ] ) at higher field strengths , as the simulation predicts a smaller lattice constant than that observed in the experiments .
however , for the @xmath74 nm system in particular , in fact equation ( [ equ ] ) provides a more accurate description of the experimental data . apparently some of our assumptions in the simulations , perhaps that of pairwise additivity , begin to break down at high field strengths . according to equation ( [ equ ] ) ,
the potential at the minimum of the attractive well is some 250 @xmath68 for an applied field of 50 kvm@xmath65 for @xmath74 nm . thus we conclude that for moderate interaction strengths , equation ( [ equ ] ) provides a reasonable description of the system .
we now consider the dependence of the system upon the colloid diameter @xmath8 .
firstly , we see from fig .
[ figkappasigma](a ) that the absolute value of the ( fitted ) debye length @xmath64 does not change significantly for all the four particle sizes studied .
thus , the reduced inverse debye length @xmath81 is linear in @xmath8 .
the contact potential @xmath82 has an approximately inverse cubic dependence on @xmath8 for @xmath81@xmath83 , equation ( [ eqepsilonyuk ] ) , but also depends upon the ( effective ) charge @xmath16 .
plotting @xmath82 as a function of @xmath8 in fig . [
figkappasigma](b ) we find an approximate power law dependence with an exponent of @xmath84 .
as noted above , the prefactor of the dipolar attraction , @xmath77 [ equation ( [ eqepsilondip ] ) ] , has a @xmath85 dependence .
although the electrostatic interactions are non - negligible , we nonetheless expect from equation ( [ eqepsilondip ] ) that upon decreasing @xmath8 we find a relatively larger lattice constant for a given field strength , and that larger field strengths are required to provide sufficient interactions that the crystalline rafts form .
this we indeed find , as shown in figs .
[ figsmallexpmc ] and [ figbigexpmc ] , and also in the next section .
now we have noted that this system has optical display applications , due to the possibility of externally tuning the lattice parameter @xmath2 .
these applications may be most usefully realised when the possibility to tune the system is maximised , ie when the range of @xmath2 is maximised .
we are therefore motivated to consider a range of particle sizes , and calculate the range of tunability of @xmath2 .
now @xmath8 sets a lower bound to @xmath2 while the upper bound is set by the melting transition which can be determined by monte - carlo simulation .
hitherto , we have used the debye length @xmath64 as a fitting parameter . however , as fig .
[ figkappasigma](a ) shows , there is relatively little change in the absolute value of the debye length for the range of colloid diameters investigated .
we therefore fix @xmath86 nm and vary the colloid size , and apply the same methodology as that outlined in sections [ theory ] and [ simulation ] , in calculating the response for @xmath87 nm colloids to an external electric field .
while an accurate determination of the melting transition is complicated by the small system size , melting is approximately determined by the splitting of the second peak in the radial distribution function @xmath49 @xcite .
we decrease the electric field to identify the weakest field at which the second peak in @xmath49 exhibits clear splitting [ fig .
[ figtuning](a ) , inset ] , thus yielding @xmath88 , the lattice spacing just prior to melting .
the melting value of the lattice parameter @xmath89 is taken as @xmath90 where @xmath91 is the amount by which the field strength is varied between simulations , typically @xmath92 @xmath56 .
the error in @xmath89 is then taken as @xmath2@xmath93 . in terms of the particle diameter
, @xmath89 grows considerably at small sizes , as the relatively larger debye length leads to a minimum in the effective potential at relatively larger distances .
however , we recall that our analysis suggests a roughly @xmath94 dependence of @xmath77 the strength of the dipolar attraction .
this suggests that we might expect a stronger electric field to be required for relatively small colloids , and indeed we find that a higher field is required to provide sufficient cohesive energy to hold the colloids together in the crystal - like ` rafts ' , so at 100 @xmath56 , a value we take as a reasonably accessible maximum field strength , the lattice spacing is still around @xmath95@xmath8 for @xmath96 @xmath97 , where @xmath98 is the lattice spacing corresponding to a field strength of 100 @xmath56 .
conversely , at larger particle sizes , the debye length is relatively small , so we find , even at melting , that the lattice constant only approaches around @xmath99@xmath8 .
we note that , at high field strengths , the assumptions of section [ theory ] will ultimately break down .
the results presented in figs . [ figsmallexpmc ] and [ figbigexpmc ] show that @xmath100 @xmath56 is reasonable , at least for @xmath101 nm , supporting our conclusion of a maximum in tuneability around @xmath99@xmath8 .
we show the values of the well depth of equation ( [ equ ] ) at melting for the different experimental system in table [ tablekappasigma ] , which indicates a trend towards deeper wells for larger particles , i.e. shorter ranges of the repulsive interaction relative to the particle size .
larger field strengths may be expected to yield greater tunability for the small particle sizes . since the overall minimum lattice spacing is @xmath102 , in principle smaller particles have greater tunability
, however experimental observations reveal intense fluid turbulance disrupts crystal formation at higher field strengths .
this is caused by large field gradients at the electrode edges that induce strong dielectrophoretic forces .
in addition , higher field strengths can cause bubble formation by electrolysis and electrochemical degradation of the electrodes .
the result is thus that intermediate particle sizes have the highest degree of tuneability , for a given maximum field strength .
here we have considered 100 @xmath56 , which leads to a maximum tuneability of @xmath103 in the range 1000 nm @xmath1041500 nm .
2d colloidal crystalline ` rafts ' with externally controllable properties have been modeled with monte - carlo simulations , and the resulting lattice constant , both experimental and simulated , has been compared to the minimum in the effective interaction potential . in treating the effective colloidal interactions , we find that a straightforward , pairwise approach provides a reasonable description in describing the lattice parameter .
the minimum in the effective interaction formed from combining the electrostatic repulsions and dipolar attractions is a useful means to approximate the lattice parameter , although close to melting fluctuations lead to a larger value in the lattice parameter than this minimum .
conversely , at high field strengths , our treatment appears to be less accurate .
apparently , higher order terms which we have not accounted for may become important , such as non - linear poisson - boltzmann contributions to the electrostatic interactions , leading to a deviations from the yukawa form [ equation ( [ eqyuk ] ) ] @xcite or limitations in our pairwise treatment of the dipolar attractions @xcite .
other possibilities include inaccuracies in our assumptions for the effective colloid charge [ equation ( [ eqzeff ] ) ] , limitations of the 2d behaviour of the experimental system . perhaps due to a cancellation of errors , at high field strengths ,
the minimum in the effective interaction can sometimes provide a more accurate value of the lattice parameter than the simulations .
another important assumption lies in our derivation on the electrostatic interactions .
while their yukawa - like form is well - accepted @xcite , the values the effective colloid charge comes from equation ( [ eqzeff ] ) .
we note that the debye length we determine seems rather constant across the four particle sizes considered , [ fig.[figkappasigma ] ( a ) ] , which suggests that out approach is at least consistent .
however challenging it may be , a more quantitative measurement of the local ion concentration in the vicinity of the crystalline rafts is desirable and will be considered in the future .
furthermore , there may be some variation in the value of @xmath16 .
in fact this should affect all state points in a similar manner ( for each particle size ) , and therefore we anticipate a similar outcome , but that a different value might be arrived at in the fitting of the debye length , however , owing to the relatively short - ranged nature of the interaction , the effect of a different value for the effective charge is unlikely to impinge significantly the debye length and should have a negligible effect on our main results .
we have also assumed that a given system is described by _ _
one single__debye length .
in fact , in the counterion - dominated regime , the debye length is in fact a function of colloid concentration @xcite , an effect we have neglected .
this might lead to a tightening at higher colloid concentration ( i.e. high field strength ) , which may be expected to lead to an increase in deviation with the experimental results ( figs .
[ figsmallexpmc ] and [ figbigexpmc ] ) .
nevertheless , varying the debye length as a function of local colloid concentration would be worth considering in the future .
we have extended our approach a range of colloid size , to optimise the lattice tuneability for display applications .
tunable 2d crystal rafts can behave as tunable diffraction gratings capable of filtering white light into visible colours@xcite .
similar tunable diffraction has been proposed for display devices @xcite.the range of colours obtainable for a given geometry is governed by the lattice tunability .
fig 9b demonstrates that lattice tunability in our system is maximised for particles of approximately 1 to 1.5 microns .
only higher field strengths applied to smaller particles can increase the lattice tunability .
as already indicated , higher field strengths are not practical in the current experimental system , however in future experiments metal coated colloids would exhibit much stronger dipolar interactions , enabling a reduction in field strength for a given attraction .
regardless , our methodology illustrates using the well - developed machinery of effective colloidal interactions , as a means to model potentially complex interactions , useful for engineering purposes .
* acknowledgments * we thank jeremy baumberg , martin cryan , daan frenkel and junpei yamanaka for helpful discussions .
cpr acknowledges the royal society for funding and hajime tanaka for kind provision of lab space and computer time .
des thanks kodak european research centre , cambridge and epsrc ( ep / do33047/1 ) for funding . )
.fitted parameters of the debye screening length used in the mc simulations for systems with different colloid diameters .
errors in the diameter are standard deviations in the data obtained found from electron microscopy measurements , while those in @xmath105 , @xmath81 , and the ionic strength are estimated from the fitting of @xmath2 from mc data.@xmath15 potential measurements were made in a 0.01 mmol kcl solution [ tablekappasigma ] [ cols="^,^,^,^,^,^ " , ] | we compare the behaviour of a new 2d aqueous colloidal model system with a simple numerical treatment . to first order
the attractive interaction between the colloids induced by an in - plane rotating ac electric field is dipolar , while the charge stablisation leads to a shorter ranged , yukawa - like repulsion . in the crystal - like ` rafts ' formed at sufficient field strengths , we find quantitative agreement between experiment and monte carlo simulation , except in the case of strongly interacting systems , where the well depth of the effective potential exceeds 250 times the thermal energy .
the ` lattice constant ' of the crystal - like raft is located approximately at the minimum of the effective potential , resulting from the sum of the yukawa and dipolar interactions .
the experimental system has display applications , owing to the possibility of tuning the lattice spacing with the external electric field .
limitations in the applied field strength and relative range of the electrostatic interactions of the particles results in a reduction of tunable lattice spacing for small and large particles respectively . the optimal particle size for maximising the lattice spacing tunability
was found to be around 1000 nm . |
functional data are more and more frequently involved in statistical problems . developping statistical methods in this special framework
has been popularized during the last few years , particularly with the monograph by ramsay & silverman ( 2005 ) .
more recently , new developments have been carried out in order to propose nonparametric statistical methods for dealing with such functional data ( see ferraty & vieu , 2006 , for large discussion and references ) .
these methods are also called doubly infinite dimensional ( see ferraty & vieu , 2003 )
. indeed these methods deal with infinite - dimensional ( i.e. functional ) data and with a statistical model which depends on an infinite - dimensional unknown object ( i.e. a nonparametric model ) .
this double infinite framework motivates the appellation of nonparametric functional statistics for such kind of methods .
our paper is centered on the functional regression model : @xmath0 where @xmath1 is a real random variable , @xmath2 is a functional random variable ( that is , @xmath2 takes values in some possibly infinite - dimensional space ) and where the statistical model assumes only smoothness restriction on the functional operator @xmath3 . at this point , it worth noting that the operator @xmath3 is not constrained to be linear .
this is a functional nonparametric regression model ( see section [ notations ] for deeper presentation ) .
the aim of this paper is to extend in several directions the current knowledges about functional nonparametric regression estimates presented in section [ notations ] . in section [ mse ]
we give asymptotic mean squared expansions , while in section [ asnorm ] the limiting distribution is derived .
the main novelty / difficuly along the statement of these results relies on the exact calculation of the leading terms in the asymptotic expressions .
section [ exsbp ] points out how such results can be used when the functional variable belongs to standard families of continuous time process .
the accuracy of our asymptotic results leads to interesting perspectives from a practical point of view : minimizing mean squared errors can govern automatic bandwidth selection procedure while the limiting distribution of the error is a useful tool for building confidence bands . to this end
, we propose in section [ computfeatures ] a functional version of the wild bootstrap procedure , and we use it , both on simulated and on real functional datasets , to get some automatic rule for choosing the bandwidth .
the concluding section [ conc ] contains some important open questions which emerge naturally from the theoretical results given in this paper , such as the theoretical study of the accuracy of the functional wild bootstrap procedure used in our applications .
the model is defined in the following way .
assume that @xmath4 is a sample of @xmath5 i.i.d .
pairs of random variables . the random variables @xmath6 are real and the @xmath7 s are random elements with values in a functional space @xmath8 . in all the sequel
we will take for @xmath8 a separable banach space endowed with a norm @xmath9 .
this setting is quite general since it contains the space of continuous functions , @xmath10 spaces as well as more complicated spaces like sobolev or besov spaces .
separability avoids measurability problems for the random variables @xmath7 s .
the model is classically written : @xmath11 where @xmath3 is the regression function mapping @xmath8 onto @xmath12 and the @xmath13 s are such that for all @xmath14 , @xmath15 and @xmath16 .
estimating @xmath3 is a crucial issue in particular for predicting the value of the response given a new explanatory functional variable @xmath17 .
however , it is also a very delicate task because @xmath3 is a nonlinear operator ( from @xmath8 into @xmath12 ) for which functional linear statistical methods were not planned . to provide a consistent procedure to estimate the nonlinear regression operator @xmath3
, we propose to adapt the classical finite dimensional nadaraya - watson estimate to our functional model .
we set @xmath18 several asymptotic properties of this estimate were obtained recently .
it turns out that the existing literature adresses either the statement of upper bounds of the rates of convergence without specification of the exact constants ( see chapter 6 in ferraty & vieu , 2006 ) , or abstract expressions of these constants which are unusable in practice ( as for instance in the recent work by masry , 2005 , which has been published during the reviewing process of this paper ) .
our aim in this paper is to give bias , variance , means square errors and asymptotic distribution of the functional kernel regression estimate with exact computation of all the constants ( see section [ theorie ] ) .
we will focus on practical purposes in section computfeatures .
several assumptions will be made later on the kernel @xmath19 and on the bandwidth @xmath20 .
remind that in a finite - dimensional setting pointwise mean squared error ( at @xmath21 ) of the estimate depends on the evaluation of the density ( at @xmath21 ) w.r.t .
lebesgue s measure and on the derivatives of this density .
we refer to schuster ( 1972 ) for an historical result about this topic . on infinite - dimensional spaces ,
there is no measure universally accepted ( as the lebesgue one in the finite - dimensional case ) and there is need for developping a free - densityapproach . as discussed along section [ exsbp ] the problem of introducing a density for @xmath2 is shifted to considerations on the measure of small balls with respect to the probability of @xmath22 .
only pointwise convergence will be considered in the forthcoming theoretical results . in
all the following , @xmath21 is a fixed element of the functional space @xmath8 .
let @xmath23 be the real valued function defined as @xmath24 , \ ] ] and @xmath25 be the c.d.f
. of the random variable @xmath26 : @xmath27 note that the crucial functions @xmath23 and @xmath25 depends implicitely on @xmath28 consequently we should rather note them by @xmath29 and @xmath30 but , as @xmath21 is fixed , we drop this index once and for all .
similarly , we will use in the remaining the notation @xmath31 instead of @xmath32 .
let us consider now the following assumptions .
* h0 :* @xmath3 and @xmath33 are continuous in a neighborhood of @xmath21 , and @xmath34 . * h1 * * :* @xmath35 exists . * h2 :* the bandwidth @xmath20 satisfies @xmath36 and @xmath37 , while the kernel @xmath19 is supported on @xmath38 $ ] , has a continuous derivative on @xmath39 , @xmath40 and @xmath41
. assumptions * h0 * and * h2 * are clearly unrestrictive , since they are the same as those classically used in the finite - dimensional setting . much more should be said on assumption * h1*. note first that , obviously , @xmath42 .
it is worth noting that , whereas we could expect assumptions on the local regularity of @xmath3 ( as in the finite - dimensional case ) , hypothesis * h1 * skips over that point and avoids to go into formal considerations on differential calculus on banach spaces . to fix the ideas ,
if we assumed differentiability of @xmath3 , we would get by taylor s expansion that @xmath43 where @xmath44 , @xmath45 being the conjugate space of @xmath8 and @xmath46 being the duality bracket between @xmath8 and @xmath47 in this context , a non trivial link would appear between @xmath35 and @xmath48 through the following relation : @xmath49 indeed , even if the link between the existency of @xmath48 and of @xmath35 is strong , one can build counter - examples for their non equivalence ( these counter - examples are available on request but they are out of the main scope of this paper ) . in the perspective of estimating the constants given in theorem [ theoremmse ] , it will be easier to estimate @xmath35 ( for instance by using @xmath50 ) than the operator @xmath48 .
therefore , we prefer to express computations by mean of @xmath35 instead of @xmath48 .
this has the additional advantage to produce more readable writings .
consequently , the differentiability of @xmath3 is not needed .
let us now introduce the function @xmath51 defined for all @xmath52 $ ] as : @xmath53 for which the following assumption is made : * h3 :* for all @xmath52 , $ ] @xmath54 as @xmath55 .
note that the function @xmath51 is increasing for all @xmath56 the measurable ( as the pointwise limit of the sequence of measurable functions @xmath51 ) mapping @xmath57 is non decreasing .
let us finally mention that this function @xmath57 will play a key role in our methodology , in particular when we will have to compute the exact constant terms involved in our asymptotic expansions .
for the sake of clarity , the following proposition ( whose a short proof will be given in the appendix ) will explicit the function @xmath58 for various cases . by @xmath59 0,1\right ] } ( \cdot)$ ] we denote the indicator function on the set @xmath60 0,1\right ] $ ] and @xmath61 stands for the dirac mass at @xmath62 .
[ prop1 ] i ) if @xmath63 for some @xmath64 then @xmath65.ii ) if @xmath66 with @xmath64 and @xmath67 then @xmath68 iii ) if @xmath69 for some @xmath70 and some @xmath64 then @xmath71 iv ) if @xmath72 then @xmath73 0,1\right ] } ( s)$ ] . a deeper discussion linking the above behavior of @xmath25 with small ball probabilities notions will be given in section [ exsbp ] .
in both following subsections we will state some asymptotic properties ( respectively mean squared asymptotic evaluation and asymptotic normality ) for the functional kernel regression estimate @xmath74 .
it is worth noting that all the results below can be seen as extensions to functional data of several ones already existing in the finite - dimensional case ( the literature is quite extensive in this field and the reader will find in sarda & vieu ( 2000 ) deep results as well as a large scope of references ) . with other words , our technique for proving both theorem theoremmse and theorem [ theoremasnorm ] is also adapted to the scalar or vector regression model since the abstract space @xmath8 can be of finite dimension ( even , of course , if our main goal is to treat infinite - dimensional cases ) .
moreover , it turns out that the transposition to finite - dimensional situations of our key conditions ( see discussion in section [ exsbp ] below ) becomes ( in some sense ) less restrictive than what is usually assumed . with other words , the result of theorem theoremmse and theorem [ theoremasnorm ]
can be directly applied to finite - dimensional settings , and will extend the results existing in this field ( see again sarda & vieu , 2000 ) to situation when the density of the corresponding scalar or multivariate variable does not exist or has all its successive derivatives vanishing at point @xmath21 ( see discussion in section [ dimfinie ] ) . all along this section
we assume that assumptions * h0-h3 * hold .
let us first introduce the following notations : @xmath75 the following result gives asymptotic evaluation of the mean squared errors of our estimate .
the asymptotic mean squared errors have a standard convex shape , with large bias when the bandwidth @xmath20 increases and large variance when @xmath20 decays to zero .
we refer to the appendix for the proof of theorem [ theoremmse ] .
[ theoremmse ] when * h0-h3 * hold , we have the following asymptotic developments : @xmath76 and@xmath77 let us denote the leading bias term by : @xmath78 before giving the asymptotic normality , one has to be sure that the leading bias term does not vanish .
this is the reason why we introduce the following additional assumption : * h4 :* @xmath79 and @xmath80.the first part of assumption * h4 * is very close to what is assumed in standard finite - dimensional literature .
it forces the nonlinear operator @xmath3 not to be too smooth ( for instance , if @xmath3 is lipschitz of order @xmath81 , then @xmath82 ) .
the second part of assumption * h4 * is specific to the infinite - dimensional setting , and the next proposition pr2 will show that this condition is general enough to be satisfied in some standard situations .
this proposition will be proved in the appendix .
[ pr2 ] i ) if @xmath83 0,1\right ] } ( s)$ ] and @xmath58 is continuously differentiable on @xmath84 , then @xmath80 for any kernel @xmath19 satisfying * h2*.ii ) if @xmath85 , then @xmath80 for any kernel @xmath19 satisfying * h2*. to emphasize the interest of these results , they should be combined with those of proposition [ prop1 ] .
note that the result @xmath86 includes the well - known family of processes for which @xmath65 ( see proposition [ prop1]-@xmath14 ) , that is those whose distributions admit fractal dimensions ( see section [ fractal ] ) . the second case when @xmath85 ( for which a particular case is given in proposition [ prop1]-@xmath87 ) corresponds to nonsmooth processes ( see section [ nonsmooth ] ) .
these two cases cover a large number of situations . however , if a more general function @xmath58 has to be used , one can make additional hypotheses on the kernel @xmath19 .
in particular , if @xmath19 is such that @xmath88 $ ] , @xmath89 and @xmath83 0,1\right ] } ( s)$ ] , then @xmath80 , which covers the case of the uniform kernel .
more complicated kernel functions @xmath19 would lead to more technical assumptions linking @xmath19 with @xmath58 .
it is out of purpose to give these tedious details ( available on request ) but let us just note that the key restriction is the condition @xmath83 0,1\right ] } ( s)$ ] ( else we have @xmath90 ) .
moreover , since the rate of convergence depends on the function @xmath91 and for producing a reasonably usable asymptotic distribution it is worth having some estimate of this function .
the most natural is its empirical counterpart : @xmath92 the pointwise asymptotic gaussian distribution for the functional nonparametric regression estimate is given in theorem [ theoremasnorm ] below which will be proved in the appendix .
note that the symbol @xmath93 stands for convergence in distribution .
[ theoremasnorm ] when * h0-h4 * hold , we have @xmath94 a simpler version of this result is stated in corollary [ coro1 ] below whose proof is obvious .
the key - idea relies in introducing the following additional assumption : * h5 :* @xmath95 which allows to cancel the bias term . [ coro1 ] when * h0-h5 * hold , we have @xmath96 in practice , the constants involved in corollary [ coro1 ] need to be estimated . in order to compute explicitely both constants @xmath97 and @xmath98 , one may consider the simple uniform kernel and get easily the following result : under assumptions of corrollary [ coro1 ] , if @xmath99}(.)$ ] and if @xmath100 is a consistent estimator of @xmath31 , then we have : @xmath101 there are many possibilities for constructing a consistent conditional variance estimate .
one among all the possibilities consists in writing that @xmath102 and , by estimating each conditional expectation with the functional kernel regression technique .
[ examples ] the distribution function @xmath25 plays a prominent role in our methodology .
this appears clearly in our conditions ( through the function @xmath58 ) and in the rates of convergence of our estimate ( through the asymptotic behavior of the quantity @xmath103 ) .
more precisely , the behaviour of @xmath25 around @xmath104 turns out to be of first importance . in other words , the small ball probabilities of the underlying functional variable @xmath2
will be determining . in order to illustrate our ideas and to connect with existing probabilistic knowledges in this field ,
let us now just discuss how @xmath25 ( and hence @xmath58 ) behave for different usual examples of processes @xmath2 valued in an infinite - dimensional space .
calculation of the quantity @xmath105 for small @xmath106 ( i.e. for @xmath106 tending to zero ) and for a fixed @xmath21 is known as a `` small ball problem '' in probability theory .
this problem is unfortunately solved for very few random variables ( or processes ) @xmath107 even when @xmath108 in certain functional spaces , taking @xmath109 yield considerable difficulties that may not be overcome .
authors usually focus on gaussian random elements .
we refer to li & shao ( 2001 ) for a survey on the main results on small ball probability .
if @xmath2 is a gaussian random element on the separable banach space @xmath8 and if @xmath21 belongs to the reproducing kernel hilbert space associated with @xmath2 , then the following well - known result holds : @xmath110 so , the small ball problem at any point @xmath21 may be shifted to a small ball problem at @xmath111 moreover , ( [ pb ] ) can be precised in a few situations .
for instance , mayer - wolf & zeitouni ( 1993 ) investigate the case when @xmath2 is a one - dimensional diffusion process and @xmath21 satisfies some conditions ( see mayer - wolf & zeitouni , 1993 , p15 ) .
they also briefly mention the non gaussian case ( see mayer - wolf & zeitouni , 1993 , remark 3 , p19 ) but many other authors have considered different settings ( see ferraty et _ al . _ , 2005 , for a large discussion and references therein ) . as far as we know , the results which are available in the literature are basically all of the form : @xmath112 where @xmath113 and @xmath114 are positive constants and @xmath115 may be a @xmath116 , a @xmath10 , a besov norm ... the next remark is a direct consequence of proposition [ prop1 ] .
it proves that non - smooth processes may satisfy the assumptions needed to get the asymptotic expansions of previous sections .
[ rem1 ] in the case of `` non - smooth '' processes defined by ( [ ex1 ] ) we have @xmath117 .
in addition , condition @xmath118 ( in * h2 * ) is checked as soon as @xmath119 for @xmath120 large enough .
another family of infinite dimensional processes is the class of fractal processes for which the small ball probabilities are of the form @xmath121 where @xmath122 and @xmath123 are once again positive constants .
like above , it is elementary to get the following result from proposition [ prop1 ] .
[ rem2 ] under ( [ ex2 ] ) , we have @xmath65 while the condition @xmath124 ( in * h2 * ) is satisfied as soon as @xmath125 for @xmath126 small enough . finally , it is important to note that a special case of fractal processs is given by the usual multivariate case ( that is , by the case when @xmath127 ) .
the following result is obvious for the uniform norm on @xmath128 and extends directly to any norm , since all of them are equivalent in finite dimension .
[ rem3 ] if @xmath127 , then any random variable @xmath2 on @xmath128 which has a finite and non zero density function at point @xmath21 satisfies ( [ ex2 ] ) with @xmath129 . from remarks [ rem2 ] and [ rem3 ] ,
it is clear that all the results of section [ theorie ] apply in a finite dimensional setting . besides
, the assumptions needed for theorems [ theoremmse ] and [ theoremasnorm ] are weaker than those described in remark [ rem3 ] since there is no need to assume the existence of a density for @xmath2 . in this sense , our results extend the standard multivariate literature ( see the discussion at the beginning of section [ theorie ] ) .
the asymptotic results presented in the previous section [ theorie ] are particularly appealing because , in addition to the specification of the rates of convergence , the exact constants involved in the leading terms of each result are precised .
this is particularly interesting in practice .
let us focus now on the mean squared errors expansion given in section [ mse ] .
in fact , theorem [ theoremmse ] could give clues for possible automatic bandwidth choice balancing the trade - off between variance and squared - bias effects .
however , the constants are unknown in practice which could seem to be a serious drawback for practical purposes .
this general problem is well - known in classical nonparametric statistics , but in our functional context this question gets even more intricate because of the rather complicated expression of @xmath130 , @xmath97 and @xmath98 . an appealing way to attack
the problem is to use bootstrap ideas . in section [ funcwildboot ]
we propose a track for building a functional version of the so - called wild bootstrap .
we will show in section [ funcwildbootsim ] , through some simulated examples , how this functional wild bootstrap procedure works on finite sample sizes for choosing automatically an optimal bandwidth .
a case study , based on spectrometric functional data coming from the food industry , will be shortly presented in section [ funcwildbootspect ] . at this stage
it is worth noting that we have no asymptotic support for this functional bootstrapping procedure
. this open question will be one of the main point discussed in the concluding section [ conc ] . basically , when using bootstrapping techniques one expects to approximate directly the distribution of the error of estimation without having to estimate the leading terms involved in some asymptotic expansion of this error . in standard finite - dimensional problems ( that is , when the variable @xmath2 is valued in @xmath128 ) ,
a so - called wild bootstrap has been constructed for approximating the distribution of the error of estimation in kernel nonparametric regression .
we refer to hrdle ( 1989 ) and hrdle & marron ( 1991 ) for a previous presentation of the wild bootstrap in nonparameric regression
. a selected set of additional references would include mammen ( 2000 ) for the description of the state of art on nonparametric bootstrapping , mammen ( 1993 ) for a large study of wild bootstrap , and hrdle , huet & jolivet ( 1995 ) for specific advances on wild bootstrap in ( finite - dimensional ) nonparametric regression setting .
the main interest of this kind of bootstrap relies on a resampling procedure of the residuals which makes it easily adaptable to our functional setting .
precisely , an adaptation to our functional setting could be the following functional wild bootstrap procedure : * given the estimate @xmath74 constructed with a bandwidth @xmath20 , compute the residuals @xmath131 and construct a sequence of bootstrapped residuals such that each @xmath132 is drawn from a distribution @xmath133 which is the sum of two dirac distributions : @xmath134 such a distribution ensures that the first three moments of the bootstrapped residuals are respectively @xmath104 , @xmath135 and @xmath136 ( see hrdle & marron , 1991 , for details ) . * given the bootstrapped residuals @xmath137 , and using a new kernel estimate @xmath138 which is defined as @xmath74 but by using another bandwidth @xmath139 , construct a bootstrapped sample @xmath140 by putting @xmath141 * given the bootstrapped sample @xmath140 , compute the kernel estimate @xmath142 which is defined as @xmath74 ( with the same bandwidth @xmath20 ) but using the bootstrapped sample @xmath140 instead of the previous sample @xmath143 .
we suggest to repeat several times ( let say @xmath144 times ) this bootstrap procedure , and to use the empirical distribution of @xmath145 for bandwidth selection purpose .
precisely , the bootstrapped bandwidth is defined as follows : [ defbb ] given @xmath144 replications of the above described bootstrapping scheme , and given a fixed set h of bandwidths , the bootstrapped bandwidth @xmath146 is defined by : @xmath147 of course , this procedure has still to be validated theoretically ( see discussion in section [ conc ] ) , but we will see in the next sections [ funcwildbootsim ] and [ funcwildbootspect ] how it behaves both on simulated on and real data samples . the aim of this section is to look at how the automatic bootstrapped bandwidth constructed in definition [ defbb ] behaves on simulated samples .
we construct random curves in the following way : @xmath148 where @xmath149 and @xmath150 ( respectively @xmath151 ) are r.r.v . drawn from a uniform distribution on @xmath84 ( respectively on @xmath152 ) .
some of these curves are presented in figure [ curves ] below .
the real response is simulated according to the following regression relation : @xmath153 where @xmath154 and where @xmath155 is drawn from a @xmath156 distribution .
our experience is based on the following lines . for each experience
, we simulated two samples : a sample of size @xmath157 on which all the estimates are computed and a testing sample of size @xmath158 which is used to look at the behaviour of our method .
also , for each experience , the number of bootstrap replications was taken to be @xmath159 .
other values for @xmath160 and @xmath144 were also tried without changing the main conclusions . to improve the speed of our algorithm , the bandwidth @xmath20 is assumed to belong to some grid in terms of nearest neighbours , that is @xmath161 where @xmath162 is the radius of the ball of center @xmath21 and containing exactly @xmath163 among the curves data @xmath164 . concerning the other parameters of our study , the kernel function @xmath19 was chosen to be @xmath165 and the norm @xmath166 was taken to be the @xmath167 one between the first order derivatives of the curves . we computed , for the @xmath168 different values of @xmath20 , the average ( over the @xmath21 s belonging to the second testing data sample ) of the true error @xmath169 and of its bootstrap approximation @xmath170 .
finally , this monte carlo scheme was repeated @xmath171 times and the results are reported in figure compar - errors ( only @xmath172 among the @xmath173 curves are presented to make the plot clearer ) .
it appears clearly that both the theoretical quadratic loss and its data - driven bootstrapped version have the same convex shape .
this convex shape is directly linked with the asymptotic expansion given in theorem theoremmse before : large values of @xmath20 give high bias , while small values of @xmath20 lead to high variance .
these results are quite promising in the sense that the similarity of the shapes of both sets of curves presented in figure [ compar - errors ] let us expect that the bootstrapped bandwidths will be closed from the optimal ones . to check that point
, we computed the theroretical minimal quadratic loss ( that is , the error obtained by using the best bandwidth ) and we compared it with the error obtained by using the boostrapped bandwidth @xmath146 . this was done for each among the @xmath171 experiences , and the results are reported in figure [ bandwchoiceresults ] which gives mean , variance and density estimates of these two errors .
undoubtedly , these results show the good behaviour ( at least on this example ) of the bootstrapping method as an automatic bandwidth selection procedure .
of course , as discussed in section [ conc ] , theoretical support for this functional bootstrap bandwidth selection rule is still an open question .
let us now quickly show how our procedure is working on real data .
these data contain of @xmath174 spectra of light absorbance @xmath175 as functions of the wavelength , and observed on finely chopped pieces of meat .
we present in figure [ spectrocurves ] the plots of the @xmath174 spectra .
for each spectral cruve corresponds some real response @xmath6 which is the percentage of fatness , and our aim is to study the regression relation existing between the real variable @xmath1 and the functional one @xmath176 .
these data have been widely studied and , inspired by previous studies ( see ferraty & vieu , 2006 ) we decide to apply the functional kernel methodology to the curves @xmath177 , and by taking as norm @xmath166 between curves the usual @xmath167 norm between the second derivatives of the spectra .
the kernel function @xmath19 was chosen to be @xmath178 .
along our study we splitted the data into two subsamples .
a first subsample of size @xmath179 from which our estimates are computed , and a testing sample of size @xmath180 on which they are applied . in a first attempt
, we used the automatic bootstrapping bandwidth selection rule , where @xmath181 was defined as in ( [ hsim ] ) .
we present in figure errorbootspeccurves the shape of the bootsrapped mean square error as a function of the number of neighbours ( and thus , as function of the bandwidth ) . the same convex form as for the simulated data appears .
this form matches the theoretical results obtained in section [ mse ] , with high bias for large values of @xmath20 and high variance for small bandwidths .
these bootstrapped errors are , in this example , minimal for the value @xmath182 .
that means that , for each new curve @xmath21 to be predicted , the data - driven bootstrapped bandwidth @xmath183 is such that there are exactly @xmath184 curves - data which are falling inside of the ball of radius @xmath183 .
these bandwidths lead to completely automatic data - driven fat contents prediction .
for instance , we present in figure [ spectropredictions ] the fat content predictions for the @xmath180 spectra in our testing sample . in order to highlight the nice behaviour of our prediction algorithm , figure spectropredictions plots the predicted values as functions of the true ones .
the implementation of the method was performed by using the _ splus _ routine _
funopare.kernel _ which is included in the package _
npfda_. this package will go with the monograph by ferraty & vieu ( 2006 ) .
this _ splus _ package ( as well as a similar _ r _ package ) and the spectrometric dataset ( as well as several other curves datasets ) will be put in free access on line in the next future . by that time ,
programs and data are available on request .
this paper completes the recent advances existing in kernel nonparametric regression for functional data , by giving not only the rates of convergence of the estimates but also the exact expressions of the constant terms involved in these rates .
these theoretical results deal with mean squared errors evaluations and asymptotic normality results . as explained in section [ computfeatures ] , these new results open interesting perspectives for applications , such as for instance data - driven automatic bandwidth selection and confidence band construction .
we support the idea that bootstrap methods offers interesting perspectives for the functional context .
one of them is illustrated by our functional version of the wild bootstrap for selecting the smoothing parameter .
we have observed nice results for bandwidth selection on some simulated and real datasets .
it should be pointed out that theoretical support for this functional wild bootstrap bandwidth selection rule remains an open problem .
our guess is that it should be possible to extend to functional variables some results stated in finite dimension ( for instance those in hrdle & bowman , 1987 ) , but this has still to be proved .
another direct application of our result concerns the construction of confidence bands .
once again , the problem of estimating the constants involved in the asymptotic normal distribution can be attacked by the wild bootstrap track described before .
one possible way for that would be to try to extend standard finite - dimensional knowledge ( see for instance hrdle & marron , 1991 , or hrdle , huet & jolivet , 1995 ) to infinite - dimensional variables .
in the following , given some @xmath12-valued random variable @xmath185 , @xmath186 will stand for the probability measure induced by @xmath187 to make its treatment easier , the kernel estimate @xmath74 will be decomposed as follows : * * bias term : proof of ( [ b])*. let us write the following decomposition . @xmath191
with@xmath192 \label{a1}\ ] ] and@xmath193 .
\label{a2}\ ] ] + the first step of the proof consists in rewritting the first term in right hand side of the decomposition ( [ bias ] ) in the following way : + [ lemma - bias-1 ] we have : [ b2 ] @xmath194 where @xmath195 + in a second attempt the next lemma will provide the constant term involved in this bias expression and its limit . + [ b3]we have : @xmath196 + finally , to finish this proof it suffices to prove that both last terms at right hand side of ( [ bias ] ) are neglectible .
this is done in next lemma .
+ [ b7 ] we have : @xmath197 + so the proof of ( [ b ] ) is complete . *
* variance term : proof of ( [ v])*. the starting point of the proof is the following decomposition .
this decomposition has been obtained in earlier work by collomb ( 1976 ) ( see also sarda & vieu ( 2000 ) ) in the finite - dimensional case , but since the proof is only using analytic arguments about taylor expansion of the function @xmath198 around @xmath104 , it extends obviously to our functional setting : + @xmath199 + finally , the result ( [ v ] ) will follow directly from this decomposition together with both following lemmas .
+ [ b4bis ] we have successively : @xmath200 + [ b5 ] we have successively : @xmath201 the following lemma states a preliminary pointwise limiting distribution result .
this lemma stems from the bias and variance expressions obtained along theorem 1 ; it will be proved in the next subsection . * * proof of lemma [ lemma - bias-1 ] : * to calculate @xmath204 } { ek\left ( \frac{\left\vert \mathcal{x}-\chi\right\vert } { h}\right ) } , \label{pro1}\ ] ] note first that @xmath205 & = e\left [ \left ( r\left ( \mathcal{x}\right ) -r\left ( \chi\right )
\right ) k\left ( \frac{\left\vert \mathcal{x}-\chi\right\vert } { h}\right ) \right ] \\ & = e\left [ \varphi\left ( \left\vert \mathcal{x}-\chi\right\vert \right ) k\left ( \frac{\left\vert \mathcal{x}-\chi\right\vert } { h}\right ) \right ] .\end{aligned}\ ] ] moreover , it comes : @xmath206 \\ & = \int\varphi\left ( t\right ) k\left ( \frac{t}{h}\right ) dp^{\left\vert \mathcal{x}-\chi\right\vert } \left ( t\right ) \\ & = \int\varphi\left ( ht\right ) k\left ( t\right ) dp^{\left\vert \mathcal{x}-\chi\right\vert /h}\left
( t\right ) \\ & = h\varphi^{\prime}\left ( 0\right ) \int tk\left ( t\right ) dp^{\left\vert \mathcal{x}-\chi\right\vert /h}\left ( t\right ) + o\left ( h\right ) , \end{aligned}\ ] ] + the last line coming from the first order taylor s expansion for @xmath23 around @xmath104 .
for the denumerator in ( [ pro1 ] ) we have@xmath207 = \int k\left ( t\right ) dp^{\left\vert \mathcal{x}-\chi\right\vert /h}\left
( t\right ) .\ ] ] + finally , it appears clearly that the first order bias term is @xmath208 . *
* proof of lemma [ b3 ] : * we note that@xmath209 applying fubini s theorem we get@xmath210 similarly , we have @xmath211 so the proof of this lemma is finished by applying the lebesgue s dominated convergence theorem since the denumerator may be easily bounded above by @xmath41 ( @xmath19 being decreasing ) . * * proof of lemma [ b4bis ] * : the first assertion follows directly from ( [ pro2 ] ) , while the second one can be proved similarly according to the following lines : @xmath212 \\ & = \left ( r\left ( \chi\right ) + o(1)\right ) e\left [ k\left ( \frac{\left\vert \mathcal{x}-\chi\right\vert } { h}\right ) \right ] .\end{aligned}\ ] ] * * proof of lemma [ b5 ] * : we write the variance of @xmath213 as : + @xmath214 , \ ] ] + and note that , as for getting ( [ pro2 ] ) , it holds : @xmath215 the first assertion of lemma [ b4bis ] gives @xmath216 at last @xmath217 which finishes the proof of the first assertion of our lemma .
+ the same steps can be followed to prove the second assertion .
we write @xmath218 .\ ] ] the second term at right hand side of this expression is treated directly by using the second assertion of lemma [ b4bis ] , while the first one is treated as follows by conditioning on @xmath2 : @xmath219 the continuity of @xmath220 and of @xmath221 insure that @xmath222 combining this result with ( [ ek2 ] ) allows to finish the proof of the second assertion of our lemma .
let us deal now with the covariance term : @xmath223 .
\notag\end{aligned}\ ] ] the last two terms were computed before , while the first one is treated by conditioning on @xmath2 and using continuity of @xmath3 : @xmath224 the proof of this lemma is now finished . * * proof of lemma [ b7 ] * : both assertions of this lemma are direct consequences of lemmas [ b4bis ] and [ b5 ] . * * proof of lemma [ tcl1 ] : * on one hand , ( [ bias ] ) and lemma [ b7 ] allows us to get @xmath225 on the other hand , the following decomposition holds : @xmath226 using slutsky s theorem and theorem [ theoremmse ] , we get @xmath227 noting that @xmath228 can be expressed as an array of independent centered random variables ( and the central limit theorem applies ) .
let us remark that @xmath229 which achieves the proof of this lemma .
* * proof of proposition [ prop1]-@xmath14 and @xmath230 * : obvious . * * proof of proposition [ prop1]-@xmath87 * : we have : @xmath231 if @xmath232 we have @xmath233 , while if @xmath234 we have @xmath235 . to complete this proof
it suffices to note that , for @xmath236 , we have : @xmath237 and so we have @xmath238 as @xmath239 . *
* proof of propostion [ prop1]-@xmath240 * : for any @xmath241 and any @xmath242 ,
we have : @xmath243 and so we have @xmath244 . to complete this proof it suffices to note that @xmath245 . * * proof of propostion [ pr2]-@xmath14 * : by simple integration by parts we arrive at @xmath246 where @xmath247 .
because of @xmath248 and because @xmath58 is non decreasing , there exists some nonempty interval @xmath249\subset(0,1)$ ] such that both @xmath19 and @xmath250 do not vanish on @xmath249 $ ] , and therefore we arrive at : @xmath251 * * proof of propostion [ pr2]-@xmath14 * : obvious . | we consider the problem of predicting a real random variable from a functional explanatory variable .
the problem is attacked by mean of nonparametric kernel approach which has been recently adapted to this functional context .
we derive theoretical results by giving a deep asymptotic study of the behaviour of the estimate , including mean squared convergence ( with rates and precise evaluation of the constant terms ) as well as asymptotic distribution .
practical use of these results are relying on the ability to estimate these constants .
some perspectives in this direction are discussed .
in particular a functional version of wild bootstrapping ideas is proposed and used both on simulated and real functional datasets .
* key words : * asymptotic normality , functional data , nonparametric model , quadratic error , regression , wild functional bootstrap . |
casimir - lifshitz force in an interaction originating from the fluctuations of the electromagnetic field and existing between any couple of polarizable bodies .
it was first theoretically derived by casimir in 1948 @xcite in the idealized configuration of two perfectly conducting parallel plates at zero temperature .
later , lifshitz and collaborators generalized the calculation to the case of bodies having arbitrary optical properties and of finite temperature @xcite .
the casimir - lifshitz interaction , experimentally verified for several different geometries @xcite , results from two contributions , one originating from vacuum fluctuations and present also at zero temperature , the other one from purely thermal fluctuations . the latter becomes relevant when the distance separating the bodies is larger than the thermal wavelength @xmath0 , of the order of @xmath1 m at ambient temperature .
this explains why it has been only very recently experimentally observed at thermal equilibrium @xcite .
nevertheless , the situation completely changes out of thermal equilibrium .
it was first theoretically predicted in 2005 that the atom - surface interaction ( usually referred to as casimir - polder force ) is qualitatively and quantitatively modified with respect to thermal equilibrium @xcite .
new power - law behaviors appear , the force can turn into repulsive ( being only attractive at thermal equilibrium ) and it is strongly tunable by modifying the temperatures involved in the system .
this prediction was verified in 2007 , providing the first experimental observation of thermal effects @xcite .
these results paved the way to a renewed interest in casimir - lifshitz effects out of thermal equilibrium .
in fact , this effect was studied for two slabs @xcite and in presence of atoms @xcite , and more recently several different approaches have been developed to deal with the problem of the force out of thermal equilibrium and heat transfer between two @xcite or more @xcite arbitrary bodies .
the physics of the electromagnetic field out of thermal equilibrium has also stimulated the study of other effects , such as the manipulation of atomic populations @xcite and entanglement @xcite .
, always assumed to be positive .
the gratings , in general made of different materials , are infinite in the @xmath2 plane , and periodic in the @xmath3 direction with the same period @xmath4 .
they have corrugation depths @xmath5 ( @xmath6 ) , thicknesses @xmath7 and lengths of the upper part of the grating @xmath8 .
this defines the filling factors @xmath9.,width=302 ] in parallel with the interest in the absence of thermal equilibrium , casimir - lifshitz interactions have been studied in several different geometries , with particular interest in the sphere - plane configuration , the most studied experimentally .
more recently , nanostructured surfaces have been theoretically considered in the contexts of both force @xcite and heat transfer @xcite .
experimentally , the force have been measured between a sphere and a dielectric @xcite or metallic @xcite grating .
the problem we address here for the first time is the calculation of the casimir - lifshitz force out of thermal equilibrium in presence of dielectric gratings , in order to study the combination of non - equilibrium and geometrical effects .
in particular we consider a system made of two different gratings having different temperatures , immersed in an environmental bath at a third temperature .
our calculations can be relevant both to imagine new experiments measuring the casimir - lifshitz force out of thermal equilibrium and in the more general context of the manipulation of the force in micro- and nano - electromechanical systems @xcite .
the paper is structured as follows . in sec .
[ secintro ] we introduce our physical system and provide the notation and main definitions . in sec .
[ secfmm ] we solve the problem of the scattering upon a single 1d lamellar dielectric grating using the fourier modal method . in sec .
[ secnum ] , we apply these results in order to calculate the force out of thermal equilibrium between two different gratings .
we explore the behavior of the force as a function of the three temperatures and of the geometrical parameters of the gratings , with a specific attention to the appearance and features of repulsion .
we finally give in sec .
[ secconcl ] some conclusive remarks .
we start by describing the system studied in this paper .
we address the casimir - lifshitz force between two dielectric gratings immersed in vacuum ( @xmath10 ) in the geometrical configuration shown in fig .
we label the two gratings with an index @xmath11 taking values 1 and 2 .
the gratings are infinite in @xmath3 and @xmath12 directions , with periodicity along the @xmath3 axis .
their distance @xmath13 is defined in fig .
[ fig1 ] and can only take positive values ( i.e. a plane @xmath14 must exist separating the two bodies ) .
the gratings share the same period @xmath4 and have corrugation depth @xmath5 , permittivities @xmath15 in the homogeneous zone , permittivities @xmath16 along the grating zone having thickness @xmath7 , and filling factors @xmath9 ( @xmath8 is defined as in fig .
[ fig1 ] ) .
our physical system is considered in a configuration out of thermal equilibrium ( ote ) .
this means that each body is supposed to be in local thermal equilibrium with a constant temperature @xmath17 .
we also assume that the two gratings are immersed in a radiation bath coming from bounding walls far from the system and having temperature @xmath18 , in general different from the temperatures of the two gratings ( see also @xcite ) .
the whole system is considered in a stationary regime so that the three temperatures involved are constant in time . in @xcite , this assumption has been used to characterize the properties of the source fields ( the ones emitted by the two bodies and coming from the surrounding walls ) in terms of field correlation functions .
this procedure is based on a mode decomposition of the fields , each mode @xmath19 being identified by the direction of propagation @xmath20 along the @xmath21 axis , the polarization index @xmath22 [ assuming the values @xmath23 which respectively correspond to transverse electric ( te ) and transverse magnetic ( tm ) modes ] , the frequency @xmath24 and the transverse wavevector @xmath25 . in this description ,
the @xmath21 component of the wavevector @xmath26 is a dependent variable defined as @xmath27 based on this mode decomposition , the trace of a given operator @xmath28 is defined as @xmath29 the correlation functions of the field have been expressed as a function of the reflection and transmission operators @xmath30 and @xmath31 associated to each body ( see sec .
[ secfmm ] for more details ) . using these correlation functions
, the ote casimir - lifshitz force acting on body 1 can be cast in the following form @xcite ( the distance dependence is implicit ) : @xmath32 where @xmath33 is the force acting on body 1 at thermal equilibrium at its temperature @xmath34 .
this equilibrium contribution reads @xmath35,\end{aligned}\ ] ] while the non - equilibrium term is @xmath36\nonumber\\ & + n_{21}u^{(21)}\bigl(f_{-1}(\mathcal{r}^{(2)-})-\mathcal{t}^{(2)-}\mathcal{p}_{-1}^{\text{(pw)}}\mathcal{t}^{(2)-\dag}\bigr)u^{(21)\dag}\bigl(f_2(\mathcal{r}^{(1)+})-\mathcal{t}^{(1)-\dag}\mathcal{p}_2^{\text{(pw)}}\mathcal{t}^{(1)-}\bigr)\bigr]\bigr\}.\end{aligned}\ ] ] in the equations above we have introduced the thermal population @xmath37,\ ] ] with @xmath38 and the population differences @xmath39 .
moreover we have defined the auxiliary functions @xmath40 and the operators @xmath41 finally , in we have introduced the projection operators @xmath42 where @xmath43 is the kronecker delta and being @xmath44 [ @xmath45 the projector on the propagative ( @xmath46 ) [ evanescent ( @xmath47 ) ] sector .
in order to calculate the force , we now need to compute the reflection and transmission operators associated to a lamellar 1d grating .
this will be achieved in the framework of the fourier modal method ( fmm ) @xcite . in the following ,
we implement this method for a grating of finite size along the @xmath21 axis ( see fig .
[ fig2 ] ) in order to take into account finite - size effects on the casimir - lifshitz force .
moreover , we solve the scattering problem directly in te and tm components , in order to be coherent with the formalism presented in sec .
[ secintro ] .
let us consider a system composed of a grating like the one in fig .
the space is divided in four zones : zone 1 ( @xmath48 ) , zone 2 ( @xmath49 ) , zone 3 ( @xmath50 ) and zone 4 ( @xmath51 ) . while zones 1 , 3 and 4 are homogeneous with dielectric permittivities @xmath15 ( @xmath52 ) , zone 2 represents the grating , with a dielectric function @xmath53 , periodic in @xmath3 with period @xmath4 . in each zone , every physical quantity is independent of @xmath12 .
we first decompose the electric field in any zone with respect to frequency ( only positive frequencies will be used ) : @xmath54.\ ] ] in virtue of the translational invariance of our system along the @xmath12 axis and of the periodicity along the @xmath3 axis , we will employ a fourier decomposition of any @xmath3-dependent quantity . as a consequence
, the wavevector component @xmath55 will be replaced by a new mode variable @xmath56 with @xmath55 taking values in the first brillouin zone @xmath57 $ ] and @xmath58 assuming all integer values . , corrugation depth @xmath59 , and underlying thickness @xmath60 .
this defines four zones ( see text ) with four ( in general different ) dielectric permittivities .
the period is @xmath4 and the filling factor is defined as @xmath61.,width=321 ] in any homogeneous zone , we can use a standard rayleigh expansion for the component of the field at frequency @xmath24 @xmath62}^{(i)\phi}_p(\mathbf{k}_n,\omega)e^{(i)\phi}_p(\mathbf{k}_n,\omega),\end{aligned}\ ] ] where the wavevectors are defined as ( @xmath63 ) @xmath64 and @xmath65 is the @xmath21 component of the wavevector inside each medium @xmath66 the unit polarization vectors appearing in eq .
are defined as @xmath67}_{\text{te}}^{(i)\phi}(\mathbf{k}_n,\omega)&=\frac{1}{k_n}(-k_y\hat{\mathbf{x}}+k_{x , n}\hat{\mathbf{y}}),\\ \hat{\bbm[\epsilon]}_{\text{tm}}^{(i)\phi}(\mathbf{k}_n,\omega)&=\frac{c}{\omega\sqrt{\epsilon_i(\omega)}}(-k_n\hat{\mathbf{z}}+\phi k^{(i)}_{z , n}\hat{\mathbf{k}_n}).\end{aligned}\ ] ] for convenience , we assign from now on the following labels to the field amplitudes in the three homogeneous zones ( the dependence on @xmath22 , @xmath68 and @xmath24 is implicit ) : @xmath69 where @xmath70 , @xmath71 and @xmath72 represent the incoming , reflected and transmitted field amplitudes respectively . the amplitude @xmath73 is associated to a possible incoming field coming from the other side of the body .
its presence guarantees both the full symmetry of the calculation and the possibility to derive at the same time the reflection and transmission operators @xmath74 and @xmath75 .
the magnetic field in any zone can be easily deduced from maxwell s equations and it reads @xmath76}^{(i)\phi}_{s(p)}(\mathbf{k}_n,\omega)e^{(i)\phi}_p(\mathbf{k}_n,\omega),\end{aligned}\ ] ] where the function @xmath77 is defined as @xmath78 and @xmath79 .
we now move to the periodic region ( zone 2 ) where we write an arbitrary frequency component of the field as @xmath80 where @xmath81 .
we are now ready to write maxwell s equations ( for our system we have @xmath82 ) @xmath83 where we used @xmath84 , @xmath85 , @xmath86 , @xmath87 and defined @xmath88 . from we
can easily obtain @xmath89 e_y \end{pmatrix}\\ & = \begin{pmatrix } -\dfrac{i}{k_0}\partial_x\dfrac{1}{\varepsilon(x)}\partial_y & ik_0+\dfrac{i}{k_0}\partial_x \dfrac{1}{\varepsilon(x)}\partial_x \\
-ik_0-\dfrac{i}{k_0}\partial_y \dfrac{1}{\varepsilon(x)}\partial_y & \dfrac{i}{k_0}\partial_y\dfrac{1}{\varepsilon(x)}\partial_x \end{pmatrix } \begin{pmatrix } \widetilde{h}_x \\[5pt ] \widetilde{h}_y \end{pmatrix},\nonumber\end{aligned}\ ] ] @xmath90 \widetilde{h}_y \end{pmatrix}\\ & = \begin{pmatrix } \dfrac{i}{k_0}\partial_x\partial_y & -ik_0\varepsilon(x)-\dfrac{i}{k_0}\partial_x \partial_x \\ ik_0\varepsilon(x)+\dfrac{i}{k_0}\partial_y \partial_y & -\dfrac{i}{k_0}\partial_y\partial_x \end{pmatrix } \begin{pmatrix } e_x \\[5pt ] e_y \end{pmatrix}.\nonumber\end{aligned}\ ] ] we now employ a fourier factorization for the fields @xmath91 and @xmath92 .
correspondingly , the operator @xmath93 is replaced by @xmath94 , @xmath95 being a scalar , whereas the operator @xmath96 is replaced by @xmath97 , where @xmath98 .
these replacements allow us to rewrite maxwell s equations of our system in a more compact form : @xmath99 @xmath100 ik_0{\bigg\llbracket\dfrac{1}{\varepsilon}\bigg\rrbracket}^{\!-1}\!\!\!-\dfrac{i\beta^2}{k_0 } & \dfrac{i\beta}{k_0}\alpha \end{pmatrix } \bm{\mathcal{e}}\nonumber\\ & = \amsbb{g}\bm{\mathcal{e}},\end{aligned}\ ] ] where for an arbitrary field @xmath101 we have introduced the decomposition @xmath102 gathering @xmath3 and @xmath12 components and denoting with @xmath103 a set of scattering orders . we have also introduced the toeplitz matrix @xmath104 , defined by the relation @xmath105 , @xmath106 being the @xmath58-th fourier component of @xmath107 .
we remark that going from eqs . - to eqs .
- we have used the modified factorization rule introduced in @xcite .
of course , in order to exploit numerically the fmm , a truncation has to be made , limiting the number of diffraction orders taken into account . for a given truncation @xmath108
, this corresponds to keeping @xmath109 scattering orders @xmath110 and the size of the corresponding column vector @xmath111 is thus @xmath112 .
based on this truncation , we obtain @xmath113 where @xmath114 and @xmath115 are respectively the eigenvectors and eigenvalues @xmath116 matrices of the matrix @xmath117 @xmath118 then , from eqs . and , we obtain that fields are @xmath119 @xmath120 and @xmath121 being arbitrary constant vectors , and where @xmath122 .
based on the knowledge of the electric and magnetic fields in the four regions , we can now impose the continuity of the @xmath3 and @xmath12 components of both fields at the three interfaces @xmath123 , @xmath124 and @xmath125 . in the following boundary conditions the values of @xmath55 , @xmath126 and @xmath24 are given .
exploiting this fact we use the generic simplified expression @xmath127 to refer to the amplitude @xmath128 . before proceeding in the calculation , we introduce an additional phase factor in the expression of the fields in zones 3 and 4 .
in particular , in zone 3 we replace @xmath129 $ ] with @xmath130 $ ] , while in zone 4 we replace @xmath129 $ ] with @xmath131 $ ] .
these factors make the calculation easier and can be simply recovered at the end . at the first interface @xmath123 we have for the @xmath3 and @xmath12 components of the electric field ( repeated indices are implicitly summed over ) @xmath132 \frac{k_{x , n}}{k_n}(i_{1,n}+r_{1,n})+\frac{c}{\sqrt{\varepsilon_1}\omega}k_{z , n}^{(1)}\frac{k_y}{k_n}(i_{2,n}-r_{2,n})\end{pmatrix}= \begin{pmatrix}\amsbb{p}^{(11)}_{nm}(a_{x , m}+b_{x , m})+\amsbb{p}^{(12)}_{nm}(a_{y , m}+b_{y , m})\\[5pt ] \amsbb{p}^{(21)}_{nm}(a_{x , m}+b_{x , m})+\amsbb{p}^{(22)}_{nm}(a_{y , m}+b_{y , m } ) \end{pmatrix},\ ] ] while for the magnetic field we get @xmath133 -\frac{c}{\omega } k_{z , n}^{(1 ) } \frac{k_y}{k_n } \big ( i_{1,n } -r_{1,n } \big ) + \sqrt{\varepsilon_1}\frac{k_{x , n}}{k_n } \big(i_{2,n}+r_{2,n } \big ) \end{pmatrix}=\begin{pmatrix}{\amsbb{p}'}^{(11)}_{nm}(a_{x , m}-b_{x , m})+{\amsbb{p}'}^{(12)}_{nm}(a_{y , m}-b_{y , m})\\[5pt ] { \amsbb{p}'}^{(21)}_{nm}(a_{x , m}-b_{x , m})+{\amsbb{p}'}^{(22)}_{nm}(a_{y , m}-b_{y , m})\end{pmatrix}.\ ] ] the boundary conditions at @xmath124 give us the following equations for the electric field @xmath134 \frac{k_{x , n}}{k_n}({c}_{1,n}+c'_{1,n})+\frac{c}{\sqrt{\varepsilon_3}\omega}k_{z , n}^{(3)}\frac{k_y}{k_n}(c_{2,n}-c'_{2,n } ) \end{pmatrix}\nonumber\\ & = \begin{pmatrix}\amsbb{p}^{(11)}_{nm}\big(e^{\amsbb{d}^{(11)}_{mm}h}a_{x , m}+e^{-\amsbb{d}^{(11)}_{mm}h}b_{x , m}\big)+\amsbb{p}^{(12)}_{nm}\big(e^{\amsbb{d}^{(22)}_{mm}h}\,a_{y , m}+e^{-\amsbb{d}^{(22)}_{m}h}b_{y , m}\big)\\[5pt ] \amsbb{p}^{(21)}_{nm}\big(e^{\amsbb{d}^{(11)}_{mm}h}\,a_{x , m}+e^{-\amsbb{d}^{(11)}_{mm}h}b_{x , m}\big)+\amsbb{p}^{(22)}_{nm}\big(e^{\amsbb{d}^{(22)}_{mm}h}a_{y , m}+e^{-\amsbb{d}^{(22)}_{mm}h}b_{y , m}\big)\end{pmatrix},\end{aligned}\ ] ] and the following ones for the magnetic field
@xmath135 -\frac{c}{\omega } k_{z , n}^{(3)}\frac{k_y}{k_n}(c_{1,n}-c'_{1,n})+\sqrt{\varepsilon_3}\frac{k_{x , n}}{k_n}(c_{2,n}+c'_{2,n } ) \end{pmatrix}\nonumber\\ & = \begin{pmatrix } { \amsbb{p}'}^{(11)}_{nm}\big(e^{\amsbb{d}^{(11)}_{mm}h}\,a_{x , m}-e^{-\amsbb{d}^{(11)}_{mm}h}\,b_{x , m}\big)+{\amsbb{p}'}^{(12)}_{nm}\big(e^{\amsbb{d}^{(22)}_{mm}h}\,a_{y , m}-e^{-\amsbb{d}^{(22)}_{mm}h}\,b_{y , m}\big)\\[5pt ] { \amsbb{p}'}^{(21)}_{nm}\big(e^{\amsbb{d}^{(11)}_{mm}h}\,a_{x , m}-e^{-\amsbb{d}^{(11)}_{mm}h}\,b_{x , m}\big)+{\amsbb{p}'}^{(22)}_{nm}\big(e^{\amsbb{d}^{(22)}_{mm}h}\,a_{y , m}-e^{-\amsbb{d}^{(22)}_{mm}h}\,b_{y , m}\big)\end{pmatrix}.\end{aligned}\ ] ] finally , the boundary conditions at @xmath125 read @xmath136 \frac{k_{x , n}}{k_n}\big(e^{i k_{z , n}^{(3)}\delta}\,{c}_{1,n}+e^{-i k_{z , n}^{(3)}\delta}\,c'_{1,n}\big)+\frac{c}{\sqrt{\varepsilon_3}\omega}k_{z , n}^{(3)}\frac{k_y}{k_n}\big(e^{i k_{z , n}^{(3)}\delta}\,c_{2,n}-e^{-i k_{z , n}^{(3)}\delta}\,c'_{2,n}\big ) \end{pmatrix}\nonumber\\ & = \begin{pmatrix}-\frac{k_y}{k_n}(t_{1,n}+i'_{1,n})+\frac{c}{\sqrt{\varepsilon_4}\omega}k_{z , n}^{(4)}\frac{k_{x , n}}{k_n}(t_{2,n}-i'_{2,n})\\[5pt ] \frac{k_{x , n}}{k_n}(t_{1,n}+i'_{1,n})+\frac{c}{\sqrt{\varepsilon_4}\omega}k_{z , n}^{(4)}\frac{k_y}{k_n}\big(t_{2,n}-i'_{2,n } ) \end{pmatrix},\end{aligned}\ ] ] and the ones for the magnetic field are given by @xmath137 -\frac{c}{\omega } k_{z , n}^{(3)}\frac{k_y}{k_n}\big(e^{i k_{z , n}^{(3)}\delta}\,c_{1,n}-e^{-i k_{z , n}^{(3)}\delta}\,c'_{1,n}\big)+\sqrt{\varepsilon_3}\frac{k_{x , n}}{k_n}\big(e^{i k_{z , n}^{(3)}\delta}\,c_{2,n}+e^{-i k_{z , n}^{(3)}\delta}\,c'_{2,n}\big ) \end{pmatrix}\nonumber\\ & = \begin{pmatrix}-\frac{c}{\omega}k_{z , n}^{(4)}\frac{k_{x , n}}{k_n}(t_{1,n}-i'_{1,n})-\sqrt{\varepsilon_4}\frac{k_y}{k_n}(t_{2,n}+i'_{2,n})\\[5pt ] -\frac{c}{\omega } k_{z , n}^{(4)}\frac{k_y}{k_n}(t_{1,n}-i'_{1,n})+\sqrt{\varepsilon_4}\frac{k_{x , n}}{k_n}(t_{2,n}+i'_{2,n})\end{pmatrix}.\end{aligned}\ ] ] in the following , we are going to cast eqs .
- under the form @xmath138 the column vectors @xmath120 and @xmath121 appearing in this equation gather two vectors defined as in eq . .
on the contrary , all the six other column vectors gather the two polarizations of the field under the form @xmath139 the system of equations has to be solved for the unknowns @xmath140 , @xmath141 , @xmath120 , @xmath121 , @xmath142 , and @xmath143 . the expression of @xmath140 and @xmath141 as a function of @xmath144 and @xmath145
will provide us the desired reflection and transmission operators . the fact that for @xmath120 and @xmath121 we solve in cartesian components and not in polarization is not an issue since these appear as mute variables not participating to the scattering operators .
the explicit expression of the @xmath146 matrices appearing in can be obtained by means of algebraic manipulation of eqs . - .
the final result is @xmath147 @xmath148 @xmath149 in these expressions we have defined @xmath150 where @xmath151 the symbol @xmath152 denotes a @xmath153 diagonal matrix having diagonal elements @xmath154 , @xmath155 , , @xmath156 .
we have also defined the square matrices of dimension @xmath112 @xmath157 using we obtain the final result @xmath158 where @xmath159 having introduced the associative operation @xmath160 , which for three square matrices @xmath161 , @xmath162 and @xmath163 of dimension @xmath164 is defined as @xmath165 where each matrix have been decomposed in four square blocks of dimension @xmath112 .
equation allows to identify the four blocks of @xmath146 as the reflection and transmission operators associated to the two sides of the grating . for example , the block @xmath166 is the coefficient linking the reflected amplitudes @xmath140 to the incident ones @xmath144 : it then coincides with the reflection operator @xmath167 for a wave impinging on the grating of fig . from @xmath48 . by analog reasoning
, we write the full @xmath146 matrix as @xmath168 we now need to calculate the reflection and transmission operators associated to the two gratings represented in fig .
as far as grating 1 is concerned , the problem we need to solve is exactly the one presented in this section , with the appropriate values of the geometrical parameters .
concerning grating 2 , we need to take into account the fact that its interface is the plane @xmath169 and not @xmath123 .
the modification of the scattering operators with respect to translations has been discussed in @xcite . based on these results , and using the mode expansion used in this work ,
the @xmath170 operator of grating 2 can be expressed as a function of the @xmath171 derived from fmm as @xmath172{\displaystyle{\langle p,\mathbf{k},n,\omega|}}\widetilde{\mathcal{r}}_2 ^ -{\displaystyle{|p',\mathbf{k}',n',\omega'\rangle}}.\nonumber\end{aligned}\ ] ] as we will show in the next section , this operator is the only one associated to grating 2 appearing in the expression of the force for our configuration .
in this section we will present a numerical application concerning the force between two different gratings .
being both gratings infinite in the @xmath2 plane , we actually calculate the pressure acting on any of them , as discussed in the case of two slabs in @xcite . in the first configuration we have chosen both gratings to have period @xmath173 m , corrugation depth @xmath174 m and filling factor @xmath175 . as shown in fig .
[ fig1 ] , the transition points of the two gratings are aligned , i.e. there is no shift along the @xmath3 axis .
grating 1 is made of fused silica ( sio@xmath176 ) and has thickness @xmath177 m , while grating 2 is made of silicon and has infinite thickness . in order to take into account this point
we have imposed @xmath178 in the fmm relative to grating 2 ( see sec .
[ secfmm ] ) and removed in eq .
all the terms proportional to the transmission operators of body 2 .
physically , this can be explained by observing that because of the infinite thickness all the radiation coming from the upper side of body 2 is absorbed and does not reach the cavity between the gratings . both silicon and
fused silica have been described by means of optical data taken from @xcite .
as anticipated in sec .
[ secfmm ] , the numerical use of fmm demands to choice of a truncation order , problem that will be addressed in this section .
we noted before that by choosing a truncation order @xmath108 in the fmm we obtain as a result reflection operators which are square matrices of dimension @xmath112 , that is two polarizations times @xmath109 diffraction orders .
their typical structure is thus @xmath179 & a_{1,2}[n , n']\\\hline a_{2,1}[n , n ' ] & a_{2,2}[n , n']\end{array}\right)\end{array},\ ] ] where each block @xmath180 $ ] is a @xmath153 matrix , the indices @xmath58 and @xmath181 running from @xmath182 to @xmath108 .
it is worth stressing that , for a given @xmath108 , only the elements closer to the center of each block of the matrix ( i.e. close to @xmath183 for each couple of polarizations ) are at convergence .
thus , for a given @xmath184 , we can increase the value of @xmath108 starting from @xmath185 in order to extract a @xmath186 ( @xmath187 ) scattering operator whose elements are at convergence with a given accuracy ( in our case of the order of one percent ) .
the operators obtained following this procedure can be used to compute the force using eqs . and .
since these equations imply a trace containing also a sum over the diffraction orders @xmath58 , the series has to be replaced with a finite sum from @xmath188 to @xmath189 .
the value of @xmath189 has to be found by imposing the convergence of the series at a chosen accuracy .
also in this case , we required an accuracy smaller than one percent .
the calculation of the pressure at a given distance requires the evaluation of the traces and at several different values of the wavevector @xmath190 and the frequency @xmath24 , in order to reach the convergence on the integral on the three variables .
we have observed that a single calculation of the trace requires values of @xmath189 of the order of 2 ( with peaks going up to 7 ) and corresponding values of @xmath108 of the order of 5 ( with peaks around 20 ) . a single value of the pressure required a computation time of the order of 16 hours on three 3ghz cpus . in the configuration described above ,
we have calculated the pressure acting on grating 1 . to point out the features of our ote configuration we present in fig .
[ fig3 ] the pressure as a function of distance for different sets of the temperatures @xmath191 .
m , @xmath177 m , @xmath173 m and @xmath192 ) in front of grating 2 ( made of silicon , having @xmath193 m , infinite thickness , @xmath173 m and @xmath194 ) as a function of distance @xmath13 .
the four curves correspond to different choices of the three temperatures @xmath191 ( see legend).,width=302 ] we clearly see that the modification of the three temperatures strongly affects the value of the force .
in particular , three of the four curves show a transition from an attractive to a repulsive behavior , not realizable at thermal equilibrium for this configuration .
this qualitative difference is a well - known consequence of the absence of thermal equilibrium and it has already been predicted in the case of two parallel slabs @xcite .
we stress that the transition point between attraction and repulsion is a function of the temperatures . for the values chosen
, it roughly varies from 3 to 5@xmath195 m . to underline even more the richness of our ote configuration , we focus on the temperatures @xmath196k and compare the pressure to its equivalent at thermal equilibrium at the temperature of body 1 , i.e. @xmath197k .
this comparison is presented in fig .
[ fig4 ] . in the same figure we also plot the pressure , both at and out of thermal equilibrium , for filling factors @xmath198 ( corresponding to _ filled _ gratings , that is a 11@xmath195m - thick sio@xmath176 slab at distance @xmath13 from an infinite si slab ) and for @xmath199 ( corresponding to _ empty _ gratings , that is a 10@xmath195m - thick sio@xmath176 slab at distance @xmath200 m from an infinite si slab ) .
k , solid lines ] compared to equilibrium pressure ( @xmath201k , dashed lines ) for two gratings ( black squares ) , and two slab - slab configurations corresponding to filled gratings ( @xmath202 , green circles ) and an empty ones ( @xmath203 , red triangles).,width=302 ] apart from the transition to a repulsive behavior , this figure shows that the pressure in presence of a grating always lies between the two results corresponding to filled and empty ones . finally , a comparison between figs .
[ fig3 ] and [ fig4 ] shows that the asymptotic value of the pressure can be tuned by varying the temperatures to values comparable ( apart from their sign ) to the pressure at thermal equilibrium at much smaller distances , of the order of 3@xmath195 m . to conclude this section
, we compare the grating - grating pressure obtained using fmm to the result coming from the pfa ( proximity force approximation ) , typically used to deal with complex geometries such as sphere - plane and nanostructured surfaces . in the case of two aligned gratings with equal filling factors
@xmath204 it reduces to the following weighted sum of the pressures of simple slab - slab configurations @xcite : @xmath205 where @xmath206 is the pressure acting on a @xmath207-thick slab at a distance @xmath13 from a @xmath208-thick slab . in fig .
[ fig5 ] we plot the ratio between the exact pressure and the pfa results for the four temperature configurations used in fig .
we observe that pfa provides in our range of distances a description of the pressure with a relative error typically well below 20% .
the fact the pfa predicts a change of sign not exactly at the position predicted by the exact calculation results in the existence of a vertical asymptote of the ratio @xmath209 , clearly shown in the blue and orange curves in fig .
[ fig5 ] .
it is now interesting to understand how a modification of the geometrical parameters of the gratings is able to tune the value of the pressure . to this aim
we have chosen as a reference the pressure at a distance @xmath210 m for @xmath196k , for which the pressure is around @xmath211nm@xmath212 ( see fig .
[ fig3 ] ) . starting from this result ,
we have modified one by one the values of the filling factor @xmath213 , period @xmath4 , corrugation depth @xmath59 and calculated the ratio between the modified pressure and the reference @xmath214 .
m [ temperatures @xmath196k ] as a function of the geometrical parameters .
the reference point ( black circle ) corresponds to the set of parameters @xmath215 , @xmath216 m , @xmath177 m , infinite @xmath208 , @xmath173 m .
the three curves show the variation of pressure when changing one parameter at a time ( red diamonds for the filling factor , green triangles for the period , blue squares for the corrugation depth ) . on the @xmath12 axis ,
the pressures are normalized with respect to the reference one , while on the @xmath3 axis each varying parameter is normalized with respect to its reference value ( @xmath217 , @xmath218 m , @xmath219 m ) .
note that the plot on the right side continues the one on the left with a modified @xmath3 scale.,width=302 ] k ] for three different values of filling factor , all the other geoemtrical parameters being the reference ones.,width=302 ] the results are shown in fig .
[ fig6 ] , where the pressure ratio is plotted as a function of the ratio between the modified parameter and the reference ones ( @xmath217 , @xmath218 m and @xmath219 m ) .
first , we observe that geometrical modifications can tune the pressure by a factor going from 0.5 to 1.6 .
in particular , this region can be fully explored by varying the filling factor between the two admitted extreme values @xmath203 and @xmath202 , i.e. between the two limiting slab - slab configurations . concerning the depth @xmath59 , it also allows a wide variation of the pressure .
we remark that for @xmath59 going to zero we recover the result corresponding to @xmath202 , that is a filled grating . on the contrary , for increasing values of @xmath59
, we see that we approach to a pressure approximately equal to half the value of the pressure for @xmath202 .
this can be interpreted by noticing that roughly speaking at some point the corrugation is so deep that only the upper part ( half of the total surface , being @xmath175 ) contributes to the pressure .
differently , the dependence of the pressure on the period @xmath4 is less pronounced , and absent within our accuracy in the case of a lateral shift between the gratings , not reported in figure .
as we have shown , the filling factor is a promising tool to tailor the behavior of the pressure .
this is further pointed out in fig .
[ fig7 ] , where the distance - dependent pressure is plotted for three different values of @xmath213 .
whereas the asymptotic value of the pressure is practically the same , we note that for small distances the three curves differ visibly .
more interestingly , the attractive - repulsive transition can be tuned approximately from 2.5 to 3.5@xmath195 m by changing @xmath213 from 0.15 to 0.85 .
let us focus now on the spectral properties of the pressure , by analyzing the quantity @xmath220 , defined as the spectral component at frequency @xmath24 of the non - equilibrium contribution to the force , that is @xmath221 also in this case , we consider our reference point @xmath210 m and @xmath222 and compare its spectral distribution with the two slab - slab cases ( @xmath203 and @xmath202 ) as well as with some variations of one of the three parameters discussed above .
m [ temperatures @xmath196k ] .
the solid black line corresponds to filled gratings ( @xmath202 ) , the dot - dot - dashed red line to empty ones ( @xmath203 ) , the dotted blue line to our reference gratings , having @xmath175 . in the other curves we vary the geometrical parameters one by one with respect to our reference case : dot - dashed violet line for @xmath223 , short - dashed green line for @xmath224 m , long - dashed brown line for @xmath225m.,width=302 ] the result
is shown in fig .
we see that no striking spectral difference is present between the configurations compared . roughly speaking
, no new modes ( such as the spoof plasmons observed in metal gratings @xcite ) are observed in the spectral region of interest , that is up to @xmath24 of the order of @xmath226rads@xmath227 .
the spectral properties for any considered value of the geometrical parameters show small differences with respect to the ones of the two slab - slab configurations .
m and @xmath228 as a function of @xmath229 and @xmath18 .
the solid line corresponds to zero pressure , while the dashed lines to the other contour lines shown in legend.,width=302 ] as we have seen in sec .
[ secgeopar ] , the filling factor is a promising tool to shift the distance at which the transition between attraction and repulsion takes place .
nevertheless , from an experimental point of view it is more interesting to understand how this transition can be affected by tuning parameters which can be varied during an experiment , such as the three temperatures .
this is topic of this section , where we first consider the case in which the two gratings have a common temperature @xmath228 , in general different from the environmental one @xmath18 . for this configuration ,
we plot in fig .
[ fig9 ] the pressure acting on grating 1 in the reference configuration discussed above as a function of @xmath229 and @xmath18 .
the plot is clearly divided in two regions , corresponding to positive and negative values of the pressure , separated by a solid zero - pressure line . following this line
, we see that repulsion can be obtained only for body temperatures larger than approximately 312k , and that for larger values of @xmath229 a larger region of @xmath18 realizes repulsion .
moreover , we stress the remarkable feature that for values of @xmath229 close to 312k the pressure is approximately zero and almost independent on the environmental temperature for @xmath18 up to approximately 150k . in the same spirit of our last analysis we now fix only @xmath34 at three different values ( 200 , 300 and 400k ) and let @xmath230 and @xmath18 vary .
the pressure as a function of the two temperatures is shown in fig .
[ fig10 ] .
we see a behavior similar to the one observed in fig .
[ fig9 ] , that is the existence of a minimum temperature @xmath231 below which repulsion is impossible , as well as a region where the pressure is close to zero almost independently of @xmath18 . as manifest from fig .
[ fig10 ] , the limit temperature @xmath231 is a decreasing function of @xmath34 .
m as a function of @xmath230 and @xmath18 for three different values of @xmath34 .
same convention of fig .
[ fig9 ] for contour lines.,width=680 ] of attractive - repulsive transition of the pressure as a function of @xmath18 for three different values of @xmath228.,width=302 ] finally , we discuss how the distance @xmath232 at which the attractive - repulsive transition takes place can be tuned by changing the three temperatures .
this is shown in fig .
[ fig11 ] , where we fix @xmath233k and plot @xmath232 as a function of @xmath18 . as a general remark ,
when @xmath18 is smaller than @xmath229 the distance @xmath232 tends to a constant value , which decreases from 5.5@xmath195 m to 3@xmath195 m for @xmath229 going from 200 to 400k .
furthermore , when @xmath18 tends to @xmath229 , i.e. the system approaches thermal equilibrium , @xmath232 tends to a vertical asymptote , in accordance to the fact that the pressure is always attractive at thermal equilibrium .
we calculated the casimir - lifsthiz pressure out of thermal equilibrium acting on a 1d dielectric lamellar grating in front of another ( in general different ) dielectric grating . to this aim
, we implemented the fourier modal method in order to derive the scattering operators associated to each individual grating . using the general formalism for casimir - lifshitz force based on scattering matrices , we calculated the pressure acting on a finite fused silica grating in presence of an infinite silicon grating , and also compared our results to the proximity force approximation .
we showed that the combination of geometrical structuring of the surface and absence of thermal equilibrium offers an extremely rich domain of variation both with respect to thermal equilibrium and with respect to planar slabs out of thermal equilibrium .
as in the case of two slabs , non - equilibrium is able to produce a repulsive pressure , whose intensity can be tuned by varying the temperatures as well as the several geometrical parameters associated to each grating .
we also pointed out the presence of regimes in which the pressure is close to zero and almost independent of the environmental temperature .
remarkably , the variations of all the parameters strongly affect the distance at which the transition between attractive and repulsive pressure occurs , allowing to obtain transition distances as low as @xmath234 m .
this feature is indeed promising for the experimental observation of a repulsive force .
moreover , our results can be relevant in the context of force manipulations on micro - mechanical systems @xcite .
finally , an extension of this study to three - body configurations is also promising toward the manipulation of heat transfer @xcite .
99 h. b. g. casimir , proc .
akad . wet . *
51 * , 793 ( 1948 ) .
h. b. g. casimir - lifshitzand d. polder , phys . rev . *
73 * , 360 ( 1948 ) .
i. e. dzyaloshinskii , e. m. lifshitz , and l. p. pitaevskii , adv .
* 10 * , 165 ( 1961 ) .
_ casimir - lifshitzphysics _ , lecture notes in physics vol .
834 , edited by d. dalvit , p. milonni , d. roberts , and f. da rosa ( springer - verlag , berlin , 2011 ) .
a. sushkov et al . ,
physics * 7 * , 230 ( 2011 ) .
m. antezza , l. p. pitaevskii , and s. stringari , phys .
lett . * 95 * , 113202 ( 2005 ) .
m. antezza , j. phys .
a : math . gen . * 39 * , 6117 ( 2006 ) .
j. m. obrecht , r. j. wild , m. antezza , l. p. pitaevskii , s. stringari , and e. a. cornell , phys . rev
. lett . * 98 * , 063201 ( 2007 ) .
m. antezza , l. p. pitaevskii , s. stringari , and v. b. svetovoy , phys .
lett . * 97 * , 223203 ( 2006 ) .
m. antezza , l. p. pitaevskii , s. stringari , and v. b. svetovoy , phys .
a * 77 * , 022901 ( 2008 ) .
m. antezza , l. p. pitaevskii , s. stringari , phys .
a * 70 * , 053619 ( 2004 ) .
s. y. buhmann and s. scheel , phys .
lett . * 100 * , 253201 ( 2008 ) .
y. sherkunov , phys .
a * 79 * , 032101 ( 2009 ) .
r. o. behunin and b .-
hu , phys .
a * 82 * , 022507 ( 2010 ) .
r. behunin and b .-
hu , j. phys .
a : math . theor . * 43 * , 012001 ( 2010 ) .
r. o. behunin and b .-
hu , phys . rev .
a * 84 * , 012902 ( 2011 ) .
g. bimonte , phys . rev . a * 80 * , 042102 ( 2009 ) .
r. messina and m. antezza , europhys . lett . * 95 * , 61002 ( 2011 ) .
r. messina and m. antezza , phys .
a * 84 * , 042102 ( 2011 ) .
m. krger , t. emig , g. bimonte , and m. kardar , europhys . lett . * 95 * , 21002 ( 2011 ) .
m. krger , g. bimonte , t. emig , and m. kardar , phys .
b * 86 * , 115423 ( 2012 ) .
a. w. rodriguez , o. ilic , p. bermel , i. celanovic , j. d. joannopoulos , m. soljai , and s. g. johnson , phys .
lett . * 107 * , 114302 ( 2011 ) .
a. p. mccauley , m. t. h. reid , m. krger , and s. g. johnson , phys . rev .
b * 85 * , 165104 ( 2012 ) .
a. w. rodriguez , m. t. h. reid , and s. g. johnson , phys .
b * 86 * , 220302(r ) ( 2012 ) .
m. krger , t. emig , and m. kardar , phys .
106 * , 210404 ( 2011 ) .
a. w. rodriguez , m. t. h. reid , and s. g. johnson , phys .
b * 88 * , 054305 ( 2013 ) .
r. messina and m. antezza , phys .
rev . a * 89 * , 052104 ( 2014 ) .
b. bellomo , r. messina , and m. antezza , europhys
* 100 * , 20006 ( 2012 )
. b. bellomo , r. messina , d. felbacq , and m. antezza , phys .
a * 87 * , 012101 ( 2013 ) .
b. bellomo and m. antezza , europhys . lett . * 104 * , 10006 ( 2013 ) . b. bellomo and m. antezza , new j. phys . * 15 * , 113052 ( 2013 ) .
p. s. davids , f. intravaia , f. s. s. rosa , and d. a. r. dalvit , phys .
a * 82 * , 062111 ( 2010 ) .
f. intravaia , p. s. davids , r. s. decca , v. a. aksyuk , d. lpez , and d. a. r. dalvit , phys .
a * 86 * , 042101 ( 2012 ) .
j. lussange , r. gurout , and a. lambrecht , phys .
a * 86 * , 062502 ( 2012 ) .
r. gurout , j. lussange , h. b. chan , a. lambrecht , and s. reynaud , phys .
a * 87 * , 052514 ( 2013 ) .
j. lussange , r. gurout , f. s. s. rosa , j .- j .
greffet , a. lambrecht , and s. reynaud , phys .
b * 86 * , 085432 ( 2012 ) .
r. gurout , j. lussange , f. s. s. rosa , j .-
hugonin , d. a. r. dalvit , j .- j .
greffet , a. lambrecht , and s. reynaud , phys .
b * 85 * , 180301(r ) ( 2012 ) .
h. b. chan , y. bao , j. zou , r. a. cirelli , f. klemens , w. m. mansfield , and c. s. pai , phys .
* 101 * , 030401 ( 2008 ) .
y. bao , r. gurout , j. lussange , a. lambrecht , r. a. cirelli , f. klemens , w. m. mansfield , c. s. pai , and h. b. chan , phys .
rev . lett . * 105 * , 250402 ( 2010 ) .
f. intravaia et al . ,
nat . comm . * 4 * , 2515 ( 2013 ) .
h. b. chan , v. a. aksyuk , r. n. kleiman , d. j. bishop , and f. capasso , science * 291 * , 1941 ( 2001 ) .
j. zou , a. w. rodriguez , m. t. h. reid , a. p. mccauley , i. i. kravchenko , t. lu , y. bao , s. g. johnson , and h. b. chan , nat . comm . * 4 * , 1845 ( 2013 ) .
h. kim , j. park , and b. lee , _ fourier modal method and its applications in computational nanophotonics _
( crc press , boca raton , 2012 ) .
g. granet , b. guizal , j. opt .
a * 13 * , 1019 ( 1996 ) .
_ handbook of optical constants of solids _ , edited by e. palik ( academic press , new york , 1998 ) .
f. j. garca - vidal , l. martn - moreno , and j. b. pendry , j. opt .
a : pure appl
* 7 * , s97 ( 2005 ) .
n. yu et al . , nat .
* 9 * , 730 ( 2010 ) .
r. messina , m. antezza , and p. ben - abdallah , phys .
lett . * 109 * , 244302 ( 2012 ) . | we calculate the casimir - lifshitz pressure in a system consisting of two different 1d dielectric lamellar gratings having two different temperatures and immersed in an environment having a third temperature .
the calculation of the pressure is based on the knowledge of the scattering operators , deduced using the fourier modal method .
the behavior of the pressure is characterized in detail as a function of the three temperatures of the system as well as the geometrical parameters of the two gratings .
we show that the interplay between non - equilibrium effects and geometrical periodicity offers a rich scenario for the manipulation of the force .
in particular , we find regimes where the force can be strongly reduced for large ranges of temperatures .
moreover , a repulsive pressure can be obtained , whose features can be tuned by controlling the degrees of freedom of the system . remarkably , the transition distance between attraction and repulsion can be decreased with respect to the case of two slabs , implying an experimental interest for the observation of repulsion . |
automation of financial and legal processes requires enforcement of confidentiality and integrity of transactions . for practical integration with the existing manual systems , such enforcement should be transparent to users .
for instance , a person continually signs paper - based documents ( e.g. , bank checks ) by hand , while his embedded handwritten signature images are used to secure the digitized version of the signed documents .
such scenario can be realizable using biometric cryptosystems ( also known as bio - cryptographic systems @xcite ) by means of the offline handwritten signature images . in bio - cryptography , biometric signals like fingerprints , iris , face or signature images , etc .
, secure private keys within cryptography schemes like digital signatures and encryption .
biometric samples provide a more trusted identification tool when compared to simple passwords .
for instance , a fingerprint is attached to a person and it is harder to impersonate than traditional passwords . despite its identification power , biometrics forms a challenging design problem due to its fuzzy nature .
for instance , while it is easy for a person to replicate his password during authentication , it rarely happens that a person applies exact fingerprint each time .
the main source of variability in physiological biometrics like fingerprint , face , iris , retina , etc . is the imperfect acquisition of the traits .
on the other hand , behavioral biometrics like handwritten signatures , gait , and even voice , have intrinsic variability that is harder to cancel .
fuzzy vault ( fv ) is a reliable scheme presented mainly to enable usage of fuzzy keys for cryptography @xcite .
a fv decoder permits limited variations in the decryption key so that secrets can be decrypted even with variable keys . accordingly
, this scheme fits the bio - cryptography implementations , where biometrics are considered as fuzzy keys by which private cryptographic keys are secured .
since the fv scheme has been proposed , it has being extensively employed for bio - cryptography , where most implementations focused on physiological biometrics , e.g. , fingerprints @xcite , face @xcite and iris @xcite .
fv implementations based on the behavioral handwritten signatures are few and mostly employed online signature traits , where dynamic features like pressure and speed are acquired in real time by means of special devices as electronic pens and tablets @xcite .
static offline signature images , that are scanned after the signing process ends , however , integrate too much variability to cancel by a fv decoder @xcite .
recently , the authors have proposed the first offline signature - based fuzzy vault ( osfv ) implementation @xcite-@xcite .
this implementation is employed to design a practical digital signature system by means of handwritten signatures @xcite . in this paper , this implementation is reviewed and extended .
in particular , we propose an extension to enhance the security and accuracy of the basic osfv system by adapting cryptographic key size for individual users .
finally , system performance on the gpds public signature database @xcite , besides the private pucpr brazilian database @xcite , are presented and interpreted .
the rest of the paper is organized as follows . in the next section ,
the osfv implementation and its application to produce digital signatures by means of the handwritten signature images are reviewed .
section iii describes the signature representation and lists some aspects for enhanced representations .
section iv introduces some osfv variants for enhanced accuracy .
section v lists some variants for enhanced security .
the new variant that adapts key sizes for enhanced security and accuracy is described in section vi .
the simulation results are presented in section vii .
finally , some research directions and conclusions are discussed in section viii .
the system proposed for osfv consists of two main sub - systems : enrollment and authentication ( see figure [ fig : figure6 ] ) . in the enrollment phase , some signature templates @xmath0 are collected from the enrolling user .
these templates are used for the user representation selection , as described in section iii .
the user representation selection process results in a user representations matrix @xmath1 , where @xmath2 is the vector of indexes of the selected features , @xmath3 is a vector of indexes mapping represented in @xmath4-bits , and @xmath5 is the vector of expected variabilities associated with the selected features .
this matrix is user specific and contains important information needed for the authentication phase .
accordingly , @xmath6 is encrypted by means of a user password @xmath7 . both fv and password
are then stored as a part of user bio - cryptography template ( @xmath8 ) .
then , the user parameters @xmath9 and @xmath10 are used to lock the user cryptography key @xmath11 by means of a single signature template @xmath12 in a fuzzy vault @xmath13 . in the authentication phase ,
user password @xmath7 is used to decrypt the matrix @xmath6 .
then , the vectors @xmath14 and @xmath15 are used to decode the fv by means of user query signature sample @xmath16 .
finally , user cryptographic key @xmath11 is released to the user so he can use it to decrypt some confidential information or digitally signs some documents .
the enrollment sub - system uses the user templates @xmath0 , the password @xmath7 , and the cryptography key @xmath11 to generate a bio - cryptography template ( bct ) that consists of the fuzzy vault @xmath13 and the encrypted user representation matrix @xmath17 .
the user representation selection module generates the @xmath6 matrix as described in section iii .
the osfv encoding module ( illustrated in figure [ fig : figure7 ] ) describes the following processing steps : 1 .
the virtual indexes @xmath3 are quantized in @xmath4-bits and produces a vector @xmath18 .
2 . the user feature indexes @xmath19 are used to extract feature representation @xmath20 from the signature template @xmath12 . this representation is then quantized in @xmath4-bits and produces a vector @xmath21 .
the features are encoded to produce the locking set @xmath22 , where @xmath23 consists of @xmath24-bits fv points 4 .
the cryptography key @xmath11 of size @xmath25 where : + @xmath26 + is split into @xmath27 parts of @xmath24-bits each , that constitutes a coefficient vector @xmath28 .
a polynomial @xmath29 of degree @xmath30 is encoded using @xmath31 , where @xmath32 .
the polynomial is evaluated for all points in @xmath22 and constitutes the set @xmath33 where @xmath34 .
chaff ( noise ) points @xmath35 are generated , where @xmath36 , i \in [ 1,t ] $ ] , and @xmath37 $ ] .
a chaff point @xmath38 is composed of two parts : the index part @xmath39 and the value part @xmath40 .
two groups of chaff points are generated .
chaffs of @xmath41 have their indexes equal to the indexes of the genuine points .
the chaff points and the genuine point that have the same index part are all equally spaced by a distance @xmath42 , eliminating the possibility to differentiate between the chaffs and the genuine point .
chaffs of @xmath43 have their index part differs than that of the genuine points as the number of chaffs in @xmath41 is limited by the parameters @xmath44 and @xmath42 , so to inject higher quantity of chaffs we define @xmath45 as a chaff groups ratio , where : + @xmath46 + where @xmath47 and @xmath48 are the amount of chaff features belong to @xmath41 and @xmath43 , respectively .
@xmath43 chaffs are generated with @xmath49 indexes different than the @xmath44 genuine indexes .
hence , the fv size @xmath50 is given by : + @xmath51 + so , the total number of chaffs @xmath52 is given by : + @xmath53 7 .
the genuine set @xmath54 , and the chaff set @xmath55 are merged to constitute the fuzzy vault @xmath56 , where @xmath57 , @xmath22 , @xmath58 and @xmath59 , @xmath60 .
the authentication sub - system uses the user query sample @xmath16 and the password @xmath7 , to decode the fuzzy vault @xmath13 and restore the user cryptography key @xmath11 .
first the password @xmath7 is used to decrypt the @xmath6 matrix .
then the vectors @xmath14 , and @xmath15 are used to decode the fv by means of the query @xmath16 .
the osfv decoding module ( illustrated in figure [ fig : figure9 ] ) describes the following processing steps : 1 .
the virtual indexes @xmath3 are quantized in @xmath4-bits and produces a vector @xmath61 .
2 . the user feature indexes @xmath19 are used to extract feature representation @xmath62 from @xmath16 . this representation is then quantized in @xmath4-bits and produces a vector @xmath63 .
the features are encoded to produce the unlocking set @xmath64 , where @xmath65 .
hence , the unlocking elements are represented in a field @xmath66 .
the unlocking set @xmath67 is used to filter the chaff points from the fv .
an adaptive matching method is applied to match unlocking and locking points .
items of @xmath67 are matched against all items in @xmath68 .
this process results in a matching set @xmath69 , where @xmath70 represents the projection of the matching features on the polynomial space .
chaff filtering is done as follows .
if the feature indexes are correct , then all elements of @xmath71 will have corresponding elements in @xmath72 .
so , all of chaffs of @xmath43 will be filtered out .
then , each of the remaining fv points will be compared to corresponding points extracted from the query sample .
an adaptive matching method is applied : for every feature @xmath73 , a matching window @xmath74 is adapted to the feature modeled variability @xmath75 , where @xmath76 .
a fv point @xmath77 is considered matching with an unlocking point @xmath78 , if they reside in the same matching window .
i.e. , @xmath79 .
the matching set @xmath80 is used to reconstruct a polynomial @xmath81 of degree @xmath30 by applying the rs decoding algorithm @xcite .
the coefficients of @xmath81 are assembled to constitute the secret cryptography key @xmath82 . in @xcite ,
the osfv implementation is employed to produce digital signatures using offline handwritten signatures .
this methodology facilitates the automation of business processes , where users continually employ their handwritten signatures for authentication .
users are isolated from the details related to the generation of digital signatures , yet benefit from enhanced security .
figure [ signature ] illustrates the osfv - based digital signature framework .
the user fv that is constructed during enrollment is used to sign user documents offline as follows . when a user signs a document by hand , his handwritten signature image is employed to unlock his private key @xmath83 .
the unlocked key produces a digital signature by encrypting some message @xmath84 extracted from the document ( e.g. , check amount ) .
the encrypted message is considered as a digital signature and it is attached to the digitized document .
any party who possesses the user public key can verify the digital signature , where verification of the digital signature implies authenticity of the manuscript signature and integrity of the signed document ( e.g. , check amount did not change ) .
according to aforementioned osfv implementation , the fv points encode some features extracted from the signature images .
it is obvious that accuracy of a fv system relies on the feature representation .
representations of intra - personal signatures should sufficiently overlap so that matching errors lie within the error correction capacity of the fv decoder . on contrary , representations of inter - personal signatures should sufficiently differ so that matching errors are higher than the error correction capacity of the fv decoder .
accordingly , the authors proposed to design signature representations adapted for the fv scheme by applying a feature selection process in a feature - dissimilarity space . in this space ,
features are extracted from each pair of template and query samples and the pair - wise feature distances are used as space dimensions . to illustrate this approach , see figure [ fig : figure12 ] . in this example , three signature images are represented : @xmath85 is the template signature , @xmath86 is a genuine query sample and @xmath87 is a forgery query sample . in the left side ,
signatures are represented in the fv feature encoding space , where a fv point encodes a feature index @xmath73 and its value @xmath88 . for simplicity ,
only two features ( @xmath89 and @xmath90 ) are shown , while the full representation consists of @xmath44 dimensions .
on the right side , signatures are represented in the feature dissimilarity space . in this space ,
a feature is replaced by its distance from a reference value .
for instance , @xmath89 and @xmath90 are replaced by their dissimilarity representations @xmath91 , where @xmath92 , and @xmath93 .
accordingly , while a point in the feature encoding space represents a signature image , a point in the feature dissimilarity space represents the dissimilarity between two different signature images .
the point @xmath94 represents the dissimilarity between the genuine signature @xmath86 and the template @xmath85 , and a point @xmath95 represents the dissimilarity between the forgery signature @xmath87 and the template @xmath85 , where @xmath94= ( @xmath96 ) , and @xmath95= ( @xmath97 ) . in this example ,
@xmath98 and @xmath99 are discriminant features . for instance , for all genuine query samples like @xmath86 , @xmath100 and for all forgery query samples like @xmath87 , @xmath101 .
unfolding these discriminant dissimilarity features to the original feature encoding space produces discriminant features in the encoding feature space , where the distance between two feature instances is used to determine their similarity .
for instance , a genuine feature ( like @xmath102 ) lies close to the template feature @xmath103 , so they are similar , where closeness here implies that both features reside in a matching window @xmath76 .
features extracted from a forgery image ( like @xmath104 ) do not resemble the template feature @xmath103 , as they reside outside the matching window @xmath74 .
aforementioned description of the process to design representations is generic .
some extensions are reviewed and compared below .
shortage of user samples for training is addressed by designing a global writer - independent ( wi ) representation @xcite .
a large number of signature images from a development database are represented in the feature - dissimilarity space of high dimensionality , and feature selection process runs to produce a global space of reduced dimensionality .
such global approach permits designing fv systems for any user even who provides a single signature sample during enrollment . for performance improvement
, the global representation is specified for individual users once enough number of enrolling samples becomes available . to this end
, training samples are firstly represented in the global representation space , then an additional training step runs to produce a local writer - dependent ( wd ) representation that discriminates the specific user from others .
simulation results have shown that local representations enhanced fv decoding accuracy by about 30% , where the average error rate ( aer ) is decreased from 25% in case of global representations to 17.75% for the local ones . in @xcite , the extended shadow code ( esc ) feature extraction method is adapted for the fv implementation @xcite .
these features consist in the superposition of a bar mask array over a binary image of handwritten signatures . each bar
is assumed to be a light detector related to a spatially constrained area of the 2d signal .
this method is powerful in detecting various levels of details in the signature images by varying the extraction scale .
for instance , an image could be split to @xmath105 of horizontal and vertical cells , respectively , and shadow codes are extracted within individual cells .
the higher the number of cells , the higher the resolution of detectors .
the authors observed that designing fvs based on a single extraction scale results in varying performance for the different users .
for instance , while high resolution scales are fine with users whose signatures are easy to forge or those who have high similarities with others , the low resolutions are better for users whose signatures integrate high variabilities .
accordingly , a multi - scale feature fusion method is proposed , where different feature vectors are extracted based on different extraction scales and they are combined to produce a high - dimensional representation .
this representation is then processed through the wi and wd design phased and produces the final local representation that encodes in the fv . besides fusing feature vectors that are extracted based on different scales , it is possible to fuse different types of features . in @xcite , the directional probability density function ( dpdf )
features @xcite are fused with esc features to constitute a huge dimensional representation ( of 30,201 dimensionality ) .
this representation is reduced through the wi and wd training steps and produced a concise representation of only 20 features .
it is shown that injecting the additional feature type increased the fv decoding accuracy by about 22% ( aer is reduced from 17.75% to 13.75% ) .
the aforementioned approach provides a practical scenario to produce representations with low intra - personal and high inter - personal variabilities which is mandatory feature for fv systems .
however , the authors observed that the margin between the intra and the inter classes differs when using different signature prototypes ( templates ) for fv encoding .
accordingly , a prototype selection method is proposed @xcite .
the wd representation is projected to a dissimilarity space where distances to different user prototypes are the space constituents .
then , a feature selection process runs in the dissimilarity space and locates the best prototype .
this method has enlarged the separation between the intra and inter clusters significantly ( area under roc curve ( auc ) is increased from 0.93 to 0.97 ) .
although accuracy of an osfv system relies mainly on quality of the feature representation , the proposed implementation provides additional opportunities for enhanced accuracy by applying some other design variants as described in this section .
the results mentioned so far report accuracy of fv decoders that apply strict matching approach .
two fv points are matching only if they have identical values .
accuracy of a fv decoder is enhanced by applying the adaptive matching method , where the feature variability matrix @xmath15 is used for matching so that corresponding fv points are considered matching if their difference lies within the expected variability of their encoding feature ( see figure [ fig : figure3 ] ) .
this method increased accuracy by about 27% ( aer is reduced from 13.75% to 10.08% ) @xcite . instead of decoding a single fv token ,
it is possible to decode several fvs for enhanced performance . in case that some fvs are correctly decoded , the decrypted key is released to the user based on the majority vote rule .
this method has increased detection accuracy by about 18% ( aer is reduced from 10.08% to 8.21% ) @xcite .
the limited discriminative power of fvs is alleviated by using an additional password @xmath7 , so that the false accept rate ( far ) is reduced without significantly affecting the false reject rate ( frr ) .
for the results reported so far , it was assumed that the user password @xmath7 is compromised . however , to report the actual performance of the system we have to consider the case when an attacker neither possesses a correct password nor a genuine signature sample . in this case
, he can not decrypt the ur model and hence he randomly guesses the feature indexes .
it is shown that the additional password has increased detection accuracy by about 65% ( aer is reduced from 8.21% to 2.88% ) @xcite . using additional passwords for enhanced system accuracy
comes with the expense of the user inconvenience . in @xcite ,
a novel user - convenient approach is proposed for enhancing the accuracy of signature - based biometric cryptosystems .
since signature verification ( sv ) systems designed in the original feature space have demonstrated higher discriminative power to detect impostors @xcite , they can be used to improve the fv systems . instead of using an additional password ,
the same signature sample is processed by a sv classifier before triggers the fv decoders ( see figure [ sv - fv ] ) . using this
cascaded approach , the high far of fv decoders is alleviated by the higher capacity of sv classifiers to detect impostors .
this method has increased detection accuracy by about 35% ( aer is reduced from 10.08% to 6.55% ) . when multiple fvs are fused ,
the aer is decreased by 31.30% ( from 8.21% to 5.64% ) .
security of the osfv implementation is analyzed in terms of the brute - force attack @xcite .
assume an attacker could compromise the fv without possessing neither valid password nor genuine signature sample . in this case , the attacker tries to separate enough number of genuine points ( @xmath27 ) from the chaff points .
security of a fv is given by : @xmath106 where @xmath45 is the chaff group ratio , @xmath44 is the number of genuine points in the fv , @xmath30 is the degree of the encoding polynomial and @xmath42 is the chaff separation distance .
high value of @xmath45 implies a high number of g2 features which are compromised in case that the password is compromised .
the parameter @xmath44 should be concise as it impacts the accuracy and complexity of the fv .
accordingly , entropy of the system can be increased through using different values of the parameters : @xmath42 and @xmath30 . however , there is a trade - off between system security and its recognition accuracy that could be alleviated by applying the following approaches . in the traditional chaff generation method ,
equal - spaced chaff points are generated with a separation factor @xmath42 @xcite .
in such case , there is a trade - off between security and robustness .
for instance , with small separation , e.g. , @xmath107 , there are 40 fv points generated with the same index ( 1 genuine + 39 chaff points ) . in this case , a high number of chaffs is generated and results in high system entropy of about 68-bits and low accuracy of about @xmath108 aer .
the adaptive chaff generation method enables the injection of high number of chaff with minimal impact on the fv decoding robustness . to this end , the feature variability vector @xmath15 is used during the fv locking phase so that chaff points are generated adaptively according to feature variability . for each feature
@xmath88 , @xmath109 ( see figure [ fig : figure3 ] ) . by this method , it is less likely that an unlocking element equates a chaff element .
for or instance , the same entropy ( 68-bits ) could be achieved with a minimal impact on system robustness ( aer = 10.52% ) @xcite . according to eq.[s ] ,
the longer the cryptographic key size the higher entropy of the fv .
however , this comes with expense of the accuracy @xcite . in @xcite , different key sizes ( ks )
are tried ( 128 , 256 , 512 , 1024-bits ) and it is shown that different key sizes result in different performance for the different users .
this observation motivates adapting the key length for each user as proposed in the following section ,
in @xcite , functionality of a fv decoder is formulated as a simple dissimilarity threshold as follows : = -1 pt @xmath110 where a fv encoded by a template @xmath85 can be correctly decoded by a query @xmath16 only if the total dissimilarity between @xmath16 and @xmath85 is less than the error correction capacity @xmath111 of the fv decoder . here , @xmath112 is the dissimilarity part that results from the variability between the two samples , and @xmath113 is the dissimilarity part that results from wrong matches with chaff points .
the methods discussed so far aimed to optimize the dissimilarity parts of eq.[fvd ] .
for instance , the multi - scale and multi - type feature extraction approach results in separating intra - personal and inter - personal dissimilarity ranges .
selection of robust templates ( prototypes ) and applying adaptive matching enlarged this separation .
also , impact of the chaff error is minimized by presenting the adaptive chaff method . with applying all these methods ,
however , accuracy of a signature - based fv is still below the level required for practical applications . accordingly
, performance is increased by applying some complex and user inconvenient solutions like ensemble of fvs and using additional passwords or cascading sv and fv systems . here
we investigate a new room for enhancing fvs by optimizing the error correction capacity @xmath114 which is given by : = -1 pt @xmath115 it is obvious that this parameter relies on the fv encoding size @xmath44 and the encoding polynomial degree @xmath30 .
also , from eq.[ks ] , we see that @xmath30 determines the key size @xmath25 .
accordingly , we select user specific key sizes through changing the parameter @xmath30 so that @xmath114 for a specific user covers the range of his expected signature variability . to this end
, we set @xmath114 for a user to his maximum intra - personal variability @xmath116 .
based on the resulting user - specific error correction capacity , the parameter @xmath30 is determined using eq.[fv ] and user key size @xmath25 is computed using eq.[ks ] .
once appropriate key size is computed for a user , his key is enlarged through injecting some padding bits in the original key during fv encoding . during authentication ,
the enlarged key is reconstructed and the padding bits are removed to produce the original cryptographic key .
all aforementioned performance results are reported for the pucpr brazilian signature database @xcite . here
we test the system for the public gpds-300 database @xcite as well .
this database contains signatures of @xmath117 users , that were digitized as 8-bit greyscale at resolution of 300 dpi and contains images of different sizes ( that vary from @xmath118 pixels to @xmath119 pixels ) .
all users have 24 genuine signatures and 30 simulated forgeries .
it is split into two parts .
the first part contains signatures of the first 160 users .
a subset of this part is used to design the local representation and the remaining of this part is used for performance evaluation .
the second part contains signatures of the last 140 users and it is used to design the global representation . see @xcite for a similar experimental protocol for both databases .
table [ table : impact of using a user password ] shows results for the two databases for fixed and adaptive key sizes .
it is obvious that employing the adaptive key size approach decreased the far significantly with low impact on the frr .
for instance , the aer for the pucpr database in decreased by about 21% ( from 10.08 to 7.94 ) . also , performance of the system for the gpds database is comparable to state - of - the - art traditional sv systems ( aer is about 15% ) that employ more complex classifiers @xcite .
moreover , the proposed method also enhances system security as it is possible to increase the key size , and hence the polynomial degree @xmath30 , without much impact on the accuracy . for instance , figure [ fig : adaptive ] shows the adapted polynomial degrees for different users in the pucpr database and the corresponding user variability @xmath116 .
it is obvious that users with more stable signatures have their cryptographic keys more enlarged than users with less stable signatures . according to eq.[s ] ,
system entropy of the standard osfv implementation ( with fixed keys of size 128-bits and polynomial degree @xmath120 ) is about 45-bits . with applying the adaptive key size method
, the average @xmath30 is about 9.6-bits ( see figure [ fig : adaptive ] ) which provides an average entropy of about 51-bits .
.impact of using a user password as a second authentication measure [ cols="^,^,^,^,^ " , ]
in this paper , a recently published offline signature - based fv implementation is reviewed .
several variants of the system are listed and compared for enhanced accuracy and security . a novel method to adapt cryptography key sizes for different users
is proposed and have shown accuracy and security enhancement .
the performance is also validated on a public signature database where comparable results of complex sv in the literature is reported .
although the proposed key adaptation method sounds , there is need to propose more intelligent tuning technique taking in consideration the similarities with simulated forgeries for higher forgery detection .
this study listed many new approaches that are applied successfully to the signature based bio - cryptography .
we believe that these methods shall be investigated for other biometrics which might enhance state - of - the - art of the area of bio - cryptosystems .
, r. sabourin , and e. granger , on the dissimilarity representation and prototype selection for signature - based bio - cryptographic systems . ,
york , uk , 3 - 5 july 2013 , lncs , vol.7953 , pp.265 - 280 .
eskander . , r. sabourin , and e. granger , improving signature - based biometric cryptosystems using cascaded signature verification
fuzzy vault ( sv - fv ) approach . , crete island , greece , 1 - 4
september 2014 .
g. eskander , r. sabourin , and e. granger .
`` hybrid writer - independent writer - dependent offline signature verification system '' . _
iet - biometrics journal , special issue on handwriting biometrics _ ,
vol.2 , no.4 , pp . | an offline signature - based fuzzy vault ( osfv ) is a bio - cryptographic implementation that uses handwritten signature images as biometrics instead of traditional passwords to secure private cryptographic keys . having a reliable osfv implementation is the first step towards automating financial and legal authentication processes , as it provides greater security of confidential documents by means of the embedded handwritten signatures .
the authors have recently proposed the first osfv implementation which is reviewed in this paper . in this system ,
a machine learning approach based on the dissimilarity representation concept is employed to select a reliable feature representation adapted for the fuzzy vault scheme .
some variants of this system are proposed for enhanced accuracy and security . in particular , a new method that adapts user key size is presented .
performance of proposed methods are compared using the brazilian pucpr and gpds signature databases and results indicate that the key - size adaptation method achieves a good compromise between security and accuracy . while average system entropy is increased from 45-bits to about 51-bits , the aer ( average error rate ) is decreased by about 21% . |
this work was supported by the us national science foundation grant phy-1520976 .
inspection of fig .
1 ( b ) shows that there are two phase shift extrema as a function of @xmath9 .
the one we do not use , between the spectator and line resonances , is less sensitive to detuning , but it would increase the rate of unintentional transfer out of the qubit basis . this is partly because there are many more spectator atoms than targets ; the small amount of off - resonant transfer would get multiplied by a much larger number .
in addition , there can be atoms that see a fraction of the intensity that most line atoms see , due to misalignment of beams relative to the lattice or beam profile imperfections ; this can put their resonances closer to the phase shifting microwaves .
to ensure operation near the phase shift extremum , we find it necessary to adjust the intensities of the addressing beam in a target specific way .
this corrects for beam profile distortion by the mems mirror steering system and for the finite rayleigh range of the addressing beams near their foci .
the deviation of the cross atom resonance , 1.8@xmath6 , from 2@xmath6 ( see fig . 1 ( a ) ) occurs because the zeeman shift due to the bias b - field is not very much larger than the addressing beams vector light shifts .
thus when there is only one addressing beam , the quantization axis tilts slightly toward its propagation direction .
our rb building block .
these are the microwave pulses corresponding to a pg / cg pair .
the red pulse is the pg .
the cg is inside the dashed rectangle : the blue pulses are @xmath59 , the black pulses form the spin - echo type structure and the purple pulses are off - resonant microwaves that induce phase shifts .
all the pulses have a blackman profile . ]
we choose randomized benchmarking ( rb ) to characterize the performance of our gate instead of quantum state tomography for several reasons : it can distinguish gate errors from spam errors ; it has less resource overhead and is easily scaled to multi - qubit systems ; and it reflects the fidelity of a group of operations that is not biased by a particular kind of gate . the basic building block of an rb sequence is a pg - cg pair executed on two target atoms . a cg is a @xmath84 rotation about the @xmath85 or @xmath86axis on the bloch sphere .
it contains one phase gate @xmath37 sandwiched by two global @xmath87 rotations .
a pg is implemented as a global @xmath88 rotation ( about the @xmath85 or @xmath86axis ) or the absence of one ( for the identity or a rotation about @xmath89 ) . for the @xmath90 rotation , we frame - shift all successive microwave pulses .
we extend the spin - echo type structure to include the pauli pulses , as shown in supplementary fig .
1 . the @xmath59 and the pauli pulses are at the rephasing points in this spin - echo type structure .
while the main purpose of the echo @xmath3 pulses is to cancel the crosstalk due to addressing beams and microwave pulses , they also serve to preserve the qubit in the face of inhomogeneous broadening .
moreover , their phases can be adapted to mitigate microwave power errors , as described in the next section .
we have used separate measurements to isolate the factors that contribute to the spam error : collisions with background gas atoms ( 3@xmath34 ) , imperfect state transfer from @xmath91 to @xmath20 ( 2@xmath34 ) , and imperfect clearing ( @xmath92 ) .
the sum of these expected terms matches the measured @xmath81 from rb .
generally , there are two types of microwave transition imperfections : power and frequency errors .
our passively stabilized microwave source limits our power stability .
we effectively measure it by driving atoms with up to 65 @xmath3-pulses and find the average fractional error to be @xmath93 .
the source of frequency errors in our system is the inhomogeneous broadening of the ensemble of atoms ( @xmath23130 hz ) due to vibrational excitation .
although , this is too small to affect the fidelity of individual microwave pulses at the @xmath83 level , state evolution between pulses affects our error cancellation scheme .
various techniques have been developed in nmr for dealing with such errors , including composite pulses @xcite , and dynamical decoupling@xcite .
composite pulses from the bb1 family can be made insensitive to imperfect pulse amplitude and pulses from the corpse families can be made insensitive to frequency errors . since neither does both , and they come at the cost of considerably longer pulse durations , they are not suitable for our purpose .
we instead implement phase cycling schemes in which the errors introduced by microwave pulses are cancelled by subsequent pulses . before tackling the more complicated case of a random sequence of clifford and pauli gates ,
let us consider a simple spin echo sequence , a repeating sub - cycle of torque vector directions given by : \{@xmath94 } on the bloch sphere .
dephasing from any initial state is controlled , unlike with simpler schemes or the widely used xy class of pulses @xcite , which do nt have a perfect cancellation in the absence of inhomogeneous broadening .
inhomogeneous broadening compromises error cancellation , but phase cycling still greatly improves the performance . in supplementary fig .
2 we demonstrate the performance of this pulse scheme .
supplementary fig .
2 ( a ) plots the spin echo fringe with 100 @xmath3 pulses with phase cycling , with the dashed lines marking the maximum population .
the almost lossless fringe contrast confirms the high fidelity of this spin - echo type structure .
supplementary fig .
2 ( b ) plots this contrast versus microwave rabi frequency .
it is clear that this pulse scheme is quite robust against power fluctuation .
while this pulse scheme works nearly perfectly in a set of four , its implementation in an rb sequence is interrupted when there is a @xmath59 pulse or a pauli pulse ( see supplementary fig .
1 ) . we have empirically determined the following rules for choosing the right phase when the `` phase flow '' is interrupted : 1 .
the spin echo @xmath3 pulses around successive pauli pulses ( see supplementary fig . 1 ) should follow the cycle : \{@xmath95}. 2 .
if the pg preceding a cg has a torque vector along the same axis as the cg s @xmath59 pulse , then the torque vector of the first @xmath3 pulse inside the cg should be opposite that of pg .
otherwise , it can have the same phase as the @xmath59 pulse .
3 . the final four @xmath3 pulses in a cg should follow the cycle : \{@xmath94}. these rules are based on experimental testing with many randomized benchmarking sequences and phase schemes , but they may not be optimal .
the fidelity for data in fig .
2 . is calculated as follows .
the qubit state takes the form of @xmath96 @xmath97 , where @xmath98 and @xmath99 are the polar and azimuth angles on a bloch sphere , and @xmath100 represents a spherically symmetric shrinkage of the bloch sphere .
we measure the population in @xmath12 . after a detection @xmath101 pulse with a phase @xmath102 scanned
, the population has a sinusoidal dependence on @xmath102 : @xmath103 .
we first normalize the results by the maximum / minimum population ( 0.95/0.01 ) , which is set by the spam error , and then fit the curve to the above functional form , enabling us to reconstruct the qubit state after the gate operation . the fidelity in the fig .
inset is determined as follows .
we fix all the qubit state parameters to the theoretically perfect values ( @xmath104 , @xmath105 ) , except for @xmath99 , which is determined by the @xmath106 gate .
since both the fidelity @xmath51 and the population in @xmath12 are a function of @xmath99 , a relation between @xmath51 and @xmath107 can be established : @xmath108 .
we normalize @xmath107 as before , and then calculate @xmath51 and its associated uncertainties .
this approach is less mathematically rigorous than rb , but it is sufficient to illustrate the insensitivity of the gate to addressing beam alignment and intensity noise . | although the quality of quantum bits ( qubits ) and quantum gates has been steadily improving , the available quantity of qubits has increased quite slowly . to address this important issue in quantum computing
, we have demonstrated arbitrary single qubit gates based on targeted phase shifts , an approach that can be applied to atom , ion or other atom - like systems .
these gates are highly insensitive to addressing beam imperfections and have little crosstalk , allowing for a dramatic scaling up of qubit number .
we have performed gates in series on 48 individually targeted sites in a 40% full @xmath0 3d array created by an optical lattice . using randomized benchmarking
, we demonstrate an average gate fidelity of 0.9962(16 ) , with an average crosstalk fidelity of 0.9979(2 ) .
pacs numbers : : the performance of isolated quantum gates has recently been improved for several types of qubits , including trapped ions @xcite , josephson junctions @xcite , quantum dots @xcite , and neutral atoms @xcite .
single qubit gate errors now approach or , in the case of ions , surpass the commonly accepted error - threshold @xcite ( error per gate @xmath1 ) , for fault - tolerant quantum computation @xcite .
it remains a challenge in all these systems to execute targeted gates on many qubits with fidelities comparable to those for isolated qubits @xcite .
neutral atom and ion experiments have to date demonstrated the most qubits in the same system , 50 and 18 respectively @xcite .
the highest fidelity gates in these systems are based on microwave transitions , but addressing schemes typically depend on either addressing light beams @xcite which are difficult to make as stable as microwaves , or magnetic field gradients @xcite which limit the number of addressed qubits . in this report
, we present a way to induce phase shifts on atoms at targeted sites in a @xmath2 optical lattice that is highly insensitive to addressing laser beam fluctuations .
we further show how to convert targeted phase shifts into arbitrary single qubit gates .
these high fidelity gates are only sensitive to laser fluctuations at second order in intensity and fourth order in beam pointing .
we demonstrate average single gate errors across our array that are below 0.004 , and present a path towards reaching the fault - tolerant threshold .
in previous work @xcite we performed single site addressing in a 3d lattice using crossed laser beams to selectively ac stark shift target atoms , and microwaves to temporarily map quantum states from a field insensitive storage basis to the stark - shifted computational basis . while we use most of the same physical elements here ,
the crucial difference is that these new gates are based on phase shifts in the storage basis , and do not require transitions out of it .
non - resonant microwaves are applied that give opposite - sign ac zeeman shifts for different atoms .
a specific series of non - resonant pulses and global @xmath3-pulses on the qubit transition gives a zero net phase shift for non - target atoms and a controllable net phase shift for target atoms .
the resultant gate fidelity is much better than our previous gate because of this gate s extreme insensitivity to the addressing beam alignment and power , the insensitivity of the storage basis to magnetic fields and vector light shifts , and the independence on the phase of the non - resonant microwave pulses .
the principle of this phase gate , where optical addressing does not compromise fidelity , can be adapted to other atom and ion qubit geometries . addressing spectrum , phase shift and timing sequence .
( a ) spectra of the @xmath4 transition .
we plot the ratio of the number of detected @xmath5 atoms to the initial number of atoms as a function of the detuning from the unshifted resonance .
the green diamonds are due to spectators , the blue triangles are due to line atoms and the orange circles are due to cross atoms .
the solid lines are fits to gaussians .
the resonance frequencies of line atoms and cross atoms are marked by @xmath6 and @xmath7 .
( b ) exact calculation of the cumulative phase shift ( @xmath8 ) on a target atom as a function of the microwave detuning from the unshifted resonance .
the optimum operation point is marked by @xmath9 .
( c ) addressing pulse sequence . first row : blackman - profiled microwave pulse sequence .
the black pulses ( 80 @xmath10s ) are resonant on the @xmath11 to @xmath12 transition and affect all atoms .
the purple pulses ( 120 @xmath10s ) are addressing pulses detuned from the @xmath13 to @xmath14 transitions by @xmath9 .
second row : the addressing light intensity .
the addressing light barely affects the trapping potential , so a linear ramp is optimal .
it is effectively at full power for 252 @xmath10s .
third and fourth rows : schematic of addressing beams in two planes .
the atom color codes are as in ( a ) .
, width=302 ] detailed descriptions of our apparatus can be found in refs 15 , 21 and 22 .
we optically trap and reliably image neutral @xmath15cs atoms in a 5 @xmath16 spaced cubic optical lattice . the atoms are cooled to @xmath17 ground vibrational state occupancy and then microwave transferred into the qubit basis , the @xmath18 , @xmath19 and @xmath20 hyperfine sublevels , which we will call @xmath12 and @xmath11 , respectively .
lattice light spontaneous emission is the largest source of decoherence , with a 7 s coherence time ( @xmath21 ) that is much longer than the typical microwave pulse time of 80 @xmath22 .
we detect the qubit states by clearing atoms in the @xmath20 states and imaging the @xmath19 atoms that remain .
to target an atom , we cross two circularly polarized , 880.250 nm addressing beams ( beam waist @xmath23 2.7 @xmath16 , rayleigh range @xmath23 26 @xmath16 ) .
the addressing beams can be directed to a new target in @xmath24 5 @xmath22 using micro - electro - mechanical - system ( mems ) mirrors @xcite .
the addressing beams ac stark shift the @xmath25 levels of the atom at their intersection by about twice as much as any other atom .
this is illustrated in fig .
1(a ) , for which atoms are prepared in @xmath20 and a microwave near the @xmath26 transition is scanned . the resonances are visible for atoms at the intersection ( orange , termed `` cross '' atoms ) , atoms in one addressing beam path ( blue , termed `` line '' atoms ) , and the rest of the atoms ( green , termed `` spectator '' atoms ) .
the ac stark shift for the line atoms , @xmath6 , is chosen so that there is a region between the blue and orange peaks in which only a small fraction ( @xmath27 ) of atoms in any class ( cross , line or spectator ) makes the transition . when a microwave pulse is applied in that frequency range ,
atoms experience different ac zeeman shifts depending on their class . the addressing pulse sequence for a pair of target atoms in two planes , shown in fig 1 .
( c ) , consists of four stages @xcite .
the qubit - resonant spin - echo pulses ( the black pulses on the microwave line in fig 1 .
( c ) ) reverse the sign of the phase shifts , so whatever phase shifts ( ac zeeman or ac stark ) a non - target atom gets during the cross stages ( the first and third ) are exactly canceled by the shifts it gets during the dummy stages ( the second and fourth ) , where there is no cross atom .
in contrast , the first ( second ) target atom spends stage 1 ( 3 ) as a cross , stage 3 ( 1 ) as a spectator , and stages 2 and 4 ( 2 and 4 ) as a line atom . when the microwave frequency is chosen to be between the line and cross resonances , the change in the target atom s status from cross to line changes the sign of the ac zeeman shift .
away from the resonances , the net phase shift for the target atoms is : @xmath28 where @xmath9 is the microwave detuning from the spectator resonance , @xmath29 is the shift of cross atoms in units of @xmath6 ( see supplementary materials ) , @xmath30 is the microwave rabi frequency and @xmath31 is the pulse duration .
successive terms correspond to the integrated ac zeeman shift on the first target atom during successive stages .
the overall phase shift can be directly controlled by changing the power of the microwave field . the black curve in fig .
1(b ) shows the result of an exact calculation of the target atom s phase shift as a function of @xmath9 .
the phase shift minimum at @xmath32 khz is the preferred operating point for the gate , since the shift then quadratically depends on the change in @xmath9 and thus also on the addressing beams ac stark shift , with the coefficient of 21 @xmath33 . for example
, a 2@xmath34 change in @xmath6 gives a 8 mrad phase shift , which in turn leads to only a @xmath35 gate error .
since the intensity changes quadratically with beam alignment , the gate is sensitive to beam pointing only at fourth order .
the phase shift on target sites amounts to a rotation about the @xmath36axis ( an @xmath37 gate ) , but a universal gate requires arbitrary rotations about any arbitrary axis .
we can make an @xmath38 gate by combining the @xmath37 gate with global @xmath39 rotations : @xmath40 for non - target atoms , which see the global microwave pulses but experience no @xmath37 , @xmath41 clearly has no net effect .
it is straightforward to generalize this formula to obtain arbitrary rotations on a bloch sphere for target atoms .
the corresponding complete set of single qubit gates on target atoms all leave the non - target atoms unchanged .
interference fringe of one @xmath42 phase gate applied to 48 randomly chosen sites .
the black dots correspond to the target atoms , and the green dots correspond to non - target atoms .
solid lines are fits to data .
dashed lines mark the maximum and minimum possible populations .
after all these gates , the net @xmath43 for target atoms is @xmath44 , and for the non - target atoms it is @xmath45 .
the average error per gate for both target and non - target atoms is @xmath46 .
, width=340 ] we have demonstrated one @xmath47 gate on 48 randomly chosen sites ( in 24 gate pairs ) within a @xmath48 array . given the average initial site occupancy of @xmath49 , an average of 20 qubits experience the phase gate during each implementation .
we probe the coherence by closing the spin echo sequence with a global @xmath50 pulse whose phase we scan , as shown in fig .
2 .
the green points are due to non - target atoms , and the black points are due to atoms at the 48 target sites .
the corresponding curves are sinusoidal fits to the data .
the dashed lines mark the maximum and minimum populations one expects given perfect gate fidelity , @xmath51 , defined as the square of the projection of the measured state onto the intended state , in the face of state preparation and measurement ( spam ) errors ( see supplementary materials ) . from these curves we determine that the error per gate pair , @xmath52 , is @xmath53 for both target and non - target atoms on average . at least for this particular gate ,
most of the error is clearly common to target and non - target atoms .
the good contrast of the target atom fringe illustrates the homogeneity of the phase shifts across the ensemble .
the error per gate , @xmath54 , is thus @xmath55 on average . to graphically illustrate the flexibility of our gates ,
we have performed an @xmath56 rotation on a 32 site pattern ( see fig .
3 ) .
we rotate the superposition of target atoms from @xmath57 to @xmath58 , and then use a @xmath59 detection pulse with phase of @xmath3 to return spectator atoms to @xmath11 and target atoms to @xmath12 .
fig .
3 shows images from 5 planes summed over 50 implementations . !
h @xmath56 gates performed in a specific pattern .
( 1 - 5 ) : fluorescent images of 5 successive planes .
each image is the sum of 50 implementations . for clarity ,
the contrast is enhanced using the same set of dark / bright thresholds on all pictures to account for the shot noise .
note that there are no targeted sites in planes 2 and 4 ; the light collected there is from atoms in the adjacent planes .
the drawing to the lower right marks the targeted sites in blue , with plane shading and blue lines to guide the eye.,title="fig:",width=340 ] to confirm the robustness of this gate , we applied an @xmath47 gate to 24 targets , and measured the probability that atoms return to @xmath12 after a @xmath59 pulse as a function of the fractional change in addressing beam intensity ( see the fig
. 4 inset and supplementary materials ) .
the data confirms the theoretical prediction ( solid line ) of extreme insensitivity to addressing beam intensity ( see fig .
1(b ) ) .
semi - log plot of fidelity from a randomized benchmarking sequence .
we plot the probability of returning the two target atoms to @xmath12 after a randomized benchmarking sequence versus the sequence length . at each length
we employ 3 randomized cg sequences , each of which is combined with 3 sets of randomized pg sequences , for a total of 9 points . each data point
is averaged over @xmath23100 implementations .
a least squares fit using the @xmath60 determined from the auxiliary rb measurement on non - target atoms gives @xmath61 = @xmath62 .
inset : fidelity of an @xmath47 gate as a function of the fractional change of the addressing ac stark shift .
the points are experimental data and the curve is from the exact calculation . , width=340 ] to fully characterize gate performance
we employ the standard randomized benchmarking ( rb ) protocol @xcite , which has been used in nuclear magnetic resonance @xcite , quantum dots @xcite , ion systems @xcite and neutral atoms systems @xcite .
we implement rb by choosing computation gates ( cg ) at random from \{@xmath39 , @xmath63 } and pauli gates ( pg ) at random from \{@xmath64 } where @xmath65 is the identity @xcite .
following each randomized sequence , a detection gate is applied that in the absence of errors would return either the target qubits or the non - target qubits to @xmath12 . for the targets
, the average value of @xmath51 decays exponentially with the length of the randomized sequence : @xmath66 where @xmath67 is the spam error and @xmath68 is the number of cg - pg operations applied to the pair of target atoms . a similar expression yields @xmath69 and @xmath70 for the spectator and line atoms respectively , the unwanted changes wrought on the non - target atoms when a random series of gates is executed on the targets . in this way we can characterize crosstalk and extract errors per gate from errors per gate pair . for target atoms , @xmath71 , for spectator atoms , @xmath72 , and for line atoms , @xmath73 . first analyzing the non - target data ( see supplementary fig .
3 ) , we determine that @xmath74 , @xmath75 , and @xmath76 .
the crosstalk , defined as the average error per gate for non - target atoms , is @xmath77 .
only global microwave pulse imperfections contribute to @xmath78 .
their contribution to @xmath79 may differ since the microwave sequence is optimized for spectators .
spontaneous emission from the addressing beams adds to @xmath79 , with a calculable contribution of @xmath80 .
fig .
4 shows the rb data for the two target atoms . using the @xmath81 from the non - targets , the fit of eq .
3 to the data yields @xmath82 .
a summary of the errors for the target and non - target atoms is shown in table 1 .
the errors can mostly be traced to microwave power stability .
we expect that reaching the state of the art @xcite , and perhaps tweaking our spin echo infrastructure @xcite ( see supplementary fig .
1 ) , can bring it below @xmath83 .
the next largest contribution is from the readily calculable spontaneous emission of addressing light .
its contribution is invariant with gate times , since @xmath6 must be changed inversely with gate time . addressing with light between the d1 and d2 lines , as we do ,
locally minimizes spontaneous emission , but an eight - fold reduction could be obtained by doubling the wavelength , which would not dramatically compromise site addressing .
the required powerful addressing beams would exert a significant force on line atoms , but deeper lattices and adiabatic addressing beam turn - on would make the associated heating negligible . in our current experiment we need to wait 70 @xmath22 after the addressing beams are turned on for our intensity lock to settle .
technical improvement there could halve the time the light is on , ultimately leaving the spontaneous emission contribution to the error just below @xmath83 per gate .
that limit is based on addressing with a differential light shift .
it would disappear were the same basic scheme to be employed on a 3-level system , where only one of the qubit states is strongly ac stark shifted by the addressing light . in that case the phase shifting microwaves ( or light ) would simply be off - resonant from the qubit transition , and the error due to spontaneous emission would decrease proportional to the detuning . since the dominant target errors have the same sources as the non - target errors , improvements to the gate will correspondingly improve the crosstalk .
adapting this method to more common 1d and 2d geometries is straightforward .
only a single addressing beam is needed , and the dummy stages would be executed with half - intensity addressing beams . the same insensitivity to addressing beam power and alignment would follow . in summary
, we report on a universal single qubit gate based on targeted phase shifts .
the gate is robust to addressing light fluctuations , and can be applied to large arrays of qubits , including our demanding 3d lattice geometry .
scalable addressing is an important step towards a scalable quantum computer .
[ tb1 ] [ cols="^,^,^,^,^",options="header " , ] 30 _ et al . _ _ _ * * , ( ) .
, , & .
_ _ * * ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _ _ ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * ( ) . , , & .
_ _ * * , ( ) . .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
, , , & .
_ _ * * ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * ( ) .
_ et al . _ . _
_ * * ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _ _ * * , ( ) . , & . _
_ * * , ( ) .
, , & .
_ _ * * ( ) .
_ et al . _ . _ _ ( ) .
_ et al . _ . _
_ * * ( ) . , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * ( ) .
, , & .
_ _ * * ( ) .
, , & .
_ _ * * ( ) .
& .
_ _ * * , ( ) . , & _ _ * * , ( ) . |
quantum computing based on qubits has attracted considerable attention ( see , e.g. , @xcite ) .
there are several candidates to realize quantum computers , such as using nuclear spins in molecules , photons , trapped ions , superconducting circuit and quantum dots ( see , e.g. , @xcite ) .
however , it is still a great challenge to build a large - scale quantum computer .
quantum computers can significantly outperform classical computers in doing some specific tasks @xcite .
for example , two important quantum algorithms are shor s @xcite and grover s @xcite .
algorithm @xcite can factorize a large integer in polynomial time , offereing an exponential speed - up over classical computation .
algorithm @xcite gives a quadratic speed - up in searching database .
this search algorithm has been found to be very useful in other related problems @xcite . to date , the study of quantum algorithms is a very active area of research ( see , e.g. , @xcite ) . using three coupled harmonic oscillators ,
we have recently proposed @xcite an alternative approach ( without using qubits ) for quantum factorization .
we consider these three harmonic oscillators to be coupled together via nonlinear interactions @xcite .
to factorize an integer @xmath0 , this approach involves only three steps : initialization , time evolution , and conditional measurement . in this approach ,
the states of the first two harmonic oscillators are prepared in a number - state basis , while the state of the third oscillator is prepared in a coherent state .
the states of the first two harmonic oscillators encode the trial factors of the number @xmath0 .
the nonlinear interactions between the oscillators produce coherent states that simultaneously rotate in phase space with different effective frequencies , which are proportional to the product of two trial factors @xcite . in this way , _ all _ possible products of any two trial factors can be _ simultaneously _ computed , and then they are `` written '' to the rotation frequencies of the coherent states in _ a single step_. this saves considerable computational resources .
the resulting state of the first two oscillators is the factors state @xcite by performing a conditional measurement of a coherent state rotating with an effective frequency which is proportional to @xmath0 .
however , the probability of obtaining this coherent state becomes low when @xmath0 is large . in this paper
, we can circumvent this limitation by using an iterative method for increasing the chance of finding the states of the factors .
this amplitude - amplification method involves a number of iterations , where each iteration is very similar to the factoring approach we recently proposed @xcite .
we show that the number of iterations @xmath1 is of order of @xmath2 .
thus , using this method , _ the factors of a large integer can be obtained , with a high probability , in linear time _
the performance of this approach is even better than that of shor s algorithm , which factorizes a number in polynomial time @xcite .
now we briefly describe this amplitude - amplification method for quantum factorization using three coupled harmonic oscillators .
let us now consider the first step of our approach .
initially , the first two harmonic oscillators are in a number - state basis and the third oscillator is in a coherent state . let the three coupled harmonic oscillators evolve for a period of time .
the detection is then conditioned on a coherent state with a rotation frequency being proportional to @xmath0 .
the probability of finding this coherent state can be adjusted by choosing both an appropriate period of time evolution and magnitude of the coherent state . here
we find that this probability is not small . indeed
, the probability of finding the factors state can be increased by a factor which is the reciprocal of the probability of obtaining this coherent state .
the resulting states of the first two oscillators , after the first step , are used as new input states in the second step of our approach .
also , the state of the third oscillator is now prepared as a coherent state with the same , or higher , magnitude . by repeating the same procedure described in the first step , we can obtain the states of the factors with a much higher probability .
we then iterate these procedures @xmath1 times , until the probability of finding the factors state is close to one . as an example of how this method works ,
we show how to factorize the integer @xmath4 . here
the probabilities of obtaining coherent states , with rotation frequencies proportional to @xmath0 , are larger than 0.1 in each iteration .
the probability of finding the factors can reach nearly one after 12 iterations .
in addition , this amplitude - amplification method can be applied to search problems by suitably controlling nonlinear interactions between the harmonic oscillators and making appropriate conditional measurements .
this approach can search a `` target '' from @xmath0 possible inputs in _
linear time_. it surpasses grover s algorithm which only provides a quadratic speed - up for searching @xcite .
since np - complete problems such as 3-sat , the traveling salesman problem , etc @xcite , can be mapped into search problems @xcite .
this implies that _ np - complete problems can be exponentially sped up_. ( color online ) schematics of harmonic - oscillator quantum computation .
there are two groups of @xmath5 coupled harmonic oscillators : @xmath6 of them ( in blue ) in the left side , and @xmath7 of them ( in red ) in the right side .
this system can be `` programmed '' to find solutions of a system of functions @xmath8 in eq .
( [ nonlinear_equation ] ) , by appropriately controlling nonlinear interactions between the @xmath6 oscillators .
initially , all trial solutions are prepared for the collective state of the @xmath6 oscillators .
all possible answers of the functions @xmath8 are simultaneously computed and then are `` written '' to the rotation frequencies of the coherent states of the @xmath7 coupled oscillators . by repeating the procedures of time evolution and conditional measurement on the @xmath7 oscillators , the states of the solutions of @xmath8 can be obtained with a high probability . finally , the solutions can be obtained by measuring the resulting state of the @xmath6 oscillators .
, height=340 ] we also generalize this method of amplitude amplification by using a system of ( more than three ) coupled harmonic oscillators . in fig .
[ hoqcomputer ] , the harmonic - oscillator quantum computation is schematically depicted .
this can be used for solving a system of linear or nonlinear functions with @xmath6 integer unknowns which are subject to @xmath7 constraints .
this is very useful for integer programming @xcite which is an important tool in operational research @xcite . to obtain the solutions of the functions
, we require to perform conditional measurements of coherent states of the @xmath7 oscillators which are used for checking the @xmath7 constraints .
this paper is organized as follows : in sec .
ii , we introduce a system of coupled harmonic oscillators . in sec .
iii , we study the quantum dynamics of the coupled harmonic oscillators starting with a product state of number states and a coherent state . in sec .
iv , we propose an amplitude - amplification method to factorize an integer using three coupled harmonic oscillators .
we discuss the convergence and performance of this factoring algorithm .
for example , we show how to factor the number @xmath9 using this approach . in sec .
v , we study the application of this amplitude - amplification method to quantum searching . in sec .
vi , we generalize this amplitude - amplification method by using a system of coupled harmonic oscillators .
we apply this method to solving linear or nonlinear functions with integer unknowns .
appendix a shows details of this generalized method .
we close this paper with a discussion and conclusions .
we consider a system of @xmath5 coupled harmonic oscillators .
the hamiltonian of the @xmath10-th harmonic oscillator is written as @xmath11 where @xmath12 is the frequency of the harmonic oscillator , @xmath13 is the mass of the particle , and @xmath14 . the operators @xmath15 and @xmath16 are the position and momentum operators , which satisfy the commutation relation , @xmath17=i\hbar$ ] .
the annihilation and creation operators of the @xmath10-th harmonic oscillator are defined as @xmath18 the operators @xmath19 and @xmath20 obey the commutator @xmath21=1 $ ] .
the hamiltonian of the three harmonic oscillators can be expressed in terms of the annihilation and creation operators : @xmath22 here we have ignored the constant term .
we consider the harmonic oscillators coupled to each other via nonlinear interactions @xcite .
such nonlinear interactions can be described by the hamiltonian @xmath23 as @xmath24 where @xmath25 are linear or nonlinear - operator functions ( excluding divisions ) of the number operators @xmath26 , for @xmath27 , and @xmath28 . the total hamiltonian @xmath29 can be written as @xmath30 the hamiltonians @xmath31 and @xmath23 commute with each other , i.e , @xmath32}=0 $ ] . the total hamiltonian @xmath33 in eq . ( [ hamiltonian ] ) is exactly solvable
. the eigenstate @xmath34 of the hamiltonian @xmath33 is a product state of number states @xmath35 of the harmonic oscillators , i.e. , @xmath36 @xmath37
( color online ) schematic diagram of the time evolution in phase space of the coherent state of the harmonic oscillator @xmath38 , corresponding to the product of number states @xmath39 , where @xmath40 .
the coherent state with complex amplitude @xmath41 is here depicted as a light blue circle in phase space ( @xmath42,@xmath43 ) , where @xmath42 and @xmath43 represent the position and the momentum of the harmonic oscillator @xmath38 . after a time @xmath44 ,
the coherent state rotates about the origin with an angle @xmath45 .
, height=264 ] we study the time evolution of the @xmath38-th harmonic oscillator in phase space starting with a state @xmath46 , where @xmath39 is the product state of the number states of the harmonic oscillators , and @xmath47 is the coherent state of the @xmath38-th harmonic oscillator , for @xmath27 , and @xmath38 is a number from @xmath48 to @xmath49 . by applying the time - evolution operator @xmath50
[ @xmath33 is the hamiltonian in eq .
( [ hamiltonian ] ) ] to the initial state @xmath51 , it becomes @xmath52\prod_{j}|m_j\rangle_{j}\nonumber\\ & & \times\exp{\big\{-i[\omega_k+f_k(m_1,\ldots , m_{a})]a^\dag_ka_k{t}\big\}}|\alpha\rangle_k,\nonumber\\ \label{u_inputstate } & = & \exp\bigg[-i\bigg(\sum_j\omega_j{m}_j\bigg)t\bigg]\prod_{j}|m_j\rangle_j|\alpha_{{m_1,\ldots , m_a}}(t)\rangle_k,\end{aligned}\ ] ] where @xmath53 is a complex function , i.e. , @xmath54\alpha,\\ \alpha_{m_1,\ldots , m_a}(t)&=&\exp[-i\omega_{m_1,\ldots , m_{a}}t]\alpha,\end{aligned}\ ] ] and @xmath55 is an effective rotation frequency of the coherent state in phase space @xmath56 in eq .
( [ u_inputstate ] ) , we have used the relation @xcite , @xmath57 where @xmath58 is a phase factor .
note that the product state of the number states @xmath39 is an invariant ; namely it does not change with time .
nonlinear interactions , described by the hamiltonian @xmath23 in eq .
( [ nonlinearfunction ] ) , cause the coherent state of oscillator @xmath38 to rotate about the origin with a frequency @xmath55 in phase space . from eq .
( [ omega_f_k ] ) , the frequency @xmath55 depends on the number states @xmath39 of the @xmath6 oscillators . a schematic diagram of the time evolution of the coherent state of the harmonic oscillator @xmath38 in phase space is shown in fig .
[ phase_diag ] .
extending our proposal in ref .
@xcite , we now present an amplitude - amplification method to factor any positive integer @xmath0 . for example , let us consider three coupled harmonic oscillators for factorization by setting @xmath59 and @xmath60 . using this method , the probability of finding the factors state can reach nearly one after @xmath1 iterations . simplified schematic diagrams of the factorization approach
are shown in figs .
[ factqcircuit ] and [ qcir_iteration ] .
( color online ) quantum factoring algorithm using three coupled harmonic oscillators .
initially , the state @xmath61 of the harmonic oscillators 1 and 2 , is prepared in a number - state basis . by repeatedly applying @xmath1 times the iterations @xmath62 ( as shown in fig .
[ qcir_iteration ] ) , the resulting state of the harmonic oscillators 1 and 2 becomes the state of the factors of @xmath0 .
finally , the factors of the number @xmath0 can be obtained by measuring the state of the oscillators 1 and 2 .
, height=151 ] ( color online ) the @xmath63-th iteration , @xmath62 , in quantum factorization . in each iteration
, the third harmonic oscillator is prepared in a coherent state @xmath64 with a magnitude @xmath65 .
then , a unitary operator @xmath66 is applied . by performing a conditional measurement of the coherent state @xmath67
, the reduced density matrix @xmath68 can be obtained .
, height=170 ] now we consider the total hamiltonian @xmath69 of the system @xmath70 where the nonlinear - operator function @xmath71 for computing the product of any two trial factors is written as , @xmath72 here the parameters @xmath73 are the coupling strengths of the nonlinear interactions @xmath74 , for @xmath75 .
this operator function @xmath76 will output an eigenvalue which is a power series of the product @xmath77 , for the product state @xmath78 .
the total state of the first two harmonic oscillators , in the number state basis , is initially prepared , i.e. , @xmath79 where the @xmath80 are the probabilities of the states of the first and second oscillators , while @xmath81 .
the states of the harmonic oscillators 1 and 2 can be prepared in arbitrary states , including pure states or mixed states @xcite .
the states of oscillators 1 and 2 encode all trial factors of the number @xmath0 .
each number state @xmath82 represents each trial factor @xmath83 , for @xmath84 . in the first iteration
, the state of the third harmonic oscillator is prepared in a coherent state @xmath86 . by applying the time - evolution operator @xmath87 to the initial state @xmath88
, it becomes @xmath89 where @xmath90t_1\}p^{n'm'}_{nm}.\ ] ] here @xmath91 , for @xmath92 , in eq .
( [ rho1(t_1 ) ] ) is a complex function , @xmath93 and @xmath94 is an effective rotation frequency of the coherent state in phase space @xmath95 note that the product of the two factors of @xmath96 and @xmath97 is equal to @xmath0 : @xmath98 .
the rotation frequency @xmath99 of the coherent state is @xmath100 if the product of any two numbers is not equal to @xmath0 , then the frequencies @xmath94 are different to the frequency @xmath99 .
thus , the state of the harmonic oscillator 3 can act as a `` marker '' for the states of factors and non - factors @xcite .
recall that a state is conditioned when this state is conditional on the measurement of a certain state @xcite .
after the measurement of @xmath113 , the density matrix of the conditioned state can be written as @xcite : @xmath114},\\ \label{rho^1_c } & = & \frac{1}{c_1}\sum_{n , m , n',m'}\!\!\tilde{p}^{n'm'}_{nm}\!(t_1)\;{\epsilon}^{(1)}_{nm}\;{\epsilon}^{(1)*}_{n'm'}|n , m\rangle\langle{n',m'}|\otimes|\alpha^{(1)}_n(t_1)\rangle_3\;{}_3{\langle}\alpha^{(1)}_n(t_1)|,\end{aligned}\ ] ] where @xmath115 is a normalization constant @xmath116 note that the trace of this density matrix @xmath117 $ ] is equal to one .
after the first iteration , the probability of finding the factors of the number @xmath0 is increased by a factor @xmath118 , which is the inverse of the probability @xmath111 as seen from eqs .
( [ pr_e_1 ] ) and ( [ rho^1_c ] ) . the probability amplification [ see also eqs .
( [ varrho1 ] ) and ( [ varrho2 ] ) ] is thus inversely proportional to the probability of obtaining the coherent state @xmath111 .
after the first iteration , we now obtain the reduced density matrix @xmath120 of the first two harmonic oscillators as @xmath121 we consider the state @xmath120 in eq .
( [ rho2 ] ) of the oscillators 1 and 2 as an input state for the second iteration .
the coherent state of the third harmonic oscillator is prepared in a coherent state @xmath122 , with a magnitude @xmath123 .
the nonlinear interactions between the three harmonic oscillators are then turned on for a time @xmath124 .
the state evolves as @xmath125 where @xmath126 .
next , a conditional measurement @xmath127 is applied to the system at the time @xmath124 .
the probability of obtaining the coherent state @xmath128 becomes @xmath129,\\ & = & \frac{1}{c_1}\sum_{n , m}{p}^{nm}_{nm}\;|{\epsilon}^{(1)}_{nm}|^2\;|{\epsilon}^{(2)}_{nm}|^2.\end{aligned}\ ] ] the conditioned state can be written as @xmath130},\\ & = & \frac{1}{c_2}\sum_{n , m , n',m'}\tilde{p}^{n'm'}_{nm}(\tilde{t}_2){\epsilon}^{(1)}_{nm}{\epsilon}^{(1)*}_{n'm'}{\epsilon}^{(2)}_{nm}{\epsilon}^{(2)*}_{n'm ' } where @xmath131 is the normalization constant , @xmath132 the probability of finding the factors is enhanced by a factor @xmath133 after the second iteration .
the coefficients @xmath134 is less than one for any product @xmath135 . from eqs .
( [ c_1 ] ) and ( [ c_2 ] ) , it can be seen that @xmath136 .
therefore , the probability of finding the factors is now higher after one additional iteration .
similarly , we now iterate the procedure @xmath138 times . after @xmath138 iterations ,
the reduced density matrix of the oscillators 1 and 2 can be written as @xmath139 where @xmath140 the state of the third harmonic oscillator is now prepared in a coherent state @xmath141 , with a magnitude @xmath142 .
let the three coupled harmonic oscillators evolve for a time @xmath143 .
this gives @xmath144 by performing a conditional measurement @xmath145 , the state becomes @xmath146},\\ & = & \frac{1}{c_{l}}\sum_{n , m , n',m'}\tilde{p}^{n'm'}_{nm}(\tilde{t}_l)\prod^{l}_{l=1}{\epsilon}^{(l)}_{nm}{\epsilon}^{(l)*}_{n'm'}|n , m\rangle\langle{n',m'}|\otimes|\alpha^{(l)}_n(t_l)\rangle_3\;{}_3\langle\alpha^{(l)}_n(t_l)|.\end{aligned}\ ] ] after the @xmath1-th step , the probability of finding the factors is increased by a factor @xmath147 . from eq .
( [ c_l-1 ] ) , the probability of obtaining the coherent state @xmath145 can be written as @xmath148 the entire iterative procedure is now completed . the convergence and performance of this method
will be discussed in the following subsections .
we now study the convergence of this iterative method .
we first consider the magnitude of the coherent state @xmath65 for each iteration as @xmath149 thus , we have @xmath150 for any product of @xmath83 and @xmath151 being not equal to @xmath0 , the coefficient @xmath152 is less than one and decreasing for higher @xmath63 , and the evolution time @xmath153 is non - zero and appropriately chosen .
when the number of iterations @xmath1 tends to infinity , the product of the coefficients @xmath152 tends to zero , @xmath154 the coefficients @xmath155 are equal to one for any product of two factors @xmath96 and @xmath97 being equal to @xmath0 , i.e. , when @xmath98 .
now we consider the probability of finding a pair of factors , @xmath96 and @xmath97 , after the @xmath63-th iteration , which is @xmath156 since the coefficient @xmath157 is less than one , the normalization constant @xmath158 in eq .
( [ c_l-1 ] ) is decreasing , i.e. , @xmath159 .
therefore , the probability of finding a pair of factors increases after an additional iteration .
this shows that this iterative method is convergent . from eqs .
( [ c_l-1 ] ) and ( [ lime ] ) , it is very easy to show that @xmath160 in the limit of large number of iterations @xmath1 , we can obtain the state @xmath161 of the factors @xmath162 where @xmath163 and @xmath164 , with @xmath165 are factors of @xmath0 ( @xmath166 ) .
this shows that the state of the factors can be achieved by employing this iterative method , if a sufficiently large number of iterations is used .
we can now estimate the number @xmath1 of the iterations required to achieve a probability of order of one for factoring @xmath0 .
we investigate the amplification ratio @xmath167 of the two probabilities of finding the factors @xmath168 and @xmath169 after the @xmath170-th and the @xmath63-th iterations . from eq .
( [ pr_l ] ) , we have @xmath171 note that this ratio @xmath167 is just the reciprocal @xmath172 in eq .
( [ pr_e_l ] ) of the probability of obtaining the coherent state .
practically , this probability @xmath173 can not be too small .
for example , let the amplification ratio @xmath167 be roughly equal to @xmath174 for each iteration , and let the probability of the factors state @xmath175 before the iterations be of order @xmath176 , where @xmath177 is a positive number .
after @xmath1 iterations , the probability of finding the states of the factors can be increased by a factor @xmath178 . here
we require @xmath179 therefore , we obtain that the number @xmath1 of necessary iterations is of order of @xmath180 .
this shows that this approach can factor an integer @xmath0 with a high probability and this in the linear time @xmath180 . in the limit of large @xmath1 , the probabilities @xmath173 in eq .
( [ pr_e_l ] ) tend to one because @xmath181 approaches the sum of probabilities of the factors states in eq .
( [ limit_c ] ) .
the probability of finding the factors will be slowly increased after the number @xmath182 of iterations is reached . in this section ,
we study how to factorize an integer @xmath0 with an initial pure state using this factoring algorithm .
we consider the initial state of the first two harmonic oscillators as the superposition of number states , i.e. , @xmath183 are two normalization constants . here
we consider trial factors from 3 , @xmath184 , @xmath185 .
the probability of finding the product of two factors is of order of @xmath186 .
for example , now we show how to factor the integer @xmath4 . for simplicity
, we now take @xmath187 , which is the lowest order of nonlinearity .
the hamiltonian can be written as @xmath188 where @xmath189 is the nonlinear strength .
the stronger nonlinear strengths and high - order nonlinearity can significantly shorten the required time evolution of the system @xcite .
but the role of nonlinearity is not directly relevant to the number @xmath1 of the required iterations for the amplitude amplification .
.[table1 ] this table shows the fidelities @xmath190 , the probabilities pr@xmath191 of obtaining the coherent states @xmath192 in eq .
( [ coherent_state ] ) , and the evolution time @xmath193 for the @xmath63 iterations . here
@xmath194 is the density matrix of the coherent state @xmath192 in eq .
( [ coherent_state ] ) and @xmath193 is measured in units of @xmath195 . [
cols="^,^,^,^,^,^,^,^ " , ] we take @xmath193 as the evolution time for the @xmath63-th iteration , @xmath196 where @xmath197 is a uniformly distributed random number on the interval @xmath198 $ ] .
we now evaluate the performance of this method by investigating the fidelity @xmath190 between the reduced density matrix @xmath68 and the factor s states @xmath161 as @xcite @xmath199\big\}^2.\end{aligned}\ ] ] ( color online ) bar chart : the average of the fidelities between the reduced density matrix @xmath200 and the factors states @xmath161 is plotted versus the number @xmath63 of iterations . in each iteration , the fidelity @xmath190 is taken at the time @xmath193 , which is a uniformly distributed random number ranging from 0 to @xmath201 . here
the sample size is equal to 100 fidelities .
the bars , which indicate the mean values , are shown in light blue .
the error bars , which indicate the standard deviations , are shown in black.,height=226 ] table [ table1 ] shows the relevant fidelity , the probability of obtaining the coherent state @xmath192 with a magnitude @xmath202 , and the evolution time @xmath193 for each iteration . before starting the iteration , the initial fidelity @xmath203 is very low : @xmath204 . in the first few steps , the probabilities , pr@xmath191 , of obtaining the coherent states , are about @xmath205 .
we emphasize that now the probabilities pr@xmath191 are not extremely small , even when @xmath0 is large .
this resolves the limitation of our previous proposal @xcite .
the probability of finding the factor s state can be increased by 10 after a single iteration .
after ten iterations , the fidelity can exceed 0.5 .
the fidelity can reach nearly one after two more iterations , while the probabilities @xmath206 approach one .
also , we have explicitly shown that this method can factor a number of order of @xmath207 with 12 iterations . since the evolution time @xmath193 in eq .
( [ tr ] ) is a uniformly distributed random number from 0 to @xmath201 , it is necessary to study the effect of this randomly chosen time @xmath193 to the performance of factoring .
now we study the averages of the fidelities @xmath190 and probabilities pr@xmath191 of obtaining the coherent states .
we also examine their standard deviations in each iteration .
( color online ) bar chart : the average of the probabilities of obtaining the coherent states @xmath192 is plotted versus the number @xmath63 of iterations . in each iteration , the probability is taken at the time @xmath193 which is a uniformly distributed random number from 0 to @xmath201 .
here the sample size is taken to be 100 probabilities .
the bars , which indicate the mean values , are in red .
the error bars , which indicate the standard deviations , are shown in black .
, height=226 ] in fig .
[ figfid ] , we plot the average of the fidelities @xmath190 versus the number @xmath63 of iterations . here
we take the sample size to be equal to 100 fidelities in each iteration . as shown in fig .
[ figfid ] , the average of the fidelities is about 0.5 at the 10-th iteration .
after 12 iterations , the mean values of the fidelities reach nearly one .
the error bars , which indicate the standard deviations , are shown in the same figure .
we can see that , after the 12-@xmath208 iteration , the standard deviations are relatively small compared to the mean values in each iteration . in fig .
[ figpb ] , the average of the probabilities pr@xmath191 of obtaining the coherent states is plotted versus the number @xmath63 of iterations . in the first ten iterations , the probabilities pr@xmath191 are about 0.1 .
then , it increases and saturates around one after 13 iterations .
moreover , these standard deviations are much smaller than the mean values of pr@xmath191 .
this means that the statistical effect of the time @xmath193 is small to the performance of quantum factorization .
we have proposed a new amplitude - amplification method using three coupled harmonic oscillators to obtain , with a high probability , the factors of a number .
we have shown that this method can factor a number in linear time .
remarkably , this amplitude - amplification method can be applied to quantum searching and related problems .
we will show that this approach searches an object from @xmath0 possible inputs in linear time . since np - complete problems ( such as 3-sat and the traveling salesman problem ) can be mapped into search problems @xcite , this implies that np - complete problems can also be exponentially sped up .
( color online ) schematics diagram of the quantum search algorithm using three coupled harmonic oscillators .
initially , the state @xmath209 in eq .
( [ search_state ] ) of the harmonic oscillators 1 and 2 is prepared in a number - state basis .
the `` black box '' @xmath210 is then applied to the initial state @xmath211 .
we start the amplitude - amplification procedure by repeatedly applying @xmath1 times the iterations @xmath212 ( as shown in fig .
[ qcir_iteration_search ] ) .
the resulting states of the harmonic oscillators 1 and 2 are the solutions marked with even numbers @xmath213 . finally , all solutions can be obtained by measuring the state of the oscillators 1 . in constrast to this , for quantum factorization , the factors can be obtained by measuring the states of the oscillators 1 and 2 . , height=151 ] ( color online ) the @xmath63-th iteration , @xmath212 , in the quantum search algorithm . in each iteration , the third harmonic oscillator is prepared in a coherent state @xmath64 with a magnitude @xmath65 .
then , a unitary operator @xmath214 is applied . by performing a conditional measurement of the coherent state @xmath64
, the reduced density matrix @xmath215 can be obtained .
, height=170 ] now we briefly discuss how to apply this method to search problems by using three coupled harmonic oscillators . here
we introduce a `` black - box '' operation @xmath210 , @xmath216 where @xmath217 is an output function which depends on @xmath83 .
the output function @xmath217 is an even number if @xmath83 is a solution to the search problem , and an odd number if @xmath83 is not a solution to the search problem .
in fact , this `` black - box '' operation @xmath210 is similar to the role of the `` black - box '' function in grover s algorithm @xcite which can provide the `` yes '' and `` no '' solutions .
now , without qubits , we are considering the parity of a number ( even or odd ) instead of `` 1 '' and `` 0 '' to code the `` yes '' and `` no '' binary - decision problem . without loss of generality ,
the states of the harmonic oscillators 1 and 2 are initially prepared as @xmath218 the `` black - box '' operation @xmath210 is applied to the initial state @xmath219 .
the state becomes @xmath220 we aim to search the states `` marked '' with even @xmath217 from ( @xmath221 ) possible input states @xmath222 . for search problems ,
we consider the hamiltonian @xmath223 of the coupled harmonic oscillators as @xmath224 where the operator function @xmath225 is @xmath226 and @xmath227 is the nonlinear strength between the oscillators 2 and 3 .
we now discuss how to use this amplitude - amplification method to increase the probability of finding the states marked with even @xmath217 . in each iteration
, the harmonic oscillator 3 is prepared in a coherent state @xmath64 .
we apply the time - evolution operator @xmath228 to the initial state @xmath229 , and the state becomes @xmath230t\big\}|n , h(n)\rangle|\alpha^{(l)}_{h(n)}(t)\rangle_3,\end{aligned}\ ] ] where @xmath231\alpha^{(l)}$ ] , and @xmath232 is the effective rotation frequency of the coherent state in phase space , @xmath233 we now set the frequency of @xmath234 as an even multiple of @xmath227 . if @xmath217 is even , then the effective frequency @xmath232 is even .
otherwise , the effective frequency @xmath232 is odd .
let the system evolve for a time @xmath235 as @xmath236 for an even @xmath217 , the state of the third harmonic oscillator evolves as @xmath237 and , for an odd @xmath217 , @xmath238 the coherent states of the third harmonic oscillator are either @xmath64 or @xmath239 , which depend on the parity of @xmath217 .
a conditional measurement of the coherent state @xmath64 is made in each iteration . by repeating the procedures of the time evolution @xmath214 and the conditional measurement of the coherent state @xmath240 , the solutions [ i.e. , @xmath217 is an even number ] can be obtained , with a high probability , in linear time @xmath241 .
here we have assumed that the `` black - box '' operation can be completed in polynomial time in the worst case .
this shows that search problems can be exponentially sped up .
we schematically summarize this method for quantum searching in figs .
[ searchqcircuit ] and [ qcir_iteration_search ] .
( a ) given a system of functions @xmath8 , it is easy to evaluate them by just substituting several integer variables @xmath242 as inputs of @xmath8 .
( b ) however , it is extremely hard to do the `` opposite '' : to find out the integer variables , @xmath242 , from a known output of the functions @xmath8 .
, height=264 ] we generalize the amplitude - amplification method using more than three coupled harmonic oscillators .
this enables us to solve general linear or nonlinear problems which require to find out the non - negative integer variables @xmath243 subject to @xmath7 constraints such that @xmath244 where @xmath245 is a system of linear or nonlinear functions ( without any division of integer variables ) , while @xmath246 are real numbers for @xmath247 .
here we assume that the solutions exist in this system of functions in eq .
( [ nonlinear_equation ] ) , and each input integer variable @xmath248 is bounded by a number @xmath249 .
figure [ amproblem2 ] schematically illustrates this problem .
it is very easy to obtain the output of the functions @xmath8 by directly substituting the integer variables @xmath250 into @xmath8 .
however , it is extremely hard to do the opposite , namely , to find out the inverse solutions from the `` output '' of the functions @xmath8 .
a simple example with two arbitrary positive integers : easy to multiply , but hard to factorize .
now we provide a way to solve this problem by using a system of @xmath5 coupled harmonic oscillators .
we now consider a hamiltonian of the same form as in eq .
( [ hamiltonian ] ) , @xmath251 where @xmath25 are the linear or nonlinear - operator functions of the number operators @xmath252 in eq .
( [ nonlinear_equation ] ) ( by setting integer variables @xmath13 to the number operators @xmath252 ) for @xmath40 and @xmath253 . without loss of generality ,
the total state of a system of @xmath6 coupled harmonic oscillators is first prepared as a superposition of products of number states @xmath254 .
each of these product states , @xmath254 , represents a trial solution @xmath242 in eq .
( [ nonlinear_equation ] ) .
while a system of @xmath7 coupled harmonic oscillators is prepared in the product state of coherent states , @xmath255 we apply the time - evolution operator @xmath256 to the product state of @xmath254 and also to @xmath257 . from eq .
( [ u_inputstate ] ) , we obtain @xmath258 where the amplitude @xmath259 is @xmath260 and @xmath261 is the rotation frequency , @xmath262 by quantum parallelism , all possible functions @xmath263 can be simultaneously computed .
the value of @xmath263 has been `` written '' to the rotation frequency of the coherent state of the oscillator @xmath38 in the system of @xmath7 harmonic oscillators . by performing the conditional measurements of the coherent states of the @xmath7 harmonic oscillators depending on @xmath7 constraints ,
we can obtain the solutions @xmath264 in eq .
( [ nonlinear_equation ] ) , where @xmath265 is the number of solutions . however , the probability of obtaining these solutions is low for a large number of trial solutions .
we generalize this amplitude - amplification method to the case of using multi - harmonic - oscillator .
this can be used to efficiently increase the probability of obtaining the solutions .
in fact , this generalized method is very similar to the method we described in sec .
appendix a presents details of this amplitude - amplification method for solving general linear or nonlinear problems .
we can show that the number of iterations is of order of @xmath266 , where @xmath267 is the number of trial solutions for each nonlinear function @xmath268 in eq .
( [ nonlinear_ham ] ) and @xmath253 .
let us now provide more details on this amplitude - amplification method to solve a system of linear or nonlinear functions as in eq .
( [ nonlinear_equation ] ) by using a system of coupled harmonic oscillators .
this generalization is very similar to the method discussed in sec .
iv . for completeness , we show all the mathematical details in the following subsections
. the total state of a system of @xmath6 harmonic oscillators , in the number - state basis , is initially prepared , i.e. , @xmath273 where the @xmath274 are the probabilities of the states @xmath275 , while @xmath276 , @xmath277 and @xmath40 .
the states @xmath278 can be prepared in arbitrary states , including pure states or mixed states .
our purpose is to find the target states @xmath279 from the ensemble @xmath280 such that @xmath281 where @xmath265 is the number of solutions and @xmath282 are the functions in eq .
( [ nonlinear_ham ] ) .
initially , the probability of obtaining the target states @xmath283 is very low . using this method
, this probability can be efficiently increased after a sufficiently large number of iterations .
the amplitude - amplification method is discussed in the following subsections . in the first iteration ,
the states of a system of @xmath7 oscillators are prepared in the product state of coherent states @xmath285 , for @xmath28 . by applying the time - evolution operator @xmath286
[ @xmath33 is the hamiltonian in eq .
( [ hamiltonian ] ) ] to the initial state @xmath287 , it becomes @xmath288 where @xmath289p^{m_1,\ldots , m_a}_{m_1',\ldots , m_a'},\\ we consider the coherent states @xmath292 , corresponding to the target states @xmath283 , which have their rotation frequencies @xmath293 subject to the constraints in eq .
( [ nonlinear_equation ] ) , namely @xmath294 where @xmath63 is the number of iterations , and @xmath295 .
this means that the coherent states , corresponding to the non - target states have rotation frequencies which are larger than @xmath296 .
the coherent states of the system of @xmath7 harmonic oscillators can act as `` markers '' for the target states @xmath283 and the other states . a measurement operator @xmath297 for the conditional measurements for the @xmath38 oscillators is defined as @xcite @xmath298 where @xmath299 and @xmath63 is the number of iterations .
a conditional measurement @xmath300 of the coherent states with the rotation frequencies @xmath301 is performed on the system .
the probability of obtaining the product of coherent states @xmath302 becomes @xmath303,\\ \label{apppr_e_1 } & = & \sum_{m_1,\ldots , m_a}p^{m_1,\ldots , m_a}_{m_1,\ldots , m_a}\;\prod_{k , x}|{\epsilon}^{(k1)}_{x , m_1,\ldots , m_a}|^2,\end{aligned}\ ] ] where the coefficient @xmath304 , for @xmath92 , @xmath305\},~~~~{\rm for}~~l\geqslant{1}.\end{aligned}\ ] ] the value of the probability @xmath306 can be adjusted by appropriately choosing the evolution time @xmath105 and the magnitude @xmath307 . the density matrix of the conditioned state can be written as : @xmath308},\\ \label{apprho1_c } & = & \frac{1}{\mathcal{c}_1}\sum_{\substack{m_1,\ldots , m_a\\ m_1',\ldots , m_a'}}\tilde{p}^{m_1,\ldots , m_a}_{m_1',\ldots , m_a'}(t_1)\:\lambda^{m_1,\ldots , m_a}_{m_1',\ldots , m_a'}(k,1)\prod_{j}|m_j\rangle_j\:{}_{j'}\langle{m_j'}|\:\otimes\prod_{k , x}|\alpha^{(k1)}_{x}(t_1)\rangle_k\;{}_k{\langle}\alpha^{(k1)}_{x}(t_1)|,\nonumber\\\end{aligned}\ ] ] where @xmath309 and @xmath310 are @xmath311 from eqs .
( [ apppr_e_1 ] ) and ( [ apprho1_c ] ) , the probabilities of target states are increased by a factor @xmath312 , which is the inverse of the probability @xmath313 .
after a single iteration , we now obtain the reduced density matrix @xmath315 of the first two harmonic oscillators as @xmath316 we consider this state @xmath315 of the @xmath6 oscillators as an input state for the second iteration .
the states of the @xmath7 harmonic oscillators are prepared in the product of coherent states @xmath317 , with magnitudes @xmath318 . after applying a unitary operator @xmath319 ,
the state becomes @xmath320 where @xmath126 . at the time @xmath124 , a conditional measurement @xmath321
is then applied to the system .
the probability of obtaining the product state of the coherent states @xmath322 becomes @xmath323,\\ & = & \frac{1}{\mathcal{c}_1}\sum_{m_1,\ldots , m_a}\tilde{p}^{m_1,\ldots , m_a}_{m_1,\ldots , m_a}\;\lambda^{m_1,\ldots , m_a}_{m_1,\ldots , m_a}(k,2).\end{aligned}\ ] ] the conditioned state can then be written as @xmath324},\\ & = & \frac{1}{\mathcal{c}_2}\sum_{\substack{m_1,\ldots , m_a\\ m_1',\ldots , m_a'}}\tilde{p}^{m_1,\ldots , m_a}_{m_1',\ldots , m_a'}(\tilde{t}_2)\:\lambda^{m_1,\ldots , m_a}_{m_1',\ldots , m_a'}(k,2)\prod_{j}|m_j\rangle_j\;{}_{j'}\langle{m_j'}|\otimes\prod_{k , x}|\alpha^{(k2)}_x(t_2)\rangle_k\;{}_k\langle\alpha^{(k2)}_x(t_2)|,\nonumber\\\end{aligned}\ ] ] where @xmath325 is the normalization constant @xmath326 the probability of finding the factors is enhanced by a factor @xmath327 after the second iteration . from eqs .
( [ appc_1 ] ) and ( [ appc_2 ] ) , the normalization constant @xmath325 is larger than the constant @xmath310 because the coefficient @xmath328 is less than one .
therefore , the probabilities of the target states increase after the second iteration .
after @xmath138 iterations , the reduced density matrix of the @xmath6 oscillators can be written as @xmath330 where @xmath331 the states of the @xmath7 harmonic oscillators are prepared in the product of the coherent states @xmath332 , for @xmath247 .
let the coupled harmonic oscillators evolve for a time @xmath143 .
the state evolves as @xmath333 by performing a conditional measurement @xmath334 , the state becomes @xmath335},\\ & = & \frac{1}{\mathcal{c}_{l}}\sum_{\substack{m_1,\ldots , m_a\\ m_1',\ldots , m_a'}}\tilde{p}^{m_1,\ldots , m_a}_{m_1',\ldots , m_a'}(\tilde{t}_{l})\:\lambda^{m_1,\ldots , m_a}_{m_1',\ldots , m_a'}(k , l)\prod_{j}|m_j\rangle_j\;{}_{j}\langle{m_j'}| \otimes\prod_{k , x}|\alpha^{(kl)}_x(t_l)\rangle_k\;{}_k\langle\alpha^{(kl)}_x(t_l)|.\nonumber\\\end{aligned}\ ] ] after the @xmath1-th iteration , the probability of finding the target states is increased by a factor @xmath336 . from eq .
( [ appc_l-1 ] ) , the probability of obtaining the product of coherent states @xmath337 is @xmath338 we now analyze the performance of this amplitude amplification procedure . after the @xmath63-th iteration , the probability of finding the target state @xmath339 , @xmath340 the amplification ratio @xmath341 of the two probabilities @xmath342 and @xmath343 is @xmath344 this ratio @xmath341 is the reciprocal @xmath345 in eq .
( [ apppr_e^k_l ] ) of the probability of obtaining the coherent states @xmath346 . the amplification ratio @xmath341 is taken to be roughly equal to @xmath347 for each iteration . here
we assume that the number of trial solutions is about @xmath348 for each integer variable @xmath13 .
thus , the probability of the target state @xmath339 before the iterations are of order of @xmath349 if the uniform distribution of the initial state in eq .
( [ appinitialden ] ) is assumed .
after @xmath1 iterations , the probability of the target state @xmath339 can be increased by a factor @xmath350 . here
we require @xmath351 therefore , we obtain that the number @xmath1 of necessary iterations is of order of @xmath352 . in each iteration , it is required at most @xmath353 steps to perform the conditional measurements for @xmath7 oscillators .
this approach can solve the functions in eq .
( [ nonlinear_equation ] ) in @xmath354 .
ideal quantum algorithms usually assume that quantum computing is performed continuously by a sequence of unitary transformations
. however , there always exist idle finite time intervals between consecutive operations in a realistic quantum computing process . during these delays
, coherent errors will accumulate from the dynamical phases of the superposed wave functions .
reference @xcite explores the sensitivity of shor s quantum factoring algorithm to such errors .
those results @xcite clearly show a severe sensitivity of shor s factorization algorithm to the presence of ( the experimentally unavoidable ) delay times between successive unitary transformations .
specifically , in the presence of these coherent errors , the probability of obtaining the correct answer decreases _ exponentially _ with the number of qubits of the work register .
a particularly simple phase - matching approach was proposed in @xcite to avoid or suppress these coherent errors when using shor s algorithm to factorize integers .
the robustness of that phase - matching condition was evaluated analytically and numerically for the factorization of several integers : 4 , 15 , 21 , and 33 . in spite of these , and many other efforts , shor s algorithm has been found to be problematic , in its implementation .
thus , it is imperative to explore very different approaches to factorize integers .
this is one of the goals of this work .
note that the method for factorization presented here is totally different from the integer factoring approach in ref .
@xcite , which is based on gauss sums @xcite .
the standard gauss - sum approach can find a factor of @xmath0 by scanning through all numbers from 2 to @xmath269 .
it therefore does not provide a speed - up over classical computation .
we also note the following very recent proposal @xcite to perform quantum computation using a spin chain .
this method is also related to recent methods for hamiltonian identification @xcite using networks of spin - half particles @xcite and even general many - body systems described by quadratic hamiltonians @xcite . in this approach , quantum computation @xcite and estimating coupling strengths in a hamiltonian @xcite can be done by just locally controlling a few quantum registers .
indeed , in our approach presented here and also in their methods @xcite , very few resources are required and only a few quantum registers are necessary to be operated for quantum information processing .
the approaches described above are quite different from the method presented in our work , where the solutions are directly `` written '' to the coherent states and then the probabilities of these solutions are increased through a number of iterations . moreover , the philosophies behind these two approaches are entirely different .
moreover , let us now we briefly discuss the difference between this harmonic - oscillator - based quantum computation and other existing models of quantum computation .
our proposal here obviously differs from the conventional quantum - circuit based model , which only uses quantum gates for quantum computation @xcite .
a very different model of quantum computation , relying on measurements of single qubits , is called one - way quantum computation @xcite . to perform a specific task using a one - way quantum computer , an appropriate sequence of measurments
is performed on qubits in the cluster , which is initially prepared in a highly - entangled state @xcite .
apart from using harmonic oscillators instead of qubits , there are two main differences between this harmonic - oscillator and one - way quantum computations .
first , in this harmonic - oscillator quantum computation , the specific nonlinear interactions between oscillators are prepared for tackling the specific nonlinear problem .
second , it is only required to perform conditional measurements on the same set of harmonic oscillators throughout the entire computation .
we believe that this idea of amplitude amplification may be generalized to quantum computation using discrete quantum states such as qubits .
let us briefly summarize this approach of using the harmonic - oscillator quantum computer described here . in general , this method can be used for solving linear or nonlinear functions with integer unknowns subject to a number of constraints .
a system of @xmath6 coupled harmonic oscillators are used for computing the functions , while another system of @xmath7 coupled oscillators act as `` markers '' for solutions and non - solutions of these functions .
initially , a system of @xmath6 coupled harmonic oscillators are prepared in the number - state basis for all possible solutions , and a system of @xmath7 coupled oscillators are prepared in a product state of coherent states . by quantum parallelism , the answers of all possible solutions can be simultaneously computed .
then , all the answers are `` written '' to the rotation frequencies of the coherent states of the @xmath7 oscillators .
we can obtain the solutions by performing a conditional measurement of coherent states subject to the constraints .
however , the probability of obtaining the solutions is low .
this can be resolved by iterating the procedures of time evolution and the conditional measurement .
the number of iterations is of order of @xmath270 , where @xmath271 is the input size of the trial solutions .
the solutions can be obtained with a high probability .
we have proposed a new algorithm for amplitude amplification using only a system of coupled harmonic oscillators .
we have shown that this approach can be used for factoring integers , and the factors of an integer can be obtained , with a high probability , in linear time . as an example , using this approach , we show how to factorize an integer of order of @xmath272 within 12 iterations . in each iteration ,
the probability of obtaining this coherent state , with the rotation frequency being proportional to @xmath0 , is not less than 0.1 .
this approach can also be applied to search - based problems . moreover ,
the solutions to search problems can be achieved in polynomial time .
this implies that np - complete problems can be exponentially sped up .
we stress out that the nonlinear interactions between the coupled oscillators and conditional measurements are essential in this approach . by appropriately controlling nonlinear interactions between the coupled harmonic oscillators ,
the functions with integer inputs can be evaluated in a single operation . to implement this harmonic - oscillator quantum computation ,
it is necessary to engineer `` many - body '' interactions of the system of harmonic oscillators .
for example , to perform quantum factorization , it is required to generate `` three - body '' interactions between the harmonic oscillators .
we have briefly discussed the possible implementations in ref .
one of the promising candidates is neutral atoms or polar molecules trapped in optical lattices @xcite .
the `` three - body '' interactions can be tuned by external fields @xcite .
bearing in mind that `` two - body '' interactions is only necessary for the implementation of quantum searching .
this type of nonlinear interactions has been widely studied various physical systems such as optical kerr mediums @xcite , two - component bose - einstein condensates of atomic gases @xcite , trapped ions within a cavity @xcite , cavity qed @xcite , superconducting resonator @xcite and optomechanical systems@xcite . the `` proof - of - principle '' experiment can be realized with current technology .
however , it is challenging to engineer `` many - body '' nonlinear interactions for solving multi - variables functions in sec .
we hope that future theoretical and experimental advances can overcome these issues .
fn acknowledges partial support from the laboratory of physical sciences , national security agency , army research office , defense advanced research projects agency , air force office of scientific research , national science foundation grant no .
0726909 , jsps - rfbr contract no .
09 - 02 - 92114 , grant - in - aid for scientific research ( s ) , mext kakenhi on quantum cybernetics , and the funding program for innovative r&d on science and technology ( first ) .
h. mack , m. bienert , f. haug , m. freyberger and w. p. schleich , phys .
status solidi b * 233 * , 408 ( 2002 ) ; m. mehring , k. mller , i. sh .
averbukh , w. merkel and w. p. schleich , phys .
rev . lett . * 98 * , 120502 ( 2007 ) . | using three coupled harmonic oscillators , we have recently proposed [ arxiv:1007.4338 ] an alternative approach for quantum factorization of an integer .
however , the probability of obtaining the factors becomes low when this number is large .
here , we circumvent this limitation by using a new iterative method which can efficiently increase the probability of finding the factors .
we show that the factors of a number can be obtained , in linear time , with a high probability .
this amplitude - amplification method can be applied to search problems .
this implies that search - based problems including np - complete problems can be exponentially sped up .
in addition , we generalize the method of amplitude amplification by using a system of coupled harmonic oscillators .
this can be used for solving a system of linear or nonlinear functions with integer inputs subject to a number of constraints .
the total number of oscillators in the system is equal to the number of integer variables in the functions plus the number of constraints . |
cosmology is now becoming a precision science , and cosmological observations can give us a lot of information for our understanding of the universe .
moreover , the interplay between cosmology and particle physics in various contexts has also been discussed vigorously .
one of such examples is the effective number of neutrino species @xmath0 . although collider experiments such as lep have measured the number of light active neutrino types to be @xmath8 @xcite , it is important to cross - check this value because cosmological measurements may lead to different value
. this could be due to an extra radiation component which is predicted by some models of particle physics such as sterile neutrinos ( see ref .
@xcite and references therein ) , or due to incomplete thermalization of neutrinos in the low - scale reheating universe in which the reheating temperature @xmath9 can be as low as @xmath10mev and @xmath0 is predicted to be less than three @xcite .
if such a non - standard ingredient exists , it can affect big bang nucleosynthesis ( bbn ) , cosmic microwave background ( cmb ) , large scale structure ( lss ) and so on ; thus precise cosmological observations can probe these scenarios through the effective number of neutrino species .
constraints on @xmath0 have been investigated in the literature using the information of cmb and lss , sometimes with priors on the hubble constant , cosmic age and helium abundance @xcite .
although cmb in general can constrain various quantities severely , since the effects of @xmath0 on cmb are degenerate with some cosmological parameters , the studies so far have combined cmb data with some other observations such as lss to obtain a sensible constraint on @xmath0 . however ,
when one uses the data from lss , constraints can become different depending on how one treats non - linear correction / bias on small scales for the matter power spectrum @xcite .
furthermore , different lss data seem to give different constraints on @xmath0 @xcite . regarding the prior on the hubble constant @xmath11 , as is summarized in ref .
@xcite , it can yield some constraints on @xmath0 when combined with cmb data ( without lss data ) @xcite , but they depend on the @xmath11 prior adopted .
one may consider that we can use the usually assumed prior on the hubble constant based on the result by freedman et al .
@xmath12 @xcite , but another group reported a somewhat lower value as @xmath13 @xcite .
if the lower value for @xmath11 is adopted as the prior , a resulting constraint on @xmath0 would be different .
having these considerations in mind , it is desirable to investigate a constraint on @xmath0 without these kind of uncertainties . in this paper
, we study a constraint on @xmath0 from cmb experiments alone . by making the analysis of cmb data alone
, we can avoid such subtleties as the galaxy - bias / non - linear corrections and the value for the prior on the hubble constant .
however , as is mentioned above , the effects of @xmath0 are strongly degenerate in cmb with other cosmological parameters such as energy density of matter , the hubble constant , and the scalar spectral index , and , in fact , we could not obtain a meaningful bound only with wmap3 @xcite .
recent wmap5 alone analysis gives a better constraint but it still can not give an upper bound @xcite . as we will discuss later , the degeneracy is significant up to about the 2nd/3rd peak of the cmb power spectrum where the observation of wmap has precisely measured . to break this degeneracy to some extent , it would be helpful to have the information at higher multipoles where signals unique to relativistic neutrinos are expected to appear @xcite .
recently , the data from acbar which probes cmb at higher multipoles than those of wmap has been updated @xcite . by using this data in addition to other small scale observations such as boomerang and cbi ,
we can obtain a relatively severe constraint on @xmath0 which is comparable to that have been obtained previously with lss data .
the organization of this paper is as follows . in the next section ,
we start with the discussion how @xmath0 affects the cmb power spectrum , which helps to understand our results for the constraint on @xmath0 . in section [ sec : current ] , we study the current constraint on @xmath0 using observations of cmb alone .
we use the data from wmap5 , the recent acbar , boomerang and cbi .
furthermore , we forecast the constraint from the future planck experiment . in the final section ,
we summarize our results and discuss its implications for some models of particle physics / the early universe .
the effective number of neutrino species @xmath0 represents the energy density stored in relativistic components as @xmath14 \rho_\gamma\ ] ] where @xmath15 , @xmath16 and @xmath17 are energy densities of photons , three species of massless active neutrinos and some possible extra radiation components , respectively . in this paper
, we assume that neutrinos are massless and have no chemical potential .
for the case with the standard three neutrino flavors without an extra relativistic component , the effective number of neutrino is @xmath18 where some corrections from the incomplete decoupling due to a slight interaction of neutrinos with electrons / positrons and finite temperature qed effect to the electromagnetic plasma are taken into account @xcite .
any deviation of @xmath0 from this value implies that there exists an extra relativistic component and/or some non - standard thermal history takes place such as the low reheating temperature scenario . to illustrate the effects of @xmath0 on cmb , we plot cmb power spectra for several values of @xmath0 in fig .
[ fig : cl ] . other cosmological parameters
are assumed as the mean values of wmap5 alone analysis for a power - law flat @xmath2cdm model . as seen from the figure , as we increase the value of @xmath0 ,
the height of the 1st peak is strongly enhanced and the positions of acoustic peaks are shifted to higher multipoles . also , the amplitude on small scales ( higher multipoles ) is suppressed .
below , we discuss where these changes are coming from .
one of the main effects of @xmath0 comes from the change of the epoch of the radiation - matter equality . by increasing ( decreasing ) the value of @xmath0
, the radiation - matter equality occurs later ( earlier ) .
thus the increase ( decrease ) of @xmath0 gives an almost the same effect of the decrease ( increase ) of energy density of matter .
one of noticeable features is that the height of the first acoustic peak is enhanced by increasing the value of @xmath0 .
this is due to the early integrated sachs - wolfe ( isw ) effect in which fluctuations of the corresponding scale , having crossed the sound horizon in the radiation - dominated epoch are boosted by the decay of the gravitational potential .
thus a larger amount of relativistic species drives the first peak higher .
another effect is the shift of the position of acoustic peaks due to the change of the radiation - matter equality through the change of @xmath0 .
the position of acoustic peaks is well captured by the so - called acoustic scale @xmath19 which is inversely proportional to the peak position and written as @xmath20 where @xmath21 and @xmath22 are the comoving angular diameter distance to the last scattering surface and the sound horizon at the recombination epoch @xmath23 , respectively .
although @xmath21 almost remains the same for different values of @xmath0 , @xmath22 becomes smaller when @xmath0 is increased .
thus the positions of acoustic peaks are shifted to higher multipoles ( smaller scales ) by increasing the value of @xmath0 .
furthermore , since the position of the @xmath24-th peak can be roughly written as @xmath25 , separations of the peaks become also greater for larger @xmath0 .
another important effect is free - streaming of ultrarelativistic neutrinos @xcite .
the perturbation of ultrarelativistic neutrino propagates with the speed of light , which is faster than the sound speed of acoustic oscillations of photon - baryon fluid .
the coupled photon - baryon component behaves to oscillate like a compressional fluid ; on the other hand , ultrarelativistic neutrinos free - stream to erase their fluctuations .
these two components are coupled via gravity ; thus the fluctuations of photons can also be affected by the free - streaming of neutrinos , which results in the damping of the amplitude and the shift of the acoustic oscillations .
the effects are significant on small scales where fluctuations of a given scale enter the horizon while the energy density of ultrarelativistic neutrinos takes a significant portion of that of the universe .
although the effects of the standard cosmological parameters on the heights and positions of acoustic peaks are well known , here we discuss them in some phenomenological way including the effects of @xmath0 .
for this purpose , we calculated the response of the heights and positions of the acoustic peaks to the change of the cosmological parameters up to the 5th peak around the fiducial values .
as a fiducial parameter set , we take those of the mean value from wmap5-alone analysis for a power - law flat @xmath2cdm model .
the shifts of the positions of acoustic peaks are @xmath26 where @xmath27 means the position of the @xmath28-th peak .
@xmath29 , @xmath30 , @xmath31 , @xmath32 and @xmath33 are energy densities of baryon and matter , the scalar spectral index , the primordial helium abundance and the normalized hubble constant .
@xmath31 is defined at the wave number @xmath34mpc@xmath35 .
the positive derivatives of the peak positions with respect to @xmath0 demonstrate the decrease in @xmath36 due to the increase in @xmath0 .
the responses of the heights of acoustic peaks to the change of various cosmological parameters are @xmath37 where @xmath38 .
since @xmath33 gives only a negligible change to the height of the peaks , we omit it . by increasing the value of @xmath0
, the height of the 1st peak is enhanced due to the early isw effect , as previously mentioned . for the third and the higher peaks ,
the heights are damped more by increasing @xmath0 , which is inferred from the negative coefficients .
this is due to the effect of free streaming of neutrinos @xcite .
in addition , for later convenience , we also show the derivatives of the peak heights relative to the first peak height following refs .
@xcite . here ,
@xmath39 for @xmath405 .
they are useful quantities when we interpret degeneracies since the dependence on the overall amplitude is cancelled out .
in this section , we present our result for the constraint on @xmath42 from cmb alone .
first , we give some details of our analysis .
we use the cmb data of wmap5 @xcite , boomerang @xcite , cbi @xcite and acbar @xcite .
we performed a markov chain monte calro ( mcmc ) analysis to obtain constraints on cosmological parameters using cosmomc code @xcite with some modifications which are described in the following .
we explore a 9 dimensional parameter space which consists of @xmath29 , @xmath43 , @xmath44 , @xmath45 , @xmath31 , @xmath46 , @xmath47 , @xmath32 and @xmath0 . here
, @xmath43 is the energy density of dark matter , @xmath44 is the optical depth of reionization , @xmath45 is the acoustic peak scale @xcite , @xmath46 is the amplitude of primordial fluctuations and @xmath47 is the amplitude of thermal sunyaev - zeldovich ( sz ) effect which is normalized to the @xmath48 template from ref .
@xcite .
.the mean values , 68% and 95% limits of @xmath0 for several current cmb data sets and assumptions of @xmath32 .
[ cols= " < , > , > , > " , ] we would like in the end to comment on how our discussion so far can be affected by theoretical uncertainties in the recombination process @xcite . since the change of @xmath32 can influence the recombination process
, its uncertainties might affect the cosmological parameter determination in some way .
thus it may be worth mentioning here on the effects .
for this purpose , we proceed with the same analysis as have been done in ref .
@xcite but varying @xmath0 here .
two parameters @xmath49 and @xmath50 , which represent the uncertainties in the recombination modeling , are included among other free parameters .
( see ref .
@xcite and references therein for more details ) .
we impose top - hat priors , @xmath51 and @xmath52 , which are very conservative ones , to take into account the uncertainties in the recombination theory .
we made the analyses for the two cases where @xmath32 is given from the bbn relation and @xmath32 is treated as a free parameter . in both cases , we found that the constraints on other cosmological parameters including @xmath0 are scarcely affected even by very conservative prior on @xmath49 and @xmath50 .
the mean values are unchanged and errors increase only very slightly ( no more than 10% for any parameters other than @xmath49 and @xmath50 ) .
therefore we can say that the uncertainties parametrized with @xmath49 and @xmath50 do not change much our results of the planck forecast discussed above .
however , we would need more understanding of uncertainties in the recombination theory for the precise determination of cosmological parameters in future cmb surveys .
we discussed the issue of probing the effective number of neutrino species @xmath0 from cmb data alone .
although a constraint on @xmath0 has been investigated by many authors , in most analyses , some combinations of data set such as cmb+lss , cmb+@xmath11 , cmb+lss+@xmath11 have been used to constrain @xmath0 .
this is partly because @xmath0 has severe degeneracies in wmap with some other cosmological parameters such as @xmath30 and @xmath33 ; thus @xmath0 can be more constrained by combining some data sets . however , when we combine data from lss , some subtleties can arise : a constraint from lss data depends on how we treat non - linear corrections / bias .
furthermore , different galaxy data lead to slightly different constraints on @xmath0 .
in addition , as for the hubble prior , the prior usually adopted is @xmath12 which is from the result of freedman et al @xcite .
however , another group has reported somewhat different value as @xmath53 @xcite . since different priors on the hubble
constant can give different results , in this respect , the constraint obtained by assuming some prior on @xmath11 should be regarded taking into account the above uncertainty .
taking these issues into consideration , it may be interesting to study a constraint on @xmath0 removing such subtleties , which can be done by using cmb data alone . in this paper , first we discussed the effects of @xmath0 on cmb and the issues of degeneracies with some other cosmological parameters .
phenomenological descriptions of its effects on the heights and the positions of acoustic peaks were also given .
then , in section [ sec : current ] , a constraint on @xmath0 was studied by using cmb data alone .
we made use of the data from wmap5 , acbar , boomerang and cbi .
as discussed there , although wmap measurement is very accurate , its precision is limited up to the 2nd peak/3rd peak .
we have explicitly shown that the information up to the 2nd/3rd peak is not enough to constrain @xmath0 severely .
this was demonstrated by making the analysis with wmap data alone , in which a sensible constraint can not be obtained .
however , if we include the data on small scales , the degeneracies of @xmath0 with some other cosmological parameters can be removed to some extent ; then a stronger constraint can be obtained .
in fact , on small scales , the amplitude is suppressed due to the free streaming effect by increasing @xmath0 , which is similar to the effects of @xmath32 through the diffusion damping .
thus we have studied the constraint on @xmath0 assuming different priors on @xmath32 : adopting the bbn relation to derive @xmath32 for given @xmath0 and @xmath29 , assuming @xmath32 as a free parameter and usual fixing of @xmath54 . depending on the prior , the constraint slightly changes .
we obtained the 95 % limits as @xmath55 for the case with @xmath32 being free , @xmath56 for @xmath32 being fixed as @xmath54 and @xmath57 when the bbn relation being adopted .
it should be noted that these constraints are comparable to that obtained using cmb+lss in previous works .
one of the main purposes of constraining the effective number of neutrino species using cosmological data is to check the standard value of @xmath0 independently from particle physics experiments .
thus we primarily focus on the analysis with no prior on @xmath0 .
however , from the viewpoint of constraining extra radiation which may be motivated from some particle physics models , a constraint obtained by assuming the prior @xmath58 may be interesting since an extra radiation always increases the value of @xmath0 . in this respect
, we also made an analysis adopting this prior and obtained the constraint on the effective number of neutrino as @xmath59 and @xmath60 for the cases where the bbn relation is adopted and @xmath32 is treated as a free parameter .
on the other hand , in a scenario with low - reheating temperature , the effective number of neutrino species can be reduced . in this case , another prior may be motivated to be assumed for a simple scenario of low - reheating temperature with three relativistic neutrino species . in this regard , we have also studied the case with the prior @xmath61 and obtained the constraints as @xmath62 for the cases with @xmath32 being assumed as a free parameter .
this can be translated into the lower bound on the reheating temperature as @xmath63 mev .
we have also discussed a future constraint on @xmath0 using the expected data from planck experiment .
it was shown that the attainable constraint on @xmath0 from planck is @xmath64 at 95% c.l . when the bbn relation is adopted for @xmath32 , which is most stringent compared to the other cases . since planck experiment can probe cmb down to smaller scales than wmap
, planck alone can give a stringent constraint on @xmath0 .
the interplay between particle physics and cosmology is now becoming more important in the era of precision cosmology .
one of such examples is the number of neutrino species , which was investigated in this paper . in light of upcoming more precise observations of cosmology , research of this kind
will bring us fruitful insight for particle physics and cosmology . *
acknowledgments : * this work is supported in part by the japan society for the promotion of science ( k.i . and t.s . ) , the sumitomo foundation ( t.t . ) , and the grant - in - aid for scientific research from the ministry of education , science , sports , and culture of japan , no.18840010 ( k.i . ) and no.19740145 ( t.t . ) .
w. m. yao _ et al .
_ [ particle data group ] , j. phys .
g * 33 * ( 2006 ) 1 .
a. d. dolgov , phys .
rept . * 370 * , 333 ( 2002 ) [ arxiv : hep - ph/0202122 ] .
m. kawasaki , k. kohri and n. sugiyama , phys .
* 82 * , 4168 ( 1999 ) [ arxiv : astro - ph/9811437 ] .
m. kawasaki , k. kohri and n. sugiyama , phys .
d * 62 * , 023506 ( 2000 ) [ arxiv : astro - ph/0002127 ]
. s. hannestad , phys .
d * 70 * , 043506 ( 2004 ) [ arxiv : astro - ph/0403291 ] .
k. ichikawa , m. kawasaki and f. takahashi , phys .
d * 72 * , 043522 ( 2005 ) [ arxiv : astro - ph/0505395 ] .
j. p. kneller , r. j. scherrer , g. steigman and t. p. walker , phys . rev .
d * 64 * , 123506 ( 2001 ) [ arxiv : astro - ph/0101386 ] .
s. hannestad , phys .
d * 64 * , 083002 ( 2001 ) [ arxiv : astro - ph/0105220 ] .
r. bowen , s. h. hansen , a. melchiorri , j. silk and r. trotta , mon . not .
soc . * 334 * , 760 ( 2002 ) [ arxiv : astro - ph/0110636 ] .
p. crotty , j. lesgourgues and s. pastor , phys .
d * 67 * , 123005 ( 2003 ) [ arxiv : astro - ph/0302337 ] .
e. pierpaoli , mon . not .
soc . * 342 * , l63 ( 2003 ) [ arxiv : astro - ph/0302465 ] .
s. hannestad , jcap * 0305 * , 004 ( 2003 ) [ arxiv : astro - ph/0303076 ] . v. barger , j. p. kneller , h. s. lee , d. marfatia and g. steigman , phys .
b * 566 * , 8 ( 2003 ) [ arxiv : hep - ph/0305075 ] .
p. crotty , j. lesgourgues and s. pastor , phys .
d * 69 * , 123007 ( 2004 ) [ arxiv : hep - ph/0402049 ] .
s. hannestad , jcap * 0601 * , 001 ( 2006 ) [ arxiv : astro - ph/0510582 ] .
d. n. spergel _ et al . _
[ wmap collaboration ] , astrophys .
j. suppl .
* 170 * , 377 ( 2007 ) [ arxiv : astro - ph/0603449 ] .
u. seljak , a. slosar and p. mcdonald , jcap * 0610 * , 014 ( 2006 ) [ arxiv : astro - ph/0604335 ] . s. hannestad and g. g. raffelt , jcap * 0611 * , 016 ( 2006 ) [ arxiv : astro - ph/0607101 ] .
m. cirelli and a. strumia , jcap * 0612 * , 013 ( 2006 ) [ arxiv : astro - ph/0607086 ] . k. ichikawa , m. kawasaki and f. takahashi , jcap * 0705 * , 007 ( 2007 ) [ arxiv : astro - ph/0611784 ]
. g. mangano , a. melchiorri , o. mena , g. miele and a. slosar , jcap * 0703 * , 006 ( 2007 ) [ arxiv : astro - ph/0612150 ] .
a. friedland , k. m. zurek and s. bashinsky , arxiv:0704.3271 [ astro - ph ] .
j. hamann , s. hannestad , g. g. raffelt and y. y. y. wong , jcap * 0708 * , 021 ( 2007 ) [ arxiv:0705.0440 [ astro - ph ] ] .
k. ichikawa , arxiv:0706.3465 [ astro - ph ] .
f. de bernardis , a. melchiorri , l. verde and r. jimenez , jcap * 0803 * , 020 ( 2008 ) [ arxiv:0707.4170 [ astro - ph ] ] . l. a. popa and a. vasile , arxiv:0801.3928 [ astro - ph ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , arxiv:0803.0547 [ astro - ph ] .
j. dunkley _ et al . _
[ wmap collaboration ] , arxiv:0803.0586 [ astro - ph ] .
v. simha and g. steigman , jcap * 0806 * , 016 ( 2008 ) [ arxiv:0803.3465 [ astro - ph ] ]
. l. a. popa and a. vasile , jcap * 0806 * , 028 ( 2008 ) [ arxiv:0804.2971 [ astro - ph ] ] .
w. l. freedman _ et al .
_ [ hst collaboration ] , astrophys . j. * 553 * , 47 ( 2001 ) [ arxiv : astro - ph/0012376 ] .
a. sandage , g. a. tammann , a. saha , b. reindl , f. d. macchetto and n. panagia , astrophys . j. * 653 * , 843 ( 2006 ) [ arxiv : astro - ph/0603647 ] .
s. bashinsky and u. seljak , phys .
d * 69 * , 083002 ( 2004 ) [ arxiv : astro - ph/0310198 ] . c. l. reichardt _ et al .
_ , arxiv:0801.1491 [ astro - ph ]
. g. mangano , g. miele , s. pastor , t. pinto , o. pisanti and p. d. serpico , nucl .
b * 729 * , 221 ( 2005 ) [ arxiv : hep - ph/0506164 ] .
w. hu , m. fukugita , m. zaldarriaga and m. tegmark , astrophys .
j. * 549 * , 669 ( 2001 ) [ arxiv : astro - ph/0006436 ] .
g. hinshaw _ et al . _
[ wmap collaboration ] , arxiv:0803.0732 [ astro - ph ] .
r. s. hill _
[ wmap collaboration ] , arxiv:0803.0570 [ astro - ph ] .
m. r. nolta _ et al . _
[ wmap collaboration ] , arxiv:0803.0593 [ astro - ph ] .
w. c. jones _ et al .
_ , astrophys .
j. * 647 * , 823 ( 2006 ) [ arxiv : astro - ph/0507494 ] .
f. piacentini _ et al .
_ , astrophys .
j. * 647 * , 833 ( 2006 ) [ arxiv : astro - ph/0507507 ] .
t. e. montroy _ et al .
_ , astrophys . j. * 647 * , 813 ( 2006 ) [ arxiv : astro - ph/0507514 ] .
j. l. sievers _ et al . _ ,
arxiv : astro - ph/0509203 . a. lewis and s. bridle , phys .
d * 66 * , 103511 ( 2002 ) [ arxiv : astro - ph/0205436 ] .
a. kosowsky , m. milosavljevic and r. jimenez , phys .
d * 66 * , 063007 ( 2002 ) [ arxiv : astro - ph/0206014 ] .
e. komatsu and u. seljak , mon . not .
soc . * 336 * , 1256 ( 2002 ) [ arxiv : astro - ph/0205468 ] .
l. h. kawano , fermilab - pub-92/04-a ( 1992 ) .
g. huey , r. h. cyburt and b. d. wandelt , phys .
d * 69 * , 103503 ( 2004 ) [ arxiv : astro - ph/0307080 ] .
k. ichikawa and t. takahashi , phys .
d * 73 * , 063528 ( 2006 ) [ arxiv : astro - ph/0601099 ] .
j. hamann , j. lesgourgues and g. mangano , jcap * 0803 * , 004 ( 2008 ) [ arxiv:0712.2826 [ astro - ph ] ] .
k. ichikawa , t. sekiguchi and t. takahashi , phys .
d * 78 * , 043509 ( 2008 ) [ arxiv:0712.4327 [ astro - ph ] ] .
d. serpico , s. esposito , f. iocco , g. mangano , g. miele and o. pisanti , jcap * 0412 * , 010 ( 2004 ) [ arxiv : astro - ph/0408076 ]
. k. ichikawa , m. kawasaki and f. takahashi , phys .
b * 597 * , 1 ( 2004 ) [ arxiv : astro - ph/0402522 ] .
k. ichikawa , m. kawasaki , k. nakayama , m. senami and f. takahashi , jcap * 0705 * , 008 ( 2007 ) [ arxiv : hep - ph/0703034 ] .
l. m. krauss and b. chaboyer , science * 299 * , 65 ( 2003 ) .
l. perotto , j. lesgourgues , s. hannestad , h. tu and y. y. y. wong , jcap * 0610 * , 013 ( 2006 ) [ arxiv : astro - ph/0606227 ] .
s. seager , d. d. sasselov and d. scott , astrophys .
j. lett .
* 523 * , l1 ( 1999 ) arxiv : astro - ph/9909275 . j. chluba and r. a. sunyaev , astron .
astrophys .
* 446 * , 39 ( 2006 ) [ arxiv : astro - ph/0508144 ] .
a. lewis , j. weller and r. battye , mon . not .
373 * , 561 ( 2006 ) [ arxiv : astro - ph/0606552 ] .
j. chluba , j. a. rubino - martin and r. a. sunyaev , mon . not .
soc . * 374 * , 1310 ( 2007 ) [ arxiv : astro - ph/0608242 ] . j. chluba and r. a. sunyaev , arxiv : astro - ph/0702531 .
e. r. switzer and c. m. hirata , phys .
d * 77 * , 083006 ( 2008 ) [ arxiv : astro - ph/0702143 ] . c. m. hirata and e. r. switzer , phys .
d * 77 * , 083007 ( 2008 ) [ arxiv : astro - ph/0702144 ] .
e. r. switzer and c. m. hirata , phys .
d * 77 * , 083008 ( 2008 ) [ arxiv : astro - ph/0702145 ] .
w. y. wong , a. moss and d. scott , arxiv:0711.1357 [ astro - ph ] . | we discuss how much we can probe the effective number of neutrino species @xmath0 with cosmic microwave background alone . using the data of wmap , acbar , cbi and boomerang experiments , we obtain a constraint on the effective number of neutrino species as @xmath1 at 95%c.l . for a power - law @xmath2cdm flat universe model .
the limit is improved to be @xmath3 at 95%c.l .
if we assume that the baryon density , @xmath0 and the helium abundance are related by the big bang nucleosynthesis theory .
we also provide a forecast for the planck experiment using a markov chain monte carlo approach .
in addition to constraining @xmath0 , we investigate how the big bang nucleosynthesis relation affects the estimation for these parameters and other cosmological parameters .
july 2008 + .5 in * probing the effective number of neutrino species with cosmic microwave background * .45 in kazuhide ichikawa@xmath4 , toyokazu sekiguchi@xmath5 , and tomo takahashi@xmath6 .45 in _
@xmath5 institute for cosmic ray research , university of tokyo , kashiwa 277 - 8582 , japan + @xmath7 department of physics and astronomy , university college london , gower street , london , wc1e 6bt , u.k .
+ @xmath6 department of physics , saga university , saga 840 - 8502 , japan _
.4 in |
the information contained by an individual finite object ( like a finite binary string ) is objectively measured by its kolmogorov complexity the length of the shortest binary program that computes the object .
such a shortest program contains no redundancy : every bit is information ; but is it meaningful information ? if we flip a fair coin to obtain a finite binary string , then with overwhelming probability that string constitutes its own shortest program .
however , also with overwhelming probability all the bits in the string are meaningless information , random noise . on the other hand ,
let an object @xmath0 be a sequence of observations of heavenly bodies .
then @xmath0 can be described by the binary string @xmath1 , where @xmath2 is the description of the laws of gravity , and the observational parameter setting , while @xmath3 is the data - to - model code accounting for the ( presumably gaussian ) measurement error in the data . this way we can divide the information in @xmath0 into meaningful information @xmath2 and data - to - model information @xmath3 .
the main task for statistical inference and learning theory is to distil the meaningful information present in the data .
the question arises whether it is possible to separate meaningful information from accidental information , and if so , how . in statistical theory ,
every function of the data is called a `` statistic '' of the data . the central notion in probabilistic statistics is that of a `` sufficient '' statistic , introduced by the father of statistics r.a .
fisher @xcite : `` the statistic chosen should summarise the whole of the relevant information supplied by the sample .
this may be called the criterion of sufficiency @xmath4 in the case of the normal curve of distribution it is evident that the second moment is a sufficient statistic for estimating the standard deviation . '' for traditional problems , dealing with frequencies over small sample spaces , this approach is appropriate .
but for current novel applications , average relations are often irrelevant , since the part of the support of the probability density function that will ever be observed has about zero measure .
this is the case in , for example , complex video and sound analysis .
there arises the problem that for individual cases the selection performance may be bad although the performance is good on average .
there is also the problem of what probability means , whether it is subjective , objective , or exists at all . to simplify matters , and because all discrete data can be binary coded , we consider only data samples that are finite binary strings .
the basic idea is to found statistical theory on finite combinatorial principles independent of probabilistic assumptions , as the relation between the individual data and its explanation ( model ) .
we study extraction of meaningful information in an initially limited setting where this information be represented by a finite set ( a model ) of which the object ( the data sample ) is a typical member . using the theory of kolmogorov complexity
, we can rigorously express and quantify typicality of individual objects .
but typicality in itself is not necessarily a significant property : every object is typical in the singleton set containing only that object .
more important is the following kolmogorov complexity analog of probabilistic minimal sufficient statistic which implies typicality : the two - part description of the smallest finite set , together with the index of the object in that set , is as concise as the shortest one - part description of the object .
the finite set models the regularity present in the object ( since it is a typical element of the set ) .
this approach has been generalized to computable probability mass functions .
the combined theory has been developed in detail in @xcite and called `` algorithmic statistics . '' here we study the most general form of algorithmic statistic : recursive function models . in this setting the issue of meaningful information versus accidental information is put in its starkest form ; and in fact , has been around for a long time in various imprecise forms unconnected with the sufficient statistic approach : the issue has sparked the imagination and entered scientific popularization in @xcite as `` effective complexity '' ( here `` effective '' is apparently used in the sense of `` producing an effect '' rather than `` constructive '' as is customary in the theory of computation ) .
it is time that it receives formal treatment .
formally , we study the minimal length of a total recursive function that leads to an optimal length two - part code of the object being described .
( `` total '' means the function value is defined for all arguments in the domain , and `` partial '' means that the function is possibly not total . ) this minimal length has been called the `` sophistication '' of the object in @xcite in a different , but related , setting of compression and prediction properties of infinite sequences . that treatment is technically sufficiently vague so as to have no issue for the present work .
we develop the notion based on prefix turing machines , rather than on a variety of monotonic turing machines as in the cited papers . below we describe related work in detail and summarize our results .
subsequently , we formulate our problem in the formal setting of computable two - part codes .
kolmogorov in 1974 @xcite proposed an approach to a non - probabilistic statistics based on kolmogorov complexity .
an essential feature of this approach is to separate the data into meaningful information ( a model ) and meaningless information ( noise ) .
cover @xcite attached the name `` sufficient statistic '' to a model of which the data is a `` typical '' member . in kolmogorov
s initial setting the models are finite sets . as kolmogorov himself pointed out , this is no real restriction : the finite sets model class is equivalent , up to a logarithmic additive term , to the model class of computable probability density functions , as studied in @xcite .
related aspects of `` randomness deficiency '' were formulated in @xcite and studied in @xcite . despite its evident epistemological prominence in the theory of hypothesis selection and prediction ,
only selected aspects of the theory were studied in these references .
recent work @xcite can be considered as a comprehensive investigation into the sufficient statistic for finite set models and computable probability density function models .
here we extend the approach to the most general form : the model class of total recursive functions .
this idea was pioneered by @xcite who , unaware of a statistic connection , coined the cute word `` sophistication . ''
the algorithmic ( minimal ) sufficient statistic was related to an applied form in @xcite : the well - known `` minimum description length '' principle @xcite in statistics and inductive reasoning . in another paper @xcite ( chronologically following the present paper )
we comprehensively treated all stochastic properties of the data in terms of kolmogorov s so - called structure functions .
the sufficient statistic aspect , studied here , covers only part of these properties .
the results on the structure functions , including ( non)computability properties , are valid , up to logarithmic additive terms , also for the model class of total recursive functions , as studied here .
it will be helpful for the reader to be familiar with initial parts of @xcite . in @xcite
, kolmogorov observed that randomness of an object in the sense of having high kolmogorov complexity is being random in just a `` negative '' sense .
that being said , we define the notion of sophistication ( minimal sufficient statistic in the total recursive function model class ) .
it is demonstrated to be meaningful ( existence and nontriviality ) .
we then establish lower and upper bounds on the sophistication , and we show that there are objects the sophistication achieves the upper bound .
in fact , these are objects in which all information is meaningful and there is ( almost ) no accidental information .
that is , the simplest explanation of such an object is the object itself . in the simpler
setting of finite set statistic the analogous objects were called `` absolutely non - stochastic '' by kolmogorov .
if such objects have high kolmogorov complexity , then they can only be a random outcome of a `` complex '' random process , and kolmogorov questioned whether such random objects , being random in just this `` negative '' sense , can occur in nature .
but there are also objects that are random in the sense of having high kolmogorov complexity , but simultaneously are are typical outcomes of `` simple '' random processes .
these were therefore said to be random in a `` positive '' sense @xcite .
an example are the strings of maximal kolmogorov complexity ; those are very unsophisticated ( with sophistication about 0 ) , and are typical outcomes of tosses with a fair coin a very simple random process .
we subsequently establish the equivalence between sophistication and the algorithmic minimal sufficient statistics of the finite set class and the probability mass function class .
finally , we investigate the algorithmic properties of sophistication : nonrecursiveness , upper semicomputability , and intercomputability relations of kolmogorov complexity , sophistication , halting sequence .
a _ string _ is a finite binary sequence , an element of @xmath5 .
if @xmath0 is a string then the _ length _ @xmath6 denotes the number of bits in @xmath0 .
we identify @xmath7 , the natural numbers , and @xmath5 according to the correspondence @xmath8 here @xmath9 denotes the _
empty word_. thus , @xmath10 .
the emphasis is on binary sequences only for convenience ; observations in any alphabet can be so encoded in a way that is ` theory neutral ' . below we will use the natural numbers and the strings interchangeably .
a string @xmath11 is a _ proper prefix _ of a string @xmath0 if we can write @xmath12 for @xmath13 .
a set @xmath14 is _ prefix - free _ if for any pair of distinct elements in the set neither is a proper prefix of the other .
a prefix - free set is also called a _ prefix code _ and its elements are called _
code words_. an example of a prefix code , that is useful later , encodes the source word @xmath15 by the code word @xmath16 this prefix - free code is called _ self - delimiting _ , because there is fixed computer program associated with this code that can determine where the code word @xmath17 ends by reading it from left to right without backing up . this way a composite code message can be parsed in its constituent code words in one pass , by the computer program .
( this desirable property holds for every prefix - free encoding of a finite set of source words , but not for every prefix - free encoding of an infinite set of source words . for a single finite computer program to be able to parse a code message the encoding needs to have a certain uniformity property like the @xmath18 code . )
since we use the natural numbers and the strings interchangeably , @xmath19 where @xmath0 is ostensibly an integer , means the length in bits of the self - delimiting code of the string with index @xmath0 . on the other hand , @xmath20 where @xmath0 is ostensibly a string , means the self - delimiting code of the string with index the length @xmath6 of @xmath0 . using this code we define the standard self - delimiting code for @xmath0 to be @xmath21 .
it is easy to check that @xmath22 and @xmath23 .
let @xmath24 denote a standard invertible effective one - one encoding from @xmath25 to a subset of @xmath7 .
for example , we can set @xmath26 or @xmath27 .
we can iterate this process to define @xmath28 , and so on . for definitions , notation , and an introduction to kolmogorov complexity ,
see @xcite .
informally , the kolmogorov complexity , or algorithmic entropy , @xmath29 of a string @xmath0 is the length ( number of bits ) of a shortest binary program ( string ) to compute @xmath0 on a fixed reference universal computer ( such as a particular universal turing machine ) .
intuitively , @xmath29 represents the minimal amount of information required to generate @xmath0 by any effective process .
the conditional kolmogorov complexity @xmath30 of @xmath0 relative to @xmath11 is defined similarly as the length of a shortest program to compute @xmath0 , if @xmath11 is furnished as an auxiliary input to the computation . for technical reasons we use a variant of complexity , so - called prefix complexity , which is associated with turing machines for which the set of programs resulting in a halting computation is prefix free .
we realize prefix complexity by considering a special type of turing machine with a one - way input tape , a separate work tape , and a one - way output tape .
such turing machines are called _ prefix _ turing machines . if a machine @xmath31 halts with output @xmath0 after having scanned all of @xmath2 on the input tape , but not further , then @xmath32 and we call @xmath2 a _ program _ for @xmath31 .
it is easy to see that @xmath33 is a _
prefix code_. a function @xmath34 from the natural numbers to the natural numbers is _ partial recursive _ , or _
, if there is a turing machine @xmath31 that computes it : @xmath35 for all @xmath0 for which either @xmath34 or @xmath31 ( and hence both ) are defined .
this definition can be extended to ( multi - tuples of ) rational arguments and values .
let @xmath36 be a standard enumeration of all prefix turing machines with a binary input tape , for example the lexicographical length - increasing ordered syntactic prefix turing machine descriptions , @xcite , and let @xmath37 be the enumeration of corresponding functions that are computed by the respective turing machines ( @xmath38 computes @xmath39 ) .
these functions are the partial recursive functions of effectively prefix - free encoded arguments .
the kolmogorov complexity of @xmath0 is the length of the shortest binary program from which @xmath0 is computed by such a function .
the _ prefix kolmogorov complexity _ of @xmath0 is @xmath40 where the minimum is taken over @xmath41 and @xmath42 . for the development of the theory
we actually require the turing machines to use _ auxiliary _ ( also called _ conditional _ ) information , by equipping the machine with a special read - only auxiliary tape containing this information at the outset .
then , the _ conditional version _ @xmath43 of the prefix kolmogorov complexity of @xmath0 given @xmath11 ( as auxiliary information ) is is defined similarly as before , and the unconditional version is set to @xmath44 . from now on
, we will denote by @xmath45 an inequality to within an additive constant , and by @xmath46 the situation when both @xmath45 and @xmath47 hold .
let @xmath48 be the standard enumeration of turing machines , and let @xmath49 be a standard universal turing machine satisfying @xmath50 for all indices @xmath51 and programs @xmath2 .
we fix @xmath49 once and for all and call it the _ reference universal prefix turing machine_. the shortest program to compute @xmath0 by @xmath49 is denoted as @xmath52 ( if there is more than one of them , then @xmath52 is the first one in standard enumeration ) .
it is a deep and useful fact that the shortest effective description of an object @xmath0 can be expressed in terms of a _ two - part code _ : the first part describing an appropriate turing machine and the second part describing the program that interpreted by the turing machine reconstructs @xmath0 .
the essence of the theory is the invariance theorem , that can be informally stated as follows : for convenience , in the sequel we simplify notation and write @xmath53 for @xmath54 .
rewrite @xmath55 here the minima are taken over @xmath56 and @xmath57 .
the last equalities are obtained by using the universality of @xmath58 with @xmath59 . as consequence , @xmath60
thus , @xmath29 and @xmath61 differ by at most an additive constant depending on the choice of @xmath49 .
it is standard to use @xmath62 instead of as the definition of _ prefix kolmogorov complexity _ , @xcite .
however , we highlighted definition to bring out the two - part code nature . by universal logical principles ,
the resulting theory is recursively invariant under adopting either definition or definition , as long as we stick to one choice .
if @xmath31 stands for a literal description of the prefix turing machine @xmath31 in standard format , for example the index @xmath63 when @xmath64 , then we can write @xmath65 .
the string @xmath66 is a shortest self - delimiting program of @xmath67 bits from which @xmath49 can compute @xmath63 , and subsequent execution of the next self - delimiting fixed program @xmath68 will compute @xmath69 from @xmath63 . altogether , this has the effect that @xmath70 . if @xmath71 minimizes the expression above , then @xmath72 , and hence @xmath73 , and @xmath74 .
it is straightforward that @xmath75 , and therefore we have @xmath76 .
altogether , @xmath77 . replacing the minimizing @xmath78 by the minimizing @xmath79 and @xmath80 by @xmath81 , we can rewrite the last displayed equation as @xmath82 expression emphasizes the two - part code nature of kolmogorov complexity : using the regular aspects of @xmath0 to maximally compress .
suppose we consider an ongoing time - series @xmath83 and we randomly stop gathering data after having obtained the initial segment @xmath84 we can encode this @xmath0 by a small turing machine representing `` the repeating pattern is 01 , '' and which computes @xmath0 , for example , from the program `` 13 . '' intuitively , the turing machine part of the code squeezes out the _ regularities _ in @xmath0 .
what is left are irregularities , or _ random aspects _ of @xmath0 relative to that turing machine .
the minimal - length two - part code squeezes out regularity only insofar as the reduction in the length of the description of random aspects is greater than the increase in the regularity description . in this setup
the number of repetitions of the significant pattern is viewed as the random part of the data .
this interpretation of @xmath29 as the shortest length of a two - part code for @xmath0 , one part describing a turing machine , or _ model _
, for the _ regular _ aspects of @xmath0 and the second part describing the _ irregular _ aspects of @xmath0 in the form of a program to be interpreted by @xmath31 , has profound applications .
the `` right model '' is a turing machine @xmath31 among the ones that halt for all inputs , a restriction that is justified later , and reach the minimum description length in ( [ eq.kcmdl ] ) .
this @xmath31 embodies the amount of useful information contained in @xmath0 .
it remains to decide which such @xmath31 to select among the ones that satisfy the requirement .
following occam s razor we opt here for the shortest one a formal justification for this choice is given in @xcite .
the main problem with our approach is how to properly define a shortest program @xmath52 for @xmath0 that divides into parts @xmath85 such that @xmath2 represents an appropriate @xmath31 .
the following central notions are used in this paper .
the _ information in @xmath0 about @xmath11 _ is @xmath86 . by the symmetry of information ,
a deep result of @xcite , @xmath87 rewriting according to symmetry of information we see that @xmath88 and therefore we call the quantity @xmath89 the _ mutual information _ between @xmath0 and @xmath11 .
instead of the model class of finite sets , or computable probability density functions , as in @xcite , in this work we focus on the most general form of algorithmic model class : total recursive functions .
we define the different model classes and summarize the central notions of `` randomness deficiency '' and `` typicality '' for the canonical finite set models to obtain points of reference for the related notions in the more general model classes .
the model class of _ finite sets _ consists of the set of finite subsets @xmath90 .
the _ complexity of the finite set _
@xmath91 is @xmath92the length ( number of bits ) of the shortest binary program @xmath2 from which the reference universal prefix machine @xmath49 computes a listing of the elements of @xmath91 and then halts .
that is , if @xmath93 , then @xmath94 .
the _ conditional complexity _
@xmath95 of @xmath0 given @xmath91 , is the length ( number of bits ) in the shortest binary program @xmath2 from which the reference universal prefix machine @xmath49 , given @xmath91 literally as auxiliary information , computes @xmath0 . for every finite set @xmath96 containing @xmath0 we have @xmath97 indeed , consider the selfdelimiting code of @xmath0 consisting of its @xmath98 bit long index of @xmath0 in the lexicographical ordering of @xmath91
this code is called _ data - to - model code_. its length quantifies the maximal `` typicality , '' or `` randomness , '' data ( possibly different from @xmath0 ) can have with respect to this model .
the lack of typicality of @xmath0 with respect to @xmath91 is measured by the amount by which @xmath95 falls short of the length of the data - to - model code , the _ randomness deficiency _ of @xmath0 in @xmath91 , defined by @xmath99 for @xmath100 , and @xmath101 otherwise .
data @xmath0 is _ typical with respect to a finite set _
@xmath91 , if the randomness deficiency is small . if the randomness deficiency is close to 0 , then there are no simple special properties that single it out from the majority of elements in @xmath91 .
this is not just terminology .
let @xmath102 . according to common viewpoints in probability theory , each property represented by @xmath91 defines a large subset of @xmath91 consisting of elements having that property , and , conversely , each large subset of @xmath91 represents a property . for probabilistic ensembles we take high probability subsets as properties ;
the present case is uniform probability with finite support .
for some appropriate fixed constant @xmath103 , let us identify a property represented by @xmath91 with a subset @xmath104 of @xmath91 of cardinality @xmath105 .
if @xmath106 is close to 0 , then @xmath0 satisfies ( that is , is an element of ) _ all _ properties ( that is , sets ) @xmath107 of low kolmogorov complexity @xmath108 . the precise statements and quantifications are given in @xcite , and we do not repeat them here . the model class of _ computable probability density functions _ consists of the set of functions @xmath109 $ ] with @xmath110 .
`` computable '' means here that there is a turing machine @xmath111 that , given @xmath0 and a positive rational @xmath112 , computes @xmath113 with precision @xmath112 .
the ( prefix- ) complexity @xmath114 of a computable ( possibly partial ) function @xmath115 is defined by @xmath116 the model class of _ total recursive functions _
consists of the set of functions @xmath117 such that there is a turing machine @xmath31 such that @xmath118 and @xmath119 , for every @xmath120 .
the ( prefix- ) complexity @xmath121 of a total recursive function @xmath34 is defined by @xmath122 if @xmath123 is a shortest program for computing the function @xmath34 ( if there is more than one of them then @xmath123 is the first one in enumeration order ) , then @xmath124 . in the definitions of @xmath114 and @xmath121 ,
the objects being described are functions rather than finite binary strings . to unify the approaches
, we can consider a finite binary string @xmath0 as corresponding to a function having value @xmath0 for argument 0 .
note that we can upper semi - compute @xmath52 given @xmath0 , but we can not upper semi - compute @xmath125 given @xmath115 ( as an oracle ) , or @xmath123 given @xmath34 ( again given as an oracle ) , since we should be able to verify agreement of a program for a function and an oracle for the target function , on all infinitely many arguments .
to explain typicality for general model classes , it is convenient to use the distortion - rate @xcite approach for individual data recently introduced in @xcite . modeling the data can be viewed as encoding the data by a model : the data are source words to be coded , and models are code words for the data . as before , the set of possible data is @xmath126 .
let @xmath127 denote the set of non - negative real numbers .
for every model class @xmath128 ( particular set of code words ) we choose an appropriate recursive function @xmath129 defining the _ distortion _
@xmath130 between data @xmath0 and model @xmath131 .
the choice of distortion function is a selection of which aspects of the data are relevant , or meaningful , and which aspects are irrelevant ( noise ) .
we can think of the distortion as measuring how far the model falls short in representing the data .
distortion - rate theory underpins the practice of lossy compression .
for example , lossy compression of a sound file gives as `` model '' the compressed file where , among others , the very high and very low inaudible frequencies have been suppressed .
thus , the distortion function will penalize the deletion of the inaudible frequencies but lightly because they are not relevant for the auditory experience .
[ ex.11 ] let us look at various model classes and distortion measures : \(i ) the set of models are the finite sets of finite binary strings .
let @xmath90 and @xmath132 .
we define @xmath133 if @xmath100 , and @xmath101 otherwise .
\(ii ) the set of models are the computable probability density functions @xmath115 mapping @xmath5 to @xmath134 $ ] .
we define @xmath135 if @xmath136 , and @xmath101 otherwise .
\(iii ) the set of models are the total recursive functions @xmath34 mapping @xmath5 to @xmath7 .
we define @xmath137 , and @xmath101 if no such @xmath3 exists .
if @xmath128 is a model class , then we consider _ distortion balls _ of given radius @xmath138 centered on @xmath139 : @xmath140 this way , every model class and distortion measure can be treated similarly to the canonical finite set case , which , however is especially simple in that the radius not variable .
that is , there is only one distortion ball centered on a given finite set , namely the one with radius equal to the log - cardinality of that finite set .
in fact , that distortion ball equals the finite set on which it is centered .
let @xmath128 be a model class and @xmath3 a distortion measure .
since in our definition the distortion is recursive , given a model @xmath139 and diameter @xmath138 , the elements in the distortion ball of diameter @xmath138 can be recursively enumerated from the distortion function . giving the index of any element @xmath0 in that enumeration we can find the element .
hence , @xmath141 .
on the other hand , the vast majority of elements @xmath11 in the distortion ball have complexity @xmath142 since , for every constant @xmath103 , there are only @xmath143 binary programs of length @xmath144 available , and there are @xmath145 elements to be described .
we can now reason as in the similar case of finite set models . with data
@xmath0 and @xmath146 , if @xmath147 , then @xmath0 belongs to every large majority of elements ( has the property represented by that majority ) of the distortion ball @xmath148 , provided that property is simple in the sense of having a description of low kolmogorov complexity .
the _ randomness deficiency _ of @xmath0 with respect to model @xmath131 under distortion @xmath3 is defined as @xmath149 data @xmath0 is _ typical _ for model @xmath139 ( and that model `` typical '' or `` best fitting '' for @xmath0 ) if @xmath150 if @xmath0 is typical for a model @xmath131 , then the shortest way to effectively describe @xmath0 , given @xmath131 , takes about as many bits as the descriptions of the great majority of elements in a recursive enumeration of the distortion ball .
so there are no special simple properties that distinguish @xmath0 from the great majority of elements in the distortion ball : they are all typical or random elements in the distortion ball ( that is , with respect to the contemplated model ) . continuing example [ ex.11 ] by applying to different model classes : \(i ) _ finite sets : _ for finite set models @xmath91 , clearly @xmath151 .
together with we have that @xmath0 is typical for @xmath91 , and @xmath91 best fits @xmath0 , if the randomness deficiency according to satisfies @xmath152 .
\(ii ) _ computable probability density functions : _ instead of the data - to - model code length @xmath153 for finite set models , we consider the data - to - model code length @xmath154 ( the shannon - fano code ) .
the value @xmath154 measures how likely @xmath0 is under the hypothesis @xmath115 . for probability models
@xmath115 , define the conditional complexity @xmath155 as follows .
say that a function @xmath156 approximates @xmath115 if @xmath157 for every @xmath11 and every positive rational @xmath112 .
then @xmath158 is defined as the minimum length of a program that , given @xmath159 and any function @xmath156 approximating @xmath115 as an oracle , prints @xmath0 .
clearly @xmath160 .
together with , we have that @xmath0 is typical for @xmath115 , and @xmath115 best fits @xmath0 , if @xmath161 .
the right - hand side set condition is the same as @xmath162 , and there can be only @xmath163 such @xmath11 , since otherwise the total probability exceeds 1 . therefore , the requirement , and hence typicality , is implied by @xmath164 . define the randomness deficiency by @xmath165 altogether , a string @xmath0 is _
typical for a distribution _ @xmath115 , or @xmath115 is the _ best fitting model _ for @xmath0 , if @xmath166 .
\(iii ) _ total recursive functions : _ in place of @xmath153 for finite set models we consider the data - to - model code length ( actually , the distortion @xmath167 above ) @xmath168 define the conditional complexity @xmath169 as the minimum length of a program that , given @xmath170 and an oracle for @xmath34 , prints @xmath0 .
clearly , @xmath171 . together with , we have that @xmath0 is typical for @xmath34 , and @xmath34 best fits @xmath0 , if @xmath172 .
there are at most @xmath173- many @xmath11 satisfying the set condition since @xmath174 .
therefore , the requirement , and hence typicality , is implied by @xmath175 .
define the randomness deficiency by @xmath176 altogether , a string @xmath0 is _
typical for a total recursive function _
@xmath34 , and @xmath34 is the _ best fitting recursive function model _ for @xmath0 if @xmath177 , or written differently , @xmath178 note that since @xmath170 is given as conditional information , with @xmath179 and @xmath180 , the quantity @xmath181 represents the number of bits in a shortest _ self - delimiting _ description of @xmath3 .
we required @xmath170 in the conditional in .
this is the information about the radius of the distortion ball centered on the model concerned .
note that in the canonical finite set model case , as treated in @xcite , every model has a fixed radius which is explicitly provided by the model itself . but in the more general model classes of computable probability density functions , or total recursive functions , models can have a variable radius .
there are subclasses of the more general models that have fixed radiuses ( like the finite set models ) .
\(i ) in the computable probability density functions one can think of the probabilities with a finite support , for example @xmath182 for @xmath183 , and @xmath184 otherwise .
\(ii ) in the total recursive function case one can similarly think of functions with finite support , for example @xmath185 for @xmath186 , and @xmath187 for @xmath188 .
the incorporation of te radius in the model will increase the complexity of the model , and hence of the minimal sufficient statistic below .
a _ statistic _ is a function mapping the data to an element ( model ) in the contemplated model class . with some sloppiness of terminology
we often call the function value ( the model ) also a statistic of the data .
the most important concept in this paper is the sufficient statistic . for an extensive discussion of this notion for specific model classes
see @xcite .
a statistic is called sufficient if the two - part description of the data by way of the model and the data - to - model code is as concise as the shortest one - part description of @xmath0 .
consider a model class @xmath128 .
a model @xmath139 is a _
sufficient statistic _ for @xmath0 if @xmath189 [ lem.v2 ] if @xmath131 is a sufficient statistic for @xmath0 , then @xmath190 , that is , @xmath0 is typical for @xmath131 .
we can rewrite @xmath191 .
the first three inequalities are straightforward and the last equality is by the assumption of sufficiency . altogether , the first sum equals the second sum , which implies the lemma .
thus , if @xmath131 is a sufficient statistic for @xmath0 , then @xmath0 is a typical element for @xmath131 , and @xmath131 is the best fitting model for @xmath0 .
note that the converse implication , `` typicality '' implies `` sufficiency , '' is not valid .
sufficiency is a special type of typicality , where the model does not add significant information to the data , since the preceding proof shows @xmath192 . using the symmetry of information
this shows that @xmath193 this means that : \(i ) a sufficient statistic @xmath131 is determined by the data in the sense that we need only an @xmath194-bit program , possibly depending on the data itself , to compute the model from the data .
\(ii ) for each model class and distortion there is a universal constant @xmath103 such that for every data item @xmath0 there are at most @xmath103 sufficient statistics . _
finite sets : _ for the model class of finite sets , a set @xmath91 is a sufficient statistic for data @xmath0 if @xmath195 _ computable probability density functions : _ for the model class of computable probability density functions , a function @xmath115 is a sufficient statistic for data @xmath0 if @xmath196 for the model class of _ total recursive functions _ , a function @xmath34 is a _ sufficient statistic _ for data @xmath0 if @xmath197 following the above discussion , the meaningful information in @xmath0 is represented by @xmath34 ( the model ) in @xmath121 bits , and the meaningless information in @xmath0 is represented by @xmath3 ( the noise in the data ) with @xmath180 in @xmath198 bits .
note that @xmath199 , since the two - part code @xmath200 for @xmath0 can not be shorter than the shortest one - part code of @xmath29 bits , and therefore the @xmath3-part must already be maximally compressed . by lemma [ lem.v2 ] , @xmath201 , @xmath0 is typical for @xmath34 , and hence @xmath202 .
consider the model class of total recursive functions . a _
minimal sufficient statistic _ for data @xmath0 is a sufficient statistic for @xmath0 of minimal prefix complexity .
its length is known as the _ sophistication _ of @xmath0 , and is defined by @xmath203 .
recall that the _ reference _ universal prefix turing machine @xmath49 was chosen such that @xmath204 for all @xmath31 and @xmath3 .
looking at it slightly more from a programming point of view , we can define a pair @xmath205 to be a _ description _ of a finite string @xmath0 , if @xmath206 prints @xmath0 and @xmath31 is a turing machine computing a function @xmath34 so that @xmath180 . for the notion of minimal sufficient statistic to be nontrivial , it should be impossible to always shift , if @xmath180 and @xmath207 with @xmath208 , always information information from @xmath34 to @xmath3 and write , for example , @xmath209 with @xmath210 with @xmath211 .
if the model class contains a fixed universal model that can mimic all other models , then we can always shift all model information to the data - to-(universal ) model code .
note that this problem does nt arise in common statistical model classes : these do not contain universal models in the algorithmic sense . first we show that the partial recursive recursive function model class , because it contains a universal element , does not allow a straightforward nontrivial division into meaningful and meaningless information .
assume for the moment that we allow all partial recursive programs as statistic .
then , the sophistication of all data @xmath0 is @xmath212 .
let the index of @xmath49 ( the reference universal prefix turing machine ) in the standard enumeration @xmath48 of prefix turing machines be @xmath213 .
let @xmath214 be a turing machine computing @xmath34 .
suppose that @xmath215 .
then , also @xmath216 .
this shows that unrestricted partial recursive statistics are uninteresting .
naively , this could leave the impression that the separation of the regular and the random part of the data is not as objective as the whole approach lets us hope for . if we consider complexities of the minimal sufficient statistics in model classes of increasing power : finite sets , computable probability distributions , total recursive functions , partial recursive functions , then the complexities appear to become smaller all the time eventually reaching zero .
it would seem that the universality of kolmogorov complexity , based on the notion of partial recursive functions , would suggest a similar universal notion of sufficient statistic based on partial recursive functions .
but in this case the very universality trivializes the resulting definition : because partial recursive functions contain a particular universal element that can simulate all the others , this implies that the universal partial recursive function is a universal model for all data , and the data - to - model code incorporates all information in the data .
thus , if a model class contains a universal model that can simulate all other models , then this model class is not suitable for defining two - part codes consisting of meaningful information and accidental information .
it turns out that the key to nontrivial separation is the requirement that the program witnessing the sophistication be _
total_. that the resulting separation is non - trivial is evidenced by the fact , shown below , that the amount of meaningful information in the data does not change by more than a logarithmic additive term under change of model classes among finite set models , computable probability models , and total recursive function models .
that is , very different model classes all result in the same amount of meaningful information in the data , up to negligible differences .
so if deterioration occurs in widening model classes it occurs all at once by having a universal element in the model class .
apart from triviality , a class of statistics can also possibly be vacuous by having the length of the minimal sufficient statistic exceed @xmath29 .
our first task is to determine whether the definition is non - vacuous .
we will distinguish sophistication in different description modes : [ lem.exists ] for every finite binary string @xmath0 , the sophistication satisfies @xmath217 . by definition of the prefix complexity
there is a program @xmath52 of length @xmath218 such that @xmath219 .
this program @xmath52 can be partial .
but we can define another program @xmath220 where @xmath221 is a program of a constant number of bits that tells the following program to ignore its actual input and compute as if its input were @xmath9 . clearly , @xmath222 is total and is a sufficient statistic of the total recursive function type , that is , @xmath223 .
the previous lemma gives an upper bound on the sophistication .
this still leaves the possibility that the sophistication is always @xmath212 , for example in the most liberal case of unrestricted totality . but this turns out to be impossible .
[ h - sophi ] ( i ) for every @xmath0 , if a sufficient statistic @xmath34 satisfies @xmath224 , then @xmath225 and @xmath226 .
\(ii ) for @xmath0 as a variable running through a sequence of finite binary strings of increasing length , we have @xmath227 \(iii ) for every @xmath228 , there exists an @xmath0 of length @xmath228 , such that every sufficient statistic @xmath34 for @xmath0 that satisfies @xmath224 has @xmath229 .
\(iv ) for every @xmath228 there exists an @xmath0 of length @xmath228 such that @xmath230 .
\(i ) if @xmath34 is a sufficient statistic for @xmath0 , then @xmath231 since @xmath224 , given an @xmath194 bit program @xmath232 we can retrieve both @xmath170 and and also @xmath233 from @xmath123 .
therefore , we can retrieve @xmath234 from @xmath235 .
that shows that @xmath236 .
this proves both the first statement , and the second statement follows by ( [ eq.pd ] ) .
\(ii ) an example of very unsophisticated strings are the individually random strings with high complexity : @xmath0 of length @xmath237 with complexity @xmath238 .
then , the _ identity _ program @xmath239 with @xmath240 for all @xmath3 is total , has complexity @xmath241 , and satisfies @xmath242 .
hence , @xmath239 witnesses that @xmath243 .
this shows ( [ eq.liminf ] ) .
\(iii ) consider the set @xmath244 .
by @xcite we have @xmath245 .
let @xmath246 .
since there are @xmath247 strings of length @xmath228 , there are strings of length @xmath228 not in @xmath248 .
let @xmath0 be any such string , and denote @xmath249 .
then , by construction @xmath250 and by definition @xmath251
. let @xmath34 be a sufficient statistic for @xmath0 .
then , @xmath252 . by assumption
, there is an @xmath194-bit program @xmath232 such that @xmath253 .
let @xmath3 witness @xmath170 by @xmath180 with @xmath198 .
define the set @xmath254 . clearly , @xmath255 . since @xmath0 can be retrieved from @xmath34 and the lexicographical index of @xmath3 in @xmath256 , and @xmath257 , we have @xmath258 .
since we can obtain @xmath256 from @xmath235 we have @xmath259 . on the other hand ,
since we can retrieve @xmath0 from @xmath256 and the index of @xmath3 in @xmath256 , we must have @xmath260 , which implies @xmath261 .
altogether , therefore , @xmath262 .
we now show that we can choose @xmath0 so that @xmath263 , and therefore @xmath229 . for every length @xmath228
, there exists a @xmath264 of complexity @xmath265 such that a minimal sufficient finite set statistic @xmath91 for @xmath264 has complexity at least @xmath266 , by theorem iv.2 of @xcite .
since @xmath267 is trivially a sufficient statistic for @xmath264 , it follows @xmath268 .
this implies @xmath269 .
therefore , we can choose @xmath270 for a large enough constant @xmath271 so as to ensure that @xmath272 .
consequently , we can choose @xmath0 above as such a @xmath264 .
since every finite set sufficient statistic for @xmath0 has complexity at least that of an finite set minimal sufficient statistic for @xmath0 , it follows that @xmath263 .
therefore , @xmath229 , which was what we had to prove .
\(iv ) in the proof of ( i ) we used @xmath273 . without using this assumption , the corresponding argument yields @xmath274 .
we also have @xmath275 and @xmath276 . since we can retrieve @xmath0 from @xmath256 and its index in @xmath256 , the same argument as above shows @xmath277 , and still following the argument above , @xmath278 . since @xmath279 we have @xmath280 .
this proves the statement .
the useful ( [ eq.pcondx ] ) states that there is a constant , such that for every @xmath0 there are at most that constant many sufficient statistics for @xmath0 , and there is a constant length program ( possibly depending on @xmath0 ) , that generates all of them from @xmath52 .
in fact , there is a slightly stronger statement from which this follows : there is a universal constant @xmath103 , such that for every @xmath0 , the number of @xmath281 such that @xmath282 and @xmath283 , is bounded above by @xmath103 .
let the prefix turing machine @xmath214 compute @xmath34 . since @xmath215 and @xmath284 , the combination @xmath285 ( with self - delimiting @xmath123 ) is a shortest prefix program for @xmath0 . from @xcite ,
exercise 3.3.7 item ( b ) on p. 205 , it follows that the number of shortest prefix programs is upper bounded by a universal constant .
previous work studied sufficiency for finite set models , and computable probability mass functions models , @xcite .
the most general models that are still meaningful are total recursive functions as studied here .
we show that there are corresponding , almost equivalent , sufficient statistics in all model classes .
[ lem.explimpl ] \(i ) if @xmath91 is a sufficient statistic of @xmath0 ( finite set type ) , then there is a corresponding sufficient statistic @xmath115 of @xmath0 ( probability mass function type ) such that @xmath286 , @xmath287 , and @xmath288 .
\(ii ) if @xmath115 is a sufficient statistic of @xmath0 of the computable total probability density function type , then there is a corresponding sufficient statistic @xmath34 of @xmath0 of the total recursive function type such that @xmath289 , @xmath290 , and @xmath291 .
\(i ) by assumption , @xmath91 is a finite set such that @xmath100 and @xmath292 .
define the probability distribution @xmath293 for @xmath294 and @xmath295 otherwise . since @xmath91 is finite , @xmath115 is computable . since @xmath296 , and @xmath297
, we have @xmath298 .
since @xmath115 is a computable probability mass function we have @xmath299 , by the standard shannon - fano code construction @xcite that assigns a code word of length @xmath300 to @xmath0 .
since by ( [ eq.soi ] ) we have @xmath301 it follows that @xmath302 . hence , @xmath303 .
therefore , by ( [ eq.soi ] ) , @xmath304 and , by rewriting @xmath305 in the other way according to ( [ eq.soi ] ) , @xmath288 .
\(ii ) by assumption , @xmath115 is a computable probability density function with @xmath136 and @xmath306 .
the witness of this equality is a shortest program @xmath125 for @xmath115 and a code word @xmath307 for @xmath0 according to the standard shannon - fano code , @xcite , with @xmath308 .
given @xmath115 , we can reconstruct @xmath0 from @xmath307 by a fixed standard algorithm .
define the recursive function @xmath34 from @xmath115 such that @xmath309 .
in fact , from @xmath125 this only requires a constant length program @xmath232 , so that @xmath310 is a program that computes @xmath34 in the sense that @xmath311 for all @xmath3 .
similarly , @xmath115 can be retrieved from @xmath34 .
hence , @xmath289 and @xmath312 .
that is , @xmath34 is a sufficient statistic for @xmath0 . also , @xmath34 is a total recursive function .
since @xmath313 we have @xmath314 , and @xmath314 .
this shows that @xmath315 , and since @xmath0 can by definition be reconstructed from @xmath123 and a program of length @xmath316 , it follows that equality must hold .
consequently , @xmath317 , and hence , by ( [ eq.soi ] ) , @xmath318 and @xmath291 .
we have now shown that a sufficient statistic in a less general model class corresponds directly to a sufficient statistic in the next more general model class .
we now show that , with a negligible error term , a sufficient statistic in the most general model class of total recursive functions has a directly corresponding sufficient statistic in the least general finite set model class .
that is , up to negligible error terms , a sufficient statistic in any of the model classes has a direct representative in any of the other model classes .
let @xmath0 be a string of length @xmath228 , and @xmath34 be a total recursive function sufficient statistic for @xmath0 .
then , there is a finite set @xmath319 such that @xmath320 . by assumption
there is an @xmath194-bit program @xmath232 such that @xmath253 . for each @xmath321 ,
let @xmath322 .
define @xmath323 .
we can compute @xmath91 by computation of @xmath324 , on all arguments @xmath51 of at most @xmath325 bits , since by assumption @xmath34 is total .
this shows @xmath326 . since @xmath327 , we have @xmath328 .
moreover , @xmath329 . since @xmath100 , @xmath330 , where we use the sufficiency of @xmath34 to obtain the last inequality
we investigate the recursion properties of the sophistication function . in @xcite
, gcs gave an important and deep result ( [ eq.gacs ] ) below , that quantifies the uncomputability of @xmath29 ( the bare uncomputability can be established in a much simpler fashion ) .
for every length @xmath228 there is an @xmath0 of length @xmath228 such that : @xmath331 note that the right - hand side holds for every @xmath0 by the simple argument that @xmath332 and hence @xmath333 .
but there are @xmath0 s such that the length of the shortest program to compute @xmath29 almost reaches this upper bound , even if the full information about @xmath0 is provided .
it is natural to suppose that the sophistication function is not recursive either .
the following lemma s suggest that the complexity function is more uncomputable than the sophistication .
the function @xmath334 is not recursive .
given @xmath228 , let @xmath335 be the least @xmath0 such that @xmath336 . by theorem [ h - sophi ] we know that there exist @xmath0 such that @xmath337 for @xmath338 , hence @xmath335 exists
. assume by way of contradiction that the sophistication function is computable .
then , we can find @xmath335 , given @xmath228 , by simply computing the successive values of the function .
but then @xmath339 , while by lemma [ lem.exists ] @xmath340 and by assumption @xmath341 , which is impossible .
the _ halting sequence _
@xmath342 is the infinite binary characteristic sequence of the halting problem , defined by @xmath343 if the reference universal prefix turing machine @xmath49 halts on the @xmath51th input : @xmath344 , and 0 otherwise .
[ lem.compks ] let @xmath123 be a total recursive function sufficient statistic of @xmath0 .
\(i ) we can compute @xmath29 from @xmath123 and @xmath0 , up to fixed constant precision , which implies that @xmath345 .
\(ii ) if also @xmath346 , then we can compute @xmath29 from @xmath123 , up to fixed constant precision , which implies that @xmath347 .
\(i ) since @xmath34 is total , we can run @xmath348 on all strings @xmath349 in lexicographical length - increasing order .
since @xmath34 is total we will find a shortest string @xmath350 such that @xmath351 .
set @xmath352 . since @xmath353 , and by assumption , @xmath354 ,
we now can compute @xmath355 .
\(ii ) follows from item ( i ) .
given an oracle that on query @xmath0 answers with a sufficient statistic @xmath123 of @xmath0 and a @xmath356 as required below .
then , we can compute the kolmogorov complexity function @xmath357 and the halting sequence @xmath358 . by lemma [ lem.compks ]
we can compute the function @xmath29 , up to fixed constant precision , given the oracle ( without the value @xmath359 ) in the statement of the theorem .
let @xmath359 in the statement of the theorem be the difference between the computed value and the actual value of @xmath29 . in @xcite ,
exercise 2.2.7 on p. 175 , it is shown that if we can solve the halting problem for plain turing machines , then we can compute the ( plain ) kolmogorov complexity , and _
vice versa_. the same holds for the halting problem for prefix turing machines and the prefix turing complexity .
this proves the theorem .
[ lem.chisoph ] there is a constant @xmath103 , such that for every @xmath0 there is a program ( possibly depending on @xmath0 ) of at most @xmath103 bits that computes @xmath360 and the witness program @xmath34 from @xmath361 .
that is , @xmath362 . with some abuse of notation
we can express this as @xmath363 .
by definition of sufficient statistic @xmath123 , we have @xmath364 . by ( [ eq.pcondx ] ) the number of sufficient statistics for @xmath0 is bounded by an independent constant , and we can generate all of them from @xmath0 by a @xmath212 length program ( possibly depending on @xmath0 ) . then , we can simply determine the least length of a sufficient statistic , which is @xmath360 .
there is a subtlety here : lemma [ lem.chisoph ] is nonuniform .
while for every @xmath0 we only require a fixed number of bits to compute the sophistication from @xmath361 , the result is nonuniform in the sense that these bits may depend on @xmath0 . given a program , how do we verify if it is the correct one ? trying all programs of length up to a known upper bound , we do nt know if they halt or if they halt they halt with the correct answer .
the question arising is if there is a single program that computes the sopistication and its witness program for all @xmath0 . in @xcite
this much more difficult question is answered in a strong negative sense : there is no algorithm that for every @xmath0 , given @xmath361 , approximates the sophistication of @xmath0 to within precision @xmath365 . for every @xmath0 of length @xmath228 , and @xmath123 the program that witnesses the sophistication of @xmath0
, we have @xmath366 . for every length @xmath228 , there are strings @xmath0 of length @xmath228 , such that @xmath367 .
let @xmath123 witness the @xmath360 : that is , @xmath364 , and @xmath368 . using the conditional version of ( [ eq.soi ] ) ,
see @xcite , we find that @xmath369 @xmath370 in lemma [ lem.compks ] , item ( i ) , we show @xmath371 , hence also @xmath372 . by lemma [ lem.chisoph ] , @xmath373 ,
hence also @xmath374 .
substitution of the constant terms in the displayed equation shows @xmath375 this shows that the shortest program to retrieve @xmath123 from @xmath0 is essentially the same program as to retrieve @xmath52 from @xmath0 or @xmath29 from @xmath0 . using , this shows that @xmath376 since @xmath123 is the witness program for @xmath368 , we have @xmath377 .
a function @xmath34 from the rational numbers to the real numbers is _ upper semicomputable _ if there is a recursive function @xmath378 such that @xmath379 and @xmath380 .
here we interprete the total recursive function @xmath381 as a function from pairs of natural numbers to the rationals : @xmath382 .
if @xmath34 is upper semicomputable , then @xmath383 is _
lower semicomputable_. if @xmath34 is both upper - a and lower semicomputable , then it is _ computable_. recursive functions are computable functions over the natural numbers . since @xmath384 is upper semicomputable , @xcite , and from @xmath384 we can compute @xmath360 , we have the following : \(i ) the function @xmath360 is not computable to any significant precision .
\(ii ) given an initial segment of length @xmath385 of the halting sequence @xmath342 , we can compute @xmath360 from @xmath0 .
that is , @xmath386 .
\(i ) the fact that @xmath360 is not computable to any significant precision is shown in @xcite .
\(ii ) we can run @xmath387 for all ( program , argument ) pairs such that @xmath388 .
( not @xmath6 since we are dealing with self - delimiting programs . )
if we know the initial segment of @xmath358 , as in the statement of the theorem , then we know which ( program , argument ) pairs halt , and we can simply compute the minimal value of @xmath389 for these pairs .
`` sophistication '' is the algorithmic version of `` minimal sufficient statistic '' for data @xmath0 in the model class of total recursive functions .
however , the full stochastic properties of the data can only be understood by considering the kolmogorov structure function @xmath390 ( mentioned earlier ) that gives the length of the shortest two - part code of @xmath0 as a function of the maximal complexity @xmath68 of the total function supplying the model part of the code .
this function has value about @xmath6 for @xmath68 close to 0 , is nonincreasing , and drops to the line @xmath29 at complexity @xmath391 , after which it remains constant , @xmath392 for @xmath393 , everything up to a logarithmic addive term .
a comprehensive analysis , including many more algorithmic properties than are analyzed here , has been given in @xcite for the model class of finite sets containing @xmath0 , but it is shown there that all results extend to the model class of computable probability distributions and the model class of total recursive functions , up to an additive logarithmic term .
the author thanks luis antunes , lance fortnow , kolya vereshchagin , and the referees for their comments
. 99 a.r .
barron , j. rissanen , and b. yu , the minimum description length principle in coding and modeling , _ ieee trans . inform . theory _ , it-44:6(1998 ) , 27432760 .
cover , kolmogorov complexity , data compression , and inference , pp .
2333 in : _ the impact of processing techniques on communications _ , j.k .
skwirzynski , ed . ,
martinus nijhoff publishers , 1985 .
t.m . cover and j.a .
thomas , _ elements of information theory _ , wiley , new york , 1991 .
r. a. fisher , on the mathematical foundations of theoretical statistics , _ philosophical transactions of the royal society of london , ser .
a _ , 222(1922 ) , 309368 .
p. gcs , on the symmetry of algorithmic information , _
soviet math .
_ , 15 ( 1974 ) 14771480 .
correction : ibid .
, 15 ( 1974 ) 1480 .
p. gcs , j. tromp , and p. vitnyi , algorithmic statistics , _ ieee trans .
inform . theory _
, 47:6(2001 ) , 24432463 .
q. gao , m. li and p.m.b .
vitnyi , applying mdl to learn best model granularity , _ artificial intelligence _ , 121(2000 ) , 129 .
m. gell - mann , _ the quark and the jaguar _ , w. h. freeman and company , new york , 1994 .
grnwald and p.m.b .
vitnyi , shannon information and kolmogorov complexity , manuscript , cwi , december 2003 .
kolmogorov , three approaches to the quantitative definition of information , _ problems inform . transmission _ 1:1 ( 1965 )
complexity of algorithms and objective definition of randomness .
a talk at moscow math .
meeting 4/16/1974 .
abstract in _ uspekhi mat .
nauk _ 29:4(1974),155 ( russian ) ; english translation in @xcite .
kolmogorov , on logical foundations of probability theory , pp .
15 in : _ probability theory and mathematical statistics _ ,
notes math .
1021 , k. it and yu.v .
prokhorov , eds .
, springer - verlag , heidelberg , 1983 .
kolmogorov and v.a .
uspensky , algorithms and randomness , _ siam theory probab
. appl . _ , 32:3(1988 ) , 389412 .
m. koppel , complexity , depth , and sophistication , _ complex systems _ , 1(1987 ) , 10871091 m. koppel , structure , _ the universal turing machine : a half - century survey _ , r. herken ( ed . ) , oxford univ . press , 1988 , pp .
435452 .
m. li and p. vitanyi , _ an introduction to kolmogorov complexity and its applications _ , springer - verlag , new york , 1997 ( 2nd edition ) .
the mathematical theory of communication .
, 27:379423 , 623656 , 1948 .
coding theorems for a discrete source with a fidelity criterion . in _ ire
national convention record , part 4 _ , pages 142163 , 1959 .
shen , the concept of @xmath394-stochasticity in the kolmogorov sense , and its properties , _
soviet math .
_ , 28:1(1983 ) , 295299 .
shen , discussion on kolmogorov complexity and statistical analysis , _ the computer journal _ , 42:4(1999 ) , 340342 .
vereshchagin and p.m.b .
vitnyi , kolmogorov s structure functions and model selection , _ ieee trans .
_ , to appear .
vereshchagin and p.m.b .
vitnyi , rate distortion theory for individual data , draft , cwi , 2004 .
vitnyi and m. li , minimum description length induction , bayesianism , and kolmogorov complexity , _ ieee trans .
inform . theory _ ,
it-46:2(2000 ) , 446464 . v.v .
vyugin , on the defect of randomness of a finite object with respect to measures with given complexity bounds , _ siam theory probab . appl .
_ , 32:3(1987 ) , 508512 . v.v .
vyugin , algorithmic complexity and stochastic properties of finite binary sequences , _ the computer journal _ , 42:4(1999 ) , 294317 .
paul m.b . vitnyi is a fellow of the center for mathematics and computer science ( cwi ) in amsterdam and is professor of computer science at the university of amsterdam .
he serves on the editorial boards of distributed computing ( until 2003 ) , information processing letters , theory of computing systems , parallel processing letters , international journal of foundations of computer science , journal of computer and systems sciences ( guest editor ) , and elsewhere .
he has worked on cellular automata , computational complexity , distributed and parallel computing , machine learning and prediction , physics of computation , kolmogorov complexity , quantum computing .
together with ming li they pioneered applications of kolmogorov complexity and co - authored `` an introduction to kolmogorov complexity and its applications , '' springer - verlag , new york , 1993 ( 2nd edition 1997 ) , parts of which have been translated into chinese , russian and japanese . | the information in an individual finite object ( like a binary string ) is commonly measured by its kolmogorov complexity .
one can divide that information into two parts : the information accounting for the useful regularity present in the object and the information accounting for the remaining accidental information .
there can be several ways ( model classes ) in which the regularity is expressed .
kolmogorov has proposed the model class of finite sets , generalized later to computable probability mass functions .
the resulting theory , known as algorithmic statistics , analyzes the algorithmic sufficient statistic when the statistic is restricted to the given model class .
however , the most general way to proceed is perhaps to express the useful information as a recursive function .
the resulting measure has been called the `` sophistication '' of the object .
we develop the theory of recursive functions statistic , the maximum and minimum value , the existence of absolutely nonstochastic objects ( that have maximal sophistication all the information in them is meaningful and there is no residual randomness ) , determine its relation with the more restricted model classes of finite sets , and computable probability distributions , in particular with respect to the algorithmic ( kolmogorov ) minimal sufficient statistic , the relation to the halting problem and further algorithmic properties .
_ index terms_ constrained best - fit model selection , computability , lossy compression , minimal sufficient statistic , non - probabilistic statistics , kolmogorov complexity , kolmogorov structure function , sufficient statistic , sophistication |
particle interferometry based on the hanbury - brown twiss ( hbt ) effect has been used extensively in heavy ion collisions to extract the information on the emission source of particles @xcite . in particular , the long emission time as a result of the phase transition from the quark - gluon plasma to hadronic matter in relativistic heavy ion collisions may lead to an emission source which has a much larger radius in the direction of the total transverse momentum of detected two particles ( @xmath2 ) than that perpendicular to both this direction and the beam direction ( @xmath3 ) @xcite . since the quark - gluon plasma is expected to be formed in heavy ion collisions at rhic , it is surprising to find that the extracted ratio @xmath4 from a gaussian fit to the measured two - pion correlation function in au+au collisions at @xmath5 gev is close to one @xcite , very different from predictions of hydrodynamical models @xcite .
since particle interferometry probes the phase - space distributions of particles at freeze - out , it is natural to apply transport models to hbt .
the reason is that particle freeze - out is dynamically generated in transport models when the mean - free - path exceeds the system size at later stage of expansion , whereas freeze - out has to be imposed in hydrodynamical models . using a multi - phase transport ( ampt ) model , we have found that the small pion @xmath1 ratio could be due to a large and positive space - time correlation in the emission source @xcite .
furthermore , the pion source at freeze - out is highly non - gaussian , leading to much larger pion source radii than the radius parameters from a gaussian fit to the three - dimensional correlation function . in this study
, we extend the work of ref.@xcite by studying the kaon interferometry in central au+au collisions at rhic energies . using the ampt model ,
we find that , unlike the pion case , the kaon source radii extracted directly from the emission function are close to the fitted radius parameters extracted from a gaussian fit to the three - dimensional correlation function . our results also show that the kaon emission source has a large and positive correlation between time and position along the @xmath6-direction at freeze - out , similar to what we have found earlier for the pion emission source .
we expect that the study of kaon interferometry , as well as other observables such as the strange hadron elliptic flow , will be useful in understanding the dynamics of strange quarks and hadrons in heavy ion collisions at rhic .
the ampt model is a hybrid model that uses the minijet partons from hard processes and excited strings from soft processes in the hijing model @xcite for the initial condition of relativistic heavy ion collisions .
the time evolution of partons is then described by the zpc @xcite parton cascade model , and that of hadrons by an extended art model @xcite . in the default ampt model , only minijet partons are included in the parton cascade with a parton scattering cross section of @xmath7 mb .
after partons freeze out , they combine with their parent strings and then fragment to hadrons according to the lund string fragmentation as implemented in pythia @xcite .
the default model has been quite reasonable in describing the measured rapidity distributions of charge particles @xcite , particle to antiparticle ratios @xcite , and the spectra of low transverse momentum pions , kaons @xcite , multi - strange baryons @xcite and @xmath8 mesons @xcite in heavy ion collisions at sps and/or rhic . since the initial energy density in au+au collisions at rhic is expected to be much larger than the critical energy density at which the transition from hadronic matter to quark - gluon plasma would occur @xcite , the ampt model has been extended to allow the conversion of initial excited strings to partons at rhic energies @xcite . in this _ string melting _
scenario , hadrons that would have been produced from string fragmentation are converted instead to valence quarks and/or antiquarks .
interactions among these partons are again described by the zpc parton cascade model .
the transition from the partonic matter to hadronic matter at parton freeze - out is achieved using a simple quark coalescence model by combining two nearest partons into mesons and three nearest partons into baryons ( or anti - baryons ) @xcite . with the energy in excited strings taking part in the early partonic interactions and using quark coalescence to model hadronization ,
the extended ampt model with string melting @xcite is able to describe the observed elliptic flow at rhic @xcite , which the default ampt model failed to reproduce . at present , the zpc parton cascade @xcite in the ampt model includes only two - parton elastic scatterings .
the in - medium differential cross section is given by @xmath9 , where the effective screening mass @xmath10 in principle depends on the temperature and density of the partonic matter@xcite .
in this study , we take @xmath10 as a parameter in order to study the effect of partonic scatterings .
also , for simplicity , we assume the same scattering cross section for partons of different flavors .
we note , however , that comparisons of high - quality data on the elliptic flow of strange hadrons @xcite with theoretical predictions @xcite is expected to provide very useful information on the interactions of strange quarks in dense matter .
to evaluate the two - kaon correlation function requires the knowledge of the single kaon emission function @xmath11 . in the ampt model , it is obtained from the kaon space - time coordinate @xmath12 and momentum @xmath13 at kinetic freeze - out .
the hbt correlation function for two identical hadrons of momenta @xmath14 and @xmath15 is then given by @xcite @xmath16 } { \int d^4x_1 s(x_1,{\bf p_1 } ) \int d^4x_2 s(x_2,{\bf p_2 } ) } , \label{emission}\end{aligned}\ ] ] where @xmath17 and @xmath18 .
expecting that the emission function is sufficiently smooth in momentum space , one can evaluate the correlation function by using @xmath14 and @xmath15 for @xmath19 in the numerator of above equation .
first , the size of the emission source can be extracted from the emission function via the curvature of the correlation function at @xmath20 : @xmath21 in the above , @xmath22 with @xmath23 denotes the average energy of the two kaons ; @xmath24 are spatial coordinates of a kaon at freeze - out ; and @xmath25 with @xmath26 denoting the average value of @xmath12 . in this study
we use the usual `` out - side - long '' ( @xmath27 ) coordinate system @xcite . on the other hand ,
the measured correlation function @xmath28 is usually fitted by a gaussian function in @xmath29 , i.e. , @xmath30 we note that , for central heavy - ion collisions , the above fitted radius parameters would be identical to the source radii given by the curvature of the emission function at @xmath20 ( i.e. eq .
( [ source ] ) ) only when the emission source is gaussian in space and time .
using the program correlation after burner @xcite , we have evaluated the correlation function @xmath28 of two kaons in their longitudinally comoving frame for central ( @xmath31 fm ) au+au collisions at @xmath32 gev . in figure 1 , we show the invariant correlation function , i.e. , the correlation function as a function of @xmath33 , and its projections onto one of the @xmath34 and @xmath35 axes . in evaluating the projected correlation function ,
the other two @xmath29-components have been integrated over the range @xmath36 mev/@xmath37 .
the dash - dotted curves in figure 1 represent results from the default ampt model ( no string melting ) , while the other curves are from the extended ampt model with string melting but using different values for @xmath38 .
it is seen that the one - dimensional kaon correlation functions become narrower as @xmath39 is increased in the extended ampt model .
their dependence on @xmath39 seems , however , to be much weaker than that of the pion correlation function @xcite .
the source radii for kaons within @xmath40 and @xmath41 mev@xmath42 are shown in figure 2 ( a ) .
the results are obtained from both the default ampt model ( shown at @xmath43 ) and the extended ampt model with string melting . for the latter ,
we have used different parton cross sections of @xmath44 , 6 , 10 , and 15 mb .
it is seen that these radii have values between 2 and 5 fm , and they are much smaller than the source radii for low @xmath45 pions ( between 7 and 25 fm ) @xcite . the @xmath1 ratio ( solid curves without symbols ) from the kaon emission function
is also shown in figure 2 ( a ) , and it increases appreciably with increasing parton scattering cross section . [ cols= " < , > " , ] figure 3 ( a ) shows the extended ampt results with @xmath46 mb for kaon source radii as functions of @xmath47 in the six @xmath45-bins , @xmath48 gev@xmath42 , for au+au collisions at @xmath5 gev .
we see that for all @xmath47 the source radii ( solid ) are close to the corresponding fitted radius parameters ( filled squares ) .
we note that , before making quantitative predictions on kaon correlations , abundances of strange resonances such as @xmath49 and @xmath8 need to be checked , and the effects of other resonances not yet included in the ampt model also need to be investigated .
figure 3 ( b ) and ( c ) show the @xmath51 and @xmath52 distributions at freeze - out for @xmath53 with @xmath41 mev@xmath42 and @xmath40 .
the kaon emission source shows a positive shift in @xmath54 as in the pion case @xcite , consistent with a strong transverse flow .
the emission source also appears to be highly non - gaussian , which leads to different values of radius parameters extracted from the two methods in section [ khbt ] @xcite .
the solid curve with open diamonds in figure 3 ( c ) gives the average value @xmath55 as a function of freeze - out time @xmath56 .
the kaon emission source is seen to have a strong positive @xmath57 correlation as in the pion emission source @xcite . since @xmath58 , but @xmath59 the ratio @xmath4 contains information about the duration of emission ( in the last term ) and has been studied extensively @xcite .
we note , however , that a direct relation between @xmath2 and the emission duration exists only if the @xmath57 correlation term @xmath60 is small , which is not the case according to our results from the extended ampt .
e.g , the above equation for the kaons is numerically written as @xmath61 in units of @xmath62 .
the @xmath57 correlation term @xmath60 is thus positive ( @xmath63 @xmath62 ) and comparable to the magnitude of @xmath64 ( @xmath65 @xmath62 ) , making it difficult to extract information about the duration of emission from @xmath4 .
the situation is similar for the pion emission source , e.g. , for mid - rapidity pions with @xmath66 mev@xmath42 , numerically the corresponding eq .
( [ rout ] ) is @xmath67 in units of @xmath62 .
figure 4 shows the correlation functions for both neutral and charged kaons within @xmath40 and @xmath41 mev@xmath42 from ampt model with string melting and @xmath46 mb for @xmath5 and @xmath68 gev .
the effect due to coulomb interactions is included for @xmath69 correlation functions using the program correlation after burner @xcite .
the kaon correlation functions are found to change only slightly from @xmath70 to @xmath68 gev at rhic .
we appreciate stimulating discussions with h. huang , m. murray , s. pratt and n. xu .
this work was supported by the u.s .
department of energy grant no .
de - fg02 - 01er41190 ( z.w.l . ) and by the u.s .
national science foundation grant no .
phy-0098805 as well as the welch foundation under grant no .
a-1358 ( c.m.k . )
99 pratt s .
bertsch g , gong m and tohyama m .
pratt s , csrg , t and zimanyi j .
rischke d h and gyulassy m .
soff s , bass s a and dumitru a .
adler c _ et al _ ( star collaboration ) .
johnson s c ( phenix collaboration ) .
et al _ ( phenix collaboration ) .
lin z w , ko c m and pal s .
wang x n and gyulassy m .
li b a and ko c m .
sjostrand t .
zhang b _ et al _ .
pal s , ko c m and lin z w .
pal s , ko c m and lin z w .
zhang b _ et al _ .
lin z w and ko c m .
ackermann k h _ et al _
( star collaboration ) .
adler c _ et al _ ( star collaboration ) .
adler c _ et al _ ( star collaboration ) ; _ ibid _ * 89 * 132301 .
sorensen p .
lin z w and ko c m .
voloshin s a .
molnar d and voloshin s a .
lin z w and molnar d .
greco v , ko c m and levai p .
wiedemann u a and heinz u w .
pratt s _ et al _ .
hardtke d and voloshin s a .
bernard s _
et al _ , ; . | the two - kaon interferometry at rhic is studied in a multi - phase transport model .
similar to the pion case , we find strong space - time correlation at freeze - out for the kaon emission source , which results in a large positive @xmath0 term and tends to reduce the @xmath1 ratio .
unlike the pion case , the source radii for kaons determined from the emission function are close to the radius parameters extracted from a gaussian fit to the correlation function . . |
extensible markup language ( xml ) has reached a great success in the internet era .
xml documents are similar to html documents , but do not restrict users to a single vocabulary , which offers a great deal of flexibility to represent information . to define the structure of documents within a certain vocabulary , schema languages such as _ document type definition _
( dtd ) or _ xml schema _ are used .
xml has been adopted as the most common form of encoding information exchanged by web services @xcite . @xcite
attribute this success to two reasons .
the first one is that the xml specification is accessible to everyone and it is reasonably simple to read and understand .
the second one is that several tools for processing xml are readily available .
we add to these reasons that as xml is _ vocabulary - agnostic _ , it can be used to represent data in basically any domain .
for example , we can find the _
universal business language _ ( ubl ) in the business domain , or the standards defined by the _
open geospatial consortium _
( ogc ) in the geospatial domain .
ubl defines a standard way to represent business documents such as electronic invoices or electronic purchase orders .
ogc standards define _ web service interfaces _ and _ data encodings _ to exchange geospatial information .
all of these standards ( ubl and ogc s ) have two things in common .
the first one is that they use xml schema to define the structure of xml documents .
the second one is that the size and complexity of the standards is very high , making very difficult its manipulation or implementation in certain scenarios @xcite .
the use of such large schemas can be a problem when xml processing code based on the schemas is produced for a resource - constrained device , such as a mobile phone .
this code can be produced using a manual approach , which will require the low - level manipulation of xml data , often producing code that is hard to modify and maintain .
another option is to use an xml data binding code generator that maps xml data into application - specific concepts .
this way developers can focus on the semantics of the data they are manipulating @xcite .
the problem with generators is that they usually make a straightforward mapping of schema components to programming languages constructs that may result in a binary code with a very large size that can not be easily accommodated in a mobile device @xcite .
although schemas in a certain domain can be very large this does not imply that all of the information contained on them is necessary for all of the applications in the domain .
for example , in @xcite a study of the use of xml in a group of 56 servers implementing the _ _ ogc s sensor observation service ( sos ) specification _ _ revealed that only 29.2% of the sos schemas were used in a large collection of xml documents gathered from those servers . based on this information we proposed in @xcite an algorithm to simplify large xml schema sets in an application - specific manner by using a set of xml documents conforming to these schemas .
the algorithm allowed a 90% reduction of the size of the schemas for a real case study .
this reduction was translated in a reduction of binary code ranging between 37 to 84% when using code generators such as jaxb , xmlbeans and xbinder . in this paper
we extend the schema simplification algorithm presented in @xcite to a more complete _ instance - based xml data binding _ approach .
this approach allows to produce very compact application - specific xml processing code for mobile devices . in order to make the code as small as possible
the approach will use , similarly to @xcite , a set of xml documents conforming to the application schemas . from these documents , in addition to extract the subset of the schemas that is needed , we extract other relevant information about the use of schemas that can be utilised to reduce the size of the final code . a prototype implementation targeted to android and the java programming language has been developed .
the remainder of this paper is structured as follows .
section 2 presents an introduction to xml schema and xml data binding . in section 3 ,
related work is presented .
the _ instance - based data binding approach _ is presented in section 4 .
section 5 overviews some implementation details and limitations found during the development of the prototype .
section 6 presents experiments to measure size an execution times of the code generated by the tool in a real scenario .
last , conclusions and future work are presented .
in this section we present a brief introduction to the topics of xml schema and xml data binding .
xml schema files are used to assess the validity of well - formed element and attribute information items contained in xml instance files @xcite@xcite .
the term xml data binding refers to the idea of taking the information in an xml document and convert it to instances of application objects @xcite .
an xml schema document contains components in the form of complex and simple type definitions , element declarations , attribute declarations , group definitions , and attribute group definitions .
this language allows users to define their own types , in addition to a set of predefined types defined by the language .
elements are used to define the content of types and when global , to define which of them are valid as top - level element of an xml document .
xml schema provides a derivation mechanism to express subtyping relationships .
this mechanism allows types to be defined as subtypes of existing types , either by extending or restricting the content model of base types .
apart from type derivation , a second subtyping mechanism is provided through substitution groups .
this feature allows global elements to be substituted by other elements in instance files .
a global element e , referred to as _ head element _ , can be substituted by any other global element that is defined to belong to the e s substitution group . with _ xml data binding _
, an abstraction layer is added over the raw xml processing code , where xml information is mapped to data structures in an application data model .
xml data binding code is often produced by using code generators that use a description of the structure of xml documents using some schema language .
the use of generators potentially gives benefits such as increased productivity , consistent quality throughout all the generated code , higher levels of abstraction as we usually work with an abstract model of the system ; and the potential to support different programming languages , frameworks and platforms @xcite .
although most of the generators available nowadays are targeted to desktop or server applications , several tools have been develop for mobile devices such as xbinder and codesysnthesis xsd / e , or for building complete web services communication end - points for resource constrained environments , such as gsoap @xcite .
all of the tools mentioned before map xml schema structures to programming languages construct in a straightforward way , which is not adequate when large schemas sets are used .
problems related with having large and complex schemas have been presented in several articles @xcite .
for example , @xcite deal with problems of large schemas in schema matching in the business domain . in the context of schema and ontology mapping ,
@xcite states that current match systems still struggle to deal with large - scale match tasks to achieve both good effectiveness and good efficiency .
@xcite , the work extended here , expose the problems related to using xml data binding tools to generate xml processing code for mobile geospatial applications .
last , @xcite present an algorithm to extract fragments of large conceptual schemas arguing that the largeness of these schemas makes difficult the process of getting the knowledge of interest to users .
when considering xml processing in the context of mobile devices , literature is focused in two main competing requirements : _ compactness _ ( of information ) and _ processing efficiency _ @xcite . to achieve compactness compression techniques
are used to reduce the size of xml - encoded information @xcite . about processing efficiency
, not much work has been done in the mobile devices field .
a prominent exception in this topic is the work presented in @xcite , @xcite and @xcite .
these articles are all related to the implementation of a middleware platform for mobile devices : the _ fuego mobility middleware _ @xcite , where xml processing has a large impact . the proposed _
xml stack _ provides a general - purpose xml processing api called _ xas _
@xcite , an xml binary format called _ xebu _ @xcite , and others apis such as _ trees - with - references _ ( reftrees ) and _ random access xml store _
( raxs)@xcite . regarding the use of instance files to drive the manipulation of schemas
, @xcite presents a review of different methods that use instance files for ontology matching . in the field of schema inference ,
instance files are used as well to generated adequate schema files that can be used to assess their validity ( e.g. @xcite ) .
_ instance - based xml data binding _ , is a two - step process .
the first step , _ instance - based schema simplification _ , extracts the information about how schema components are used by a specific application , based on the assumption that a representative subset of xml documents that must be manipulated by the application is available .
the second step , _ code generation _
, consists of using all of the information extracted in the previous step to generate xml processing code as optimised as possible for a target platform .
the whole process is shown in figure [ fig : flow - xmldatabinding ] , the inputs to the first step are a set of schemas and a set of xml documents conforming to them .
the outputs will be the subset of the schemas used by the xml documents and other information about the use of certain features of the schemas that can be used to optimise the code in the following step .
the outputs of the first step are the inputs of the code generation step .
the two steps of the process are detailed in the following subsections .
+ the _ instance - based schema simplification _
step extracts the subset of the schemas used on a set of xml documents .
the algorithm used to perform this simplification was first presented in @xcite and has been extended here to extracts other information that can be used to produce more compact xml processing code .
the idea behind this algorithm , is depicted graphically in figure [ fig : simplificationalg ] .
the figure shows to the left the graph of relationships between schemas components .
the different planes represent different namespaces .
links between schema components represent dependencies between them . to the right
we have the tree of information items ( xml nodes ) contained in xml documents . for the sake of simplicity
we show in the figure only the tree of nodes corresponding to a single document .
an edge between an xml node and a schema component represents that the component describes the structure of the node . to simplify the figure we have shown only a few edges ,
although an edge for every xml node must exist .
starting from a set of xml documents and the schema files defining their structure , it is possible to calculate which schema components are used and which are not . in doing so , the following information is also recorded : * _ types that are instanced in xml documents _ : for each xml node exists a schema type describing its structure .
while xml documents are processed the type of each xml node is recorded .
this way we can know which types are instanced and which are not . * _ types and elements substitutions _ : the subtyping mechanisms mentioned in section 2.1 allow the _ real _ or _ dynamic type _ of an element to be different from its _ declared type_. elements declared as having type a , may have any type derived from a in an xml document . in this case
the real type must be specified with the attribute _ xsi : type_. something similar happens with substitution groups , although in this case the attribute _ xsi : type _ is not necessary .
the information about xml nodes whose dynamic type is different from its declared type is recorded . * _ wildcards substitutions _ : the elements used to substitute wildcards are recorded . *
_ elements occurrence constraints information _ : for all of the elements it is checked that if they allow multiple occurrences there is at least one document where several occurrences of the element are present . *
_ elements with a single child _ : all of the elements that contain a single child are also recorded .
+ a more detailed view of the code generation process is shown in figure [ fig : flow_gen ] .
the outputs of the schema simplification step are used as inputs to the _ schema processor _ , the component of the generator in charge of creating the data model that will be used later by the _ template engine_. the _ template engine _ combines pre - existing _ class templates _ with the data model to generate the final source code .
the use of a template engine allows the generation of code for other platforms and programming languages by just defining new class templates .
+ a summary of the features of the code generation process that contribute to the generation of optimised code is listed next : * _ use of information extracted from xml documents _ : the use of information about schema use allows to apply the following optimisations : * * _ remove unused schema components _ : the schema components that are not used are not considered for code generation . by removing the unused components we can substantially reduce the size of the generated code .
the amount of the reduction will depend on how specific applications make use of the original schemas .
* * _ efficient handling of subtyping and wildcards _ : the number of possible substitutions of a type by its subtypes , and a head element of a substitution group by the members of the group can be bounded with the information gathered from the instances files . in the general case , where no instance - based information is available , generic code to face any possible type or element substitution must be written .
limiting the number of possible substitutions to only a few allows the production of simpler and faster code .
the same reasoning is applied to wildcards .
* * _ inheritance flattening _ : by flattening subtyping hierarchies for a given type , i.e. , including explicitly in its type definition all of the fields inherited from base types and eliminating the subtype relationship with its parents , we can reduce the number of classes in the generated code .
the application of this technique will not necessarily result in smaller generated code , as the fields defined in base types must be replicated in all of their child types , but it will have a positive impact in the work of the class loader because a lower number of classes have to be loaded while the application is executed .
let us consider the case of the geospatial schemas introduced in section 1 .
these schemas typically present deep subtyping hierarchies with six or more levels , as a consequence when an xml node of a type in the lowest levels of the hierarchy must be processed , all of its parent types must be loaded first .
the technique of inheritance flattening has been widely explored and used in different computer science and engineering fields as is proven by the abundant literature found in the topic @xcite . * * _ adjust occurrence constraints _ : if an element is declared to have multiple occurrences it must be mapped to a data structure in the target programming language that allows the storage of the multiple instances of the elements , e.g. an array or a linked list . in practice
if the element has at most one occurrence in the xml documents that must be processed by the application it can be mapped to a single object instance .
using this optimisation the final code will make a better use of memory because instead of creating a collection ( array , linked list , etc . )
that will only contain a single object , it creates a single object instance . *
_ collapse elements containing single child elements _
: information items that will always contain single elements can be replaced directly by its content . by applying this optimisation we can reduce the number of classes in the generated code , which will have a positive impact in the size of the final code , the amount of work that has to be done by the class loader , and the use of memory during execution .
this optimization is used by mainstream xml data binding tools such as jibx and the xml schema definition tool . * _ disabling parsing / serialization operations as needed _ : some code generators always includes code for parsing and serialization even when only one of these functions is needed . for example , in the context of geospatial web services , most of the time spent in xml processing by client - side applications is dedicated to parsing , as messages received from the servers are potentially large . on the other hand , most of these services allows request to be sent to the server encoded in an http get request , therefore xml serialisation is not needed at all . * _ ignoring sections of xml documents _ : frequently , we are not interested in all of the information contained in xml files , ignoring the unneeded portions of the file will improve the speed of the parsing process and it may have a significant impact in the amount of memory used by the application .
in addition , the following features not related directly with code optimisation are also supported : * _ source code based on simple code patterns _ : the generated source code is straightforward to understand and modify in case it is necessary . * _ tolerate common validation errors : _ occasionally , xml documents that are not valid against their respective schemas must be processed by our applications . in many cases ,
the validation errors can be ignored following simple coding rules .
a detailed explanation of each of the features presented in this section can be found in @xcite .
as mentioned before , the approach presented in this paper is based on the assumption that a representative set of xml documents exists . by _ representative
we mean that these documents contain instances of all of the possible xml schema elements and types that will be processed by the application in the future . nevertheless , this subset might not always be available . in this case
, we can still take advantage of the approach by building _
synthetic _ xml documents containing relevant information .
whether xml processing code is produced manually or automatically developers typically have some knowledge of the structure of the documents that must be processed by the applications .
therefore , we can use this knowledge to build sample xml documents that can be used as input to the algorithm . in case
it were necessary , the final code can be manually modified later , or the sample files changed and used to regenerate the code . if we were using synthetic documents instead of actual documents some of the optimisations related to the information extracted from them should not be applied .
the reason for this is that we do not have enough information about how the related schema features are used .
for example , we can not apply optimisations such as the efficient handling of subtyping and wildcards , as we might not know all of the possible type substitutions .
something similar happens with the adjustment of occurrence constraints .
nevertheless , other optimisations such as inheritance flattening or removing unused schema components can be still safely applied .
_ dbmobilegen _ ( dbmg for short ) is the current implementation of the _ instance - based xml data binding _
approach @xcite .
it includes components implementing both the simplification algorithm and code generation process .
it is implemented in java and relies on existing libraries such as _ _ eclipse xsd _ _ for processing xml schemas , _ _ freemarker _ _ as template engine library , and as well as the generated code , _ _ kxml _ _ for low - level xml processing .
this tool produces code targeted to android mobile devices and the java programming language .
the current implementation has some limitations . because of the complexity of the xml schema language itself
, support for certain features and operations have been only included if it is considered necessary for the case study or applications where the tool has been used @xcite .
some of these limitations are listed next : * _ serialization is not supported yet _ : the role of parsing for our sample applications and case studies is far more important than serialization .
this is mainly because we have preferred to use http get to issue server requests wherever possible . *
_ dynamic typing using xsi : type not fully supported _ : the mechanism of dynamic type substitution by using the _ xsi : type _ attribute has not been fully implemented yet , as the xml documents processed in the applications developed so far do not use this feature .
in this section we present two experiments .
the first one tries to test how much the size of the generated code can be reduced by using dbmg .
the second one measures the execution times of generated code in a mobile phone . in this experiment
we borrow the test case presented in @xcite that implements the communication layer for an sos client .
sos is a standard web service interface defined to enhance interoperability between sensor data producers and consumers @xcite .
the sos schemas are among the most complex geospatial web service schemas as they are comprised of more than 80 files and they contain more than 700 complex types and global elements @xcite .
the client must process data retrieved from a server that contains information about air quality for the valencian community .
this information is gathered by 57 control stations located in that area .
the stations measure the level of different contaminants in the atmosphere .
a set of 2492 xml documents was gathered from the server to be used as input , along with the sos schemas , to the instance - based data binding process .
the source code generated by dbmg is compiled to the _ compressed jar _ format and compared with the final code generated by other generators : xbinder , jaxb and xmlbeans .
the last two are not targeted to mobile devices but are used here as reference to compare the size of similar code for other types of applications .
table [ generatedcode2 ] shows the size of the code produced with the different generators from the full sos schemas ( full ) and from the subset of the schemas used in the input instance files ( reduced ) .
the reduced schemas are calculated applying the schema simplification algorithm to the full sos schemas .
the last row of the table ( libs ) includes the size of the supporting libraries needed to execute the generated code in each case .
|p1.5cm| p1.1cm| p1cm| p1.6cm| p1cm| ' '' '' ' '' '' & xbinder & jaxb & xmlbeans & dbmg + full & 3,626 & 754 & 2,822 & 88 + reduced & 567 & 90 & 972 & 88 + libs & 100 & 1,056 & 2,684 & 30 + figure [ fig : codegen_full ] shows the total size of xml processing code when using the full and reduced schemas . in both cases
, we can see the enormous difference that exists between the code generated by dbmg and the code generated by other tools .
+ the size for dbmg is the same in both cases because it implicitly performs the simplification of the schemas before generating source code .
it must be noted that serialisation is not still implemented in dbmg .
we roughly estimate that including serialisation code will increase the final size in about 30% . in any case , the code generated by this tool is about 6 times smaller than the code generated by xbinder from the reduced schemas and about 30 times smaller than the code generated from the full schemas .
one of the reasons for this difference in size is the lack of serialisation support in dbmg .
another reason is that xbinder generates code to ensure all of the restrictions related to user - defined simple types .
this is an advantage if we parse data obtained for a non - trusted source and the application requires the data to be carefully validated , but it is a disadvantage in the opposite case , as unneeded verification increase processor usage and memory footprint . in the case of dbmg , as it aggressively tries to lower final code size , these simple type restrictions are ignored and these types do not even have a counterpart in the generated code .
when compared to jaxb , using the reduced schemas , the main difference in size is in the supporting libraries , as the code generated by jaxb is very simple .
still , the code generated by dbmg is slightly smaller because the step of removing elements with single child elements and inheritance flattening eliminates a large number of classes . in all of the cases ,
xmlbeans has the largest size .
this tool is mostly optimised for speed at the expense of generating a more sophisticated and complex code and the use of bigger supporting libraries . to test the performance of the generated code
we will parse a set of 38 capabilities files obtained from different sos servers .
the code needed to parse these files is generated and deployed to a htc desire android smartphone with a 1 ghz qualcomm quaddragon cpu and 576 mb of ram .
the 38 files have sizes ranging from less than 4 kb to 3.5 mb , with a mean size of 315 kb and a standard deviation of 26.7 kb . as
the size range is large and with the purpose of simplifying presentation we divide the files in two groups , those with a size below 100 kb , caps - s ( 30 files ) , and those with size equal to or higher than 100 kb , caps - l ( 8 files ) . to obtain accurate measures of the execution time for the code we selected the methodology presented in @xcite .
this methodology provides a statistically rigorous approach to deal with all of the non - deterministic factors that may affect the measurement process ( multi - threading , background processes , etc . ) .
as our goal is only to measure the execution times of xml processing code , we stored the files to be parsed locally to avoid interferences related to network delays . besides , to minimise the interference of data transfer delays from the storage medium all of the files below 500 kb were read into memory before being parsed .
it was impossible to do the same for files with sizes above 500 kb because of the device memory restrictions .
figures [ mobile_execution_small ] and [ fig : mobile_execution ] shows the execution times of code generated by dbmg .
the figures also include the execution times needed by _ kxml _ , the underlying parser used by dbmg , to process the same group of files .
the execution times for _ kxml _ were calculated by creating a simple test case where files are processed using this parser , but no action is taken when receiving the events generated by it .
+ + when files below 100 kb are processed it can be observed that the overhead added by the generated code is not high ( figure [ mobile_execution_small ] ) .
nevertheless , we can see in figure [ fig : mobile_execution ] that when file size is above 1 mb , the overhead starts to be important ( > 1s ) .
this happens because the large amount of memory that is required to store the information that is being processed forces the execution of the garbage collector with a high frequency
. we have to keep in mind that code produced manually can have similar problems if it were necessary to retain most of the information read from the xml files in memory . the experiment described above
was extended in @xcite to compare the code generated by dbmg with other data binding tools and to measure also the performance of this code when executed in a windows pc .
the experiments showed that the execution times for the mobile devices were around 30 to 90 times slower than those for the personal computer .
the experiments also showed that the code generated by dbmg was as fast as code generated by other data binding tools for the android platform .
in this paper we have presented an approach to generate compact xml processing code based on large schemas for mobile devices .
it utilises information about how xml documents make use of its associated schemas to reduce the size of the generated code as much as possible .
the solution proposed here is based on the observation that applications that makes use of xml data based on large schemas do not use all of the information included in these schemas . a code generator implementing the approach that produces code targeted to android mobile devices and the java programming language has been developed .
this tool has been tested in a real case study showing a large reduction in the size of the final xml processing code when compared with other similar tools generating code for mobile , desktop and server environments .
nevertheless , this result must be looked at with caution as the magnitude of the reduction will depend directly from the use that specific applications make of their schemas .
this work has been partially supported by the `` espaa virtual '' project ( ref .
cenit 2008 - 1030 ) through the instituto geogrfico nacional ( ign ) ; and project geocloud , spanish ministry of science and innovation ipt-430000 - 2010 - 11 .
d. beyer , c. lewerentz , and f. simon .
impact of inheritance on metrics for size , coupling , and cohesion in object - oriented systems . in _ proceedings of the 10th international workshop on new approaches in software measurement _ ,
iwsm 00 , pages 117 , london , uk , 2000 .
springer - verlag .
c. bogdan chirila , m. ruzsilla , p. crescenzo , d. pescaru , and e. tundrea . towards a reengineering tool for java based on reverse inheritance .
in _ in proceedings of the 3rd romanian - hungarian joint symposium on applied computational intelligence ( saci 2006 ) _ , pages 9637154 , 2006 .
bungartz , w. eckhardt , m. mehl , and t. weinzierl . .
in _ proceedings of the 8th international conference on computational science , part iii _ , iccs 08 , pages 213222 , berlin , heidelberg , 2008 .
springer - verlag .
a. cicchetti , d. d. ruscio , r. eramo , and a. pierantonio .
automating co - evolution in model - driven engineering . in _ proceedings of the 2008 12th international ieee enterprise distributed object computing conference _ , pages 222231 , washington , dc , usa , 2008 .
ieee computer society .
a. georges , d. buytaert , and l. eeckhout .
statistically rigorous java performance evaluation . in _ proceedings of the 22nd annual acm sigplan conference on object - oriented programming systems and applications _ , oopsla 07 , pages 5776 , new york , ny , usa , 2007 .
acm .
s. kbisch , d. peintner , j. heuer , and h. kosch . .
in _ proceedings of the 24th international conference on advanced information networking and applications workshops , waina 10 _ , volume 0 , pages 508513 , los alamitos , ca , usa , 2010 .
ieee computer society .
j. kangasharju , s. tarkoma , and t. lindholm .
xebu : a binary format with schema - based optimizations for xml data . in a.
ngu , m. kitsuregawa , e. neuhold , j .- y .
chung , and q. sheng , editors , _ web information systems engineering - wise 2005 _ , volume 3806 of _ lecture notes in computer science _ , pages 528535 .
springer berlin / heidelberg , 2005 .
e. rahm . towards large - scale schema and ontology matching . in z.
bellahsene , a. bonifati , and e. rahm , editors , _ schema matching and mapping _ , data - centric systems and applications , pages 327 .
springer berlin heidelberg , 2011 .
a. tamayo , c. granell , and j. huerta . .
in _ proceedings of the 2nd international conference and exhibition on computing for geospatial research and application ( com.geo 2011 ) _ , pages 17:117:9 , new york , ny , usa , 2011 .
a. tamayo , c. granell , and j. huerta . .
in _ proceedings of the 2nd international conference and exhibition on computing for geospatial research and application ( com.geo 2011 ) _ , pages 16:116:9 , new york , ny , usa , 2011 .
a. tamayo , p. viciano , c. granell , and j. huerta . . in s. geertman , w. reinhardt , and f. toppen , editors ,
_ advancing geoinformation science for a changing world _ , volume 1 of _ lecture notes in geoinformation and cartography _ , pages 185209 .
springer berlin heidelberg , 2011 .
r. a. van engelen and k. a. gallivan . .
in _ proceedings of the 2nd ieee / acm international symposium on cluster computing and the grid , ccgrid 02 _ , pages 128 , washington , dc , usa , 2002 .
ieee computer society .
a. villegas and a. oliv . a method for filtering large conceptual schemas . in _ proceedings of the 29th international conference on conceptual modeling _ ,
er10 , pages 247260 , berlin , heidelberg , 2010 .
springer - verlag . | xml and xml schema are widely used in different domains for the definition of standards that enhance the interoperability between parts exchanging information through the internet . the size and complexity of some standards , and their associated schemas ,
have been growing with time as new use case scenarios and data models are added to them . the common approach to deal with the complexity of producing xml processing code based on these schemas
is the use of xml data binding generators .
unfortunately , these tools do not always produce code that fits the limitations of resource - constrained devices , such as mobile phones , in the presence of large schemas .
in this paper we present _ instance - based xml data binding _ , an approach to produce compact application - specific xml processing code for mobile devices .
the approach utilises information extracted from a set of xml documents about how the application make use of the schemas .
[ languages and system , standards ] |
consider the lorentzian metric of the five - dimensional space written in a gaussian form @xmath9 where @xmath10 , and latin indices run from @xmath11 to @xmath12 while the greek ones go from @xmath11 to @xmath13 by splitting the vacuum einstein equations in terms of the extrinsic and intrinsic curvatures of the slices @xmath14 , it can be shown that the equations have the following structure : @xmath15 where @xmath16 are analytic functions of their arguments , @xmath17 is the covariant derivative with respect to the induced metric @xmath18 ; @xmath19 and @xmath20 denote , respectively , the scalar curvature and the extrinsic curvature of the hypersurface @xmath21 and @xmath22 . recall that in the coordinates adopted the extrinsic curvature assumes the simple form : @xmath23 it is well known that , owing to the bianchi identities , the second and third equations need to be imposed only on the hypersurface , since they are propagated by the first one . in this sense , it is said that the einstein equations consist of the _ dynamical _ equation ( [ dyn ] ) plus _ constraint equations _
( [ c1 ] ) and ( [ c2 ] ) for @xmath20 and @xmath24 .
let now consider the hypersurface @xmath25 according to the cauchy - kowalewski theorem , for any point in this hypersurface , say the origin , there is an open set in five dimensions containing that point , where the equation ( [ dyn ] ) always has a unique analytic solution @xmath26 provided that the following analytic initial conditions are specified : @xmath27 from the perspective of the embedding problem these initial conditions represent , respectively , the metric and the extrinsic curvature of the hypersurface @xmath28 , whereas the solution of equation ( [ dyn ] ) gives the metric of the @xmath29dimensional space . thus , if there is a solution for the constraint equations for any given metric @xmath30 , then the theorem is proved , since the solution found @xmath31 substituted in ( [ hds2 ] ) will give rise to a metric that satisfies the vacuum einstein equation @xmath32 .
clearly , the embedding map is then given by the equation @xmath33 .
it turns out , as magaard has proved @xcite , that the constraint equations always have a solution .
indeed , by simple counting operation we can see that there are @xmath34 unknown functions ( the independent elements of extrinsic curvature ) and @xmath35 constraint equations .
the metric @xmath36 must be considered as a given datum . for @xmath7 ,
the number of variables is equal or greater than the number of equations .
magaard has shown that after the elimination of equation ( [ c2 ] ) , the first - order differential equation ( [ c1 ] ) can be written in a canonical form ( similar to ( [ dyn ] ) ) with respect to @xmath0 components of @xmath20 conveniently chosen .
taking initial conditions for these components in such a way that the right - hand side of the mentioned equation is analytic at the origin , the cauchy - kowalewski theorem can be applied once more to guarantee the existence of a solution for the constraint equations .
from the above we see that magaard s proof of the cm theorem is formulated in terms of an initial value problem . roughly speaking we can say that a @xmath37-spacetime is taken as part of the initial data and it is propagated along a spacelike extra dimension by the dynamical part of the einstein vacuum equations to generate the higher - dimensional space .
nevertheless , it is enough clear that , despite some similarities , the cm theorem is not concerned with real dynamical propagation since the initial data evolve along a spacelike direction .
therefore there is no reason why we should expect a causality relation between different slices of the higher dimensional space .
however , we can look at this picture from a different perspective .
indeed , supported by the cm theorem , we know that given any point @xmath38 @xmath1 there is a five - dimensional vacuum space @xmath39 into which a neighborhood of @xmath6 in @xmath1 is embedded .
now we can determine an open subset of @xmath40 , say , @xmath41 , containing the point @xmath6 in which there exists a four - dimensional hypersurface @xmath42 , which is spacelike everywhere , acausal and that contains the point @xmath6 ( see appendix i ) .
the embedding of @xmath42 into @xmath43 induces a positive definite metric @xmath44 in the hypersurface .
let @xmath45 be the extrinsic curvature of @xmath42 in @xmath46 the metric and the extrinsic curvature are analytic fields in the hypersurface @xmath42 , thus they belong to the local sobolev space we mean that the norm of @xmath44 together with its covariant derivatives of order equal or less than @xmath47 are square integrable in any open set @xmath48 of @xmath42 with compact closure . for the sake of simplicity , we shall assume that the norm and derivatives are calculated with respect to an euclidean metric .
here we are adopting the notation of sobolev spaces used in @xcite . ] _ _
@xmath49 for any @xmath47 .
the set @xmath50 constitutes appropriate initial data for the einstein vacuum equations , since @xmath44 and @xmath45 satisfy the vacuum constraint equations in the hypersurface @xmath42 and fulfill the required condition of regularity ( see @xcite , page 248 - 249 , and @xcite , for instance ) .
consider now @xmath51 , the domain of dependence of @xmath42 relative to @xmath43 @xcite . since @xmath42 is an acausal hypersurface of @xmath41 ,
then @xmath52 is open in @xmath41 ( see @xcite , page 425 ) .
of course @xmath53 is a non - empty set , since @xmath54 by construction , the five - dimensional manifold @xmath55 is a solution of the einstein vacuum equations , hence @xmath56 is a cauchy development for the einstein vacuum equations of the initial data @xmath57 as we have mentioned before @xmath53 is open , and thus the non - empty set @xmath58 is a neighbourhood of @xmath6 in @xmath1 contained in @xmath51 .
therefore @xmath58 is embedded in @xmath51 , i.e. , in a cauchy development of @xmath59 ( see figure 1 ) .
in other words the dynamical evolution of the initial data @xmath59 generates a five - dimensional vacuum space into which the spacetime is locally embedded . in this sense
, we can say that this local embedding is dynamically generated by the physical propagation of those initial data .
more precisely this result can be stated as follows : _ consider an analytic spacetime _
@xmath60 _ _ for any _ _
_ there are initial data _
_ @xmath59 _ _ whose cauchy development for the einstein vacuum equations is a five - dimensional vacuum space into which a neighbourhood of _ _ @xmath6 _ in _ @xmath1 _ _ is analytically and isometrically embedded . _ _
furthermore , the einstein vacuum equations admit a well - posed initial value formulation with respect to the data @xmath59 ( see , for example,@xcite ) . therefore the general properties of solutions of the vacuum einstein equations , related to the hyperbolic character of the differential equations , are applicable to our solution @xmath62 .
this ensures that the dependence of the solution @xmath63 on the initial data @xmath59 is continuous ( cauchy stability ) . as a consequence
the spacetime embedded into @xmath64 is stable in a similar sense too , as we describe in the next section .
another important property is that any change of data outside @xmath42 does not affect the solution in the future domain of dependence ( causality ) .
thus it follows that any perturbation outside @xmath42 will not disturb the embedding of spacetime in @xmath51 .
consider an analytic spacetime @xmath3 and let @xmath65 be a set of analytic initial data with a cauchy development @xmath62 in which the spacetime is locally embedded , around @xmath66 additionally let us admit that this initial data set satisfies the following property : the image of @xmath6 through the embedding lies in @xmath42 .
in other words , corresponding to the set @xmath59 there are some neighbourhood @xmath67 of @xmath6 in @xmath1 and a map @xmath68 @xmath51 which is an embedding , with @xmath69 .
was shown in the previous section .
possibly the initial data set is not unique .
the results obtained in this section are applicable separately to each one of all possible initial data set . ]
now we denote by @xmath70 a new set of initial data which satisfies the vacuum einstein constraint equation in @xmath42 . for
the sake of simplicity let us assume that the fields @xmath71 and @xmath72 are @xmath73 in @xmath42 . in this case
, the new generated metric @xmath74 is a @xmath73 field .
let @xmath75 be an open set of @xmath76 , the causal future of @xmath42 in @xmath62 , with compact closure and @xmath48 @xmath77 some neighbourhood of @xmath78 , the causal past of @xmath79 ( the closure of @xmath75 ) in @xmath42 , with compact closure in @xmath42 . according to the cauchy stability theorem ( see @xcite , page 253 , and @xcite ) , for any @xmath80
there is some @xmath81 such that any initial data @xmath70 on @xmath42 close to @xmath82 in @xmath48 with respect to the local sobolev norm , i.e. , @xmath83 and @xmath84 ( @xmath85 , . ]
give rise to a new metric @xmath74 which is near to the old one @xmath86 in @xmath75 , i.e. , @xmath87 .
now let @xmath75 be such that @xmath75 @xmath88 is a non - empty set , where @xmath58 means @xmath89 the image of @xmath67 through the embedding . and
let @xmath90 be the induced metric on @xmath91 by the embedding of @xmath91 in @xmath92 .
we shall see that if @xmath93 is sufficiently small then @xmath90 will be a lorentzian metric and it will be close to the spacetime metric @xmath4 in @xmath91 . for
the sake of simplicity , let us make some assumptions .
first , we assume that @xmath51 is covered by gaussian coordinates ( [ hds2 ] ) adapted to the embedding , in which the embedding map is @xmath94 .
if this was not the case , we could find a neighbourhood @xmath95 of @xmath6 in @xmath42 and a neighbourhood @xmath41 of @xmath6 in @xmath52 such that the domain of dependence of @xmath95 relative to @xmath43 , @xmath96 is covered by ( [ hds2 ] ) .
we would proceed in the following manner .
since the embedding exists , we know that @xmath97 is a timelike hypersurface of @xmath98 by the usual procedure we construct , from the geodesics that cross @xmath58 orthogonally , gaussian normal coordinates in a neighbourhood @xmath41 of @xmath6 in @xmath99 now make @xmath100 .
we then concentrate our analysis in the region @xmath101 .
second , let us assume that the sobolev norm is evaluated with respect to an euclidean metric defined on @xmath51 and that in coordinates ( [ hds2 ] ) the euclidean metric has the canonical form , i.e. , @xmath102 . then , for example , @xmath103 ^{\frac{1}{2}}$ ] where @xmath104 ^{\frac{1}{2}}%\ ] ] as we have mentioned , in the given coordinates , the embedding map is @xmath94 .
thus the induced metric in @xmath91 by the new solution @xmath74 is given by @xmath105 .
let us show that if @xmath106 is sufficiently small the induced metric @xmath90 in @xmath91 is lorentzian .
metrics which are @xmath73 in the whole manifold belong to local sobolev spaces @xmath107 for any @xmath47 .
thus they obey some important inequalities which hold on sobolev spaces .
for example , according to lemma 7.4.1 in ref .
@xcite ( page 235 ) , we have that for @xmath108 , @xmath109 on @xmath75 , where @xmath110 is a positive constant ( depending on @xmath75 ) . thus if @xmath111 , it follows that all components satisfy the inequality @xmath112 on @xmath75 .
the gradient @xmath113 ( @xmath114 is the covariant derivative compatible with @xmath74 ) determines normal direction of the hypersurface @xmath28 in respect to the new solution @xmath74 .
the induced metric will be lorentzian if @xmath113 is spacelike with respect to the metric @xmath74 . in coordinates ( [ hds2 ] ) , we have @xmath115 now
, this metric component can be written as@xmath116 since @xmath117 on @xmath75 , @xmath118 will be positive if @xmath119 .
we know that the determinant is a continuous function with respect to the matrix components in the usual norm .
so there is some @xmath120 which depends only on @xmath86 and @xmath75 ( an estimate of @xmath121 is given in appendix ii ) such that for @xmath122 we have @xmath123 this means that @xmath124 for all points in @xmath75 , since @xmath125 is negative in @xmath75 .
this shows that @xmath90 is lorentzian for @xmath122 .
now let us compare the induced metric @xmath90 with the original spacetime metric @xmath4 .
it is known that for a field in a sobolev space the norm of the restriction of that field to a hypersurface is related to its norm in the manifold ( see @xcite , lemma 7.4.3 , page 235 ) .
according to the mentioned lemma 7.4.3 , there is a positive constant @xmath126 ( depending on @xmath91 and @xmath75 ) such that @xmath127 , where @xmath128 is defined on @xmath91 with the induced euclidean metric .
thus , for example,@xmath129 on the other hand @xmath130 ( by @xmath131 we mean that derivatives are taken only in directions tangent to @xmath91 ) . now consider @xmath132 the induced metric on @xmath133 .
in this special coordinates it is easy to see @xmath134 @xmath135 , since the induced metric has less components than the higher dimensional metric .
therefore @xmath136 . from the results obtained above we
can now show that if the initial data are sufficiently close to @xmath82 on a neighbourhood of the causal past of @xmath91 in @xmath42 then the induced metric by the new solution is lorentzian and close to @xmath4 in @xmath91 .
more precisely , _ let _
@xmath91 _ be an open set in _ @xmath137 _ with compact closure in _
@xmath1 _ and _ @xmath138 _ _ , there exist some neighborhood _
_ @xmath48 _ of _ @xmath139 _ of compact closure in _
@xmath42 _ and some _ @xmath81 _ such that _ @xmath73 _ initial data _ @xmath70 _ close to _ @xmath82 _ in _ @xmath48 _ _ , i.e. , _
_ @xmath140 _ and _ @xmath141 _ _ , give rise to a metric _
_ @xmath74 _ which induces a lorentzian metric in _
@xmath91 _ which is near to _ @xmath4 _ _ , i.e. , _ _
_ in order to see this it suffices to take @xmath143 ( @xcite ) and to make appropriate choices for @xmath106 and @xmath75 in the cauchy stability theorem .
indeed , let @xmath75 be any neighbourhood of @xmath91 in @xmath144 with compact closure . since @xmath145 then @xmath146 .
thus , taking @xmath48 to be some neighbourhood of @xmath147 of compact closure in @xmath42 , the cauchy stability theorem ensures that there exists @xmath93 such that the new metric generated @xmath74 satisfies @xmath148 for any @xmath80 . now
if we take @xmath149,@xmath150 , we guarantee that the induced metric @xmath151 is lorentzian in @xmath91 and that @xmath152 .
from this analysis we can conclude that for each ambient space whose existence is guaranteed by cm theorem there corresponds an initial data set _ _
@xmath59 _ _ in respect to which it possess , in a certain domain , the desirable physical properties of causality and stability .
this can not be ensured by cm theorem itself , but by an indirect reasoning as discussed above . as a direct consequence of this result
we have found that for any analytic spacetime there exist initial data in whose cauchy development for the vacuum einstein equations it can be locally embedded .
the embedding is cauchy stable and obeys the domain of dependence property with respect to @xmath59 .
the extension of the above result to the case where a cosmological constant is included in the field equations can easily be done . on the other hand
the same analysis can not be applied to the case of a brane in a straightforward way .
the problem is that in this case we can not guarantee the existence of a smooth spacelike hypersurface in an open set of @xmath153 at least by the method employed here , because , as it is known , the brane must be embedded in a space whose metric has a discontinuity in the derivative along the normal direction with respect to the brane .
thus this metric is not analytic in any open set containing the point @xmath6 and for this reason we can not use the cauchy - kowalewski theorem to guarantee the existence of a function whose gradient is everywhere timelike as we have done in the appendix i. nevertheless , this question deserves further investigation .
the analyticity of the initial data in @xmath42 is a restriction imposed by cm theorem .
however , they are not to be considered an unphysical condition .
indeed , it must be realized that a great part of the physical solutions , even in the relativistic regime , are analytic in a certain domain .
the crucial point here is the domain of convergence .
what seems to be an unrealistic is a field which is analytic in the whole manifold .
thus we can say that if we could handle the initial data in the spacelike four - dimensional hypersurface , it would be physically feasible to prepare initial data which are analytic in the interior of a compact domain @xmath154 containing the point @xmath6 in order to generate the desired embedding in the interior of @xmath155 .
moreover we have seen that @xmath73 initial data sufficiently close to an analytic initial data set give rise to embeddings of @xmath73spacetimes which are near to the original analytic spacetime . _
_ this may suggest the possibility of extending the cm theorem to a less restrictive differential class of embedddings .
we are currently investigating this possibility .
consider the five - dimensional vacuum space@xmath156 into which the spacetime @xmath3 is locally embedded around the point @xmath5 .
let us take the following equation for the unknown function @xmath157@xmath158 in gaussian coordinates this equation can be written in the following form @xmath159 let us consider the equation with the positive sign . according to the cauchy - kowalewski theorem
, there is an open set @xmath41 in @xmath40 containing the point @xmath6 where the equation has a solution provided the initial condition @xmath160 ensures that the right - hand side of that equation be an analytic function of its arguments @xmath161 at the point @xmath162 this can be achieved by choosing @xmath163 in such a way that the following inequality be satisfied at the point @xmath164@xmath165 this can always be done .
for example , take @xmath166 , where @xmath167 and @xmath168 is the component of a unit spacelike vector with respect to the spacetime metric at the point @xmath6 .
therefore , the solution @xmath169 is a function whose gradient is everywhere timelike in @xmath43 , and this means that stable causality condition holds on @xmath170 .
now let us assume without loss of generality that @xmath171 considering that the gradient of @xmath169 does not vanish in @xmath41 , we know that the inverse image @xmath172 is a hypersurface @xmath42 of @xmath41 . since the gradient of @xmath169 which is orthogonal to @xmath42 is everywhere timelike , then we can conclude that @xmath42 is a spacelike hypersurface .
moreover , @xmath42 is achronal .
indeed , every future directed timelike curve which leaves @xmath42 can not return to @xmath42 since @xmath169 does not change sign along these curves .
it happens that an achronal spacelike hypersurface is acausal @xcite .
therefore @xmath42 is an acausal spacelike hypersurface in @xmath173 .
in this section we want to determine an estimate of how much near to @xmath86 the new metric @xmath74 must be in order to induce a lorentzian metric in @xmath91 .
as described above , this is achieved if the condition@xmath174 holds on @xmath75 .
as we have seen , for @xmath108 , @xmath175 implies that @xmath176 on @xmath75 .
then , there exist a continuous field @xmath177 , with @xmath178 on @xmath179 such that @xmath180 , where @xmath181 .
we want to evaluate @xmath182 .
first let us define @xmath183 as a totally anti - symmetric four index ` tensor ' with @xmath184 .
thus , we can write@xmath185 this expression can be rewritten in the following form@xmath186 where@xmath187 therefore , we have @xmath188 in order to make estimates of @xmath189 it is necessary to eliminate @xmath190 . to this end ,
let us introduce @xmath191 a @xmath192 matrix with all elements equal to unity . using this matrix ,
we get the following estimate@xmath193 similar inequalities hold for every term that appear in @xmath189 .
then we have @xmath194 , for @xmath195 , where @xmath196 is obtained from @xmath197 by substituting each one of its term for the corresponding estimate as above .
it happens that @xmath196 depends only on @xmath198 .
indeed , it is a function constituted of sum and multiplication of the components of @xmath199 .
now for the sake of simplicity let us consider that @xmath200 , ( i.e. , @xmath201 ) , then we have@xmath202 recalling that @xmath203 . since @xmath199 is continuous in @xmath40 , then the right hand side is bounded in any compact set such as @xmath79 .
the same is valid for @xmath204 .
now define @xmath205 as the positive quantity@xmath206 } % \ ] ] the value of @xmath205 depends only on @xmath86 and @xmath75 and the chart ( possibly one can give a definition independent of coordinates , but this definition is sufficient for our purpose ) .
now define @xmath207 taking @xmath208 we get@xmath209 considering that @xmath210 on @xmath211 we conclude that @xmath124 on @xmath75 and hence the induced metric in @xmath91 is lorentzian .
the authors thank cnpq for financial support . this work was partially financed by cnpq / fapesq ( pronex ) . | we discuss how embeddings in connection with the campbell - magaard ( cm ) theorem can have a physical interpretation .
we show that any embedding whose local existence is guaranteed by the cm theorem can be viewed as a result of the dynamical evolution of initial data given in a four - dimensional spacelike hypersurface . by using the cm theorem
, we establish that for any analytic spacetime , there exist appropriate initial data whose cauchy development is a five - dimensional vacuum space into which the spacetime is locally embedded .
we shall see also that the spacetime embedded is cauchy stable with respect these the initial data .
the development of the braneworld scenario @xcite , in which our ordinary spacetime is viewed as a hypersurface of a higher - dimensional space , has greatly contributed to increase recent interest on embeddings theorems .
the so - called non - compact kaluza - klein ( nkk ) models @xcite have also motivated the study of embedding problems of the spacetime . in this context ,
the campbell - magaard ( cm ) theorem @xcite and its variants @xcite are of special interest when the embedding spaces possess only one extra dimension as is the case of the randall - sundrum braneworld scenario @xcite and the nkk models @xcite .
the cm theorem ensures the existence of local and analytic isometric embedding of any @xmath0-dimensional analytic manifold @xmath1 into a ricci - flat @xmath2-dimensional space .
it can be stated as follow : given a @xmath0-dimensional analytic space @xmath3 , where @xmath4 stands for the metric , for any point @xmath5 there exists a vacuum solution of einstein equations in @xmath2-dimensions into which a neighborhood of @xmath6 in @xmath1 can be embedded .
the cm theorem is valid for any analytic metric @xmath4 irrespective of its signature and for dimension @xmath7 .
however , having in mind applications to physics , we shall consider @xmath8 , throughout the paper , and assume that the metric is lorentzian .
the cm theorem is local , but the extension to the case of global embedding has been studied @xcite__. _ _ recently some interesting issues concerning the physical interpretation of embeddings in the light of the cm theorem have been arisen @xcite .
it has been argued that , since the cm theorem is based on the cauchy - kowalewski theorem it disregards causality and can not guarantee a continuous dependence between the metric of ambient space and spacetime metric .
therefore , by lacking these important properties , the _ _ embeddings _ _ obtained by employing the cm theorem would have no physical meaning . in this paper
we intend to show that , as a matter of fact , ambient spaces whose local existence is guaranteed by the cm theorem possess , in a certain domain , those desirable physical properties with respect to appropriate initial _ _ data . indeed , we shall show that each five - dimensional vacuum space into which the spacetime can be locally embedded may be locally viewed as the cauchy development of some initial data given on a spacelike four - dimensional hypersurface and hence satisfies the stability and causality conditions with respect to these initial data .
further , we shall see that the region of @xmath1 which is embedded in the cauchy development of an initial data set possesses also cauchy stability and domain of dependence property with respect to these initial data . in order to appropriately address these issues ,
let us begin our discussion by considering a brief sketch of magaard s proof of the cm theorem . |
understanding the stellar content of early - type galaxies is fundamental to understanding their star formation and chemical evolution history .
most early - type galaxies are too distant to resolve their individual stars with current technology , rendering the direct study of their stellar populations impossible .
thus , their stellar populations must be studied using indirect methods . in recent decades
, significant effort has gone into trying to better constrain the stellar contents for early - type galaxies using optical spectroscopic data .
the most commonly studied features have been ca i h and k 0.38 @xmath0 m , h@xmath4 , mgb 0.52 @xmath0 m , fe @xmath0 m 0.53 , na 0.82 @xmath0 m , and cat 0.86 @xmath0 m .
interpretation of all such spectral features is intrinsically complicated by their blended nature
each feature is really the super - position of many spectral lines , usually from several different elements , blurred together by the line - of - sight velocity dispersion within each galaxy .
there is no way to overcome this problem it must simply be taken into account during analysis .
as population synthesis models have become more sophisticated and digital stellar libraries more complete , this problem has become more tractable over time .
another challenge arises from the composite nature of galaxies : each observed feature is the luminosity - weighted integrated sum of that feature from all stars in the observed line - of - sight .
naturally , luminosity - weighted does not imply mass - weighted .
a relatively small fraction of the mass can dominate the observed luminosity and mask the underlying stellar population ( e.g. as happens during a starburst event within a pre - existing galaxy ) . even in relatively quiescent galaxies , light from stars at several important evolutionary stages contribute roughly equally to the observed spectral features between 0.4 1 @xmath0 m range .
hence , a feature depth change could be due to ( e.g. ) a change near the ( mostly ) age - driven main - sequence turnoff or the ( mostly ) metallicity - driven red giant branch .
the details can become quite complicated , as illustrated by the long standing controversy about whether observed changes in balmer line strength arise from the presence of younger main sequence stars , more metal - poor main sequence stars , or an extended horizontal giant branch ( for recent discussions of this debate , see maraston & thomas 2000 and trager et al .
a similar controversy surrounds na 0.82 @xmath0 m feature : is it driven by metallicity - driven red giant branch changes , initial mass function related differences in the relative number of cool dwarf and giant stars or both ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? however , the properties of the rgb component can be isolated by observing in the k - band ( centered near 2.2 @xmath0 m ) . at those wavelengths , cool giants near the tip of the first - ascent red giant branch ( rgb )
dominate the integrated light in old ( @xmath5 3 gyr ) stellar populations . in combination with optical observations
, k - band observations should facilitate the separation of msto and rgb light contributions .
there are two possible complications to this scenario .
first , a very young stellar population containing red supergiants will contribute a significant fraction k - band light .
fortunately , such a population is obvious from the presence of h ii region emission lines at shorter wavelengths .
second , a somewhat older population ( 1 2 gyr , i.e. an intermediate - age population ) may contain bolometrically bright carbon stars that can contribute a detectable amount of k - band light ( see discussions in silva & bothun 1998a , b ) .
such a population may or may not be connected to increased h@xmath4 strength .
initial development of these ideas can be found in silva et al .
( 1994 ) , mobasher & james ( 1996 ) , james & mobasher ( 1999 ) , mobasher & james ( 2000 ) , all of whom focused on the co 2.36 @xmath0 m feature .
origlia et al .
( 1997 ) observed a si dominated feature at 1.59 @xmath0 m as well as co dominated features .
these observational studies were limited by small - format detectors to relatively low resolving powers and/or small wavelength ranges per observation . in the cases of silva
et al . and origlia et al . , only small , heterogeneous samples of galaxies were observed .
a general conclusion of the james & mobasher studies was that changes in co strength between early - type galaxies in high - density and low - density regions were statistically consistent with different fraction contributions of intermediate - age agb light and hence galaxies in low - density regions had younger luminosity - weighted mean ages .
origilia et al .
argued that [ si / fe ] was super - solar in the four elliptical galaxies they observed . to further develop these ideas and investigate the usefulness of other k - band spectral indices in the study of early - type galaxies ,
new data have been obtained for eleven e / s0 galaxies in the nearby fornax cluster .
only measurements in the central regions of these galaxies are discussed here . in section [ sec : data ] , the galaxy sample and its observations are discussed , while in section [ sec : proc ] the data processing methodology is described . the measurement of spectral feature strength is explained in section [ sec : lines ] while basic observation results are presented in section [ sec : results ] .
the broader astrophysical implications of our observational results are discussed in section [ sec : disc ] .
a summary is provided at the end .
long - slit spectroscopic data obtained with isaac at the eso very large telescope ( vlt ) have been combined with data obtained with sinfoni at the vlt to study a small sample of early - type galaxies in the nearby fornax cluster .
details about the sample as well as the instrumental setups for these two instruments are presented in this section .
the full sample observed consists of eleven early - type galaxies in the nearby fornax cluster , sub - divided into six elliptical galaxies and five s0 galaxies . the overall sample was selected to cover a significant range in luminosity weighted age , metallicity and mass .
basic galaxy properties are summarized in table [ tab : galsample ] .
optical long - slit spectroscopy is available for all observed galaxies ( kuntschner 2000 , kuntschner et al .
2002 ) allowing comparison between the k - band line - strength indices derived in this paper with optical indices . when compared to simple stellar population models , three galaxies ( ngc1316 , ngc1344 and ngc1375 ) appear to have a significant contribution from young stars resulting in luminosity weighted ages between 1 and 3gyrs ( see figure [ fig : fornax_pops ] ) .
the remaining galaxies are consistent with being dominated by old stellar populations with a range in average metallicity .
multi - color hst advanced camera for survey images ( courtesy of a. jorden ) of all the galaxies in our sample were inspected for the presence of dust .
three galaxies ( ngc 1316 , ngc 1344 , and ngc 1380 see details below ) have clear central dust features .
the other galaxies in our sample have no visible dust feature within the investigated aperture .
a few additional object specific remarks seem warranted : * _ ngc 1316 _ fornax a , s0(3)pec .
this galaxy has large scale , irregular , central dust features ( based on inspection of hst - acs images ) . *
_ ngc 1344 _ this s0 galaxy contains morphological shells in its outer regions @xcite as well as optical spectral signatures of a younger central population ( e.g. kuntschner et al .
inspection of hst - acs images reveals patchy dust within the central 5 arcsec .
in combination , these characteristics suggest ( 1 2 gyr ) dynamical interaction or merger event that triggered a central star formation episode . * _ ngc 1375 _ kuntschner ( 2000 ) concluded that the mean age of the central population of this s0 galaxy is significantly lower than for the other galaxies in our sample . * _ ngc 1380 _ contains a kinematically decoupled core ( kdc ) @xcite .
it also contains an optically thick , well - organized central dust lane , easily visible in available hst images . * _ ngc 1381 _ this s0 galaxy contains x - bulge isophotal pattern , often attributed to a recent dynamical encounter or merger event . *
_ ngc 1399 _ cd , e0 * _ ngc 1419 _ this e0 galaxy has anomalously bright k - band surface brightness fluctuations ( sbf ) relative to its i - band surface brightness fluctuations , suggesting the presence of extended giant branches , perhaps due to the presence of an intermediate - age population @xcite . * _ ngc 1427 _ this e4 galaxy also has anomalously bright k - band sbf ( see ngc 1419 sbf references above ) .
it also contains a kdc @xcite .
llcccccccl ngc 1316 & so(3)pec & 0.0059 & 226 & 132.2 & 90.5 & 50 & & 168 & fornax a ; sinfoni data + ngc 1344 & e5 & 0.0039 & 171 & 26.7 & 15.6 & 165 & 3.3 & 121 & type from rc3 + ngc 1374 & e0 & 0.0043 & 196 & 26.6 & 24.2 & 118 & 3.3 & 124 & kdc + ngc 1375 & s0(cross ) & 0.0025 & 70 & 24.0 & 9.3 & 91 & 3.0 & 48 & central starburst + ngc 1379 & e0 & 0.0044 & 138 & 23.9 & 23.3 & 182 & 2.9 & 78 & + ngc 1380 & s0/a & 0.0063 & 213 & 52.4 & 26.6 & 7 & 3.4 & 213 & + ngc 1381 & s0(9)(boxy ) & 0.0057 & 169 & 23.5 & 8.0 & 139 & 2.9 & 135 & + ngc 1399 & e0 & 0.0048 & 360 & 134.1 & 120.8 & 110 & & 111 & cd ; sinfoni data + ngc 1404 & e2 & 0.0065 & 221 & 26.0 & 23.0 & 157 & 157 & 280 & + ngc 1419 & e0 & 0.0071 & 128 & 9.4 & 9.3 & 188 & 1.2 & 111 & k - band sbf outlier + ngc 1427 & e4 & 0.0046 & 186 & 39.2 & 27.1 & 76 & 4.9 & 104 & k - band sbf outlier , kdc + a small sample of galactic open cluster stars were also observed .
their basic properties are summarized in table [ tab : starsample ] .
cluster parameters ( age , metallicity ) and star photometry come from houdashelt , frogel , & cohen ( 1992 ) .
age has units of gyr .
spectral types ( spt ) have been assigned based on their @xmath6 colors following @xcite for k giant stars and @xcite for m giant stars .
these stars have several roles : velocity dispersion templates , co 2.36 @xmath0 m index comparisons , and small independent framework for interpreting k - band line strength variations as functions of basic atmospheric parameters .
each of these roles will be discussed in turn below .
llllccl ngc 2204 & 2.8 & 0.38 & 1146 & 5.71 & 1.12 & m4 + & & & 3304 & 3.10 & 0.79 & k3 + & & & 3325 & 3.84 & 0.94 & m0.5 + & & & 4132 & 5.03 & 1.11 & m3.5 + ngc 2477 & 1.0 & 0.02 & 1069 & 4.07 & 0.95 & m1 + & & & @xmath7 & 2.96 & 0.71 & k3 + & & & 2117 & 4.18 & 0.96 & m2.5 + & & & 6053 & 2.81 & 0.68 & k2 +
ngc 2506 & 3.5 & 0.52 & 2401 & 3.65 & 0.91 & k5 + & & & 4228 & 4.07 & 0.99 & m1 + finally , several bright hot stars were observed to measure and correct the telluric absorption features in the k - band .
b stars were selected by color from the hipparcos tycho catalog ( esa , 1997 ) . in the spectral range of interest ,
these stars have no spectral features ; therefore , that any observed features must be telluric absorption lines .
in addition to color , proximity to the target objects and brightness were used as secondary selection criteria .
the actual stars observed are listed in table [ tab : telluric ] .
the _ v _ magnitude and _ b v _ color are from the tycho catalog ( esa , 1997 ) .
llcc tycho2 7034 01323 & fornax galaxies & 6.07 & 0.11 + tycho2 5942 02406 & ngc 2204 & 5.29 & 0.19 + tycho2 7645 03250 & ngc 2477 & 6.13 & 0.17 + tycho2 5425 00772 & ngc 2506 & 7.87 & 0.14 + most of the observations discussed in this paper were obtained at the eso very large telescope using the isaac near - ir imaging spectrometer mounted at one of the nasmyth foci of the 8.2 m ut1/antu telescope .
galaxy observations were obtained in 2002 january and 2002 november .
a summary observations log is given in table [ tab : obslog ] .
for each galaxy , the following information is provided : ut date of observation , integration time per frame , number of frames , and total integration time . since ngc 1380 and ngc 1404
were observed multiple times , each observation sequence is assigned a sequence identifier ( a , b , c , ... ) .
llccccl 2002 jan 02 & ngc 1404 & 500 & 6 & 3000 & 89.45 & a + 2002 jan 14 & ngc 1404 & 500 & 6 & 3000 & & b + 2002 jan 14 & ngc 1404 & 500 & 6 & 3000 & & c + 2002 nov 12 & ngc 1404 & 500 & 6 & 3000 & & d + 2002 nov 13 & ngc 1404 & 500 & 6 & 3000 & & e + & & & & & + 2002 nov 13 & ngc 1380 & 400 & 8 & 3200 & 89.45 & a + 2002 nov 13 & ngc 1380 & 400 & 8 & 3200 & & b + 2002 nov 12 & ngc 1380 & 400 & 8 & 3200 & & c + & & & & & + 2002 nov 12 & ngc 1344 & 400 & 8 & 3200 & 164.45 & + 2002 nov 13 & ngc 1374 & 400 & 8 & 3200 & 89.45 & + 2002 nov 12 & ngc 1375 & 400 & 8 & 3200 & 89.45 & + 2002 nov 13 & ngc 1379 & 400 & 8 & 3200 & 89.45 & + 2002 nov 13 & ngc 1381 & 400 & 8 & 3200 & 138.45 & + 2002 nov 12 & ngc 1419 & 400 & 8 & 3200 & 89.45 & + 2002 nov 12 & ngc 1427 & 400 & 8 & 3200 & 77.45 & + cccccc 2004 nov 25 & ngc1399 & 300 & 4@xmath84 & 1200 & b + 2004 nov 26 & ngc1399 & 300 & 4@xmath84 & 1200 & c + 2004 nov 26 & ngc1399 & 300 & 4@xmath84 & 1200 & d + 2004 nov 26 & ngc1399 & 300 & 4@xmath84 & 1200 & e + & & & & & + 2005 oct 13 & ngc1316 & 600 & 3@xmath81 & 1800 & a + 2005 oct 13 & ngc1316 & 600 & 4@xmath82 & 2400 & b + isaac was used in short - wavelength spectroscopic mode at medium - resolution ( sws - mr ) . in this mode ,
the array used is a 1024@xmath91024 hawaii rockwell hgcdte array with a spatial plate scale of 0148 pix @xmath10 .
a 120@xmath91 slit was used .
the dispersion and spectral resolution measured from our arc lamp data were 1.21 pix@xmath10 and 7.7 ( fwhm , @xmath11 ) , respectively .
the approximate wavelength range was 2.12 2.37 @xmath0 m but the central wavelength was always corrected for object redshift to ensure that all relevant spectral features were observed . galaxy observations were obtained using the standard nod - on - slit mode . at the start of each observational sequence ,
the galaxy was centered on the slit near one end and a individual integration was executed .
the galaxy was then moved approximately 60 towards the other end of the slit where two more integrations were executed . the galaxy was then returned to the original slit position where another integration was obtained .
this abba pattern was repeated a number of times , resulting in multiple individual two - dimensional spectroscopic images . including time for telescope offsetting and array readout ,
each sequence took roughly one hour to execute .
after each set of galaxy sequences was completed , telluric standard tycho2 7034 01323 was observed ( see table [ tab : telluric ] ) at five different positions along the slit . at each position ,
a short ( @xmath12 10 secs ) integration was obtained . at the end of the telluric integrations ,
a xear lamp - on / lamp - off sequence was obtained followed by a quartz lamp - on / lamp - off sequence .
the spectrograph setup was not changed during the entire galaxy - standard - lamps sequence , ensuring that the grating did not move .
thus , each star - lamp sequence had the same central wavelength as the corresponding galaxy sequence .
this is critical for proper correction of telluric features , dispersion , and array pixel - to - pixel response . on various dates , stars from the following open clusters
were observed : ngc 2204 ( 2002 november 12 ) , ngc 2477 ( 2001 december 23 ) , and ngc 2506 ( 2002 november 13 ) .
the actual stars observed are listed in table [ tab : starsample ] .
each star was observed at five ( 5 ) positions along the slit . at each position , a short ( @xmath12 10 secs )
integration was obtained .
after each set of cluster star observations , a telluric standard was observed , followed by arc and flat lamp observations ( as described above ) .
the exact telluric standards observed are listed in table [ tab : telluric ] . as part of the vlt calibration program ,
calibration data was also obtained during the day following each set of observations .
these data include observations of quartz and arc lamps ( for flat - field and wavelength dispersion corrections ) as well as dark frames with the same detector integration times ( dits ) as the night - time science observations .
for this project , the daytime quartz and arc lamps were not useful due to grating angle positioning inaccuracy .
observations of ngc1316 ( fornax a ) and ngc1399 ( the central cd galaxy ) were obtained with sinfoni , a near - ir integral field unit ( ifu ) spectrograph mounted at the cassegrain focus of vlt ut4 ( yepun ) .
a spatial scale of the 0.1@xmath13 scale was used , corresponding to a total field of view of approximately @xmath14 .
the spectra cover the wavelengths between 1.95 2.4 @xmath0 m at 2.45 / pix .
the spectral resolution is 6.1 fwhm ( @xmath15 ) .
ngc1399 was observed 2004
november 25 and 29 during the sinfoni science verification program ( eisenhauer et al .
2003 ; bonnet et al .
the ngc1316 data come from the eso science archive ( program 076.b-0457a , pi ralf bender ) .
the original observations were made on 2005 october 13 .
details of these observations are given in table [ tab : obssinfoni ] .
ngc1399 was observed in a sequence ossoosso ( o = object integration , s = sky / background integration ) to enable accurate background correction .
ngc1316 was observed with a slightly different pattern osooso . as with the isaac observations discussed above
, telluric and velocity template standard stars were observed periodically .
the former were hot stars while the latter covered the range k5iii to m1iii .
data processing was done using a combination of shell scripts , eclipse @xcite , and iraf . as described above
, each isaac observing sequence consists of a set of galaxy or cluster star observations followed immediately by associated telluric star and lamp exposures .
the lamps exposures were processed first .
the resultant flat - field and xear arc lamp frames were used in turn to process their associated telluric star and galaxy or cluster star sequences . a raw flat - field frame
was constructed by subtracting the lamp - off exposure from the lamp - on exposure to remove thermal background and dark current .
the difference frame was used to define the illuminated region of the isaac sw detector ( the entire detector is not illuminated in sws mode ) .
the difference frame was trimmed to remove the unilluminated part .
the trimmed frame was then collapsed in the spatial dimension into a one - dimensional vector .
a one - dimensional polynomial was fit to this vector .
this polynomial was then divided into every row of the original background - corrected flat - field frame frame .
the resultant frame is suitable for removing pixel - to - pixel sensitivity variations .
these flat - field frames obtained immediately after all telluric sequences are preferred to evening or morning flats since tests demonstrated that such so - called daytime calibrations did not accurately remove fringing .
this is presumably due to grating re - positioning imprecision .
the xear exposures were processed by first subtracting the arc - off frame from the arc - on frame , again to remove any background illumination not associated with the xear lamps .
a two - dimensional wavelength dispersion correction was derived , using low - order polynomials in both dimensions . for each telluric absorption correction star sequence ,
the five individual frames were first trimmed to correspond to the illuminated part of the detector ( see above ) and then divided by their associated flat - field .
a background@xmath8dark frame was determined by median combining the individual flattened exposures .
this background frame was then subtracted from those exposures .
the five flat - field and background corrected frames were summed into a single frame .
a spectral tilt correction was determined from the resultant frame .
spectral tilt and wavelength dispersion corrections were then applied to the summed , flat - field corrected frame .
the five individual spectra were extracted and averaged into a single spectrum using a sigma - clipping algorithm for bad pixel rejection .
this combined spectrum was used to produce two things : a normalized telluric absorption line correction spectrum and a instrument sensitivity correction .
building the normalized telluric correction spectrum started by fitting a one - dimensional bi - cubic spline to the telluric star spectrum .
care was taken to reject high outliers ( e.g. cosmic rays , bad pixels ) but not too many low outliers ( likely to be the telluric absorption features of interest ) .
the derived vector was divided back into the original spectrum to normalize it and remove the continuum shape .
this bi - cubic spline fit was then re - used to determine an instrument sensitivity curve . since the color of the telluric stars is known from the hipparcos catalog , a reasonable estimate of their effective temperature can be made .
furthermore , since these are hot stars , we know that their continuum in the k - band is well approximated by the rayleigh - jeans part of the blackbody spectrum associated with their effective temperature . by computing the appropriate blackbody curve in the relevant wavelength range and then dividing it by the bi - cubic spline fit ( i.e. a representation of the observed stellar continuum ) , the instrument sensitivity function can be estimated .
of course , this is not an absolute correction but an absolute correction is not necessary for the spectral line strength measurements described below .
cluster star sequences were processed into single combined spectra in a similar manner as telluric star sequences . at that point ,
the combined spectrum of each star was divided by its associated normalized telluric absorption line correction spectrum . to remove the instrument response signature , the telluric corrected cluster star spectrum
is then multiplied by the instrument sensitivity function . to achieve the optimal telluric absorption line correction , it was typically necessary both to shift the telluric star spectrum in the dispersion direction ( usually by less than 1 pixel ) and scale it ( usually by a few percent ) before the division .
the physical cause of the physical shift in the dispersion direction is unclear .
however , the need to scale is simply due to the slight changes in airmass and time between object and telluric correction standard observations .
it is also clear that the strength of every telluric absorption feature in this wavelength window is not changing in concert .
thus , the telluric correction is not perfect at all wavelengths . in principle
, a more sophisticated correction process is possible . in practice ,
the current level of correction is good enough for this application .
the final processed spectra for all observed cluster stars are shown in figure [ fig : clusterstars ] .
the main technical goal of this paper is to make accurate line strength measurements in the central regions of our target galaxies .
however , in the future , the radial line strength gradients will also be studied . since each ab galaxy observation
is only separated by @xmath16 100 , a simple ab subtraction ( i.e. the typical procedure in the near - ir ) could cause inaccurate dark and background correction in the lower surface brightness parts of the galaxy luminosity profiles .
therefore , a procedure derived from the processing of _ optical _ long - slit spectra was applied .
the basic data processing challenge is shown in figure [ fig : galprocexample ] where a central 3 diameter extraction of ngc1404 ( sequence d ; see table [ tab : galsample ] ) is shown at three stages : before background correction , after background correction but before telluric absorption line correction , and after all processing .
one of the major advantages of isaac over other near - ir spectrometers is the ability to use a high - enough spectral resolution to resolve telluric features so that they can be accurately removed .
as shown in table [ tab : galsample ] , each galaxy observation consists of a series of 6 or 8 individual exposures taken in abba sequences . to begin ,
each galaxy frame was trimmed and the appropriate median dark frame was subtracted . following the advice of the _ isaac data reduction guide _
( available on - line from the eso isaac web page ) , electronic ghosting was removed using the eclipse recipe _
isaacp ghost_. recall that each galaxy sequence had an associated telluric@xmath8lamp sequence .
as described in the telluric star processing section above , each telluric@xmath8lamp sequence was processed into a series of calibrations : flat - field , two - dimensional wavelength dispersion solution , spectral tilt correction , normalized telluric absorption correction vector , and instrument sensitivity function .
after electronic ghost correction , each individual galaxy frame was divided by this flat - field and then geometrically rectified using the derived dispersion and spectral tilt solutions . in short , the galaxy sequence and its associated telluric star are calibrated using the same frames and vectors .
a slit illumination correction was applied next .
since twilight sky exposures were not available , the correction function had to be reconstructed from the actual observations . in each individual galaxy observation , the spatial dimension can be divided into a galaxy sub - section and a sky sub - section . for each ab pair ,
the sky sub - section is on opposite sides of the slit center .
the a and b sky sub - sections were extracted and merged into a single a@xmath8b sky frame , after applying a small additive offset to the b sky sub - section to align it better with the a sky sub - section .
this offset is caused by temporal changes in the background between the a and b exposures .
this process was repeated for all a and b galaxy exposures on a given night .
the resultant frames were then scaled to a common mean value and averaged using a sigma clipping bad pixel rejection scheme . that average sky frame was collapsed in the dispersion direction into a one - dimensional vector .
a nightly illumination correction function was derived by fitting a second - order bi - cubic spline to that vector and normalizing the result .
this illumination correction was then divided into all galaxy frames from that night .
the next step was to remove the background .
the standard procedure would be to select a radial position dominated by the background , determine the mean background at this position , and subtract that value at every pixel in the spatial dimension . as applying this procedure to the current dataset
did not produce satisfactory results , an alternative two - step approach was developed .
first , for each ab pair , an illumination corrected a@xmath8b sky frame was produced as described above and then subtracted from the parent a and b frames .
second , on each frame , the standard procedure was applied .
this two - step procedure produced a much better background correction than the standard procedure alone .
each background corrected frame was then corrected for telluric absorption lines , using the associated normalized telluric absorption correction vector .
as described in the cluster star processing section above , small shifts in the dispersion direction as well as small multiplicative scaling were necessary to achieve the best possible correction . finally , the six ( or eight ) individual frames combined into a single frame
before this combination , the individual frames were shifted to a common center ( defined by the maximum intensities in the galaxy spectra ) with sub - pixel accuracy using linear interpolation .
given the pixel scale ( 0148 pix@xmath10 ) vs. the natural seeing ( estimated to be @xmath16 08 from the profiles of the telluric star observations ) , alignment accuracy is not critical .
the frames were then scaled to a common median value at a fixed point in the galaxy spectra and then averaged using a sigma - clipping rejection scheme .
the final step was to multiple each combined frame by the sensitivity function derived from its associated telluric standard ( see above ) .
basic post - processing data quality can be assessed from figure [ fig : galcenters ] shows the central 3 diameter extraction for all the galaxies observed in 2002 november .
the spectra are ordered by velocity dispersion ( @xmath1 ) which corresponds essentially to decreasing signal - to - noise per pixel due to decreasing central intensity .
the sharper spectral features seen in the galaxies with lower @xmath1 are intrinsic to the galaxies they are not caused by less accurate telluric line correction in these fainter galaxies . to first order , measurement repeatability can be assessed from figure [ fig : repeatcenters ] where the central 3 diameter extraction of ngc1380 and ngc1404 are shown from two and three different nights , respectively . in the case of ngc1404 ,
these observations were separated by months . clearly , the gross properties of the extracted spectra are the same .
repeatability of line strength measurements is discussed in the next section .
the central spectra used for the measurement of spectroscopic features were extracted within a radius corresponding to 1/8 r@xmath17 at the observed position angle for each galaxy ( see also tables [ tab : galsample ] and [ tab : obslog ] ) . for each extracted spectra , an empirical signal - to - noise ratio ( snr )
was derived following stoehr et al .
( 2007 ) and listed in table [ tab : galsample ] . from these extracted spectra ,
recession velocities and velocity dispersions were measured using the penalized pixel - fitting method ( ppxf ) developed by cappellari & emsellem ( 2004 ) . for simplicity , purely gaussian line shapes were assumed .
the resulting measurements are needed as input for the line - strength determinations ( see section [ sec : lines ] ) .
the velocity dispersions are tabulated in table [ tab : galsample ] and are used throughput this paper .
errors on the kinematics parameters are calculated within the code based on error frames which are produced by our data reduction scripts .
see also section [ sec : proc - optical ] for further analysis of the kinematic errors .
sinfoni observations of ngc 1316 and ngc 1399 were processed and calibrated using the sinfoni pipeline ( v. 1.6 ) .
the pipeline transforms raw science data frames into flat - field corrected , wavelength calibrated and background subtracted three - dimensional data cubes for the galaxies and the calibration stars , including the extraction of one - dimensional spectrum from the central 10 pixels of the observed galaxies and stars . for ngc1399 ,
the telluric stars were solar - type . in order to obtain a telluric correction spectrum ,
the star spectra were divided by a scaled and gaussian broadened solar spectrum .
this step removes intrinsic stellar features and continuum shape .
a correction for telluric features was then applied as described above for the isaac observations .
the ngc1316 telluric stars were hotter b stars that have no prominent spectral features in the wavelength range of interest .
the telluric correction spectrum was obtained by dividing the stellar spectrum by a scaled black body spectrum with the same temperature as the star .
a correction for telluric features was then applied as described above for the isaac observations .
the individual data cubes for the two galaxies were then combined to achieve better signal - to - noise ratios .
the four ngc1399 data cubes were averaged using a sigma - clipping pixel reject algorithm .
since there were only two ngc1316 observations , pixel reject during combination was not possible . since the field - of - view of the vlt - sinfoni observations
was limited to approximately 3@xmath93 , an extraction radius of 1/8 r@xmath18 is not possible .
therefore , a slit of 2@xmath93 was extracted from the datacube to match the slitwidth of the optical data for these galaxies ( see section [ sec : proc - optical ] ) . as for the isaac data , the recession velocities and velocity dispersions
were measured with ppxf ( see also table [ tab : galsample ] ) and errors determined from noise statistics . the long - slit spectroscopy from kuntschner ( 2000 ) and kuntschner et al .
( 2002 , ngc1344 ) was used to extract a central spectrum for each galaxy .
the extraction aperture was set to match the effective area covered on the sky by the central isaac data .
the position angle of the isaac observations was set to agree with the optical data .
due to the rather different slit widths involved ( isaac 1 , optical fornax data 23 and 2 ) the extraction ranges vary .
for the two galaxies observed with vlt - sinfoni ( ngc1316 and ngc1399 ) the extraction aperture is limited by the field - of - view of sinfoni . in order to best
match the optical and near - ir data we extracted a central 23 @xmath9 3slit from the optical observations .
again ppxf was used to derive recession velocities and velocity dispersions .
good fits were achieved for all galaxies but ngc1375 where the formal result was @xmath19@xmath3 .
this is well below the resolution limit of the optical fornax data ( kuntschner 2000 ; 4.1 fwhm , r @xmath20 ) and is therefore unreliable . in this paper ,
the velocity dispersion derived from the higher spectral resolution isaac data was adopted and assigned a relatively large error of @xmath2120@xmath3 to reflect the uncertainty of this procedure . a comparison between optical and near - ir measurements of the central velocity dispersions is shown in figure [ fig : sigmas ] .
the determinations are in good agreement with each other for all galaxies , including the two galaxies with significant contributions of young stellar populations ( ngc1344 and ngc1316 ) .
taking the difference between the optical and near - ir measurements as a guideline for our true errors ( including differences in seeing and remaining aperture differences ) we conclude that our internal velocity dispersion error needs to be scaled by a factor of 2.2 . in the following analysis and figures , these conservative scaled errors are used .
the spectral features investigated are listed in table [ tab : features ] . for all but one feature , absorption - line strength is measured as equivalent width via indices , where a central feature bandpass is flanked to the blue and red by pseudo - continuum band - passes .
this method is identical to what is used to measure optical lick indices ( worthey et al .
the index definitions for and were taken from frogel et al .
revised , , and index definitions were used since earlier definitions ( e.g. frster - schreiber 2000 ) did not incorporate continuum band definitions optimized for galaxy spectra with significant velocity broadening . in order to measure the strength of the @xmath22co(2,0 ) absorption we use the definitions of frogel et al .
2001 which provide four pseudo - continuum bandpasses to the blue of the co feature and none to the red .
therefore the continuum used to derive the equivalent width is an extrapolation from the blue wavelength region in contrast to the lick indices where the continuum is based on an interpolation between two flanking bandpasses .
the feature bandpasses of all measured indices are illustrated in figure [ fig : features ] which shows how the absorption features change as a function of temperature in red giant stars .
the spectra shown ( taken from wallace & hinkle 1997 ) are similar in spectral resolution to our spectra ( see figure [ fig : clusterstars ] above ) .
it is important to remember that all of these features are really blends of distinct atomic and molecular features , especially at cooler temperatures .
this point is richly illustrated by the r @xmath16 40 000 spectra presented in wallace & hinkle ( 1996 ) ( see , in particular , their figures 2 and 3 for the and features , respectively ) .
lccccl nai & 21910 21966 & 22040 22107 & 22125 22170 & & frogel et al .
2001 + feia & 22133 22176 & 22250 22299 & 22437 22497 & & this paper@xmath23 + feib & 22133 22176 & 22368 22414 & 22437 22497 & & this paper@xmath23 + cai & 22450 22560 & 22577 22692 & 22700 22720 & & frogel et al .
2001 + mgi & 22700 22720 & 22795 22845 & 22850 22874 & & this paper@xmath23 + index & & feature bandpass & units & source + @xmath22co(2,0 ) & & 22910 23020 & & frogel et al .
2001 + in order to allow a meaningful comparison between observations of stars and galaxies ( with different internal dynamics ) to themselves and to population models , index measurements need to be carefully calibrated to a common system .
there are generally three effects to account for : ( a ) the differences in spectral resolution of the instruments used to obtain the index measurements ; ( b ) the internal velocity broadening of galaxies and ( c ) possibly small systematic offsets caused by continuum shape differences ( see also kuntschner 2000 ) .
the resolution of our isaac observations ( 6.9 fwhm , or r @xmath24 ) was adopted as the nominal resolution .
the vlt - sinfoni spectra were broadened to match this resolution before the measurement of the line - strengths indices .
the observed spectrum of a galaxy is the convolution of the integrated spectrum of its stellar population(s ) by the instrumental broadening and line - of - sight velocity distribution ( losvd ) of the stars .
these effects typically broaden the spectral features , in general reducing the observed line - strength compared to the intrinsic values . in order to compare the index measurements for galaxies we calibrate the indices to zero velocity dispersion and our nominal resolution .
although non - gaussian deviations from the losvd can have significant effects on the losvd correction of indices ( see e.g. , kuntschner 2004 ) , in this study , only the first moments , @xmath25 and @xmath1 , are considered .
this is a reasonable approximation for central spectra studied in this study since expected non - gaussian deviations in this region are negligible .
observations of k and m giants are used to determine the corrections for each index .
these spectra are a reasonable match to the observed galaxy spectra ; however , once stellar population models become available the corrections should be re - derived . by broadening the stellar spectra to velocity dispersions of up to 400@xmath3 in steps of 20@xmath3 a median correction factor is computed for each index , i.e. : @xmath26 and @xmath27 is the median index measured from the stars convolved with the losvd given in brackets .
a losvd corrected index is then @xmath28 .
the actual correction functions computed here are shown in figure [ fig : veldispcorr ] . for the optical data the velocity dispersion corrections from kuntschner ( 2004 ) were used .
the errors on the line - strength indices are evaluated by using error estimates from the spectra and taking into account the errors from the losvd correction procedure .
the optical data was calibrated to the lick / ids system . for a more detailed description of the procedure see kuntschner ( 2000 ) .
multiple isaac observations of ngc1380 and ngc1404 allow us to constrain possible systematic errors in our line strength measurements .
figure [ fig : repeatability ] is plotted with the same y - axis range as figure [ fig : gallinesigma ] but with an smaller @xmath1 range . for the , , and indices our error estimate are comparable to the observed repeatability uncertainty . however , in the case of the index the repeatability uncertainty is significantly larger and we adopt the internal dispersion of the ngc 1380 and ngc 1404 measurements ( 0.3 ) as the formal uncertainty for all subsequent analysis .
line - strength indices can be sensitive to continuum shape calibration errors . in particular index definitions spanning a wide wavelength range
are prone to suffer from this . in our study
the wide band - pass of the co index ( see table [ tab : features ] ) falls in this regime and we consider this to be the explanation for the larger spread between repeat observations .
the origin and behavior of the , , and features as a function of effective temperature in k and m giant stars has been extensively reviewed and discussed by ramrez et al .
( 1997 ) and frster schreiber ( 2000 ) ( who also studied the fe and mg features ) .
most of the stars used in those studies were field stars with bright apparent magnitudes and hence drawn from a narrow metallicity range biased towards solar metallicity .
the behavior of the , , and features as a function of metallicity for galactic globular cluster k and m giants has been investigated by frogel et al .
the metallicity range of that study covers both the metallicity range of the clusters investigated here , as well as the luminosity - weighted mean metallicities of the fornax early - type galaxies . in figure
[ fig : starlinecolor ] , the index strength ( in ) for all six measured spectral features is plotted against @xmath6 color as tabulated by houdashelt , frogel , & cohen ( 1992 ) ( hereafter hfc92 ) for the open cluster stars listed in table [ tab : starsample ] .
these stars are all k or m giant stars . in this context
, redder @xmath6 implies cooler effective temperature and lower surface gravity . within a single cluster ,
it is reasonable to assert that this lower effective temperature is not metallicity - driven .
as previously shown in the studies discussed above , all features increase in depth with redder color ( and hence decreasing @xmath29 ) except , which appears to have a more or less constant value for each cluster regardless of effective temperature ( see also frster schreiber 2000 ) .
the , , and features are highly correlated with @xmath6 and show no dependency on cluster metallicity . for the co feature , this behavior is consistent with the results of frogel et al .
( 2001 ) over a similar cluster metallicity range .
no analogous study exists for the fe features .
however , as shown in figure [ fig : starfe ] , the and index strengths are highly correlated with each other .
hence , it is convenient to define a mean index : = ( fei a + fei b ) / 2 that has a smaller formal uncertainty than either individual fe index alone .
this mean index is used throughout the rest of this paper . as anticipated by the results of frogel et al .
( 2001 ) , at a given @xmath6 ( i.e. @xmath29 ) , giant stars in more metal - rich clusters have stronger and features . the mg feature ( not included in the frogel et al .
study ) appears to have similar behavior , although we note again that within a given cluster ( and hence for a given metallicity ) , the mg feature strength is approximately constant with @xmath6 . in figure
[ fig : starlineline ] , various combinations of k - band line indices are shown . , and are highly correlated with each other ( linear correlation coefficient , @xmath30 0.9 ) . there is a hint that is stronger in the solar - metallicity stars at equivalent and than in the more metal - poor stars .
it is much clearer that at equivalent and , is stronger in the solar - metallicity stars than in the more metal - poor stars . in figure
[ fig : gallinesigma ] , central k - band spectral feature measurements are plotted against central velocity dispersion measured from our data . has not been shown here ( and is ignored for the rest of the paper ) because this line is relatively weak and a reliable velocity dispersion correction could not be derived .
relative to the stellar spectral line strengths shown in figure [ fig : starlinecolor ] , the galaxy measurements ( except ) fall in the middle of the ranges of the stellar measurements .
taken at face value , this is consistent with saying that the luminosity - weighted spectral type for all these galaxies is similar to a solar - metallicity m0 iii star . for most of our galaxies , the feature is stronger than measured in any of our observed stars or any of the field disk giants observed by ramirez et al .
only the supergiants in the fster schreiber have such strong na features .
the unlikely possibility of a supergiant - dominated population can be ruled out by index values that are completely compatible with first - ascent giant stars .
the feature is too strong for these galaxies to be dwarf dominated . at equivalent @xmath1 ,
the galaxies containing younger populations are separated from purely old galaxies . in the galaxies with purely old populations , is highly correlated with @xmath1 , while and @xmath1 are less well - correlated . in contrast , neither nor are well - correlated with @xmath1 reaches a maximum value for @xmath31 while has significant scatter at any given @xmath31 . and in purely old galaxies if ngc 1399 is removed from consideration .
however , there is no sound reason for such data filtering at this time .
obviously , a larger sample is needed to investigate this issue . ]
figure [ fig : gallinemgfeprime ] shows the relationship between the optical index and our new k - band index measurements . is an abundance ratio insensitive optical index combination in the lick system , designed to track total metallicity @xcite .
it is known to be well correlated with @xmath1 ; hence , it is not surprising that figure [ fig : gallinesigma ] and figure [ fig : gallinemgfeprime ] look very similar for the purely old galaxies .
again , the galaxies that contain a young population are separated from the purely old galaxies at equivalent and @xmath1 . figure [ fig : gallineline ] shows the relationships between various pairs of k - band spectral indices ( compare to figure [ fig : starlineline ] ) .
this figure captures several important observational conclusions already foreshadowed above : * relative to the observed cluster stars , the measured values of are significantly stronger in the galaxies .
* in all cases , the galaxies that contain signatures of younger populations ( i.e. stronger and weaker ) have stronger and fe features than comparable purely old galaxies .
* saturates circa @xmath32 150 @xmath3 ( @xmath16 3.2 ) in the same galaxies .
no self - consistent theoretical spectral synthesis models for the interpretation of integrated k - band spectra of early - type galaxies are widely available . development of such models is on - going ( e.g. marmol queralt , 2007 , in preparation ) . until such models exist , we must rely on basic ( and perhaps imperfect ) astrophysical intuition and available tools , such as the high - resolution near - ir spectral atlas of wallace & hinkle ( 1996 ) ( hereafter wk96 ) .
three galaxies in our sample ( ngc 1316 , ngc 1344 , and ngc 1375 ) have stronger and weaker features than the rest of our sample ( see figure [ fig : fornax_pops ] ) .
this suggests a simple , two component stellar population model .
one stellar component is cold in the luminosity - weighted mean , it is presumably old ( @xmath33 8 gyr ) and metal - rich ( [ fe / h ] @xmath33 0.3 ) with spectral properties consistent with the observed central velocity dispersion . in other words , it has properties similar to the purely old galaxies in our sample ( see figure [ fig : fornax_pops ] ) .
the other component is warm it contains a significant number of a / f dwarf or sub - giant stars .
this component could have several origins but only two are considered here .
each has different consequences for k - band spectral index behavior .
first , the warm component could be associated with a young population with a warm ( @xmath162 ) main sequence turnoff ( msto ) .
the msto stars are tied to thermally pulsating asymptotic giant branch ( tp - agb ) stars with bolometric magnitudes that place them above the tip of the first - ascent red giant branch ( trgb ) .
the tp - agb stars are cooler than stars on the first - ascent rgb and hence have an m spectral type . in the underlying luminosity function ( spectral type vs. number of stars ) ,
the m - star bins are relatively more populated than in a purely old galaxy .
the net effect is that integrated 2.2 @xmath0 m spectra should become more m - like .
second , the warm component could be associated with a metal - poor population with a warm msto .
these msto stars are tied to rgb stars with similar luminosity as the corresponding metal - rich rgb stars but with warmer . in the underlying luminosity function
, the k - star bins will be relatively more populated than in a purely old population .
the net effect is to make the integrated 2.2 @xmath0 m spectra more k - like .
can the observations presented here distinguish between these two scenarios ?
consider figure [ fig : galmodel ] the galaxies with relatively strong and weak have relatively stronger ( i.e. more m - like ) near - ir features .
qualitatively , this is consistent with the first scenario .
can this conclusion be better quantified ?
what is the relative ratio of young to old stellar mass ?
are the changes in optical and near - ir spectral features _
quantitatively _ consistent ? to begin to answer these questions , the thomas et al .
( 2003 ) models were used to create very simple two - component models that produced optical index values that matched the observed values in the three galaxies with strong .
one component was old ( 11 gyr ) with metallicity set to [ z / h ] @xmath34 0.35 for ngc 1316 and ngc 1344 and 0.0 for ngc 1375 .
the other component had the same metallicity but a young ( 1 gyr ) age .
the relative mass fractions of these components were varied until the model strength matched the observed strength .
the resultant young mass fractions ( where total mass @xmath34 1 ) were 0.135 , 0.075 , and 0.140 for ngc 1316 , ngc 1344 , and ngc 1375 , respectively .
the implied differential correction between the purely old model and the two - component model was then applied to the observed data ( see figure [ fig : galmodel ] ) .
next , observed strength was adjusted manually until all three galaxies lay within the locus of purely old galaxies with similar @xmath1 in the vs. panel .
as part of this adjustment , @xmath35()/@xmath35 ( ) was forced to agree with the value ( 0.62 ) determined for the observed galactic cluster giant stars ( see figure [ fig : starlineline ] ) .
this is a trend in effective temperature that is relatively insensitive to metallicity .
cooler giant stars ( more late k and early m like ) have strong k - band spectral features .
in addition to the @xmath35()/@xmath35 ( ) measured from the galactic open cluster stars , two other key numerical trends can be computed : @xmath36$^\prime$ } } ) & = & 0.70 ( 1.10 ) \\
\delta({\mbox{$<$fe\,{\small i}$>$}})/\delta({\mbox{[mgfe]$^\prime$ } } ) & = & 0.28 ( 0.44)\end{aligned}\ ] ] ngc 1316 and ngc 1344 could be forced to have consistent trajectories
. however , ngc 1375 ( values shown in parenthesis ) appears to follow somewhat different trajectories .
obviously , this kind of cartoon model is illustrative only and surely hides a plethora of details .
indeed , the real trajectories are unlikely to be linear .
nevertheless , a clear astrophysical conclusion emerges : for the -strong galaxies , the observed optical and near - ir features all change in concert , consistent with a warm msto tied to an extended rgb , which is in turn consistent with the presence of a young stellar component , not a metal - poor stellar component .
liu , graham , & charlot ( 2002 ) found that relative i - band and k - band surface brightness fluctuation strength ( expressed as an sbf color , @xmath37 ) was essentially constant for the majority of the early - type galaxies they studied in fornax .
however , two galaxies in common with our sample ( ngc 1419 and 1427 ) were found to have brighter ( larger ) k - band sbf relative to the mean relationship ( see also mei et al .
a larger @xmath38 is thought theoretically to arise from the presence of an extended giant branch , i.e. more cool , bright , m giants . in the spectroscopic data discussed here ,
there are no indications of such extended giant branches in these galaxies their central line indices are consistent with an integrated stellar populations dominated by an old , metal - rich population . of course , the sbf measurements were made in the outer regions of these galaxies , as opposed to the optical and near - ir observations of the central regions discussed here .
it may be that the integrated luminosity of such a component ( if present at all ) is not high enough in the central region for detection by our method . in purely old galaxies
, we have seen that : * is stronger than in galactic open clusters and is highly correlated with @xmath1 and . * saturates for @xmath31 @xmath3 .
* is somewhat correlated with @xmath1 and , while is not ( unless ngc 1399 is not included in the regression fits , see footnote 2 above ) .
the observed k - band spectral features are named for the dominant elemental species in the _ solar _ spectrum . however , at the effective temperatures of interest here , the on - band index definition and their companion off - band continuum bands contain lines from other elemental species as well ( as ramirez et al .
1997 discuss comprehensively ) . by referring to the high resolution spectra of wk96
, the contribution from these other absorbers can be investigated and used to explain the overall spectral feature behavior . in turn , underlying astrophysical parameters related to galaxy formation and evolution are revealed . in the current galaxy sample , is significantly stronger than observed in the galactic open cluster stars .
the off - bands defined for the feature are relatively line - free , so we can focus on the on - band spectral region ( see the @xmath7 dra , m0 iii spectrum in wk96 , p. 352 ) .
in addition to sodium , sc , si , and ( to a lesser extend ) v are significant absorbers in this region .
origlia et al .
( 1997 ) argued that [ si / fe ] is super - solar in a small sample of early - type galaxies based in measurements of the si i 1.59 @xmath0 m feature .
trager et al .
( 2000 ) have argued that both [ na / fe ] and [ si / fe ] are super - solar in early - type galaxies based on models of optical spectral feature behavior .
the k - band measurements presented here are consistent with enhanced silicon in the observed fornax galaxies and hence indirectly confirm the conclusions of origlia et al . and trager et al .
the feature contains approximately equal absorption contributions from fe , sc and ti while is dominated by fe absorption features with some additional sc absorption ( in wk96 , see the @xmath7 dra , m0 iii spectra on p. 346 and 348 ) .
hence , the index ( a combination of and ) is dominated by absorption lines from fe - peak elements ) . for purely old galaxies with @xmath39 150 @xmath3 ,
is observed to remain constant while ( an index dominated by absoprtion lines from @xmath40-elements ) and ( a total metallicity [ z / h ] indicator ) become stronger as @xmath1 increases .
it is tempting to conclude that above @xmath39 150 @xmath3 luminosity - weighted mean total metallicity [ z / h ] continues to increase as a function of central velocity dispersion , driven by a relative increase of [ @xmath40/h ] , while [ fe / h ] remains constant .
however , within the narrow [ fe / h ] range of observed galactic open cluster stars , is more strongly correlated with effective temperature than cluster [ fe / h ] ( see figure [ fig : starlinecolor ] ) . increasing mean total metallicity [ z / h ] in the galaxies should correspond to cooler mean rgb effective temperature and hence increased . recall that the relationship between and mean rgb effective temperature has already been exploited above to explain the observed behavior of ( and ) in galaxies with young stellar components
the dilemma is clear : how can a temperature - sensitive index like remain constant while mean rgb effective temperature decreases due to increasing [ z / h ] ? without appropriate population synthesis models or stellar spectra , only a few speculative thoughts can be offered . in their detailed comparison of non - solar aboundance population synthesis models with measurements of spectral indices in the optical lick system , trager et al .
( 2000 ) concluded that their best - fit models included enhancements in c ,
n , na , and si ( among others ) .
as [ @xmath40/h ] increases with @xmath1 ( as suggested by increasing ) , stronger cn bands in the and on - band and off - band windows ( see the wk96 spectra referenced above and the band definitions in table [ tab : features ] ) may have the effect of depressing the local continuum and hence decreasing the value of @xmath34 ( @xmath8 ) /2 .
for example , trager et al . noted a similar effect from c@xmath41 bands in the optical mg _ b _ index .
therefore , it may be that above some critical mean [ @xmath40/h ] and below some critical mean effective temperature , stays within a narrow range and provides no useful information about mean [ z / h ] , [ fe / h ] , or effective temperature . a conclusive astrophysical explanation of in early - type galaxies with @xmath39 150 @xmath3 awaits larger galaxy samples , observations of appropriate stars ( e.g. in the galactic bulge ) , and detailed population synthesis models .
the feature is dominated by the @xmath42co(2,0 ) bandhead and has no significant absorption components from other elements .
like above , is highly correlated with @xmath29 in the observed galactic open cluster stars and reaches larger values than observed in the galaxies . as a function of central velocity dispersion @xmath1 in the observed galaxies , does not appear to saturate ( or at least not as definitively as ) .
this adds credence to the suspicion that is not tracing mean rgb effective temperature in galaxies with @xmath39 150 @xmath3 .
based on the current measurements , no conclusions can be reached about relative carbon or oxygen abundance in these galaxies .
no obvious explanation presents itself for the behavior of in this galaxy sample .
the feature is intrinsically complex , consisting of contributions from many absorbers : ca , s , si , and ti ( all @xmath40-elements ) as well as sc and fe ( see @xmath7 dra , mo iii spectrum in wk96 , p. 342 ) . as a function of in the range of interest ,
some of the contributing absorption lines get stronger ( sc and ti , the former faster than the latter ) , some features get weaker ( si ) , and some remain roughly constant ( ca , s , and fe ) ( cf .
ramirez et al .
relative to solar abundance ratios , trager et al .
( 2000 ) have argued that si and s are over - abundant and ca is under - abundant in early - type galaxies while [ ti / fe ] and [ sc / fe ] have their solar values . in the galactic open cluster stars observed here , the feature is stronger at a given @xmath6 ( @xmath43 ) in the solar - metallicity stars than in the more metal - poor stars . within the galaxies
observed here , appears to have significant scatter at any given @xmath1 or .
no obvious explanation for this scatter ( or the global behavior of with @xmath1 and ) presents itself at this time .
using new , moderate resolution ( @xmath44 ) k - band spectra , spectral indices have been measured in the central regions of eleven early - type galaxies in the nearby fornax cluster .
based on these measurements , the following conclusions were reached : 1 .
the feature is much stronger in these fornax early - type galaxies than observed in solar - metallicity galactic open cluster stars .
this is attributed to relative [ si / fe ] ( and possible [ na / fe ] ) differences between the open cluster stars and the fornax galaxies , i.e. both are larger in the fornax galaxies than in the cluster stars .
2 . in various near - ir
diagnostic diagrams , galaxies with optical indices indicative of a warm stellar component are clearly separated from galaxies dominated by colder , presumably old ( @xmath5 8 gyr ) stellar populations .
changes in the near - ir spectra features are consistent with the presence of an cool component dominated by late k and/or early m giants stars . in combination ,
the optical and near - ir observations are consistent with the presence of a young stellar component with a warm msto and a significant extended giant branch consisting of tp - agb stars .
3 . for detecting a young stellar component
, the vs. @xmath1 or vs. @xmath1 diagnostic diagram seems as efficient as using h@xmath4 vs. ( or other similar combinations of optical indices ) .
the near - ir features have the additional advantage that no emission - line correction is needed ( as it is necessary for the stronger optical balmer lines ) .
the index saturates in galaxies with central velocity dispersion @xmath1 @xmath33 150 @xmath3 dominated by old ( @xmath5 8 gyr ) stellar populations . for @xmath39 150 @xmath3
, these fe features are unlikely to be useful for investigating stellar population differences between early - type galaxies . above @xmath39 150 @xmath3 ,
the continued increase in ( and ) strength with @xmath1 coupled with constant presents an astrophysical challenge .
although it is tempting to conclude that [ fe / h ] reaches a maximum value , while [ @xmath40/h ] ( and hence [ z / h ] ) continues to increase , more observational and population synthesis work is needed to understand conclusively the behavior of in these high - mass early - type galaxies . adding these near - ir indices to the standard diagnostic toolkit for analyzing the integrated light of early - type galaxies clearly has great potential
to develop this potential , three obvious steps are needed : observe more galaxies over a larger range of central velocity dispersion ( including field galaxies with existing optical data ) and extend current population synthesis models to the near - ir .
we know that various groups are working on both steps
. it would also be useful to study the radial behavior of these indices within individual galaxies to compare and contrast index behavior between galaxies .
for the foreseeable future , the study of the central populations in early - type galaxies will remain the study of integrated light . as distance increases ,
our ability to study the central regions of early - type galaxies within metric apertures equivalent to nearby galaxies relies on the high _ spatial _ resolution spectroscopy achievable from space or with adaptive optics on the ground . in the former case , the james webb space telescope represents the frontier yet , no optical spectrograph that works below 0.8 @xmath0 m is currently planned . in the latter case , the frontier will be shaped by further development of adaptive optics systems and these systems will achieve their best performance beyond 1 @xmath0 m . hence , understanding how to interpret near - ir spectral indices in nearby galaxies is key to facilitating the kind of investigation and characterization of more distant early - type galaxies that will not be possible at optical wavelengths .
we thank the anonymous referee for a quick , enthusiastic , and helpful review that drove us to clarify several key points .
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 .
this research has also made use of the simbad database , operated at cds , strasbourg , france .
thanks to valentin ivanov and olivier marco for observing assistance during the acquisition of the isaac data .
drs warmly thanks eso santiago for hospitality during july 2003 science leave , especially danielle alloin , during which much of these data were processed .
we acknowledge various helpful discussions with andrew stephans , daniel thomas , claudia maraston , scott trager and guy worthey on topics related to stellar populations in early - type galaxies .
thanks to chris lidman and rachael johnson for various discussions about reducing isaac spectroscopic data .
many of the science and technology concepts in this paper grew out of late night conversations between drs and richard elston ( 1961 2004 )
friend , colleague , fellow observer while we were both postdocs at kitt peak .
richard thanks . | the strong spectral features near 2.2 @xmath0 m in early - type galaxies remain relatively unexplored . yet , they open a tightly focused window on the coolest giant stars in these galaxies a window that can be used to explore both age and metallicity effects . here , new measurements of k - band spectral features are presented for eleven early - type galaxies in the nearby fornax galaxy cluster .
based on these measurements , the following conclusions have been reached : ( 1 ) in galaxies with no signatures of a young stellar component , the k - band index is highly correlated with both the optical metallicity indicator and the central velocity dispersion @xmath1 ; ( 2 ) in the same galaxies , the k - band fe features saturate in galaxies with @xmath2 @xmath3 while ( and ) continues to increase ; ( 3 ) [ si / fe ] ( and possibly [ na / fe ] ) is larger in all observed fornax galaxies than in galactic open clusters with near - solar metallicity ; ( 4 ) in various near - ir diagnostic diagrams , galaxies with signatures of a young stellar component ( strong , weak ) are clearly separated from galaxies with purely old stellar populations ; furthermore , this separation is consistent with the presence of an increased number of m - giant stars ( most likely to be thermally pulsating agb stars ) ; ( 5 ) the near - ir vs. @xmath1 or vs. @xmath1 diagrams discussed here seem as efficient for detecting putatively young stellar components in early - type galaxies as the more commonly used age / metallicity diagnostic plots using optical indices ( e.g h@xmath4 vs. ) .
the combination of these spectral indices near 2.2 @xmath0 m with high _ spatial _ resolution spectroscopy from ground - based or space - based observatories promises to provide new insights into the nature of stellar populations in the central regions of distant early - type galaxies . |
the large majority of kuiper belt objects ( kbos ) contain no detectable volatile ices on their surfaces , but a small number of the largest objects have been found to have signatures of ch@xmath0 , co , or n@xmath1 , all ices with high vapor pressures at kuiper belt temperatures .
after the discovery of volatiles on the surfaces of eris @xcite , makemake @xcite , and sedna , @xcite proposed a simple method for assessing the possibility of volatile retention on kbos .
for each relevant ice , they compared the volatile loss due to jean s escape the slowest of many possible escape mechanisms to the total volatile inventory of the object and divided the kuiper belt into objects which could and could not have retained that ice over the age of the solar system .
only a handful of objects are massive enough or cold enough to be able to retain volatiles .
their model provided a compelling explanation of the low abundance of n@xmath1 on makemake @xcite , which is smaller than pluto and eris , and was also used to successfully predict the presence of methane on quaoar @xcite . to date , the volatile retention model has been completely successful predicting which objects will and which will not have detectable surface volatiles , with the unique exception being the large kbo haumea , which is the parent body of the only collisional family known in the kuiper belt @xcite and clearly had an unusual history .
we provide an update to the @xcite calculations in figure 1 .
we have used new vapor pressure data from and , where possible , have used measured sizes and masses of the largest kbos . for quaoar ,
the current measured diameter is 890 @xmath2 70 km implying a density of 4.2 @xmath2 1.3 g @xmath3 @xcite , but we assume the upper limit of size as the smaller sizes lead to physically implausible densities .
( note that even for the smaller size and higher density , however , quaoar is still expected to retain surface methane . )
the size of 2007 or10 is unmeasured , so , as will be justified below , we assume that it has an albedo identical to the 0.18 albedo of quaoar that gives the size that we assume above , though we allow albedo uncertainties of 50% in either direction .
as in @xcite , we calculate an `` equivalent temperature '' for each object by integrating the volatile loss through the object s entire orbit and determining the temperature that an object in a circular orbit would have to have to lose the volatile at that rate . for our assumed albedo range ,
2007 or10 is somewhere between the fourth and seventh largest object known in the kuiper belt .
its potential size spans the range between the small volatile poor objects and the handful of volatile rich objects .
2007 or10 is thus an excellent test object for our understanding of volatile retention in the outer solar system .
we explore the surface composition of this object below using a combination of near - ir spectroscopy and multi - wavelength photometry .
the low - resolution , near - infrared spectrum of 2007 or10 was obtained on 2010 september 20 ( ut ) using the folded - port infrared echellette ( fire ) spectrograph on the 6.5 m magellan baade telescope @xcite .
fire s prism - dispersed mode provides continuous coverage of the 0.852.45 @xmath4 band with a variable resolution of @xmath5 = 250350 .
2007 or10 was acquired and its motion confirmed using fire s @xmath6-band imaging channel .
the source was maintained on the 0@xmath76 slit by manual corrections to sidereal tracking .
two series of abba dither exposure sequences were obtained with integrations of 120 s at an average airmass of 1.04 .
these were followed by a single abba sequence of the g2 v star hd 211544 ( @xmath8=10.9 ) at a similar airmass .
exposures of a quartz flat field lamp ( set at 1.2 v and 2.2 v ) and arc lamps ( near ) were obtained for pixel response and wavelength calibration .
data were reduced using the methods described in @xcite .
the spectrum was converted into relative reflectance as well as corrected for telluric absorption and instrument response by dividing the raw spectrum of 2007 or10 by the spectrum of the solar type star hd 211544 .
photometry were obtained with the wide - field camera 3 on the hubble space telescope during cycles 17 ( go program 11644 ) and 18 ( go program 12234 ) . in cycle 17 ,
two 130 s exposures were taken in the f606w and f814w filters , and two 453 s exposures were taken in the f139 m , and f153 m filters . during the cycle 18 observations , two exposures were acquired in each of the f606w , f775w , f098 m , f110w with exposure times of 128s , 114s , 115s , and 207 s respectively . as well , four exposures of 275 s were acquired in the f127 m filter . for both the cycle 17 and 18 observations , 3 " dithers were applied between image pairs to reduce the effects of cosmic rays and pixel defects , with the exception of the f127 m observations , in which 2 images were taken at each dither position .
all observations in a cycle were acquired within a single orbit , minimizing the effect of any light curve or10 may have .
all data were processed through calwfc3 version 2.3 , the standard wfc3 image processing pipeline @xcite .
circular apertures were used to measure the photometry . _
tiny tim _ version 7.1 psfs @xcitewere used to generate infinite aperture corrections as well as interpolate over any bad pixels flagged during the image reductions .
fluxes were converted to relative reflectance by comparing to fluxes computed using the the _ calcphot _ routine for a model solar spectrum @xcite provided as part of the _ iraf _ package _ stsdas.synphot_. approximate absolute reflectances were then obtained by scaling the f606w relative reflectance to a value of 0.18 , our assumed albedo of 2007 or10 .
all are shown in table 1 . before calculating relative reflectances ,
the cycle 18 magnitudes were adjusted upward by 0.03 to account for the difference in the f606w magnitudes between the two epochs .
the small magnitude difference is an expected consequence of object rotation .
cccc cycle 17 + f606w & 21.68@xmath2 0.02 & -27.00 & 0.18 + f814w & 21.39@xmath2 0.01 & -26.54 & 0.37 + f139 m & 22.06@xmath2 0.01 & -25.34 & 0.59 + f153 m & 22.47@xmath2 0.02 & -25.09 & 0.51 + cycle 18 + f606w & [email protected] & -27.00 & 0.18 + f775w & [email protected] & -26.64 & 0.34 + f098 m & [email protected] & -26.13 & 0.52 + f110w & [email protected] & -25.78 & 0.58 + f127 m & [email protected] & -25.54 & 0.58 +
figure 2 shows the fire reflectance spectrum with the wfc3 photometric points overlaid .
we scale all point to an albedo of 0.18 in the f606w filter , though we note that the true value of the albedo has not been measured . to increase the signal - to - noise in the fire spectrum
, we also plot the median of every 32 spectral channels , oversampled by a factor of two , to simulate how the spectrum would appear to a lower resolution spectrograph .
uncertainties on these data points are obtained by calculating the median absolute deviation of eache 32 channel sample , which we multiply by 1.48 to obtain what would be the standard deviation in a normally distributed sample , and then divide by the square root of the number of spectral channels , approximating the standard deviation . the fire spectrum and the wfc3 photometry are in broad agreement in the area of overlap , though the match is imperfect .
we suspect that the differences are due to differential refraction in the fire data , which , for these early attempts at tracking a moving object , were not obtained with the slit aligned along the parallactic angle .
both data sets show , in particular , a very red optical slope and a distinct absorption around 1.5 @xmath9 m .
the fire spectrum shows an additional broad absorption feature near 2.0 @xmath9 m and , potentially , additional features redward .
absorptions at 1.5 and 2.0 @xmath9 m are the characteristics features of water ice , which is frequently found on the largest kbos @xcite .
figure 3 compares the spectrum of 2007 or10 to a modeled spectrum of a surface consisting of a mixture of water ice and a neutral material .
we place no special significance on the precise water ice surface model , as many different types of specific parameters yield similar modeled spectra , but , for concreteness we use the water ice absorption coefficients of @xcite and construct a simple hapke model @xcite with 50@xmath9 m grains at 50 k spatially mixed with equal amounts of a neutral material with an albedo of 80% .
the model spectrum is sampled at the same resolution as the smoothed spectrum of the object . even at the low signal - to - noise ratio of the data
, the water ice model provides an excellent match to the spectrum redward of 1.4 @xmath9 m . at shorter wavelengths , however , 2007 or10 is significantly redder than water ice .
most of the kbos with significant water ice absorption , such as orcus , haumea , and the haumea family members , have nearly neutral optical colors @xcite .
the most notable exception to this trend is quaoar , which has water ice absorptions nearly identical in depth to those of 2007 or10 and is almost as red .
a spectrum of quaoar @xcite over the same wavelength range is shown , for comparison , in figure 3 .
the coloring of quaoar was suggested by @xcite to be due to the effects of methane , which turns red with irradiation @xcite .
quaoar , as seen in figure 1 , is barely large enough or cold enough to retain methane on its surface and the amount left is small and detectable only at high signal - to - noise .
2007 or10 , depending on its precise size , could be in a similar regime of volatile retention as quaoar .
if true , the quaoar - like optical color of 2007 or10 could be a signature of the retention and irradiation of methane on the kbo . indeed , if 2007 or10 is assumed to have the same albedo as quaoar as the close match in visible and near - infrared spectra might suggest
2007 or10 sits in almost precisely the same volatile loss regime as quaoar .
while the perihelion distance of 2007 or10 of only 33.6 au ( compared to quaoar s of 41.6 au ) makes it significantly hotter than quaoar , the significantly larger size of 2007 or10 allows its larger gravitational pull to nonetheless potentially retain methane .
if the hypothesis that the extreme red coloration on 2007 or10 is like that of quaoar caused by the irradiation of a small amount of remaining surface methane is correct , we predict that methane absorption should be visible on 2007 or10 as it is on quaoar .
figure 3 also shows a simple model of a surface including solid methane and a neutral component .
the model is simply the water model from above but with the absorption coefficients of methane @xcite replacing those of water .
the spectrum is shifted upward by 0.3 units for clarity .
the strongest absorption feature of methane , at 2.3 @xmath9 m does indeed correspond with a large variation from the water ice model .
we conclude that while these data do not have sufficient signal - to - noise , particularly in the k - band , to positively detect methane on 2007 or10 , the existence of this volatile is plausible , and would provide a pleasing explanation for the extreme red coloration of the kbo . the large kuiper belt object 2007 or10 provides an excellent test of our understanding of volatile loss and retention on the surfaces of objects in the outer solar system . while the size of 2007 or10 has yet to be measured , the simple assumption that it has an identical albedo to quaoar the object whose spectrum its spectrum most resembles places 2007 or10 into a regime where it would be expected to retain trace amounts of methane on its surface . such an object would be expected to have red optical coloration from methane irradiation , which both quaoar and 2007 or10 do have .
in addition , such an object should have detectable signatures of methane if observed at sufficient signal - to - noise .
such methane signatures have been detected on quaoar , but require higher signal - to - noise to positively identify on 2007 or10 .
while additional measurements of the size and spectrum of 2007 or10 are clearly required , we conclude that volatile retention models @xcite appear to continue to flawlessly predict both the presence and absence of volatiles on all objects in the kuiper belt which have been observed to date . ,
r. a. , burgasser , a. j. , bernstein , r. a. , bigelow , b. c. , fishner , j. , forrest , w. j. , mcmurtry , c. , pipher , j. l. , schechter , p. l. , & smith , m. 2008 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
7014 , society of photo - optical instrumentation engineers ( spie ) conference series , r. a. , burgasser , a. j. , bochanski , j. j. , schechter , p. l. , bernstein , r. a. , bigelow , b. c. , pipher , j. l. , forrest , w. , mcmurtry , c. , smith , m. j. , & fishner , j. 2010 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
7735 , society of photo - optical instrumentation engineers ( spie ) conference series | we present photometry and spectra of the large kuiper belt object 2007 or10 .
the data show significant near - infrared absorption features due to water ice .
while most objects in the kuiper belt with water ice absorption this prominent have the optically neutral colors of water ice , 2007 or10 is among the reddest kuiper belt objects known .
one other large kuiper belt object
quaoar has similar red coloring and water ice absorption , and it is hypothesized that the red coloration of this object is due to irradiation of the small amounts of methane able to be retained on quaoar .
2007 or10 , though warmer than quaoar , is in a similar volatile retention because it is sufficiently larger that its stronger gravity can still retain methane .
we propose , therefore , that the red coloration on 2007 or10 is also caused by the retention of small amounts of methane .
positive detection will require spectra of methane on 2007 or10 will require spectra with higher signal - to - noise .
models for volatile retention on kuiper belt objects appear to continue to do an excellent job reproducing all of the available observations . |
soft x ray transients ( sxrts ) , when in outburst , show properties similar to those of persistent low mass x ray binaries containing a neutron star ( lmxrbs ; white et al .
1984 ; tanaka & shibazaki 1996 ; campana et al . 1998 ) . the large variations in the accretion rate that are characteristic of sxrts allow the investigation of a variety of regimes for the neutron stars in these systems which are inaccessible to persistent lmxrbs .
while it is clear that , when in outbursts , sxrts are powered by accretion , the origin of the low luminosity x
ray emission that has been detected in the quiescent state of several sxrts is still unclear .
an interesting possibility is that a millisecond radio pulsar ( msp ) turns on in the quiescent state of sxrts ( stella et al .
this would provide a missing link " between persistent lmxrbs and recycled msps .
aql x-1 is the most active sxrt known : more than 30 x ray and/or optical outbursts have been detected so far .
the companion star has been identified with the k1v variable star v1333 aql and an orbital period of 19 hr has been measured ( chevalier and ilovaisky 1991 ) .
the outbursts of aql x-1 are generally characterised by a fast rise ( 510 d ) followed by a slow decay , with an @xmath4folding time of 3070 d ( see tanaka & shibazaki 1996 and campana et al .
1998 and references therein ) .
type i x ray bursts were discovered during the declining phase of an outburst ( koyama et al .
1981 ) , testifying to the presence of a neutron star .
ray luminosities are in the @xmath5 range ( for the @xmath6 kpc distance inferred from its optical counterpart ; thorstensen et al . 1978 ) .
close to the outburst maximum the x ray spectrum is soft with an equivalent bremsstrahlung temperature of @xmath7 kev .
sporadic observations of aql x-1 during the early part of the outburst decay ( czerny et al .
1987 ; tanaka & shibazaki 1996 ; verbunt et al .
1994 ) showed that when the source luminosity drops below @xmath8 the spectrum changes to a power law with a photon index of @xmath2 , extending up to energies of @xmath9 kev ( harmon et al . 1996 ) .
rosat pspc observations revealed aql x-1 in quiescence on three occasions at a level of @xmath1 ( 0.42.4 kev ; verbunt et al .
1994 ) . in this lower energy band
the spectrum is considerably softer and consistent with a black body temperature of @xmath10 kev .
an outburst from aql x-1 reaching a peak luminosity of @xmath11 ( 210 kev ) was discovered and monitored starting from mid - february , 1997 with the rossixte all sky monitor ( asm ; levine et al .
six observations were carried out with the bepposax narrow field instruments ( nfis ) starting from march 8@xmath12 , 1997 ( see table 1 ) , with the aim of studying the final stages of the outburst decay .
1a shows the light curve of the aql x-1 outburst as observed by the rossixte asm and bepposax mecs .
the first part of the outburst can be fit by a gaussian with sigma @xmath13 d. this is not uncommon in sxrts ( e.g. in the case of 4u 160852 ; lochner & roussel - dupr 1994 ) .
the flux decay rate changed dramatically around mjd 50512 ( march 5@xmath12 , 1997 ) . at the time of the first bepposax observation ( which started on march 8@xmath12 , 1997 ) the source luminosity was decreasing very rapidly , fading by about 30% in 11 hr , from a maximum level of @xmath14 .
the second observation took place on march 12@xmath12 , 1997 when the source , a factor of @xmath15 fainter on average , reduced its flux by about 25% in 12 hr .
in the subsequent four observations the source luminosity attained a constant level of @xmath16 , consistent with previous measurements of the quiescent luminosity of aql x-1 ( verbunt et al . 1994 ) . the sharp decrease after mjd 50512
is well described by an exponential decay with an @xmath4folding time @xmath17 d. the bepposax lecs , mecs and pds spectra during the fast decay phase , as well as those obtained by summing up all the observations pertaining to quiescence , can be fit with a model consisting of a black body plus a power law ( see table 2 and fig .
the soft black body component remained nearly constant in temperature ( @xmath18 kev ) , but its radius decreased by a factor of @xmath19 from the decay phase to quiescence .
the equivalent radius in quiescence ( @xmath20 km ) is consistent with the rosat results ( verbunt et al .
the power law component changed substantially from the decay phase to quiescence : during the decay the photon index was @xmath21 , while in quiescence it hardened to @xmath22 .
the two values are different with @xmath23 confidence ( see table 1 ) . the ratio of the 0.510 kev luminosities in the power law and black body components decreased by a factor of five between the first bepposax observation and quiescence .
[ cols="^,^,^,^,^,^ " , ] @xmath25 spectra from the lecs and mecs ( and pds for the first observation ) detectors have been considered .
the spectra corresponding to the quiescent state have been summed up , in order to increase the statistics and an upper limit from the summed pds data was also used to better constrain the power law slope . @xmath26 unabsorbed x ray luminosity in the energy range 0.510 kev . in the case of the first observation
the pds data were included in the fit ( the unabsorbed 0.5100 kev luminosity amounts to @xmath27 ) .
the bepposax observations enabled us to follow for the first time the evolution of a sxrt outburst down to quiescence .
the sharp flux decay leading to the quiescent state of aql x-1 is reminiscent of the final evolution of dwarf novae outbursts ( e.g. ponman et al .
1995 ; osaki 1996 ) , although there are obvious differences with respect to the x ray luminosities and spectra involved in the two cases , likely resulting from the different efficiencies in the gravitational energy release between white dwarfs and neutron stars .
models of low mass x ray transient outbursts hosting an old neutron star or a black hole are largely built in analogy with dwarf novae outbursts . in particular , van paradijs ( 1996 ) showed that the different range of time - averaged mass accretion rates over which the dwarf nova and low mass x
ray transient outbursts were observed to take place is well explained by the higher level of disk irradiation caused by the higher accretion efficiency of neutron stars and black holes .
however , the outburst evolution of low mass x ray transients presents important differences .
in particular , the steepening in the x ray flux decrease of aql x-1 has no clear parallel in low mass x
ray transients containing black hole candidates ( bhcs ) .
the best sampled light curves of these sources show an exponential - like decay ( sometimes with a superposed secondary outburst ) with an @xmath4folding time of @xmath29 d and extending up to four decades in flux , with no indication of a sudden steepening ( chen et al .
in addition , bhc transients display a larger luminosity range between outburst peak and quiescence than neutron star sxrts ( garcia et al . 1998 and references therein ) .
being the mass donor stars and the binary parameters quite similar in the two cases , it appears natural to attribute these differences to the different nature of the underlying object : neutron stars possess a surface and , likely , a magnetosphere , while bhcs do not .
when in outburst accretion down to the neutron star surface takes place in sxrts , as testified by the similarity of their properties with those of persistent lmxrbs , especially the occurrence of type i bursts and the x ray spectra and luminosities . the mass inflow rate during the outburst decay decreases , causing the expansion of the magnetospheric radius , @xmath30 .
thus , accretion onto the neutron star surface can continue as long as the centrifugal drag exerted by the corotating magnetosphere on the accreting material is weaker than gravity ( illarionov & sunyaev 1975 ; stella et al .
this occurs above a limiting luminosity @xmath31 , where @xmath32 is the gravitational constant ; @xmath33 , @xmath34 , @xmath35 g and @xmath36 ms are the neutron star mass , radius , magnetic field and spin period , respectively ( here and in the following we assume @xmath37 and @xmath38 cm ) . as @xmath30 reaches the corotation radius , @xmath39 , accretion onto the surface is inhibited and a lower accretion luminosity ( @xmath40 ) of @xmath41 is released . after this luminosity gap
the source enters the propeller regime .
if the mass inflow rate decreases further , the expansion of @xmath30 can continue up to the light cylinder radius , @xmath42 , providing a lower limit to the accretion luminosity that can be emitted in the propeller regime @xmath43 .
below @xmath44 the radio pulsar mechanism may turn on and the pulsar relativistic wind interacts with the incoming matter pushing it outwards .
matter inflowing through the roche lobe is stopped by the radio pulsar radiation pressure , giving rise to a shock front ( illarionov & sunyaev 1975 ; shaham & tavani 1991 ) .
clearly these regimes have no equivalent in the case of black hole accretion . during the february - march 1997 outburst of aql x-1 , rossixte observations led to the discovery of a nearly coherent modulation at @xmath45 hz ( @xmath46 ms ) during a type i x ray burst .
a single qpo peak , with a centroid frequency ranging from @xmath47 to 830 hz , was also observed at two different flux levels , when the persistent luminosity was @xmath48 and @xmath49 ( zhang et al .
1998 ) . in the presence of a single qpo peak , the magnetospheric and sonic point beat frequency models ( alpar & shaham 1985 ; miller et al . 1997 ) interpretation is ambiguous in that the qpo peak could represent either the keplerian frequency at the inner disk boundary or the beat frequency .
moreover , the burst periodicity at @xmath45 hz may represent the neutron star spin frequency , @xmath50 , or half its value ( zhang et al . 1998 ) . in either cases ,
the possibility that accretion onto the neutron star surface takes place even in the quiescent state of aql x-1 faces serious difficulties : for a quiescent luminosity of order @xmath51 a magnetic field of only @xmath52 g would be required , in order to fulfill the condition @xmath53 .
for such a low magnetic field , aql x-1 and , by inference , lmxrbs with khz qpos can hardly be the progenitors of recycled msps .
more crucially , the marked steepening in the outburst decay that takes place below @xmath54 , is accompanied by a marked spectral hardening , resulting from a sudden decrease of the flux in the black body spectral component .
this is clearly suggestive of a transition taking place deep in the gravitational well of the neutron star , where most of the x
rays are produced .
the most appealing mechanism is a transition to the propeller regime , where most of the inflowing matter is stopped at the magnetospheric boundary ( zhang , yu & zhang 1998 ) . in fig .
1a , the luminosity at mjd 50512 is identified with @xmath55 and the lower horizontal lines indicate the luminosity interval during which aql x-1 is likely in the propeller regime . additional information on the neutron star magnetic field ( and spin ) can be inferred as follows .
the observed ratio of the luminosity , @xmath56 , when the qpo at @xmath57 hz were detected and the luminosity @xmath58 when the centrifugal barrier closes is @xmath59 . at @xmath58 the keplerian frequency of matter at the magnetospheric boundary is , by definition , equal to the spin frequency , i.e. @xmath60 or @xmath61 hz for aql x-1 .
based on beat - frequency models , at @xmath56 the keplerian frequency at the inner disk boundary can be either @xmath62 hz or @xmath63 hz , depending on whether the single khz qpos observed corresponds to the keplerian or the beat frequency . in the magnetospheric beat - frequency models , simple theory predicts that the keplerian frequency at the magnetospheric boundary is @xmath64 ; in the radiation pressure - dominated regime relevant to the case at hand , the ghosh and lamb ( 1992 ) model predicts instead @xmath65 .
therefore we expect @xmath66 and @xmath67 , in the two models , respectively .
such a low ratio clearly favors the interpretation in which @xmath62 hz and @xmath68 hz . in the sonic point beat - frequency model ( miller et al .
1997 ) , the innermost disk radius is well within the magnetosphere , implying that the keplerian frequency at the magnetospheric boundary is @xmath69 . in this case
all possible combinations of @xmath50 and @xmath70 are allowed . by using the observed @xmath58 , a neutron star magnetic field of @xmath71 g ( depending on the adopted model of the disk - magnetosphere interaction )
is obtained in the case @xmath68 hz and @xmath72 g in the case @xmath73 hz .
once in the propeller regime , the accretion efficiency decreases further as the magnetosphere expands for decreasing mass inflow rates ( @xmath74 ) .
the @xmath75 d exponential - like luminosity decline observed with bepposax is considerably faster than the propeller accretion luminosity extrapolated from the first part of the outburst ( e.g. the gaussian profile shown by the dashed line in fig .
we note here that the spectral transition accompanying the onset of the centrifugal barrier may also modify the irradiation properties of the accretion disk , contributing to x ray luminosity turn off .
alternatively , an active contribution of the `` propeller '' mechanism or the neutron star spin - down energy dissipated into the inflowing matter can not be excluded .
it is unlikely that the quiescent luminosity of aql x-1 is powered by magnetospheric accretion in the propeller regime . as shown in fig .
1a , the quiescent x ray luminosity is probably lower than the minimum magnetospheric accretion luminosity @xmath76 allowed in the propeller phase ( this remains true for @xmath77 g if @xmath68 hz , and for @xmath78 g if @xmath73 hz ) .
moreover the bepposax x ray spectrum shows a pronounced decrease in the power law to black body flux ratio together with a flattening of the power law component between the fast decay phase and quiescence , suggesting that a transition to shock emission from the interaction of a radio pulsar wind with the matter outflowing from the companion star has taken place . note that an x ray spectrum with a slope of @xmath79 has been observed from the radio pulsar psr b125963 immersed in the wind of its be star companion .
models of this interaction predict that a power law x ray spectrum with a slope around @xmath79 should be produced for a wide range of parameters ( tavani & arons 1997 ) . the additional soft x
ray component observed during the outburst decay ( see table 2 ) might be emitted at the polar caps as a result of the residual neutron star accretion in the propeller phase .
note that the equivalent black body radius decreases for decreasing x ray luminosities , just as it would be expected if the magnetospheric boundary expanded .
alternatively , the black body - like spectral component observed in quiescence could be due to the streaming of energetic particles that hit the polar caps , in close analogy to the soft x ray component observed , in msps , at the weaker level of @xmath80 ( becker & trmper 1997 ) . assuming a magnetic field in the range derived in section 3.1 ( i.e. @xmath81 g for @xmath82 hz and @xmath83 g for @xmath84 hz )
, we can consistently explain the @xmath1 quiescent x ray luminosity as powered by a radio pulsar enshrouded by matter outflowing from the companion star , if the conversion efficiency of spin - down luminosity to x ray is @xmath85% .
this is consistent with modeling and observations of enshrouded pulsars ( tavani 1991 ; verbunt et al .
1996 ) .
there are chances of observing a msp ( a simple scaling from msps implies a signal at 400 mhz of @xmath86 mjy ; see kulkarni et al .
1990 ) , even though the emission would probably be sporadic , like in the case of the pulsar psr 174424a due to the large amount of circumstellar matter ( see lyne et al . 1991 ; shaham & tavani 1991 ) . in summary aql x-1 appears to provide the first example of an old fast rotating neutron stars undergoing a transition to the propeller regime at first , followed by a transition to the radio pulsar regime , as the transient x ray emission approaches its quiescent level .
therefore , aql x-1 ( and possibly sxrts in general ) likely represents the long - sought `` missing link '' between lmxrbs and recycled msps . | we report on the march - april 1997 bepposax observations of aql x-1 , the first to monitor the evolution of the spectral and time variability properties of a neutron star soft x
ray transient from the outburst decay to quiescence .
we observed a fast x ray flux decay , which brought the source luminosity from @xmath0 to @xmath1 in less than 10 days .
the x ray spectrum showed a power law high energy tail with photon index @xmath2 which hardened to @xmath3 as the source reached quiescence .
these observations , together with the detection by rossixte of a periodicity of a few milliseconds during an x ray burst , likely indicate that the rapid flux decay is caused by the onset of the propeller effect arising from the very fast rotation of the neutron star magnetosphere . the x ray luminosity and hard spectrum that characterise the quiescent emission can be consistently interpreted as shock emission by a turned - on rotation - powered pulsar . |
Subsets and Splits