Generalized classical and quantum dynamics within a nonextensive approach

Generalized classical and quantum dynamics within a nonextensive approach

A. Lavagno

Dipartimento di Fisica, Politecnico di Torino, I-10129 Torino, Italy, and Istituto Nazionale di Fisica Nucleare, Sezione di Torino, Italy

ABSTRACT

On the basis of generalized classical kinetic equations, reproducing the stationary distribution of the Tsallis nonextensive thermostatistics, we formulate two generalized Schr¨odinger equations which satisfy the basic assumptions of the quantum mechanics under an appropriate generalization of the operator properties. Moreover, we study the generalization of the previously introduced dynamic equations in a relativistic regime and we apply our results to the study of the rapidity distribution in the relativistic heavy-ion collisions.

1 Introduction

Let us review some basic assumptions of the nonextensive thermostatistics that will be useful in view of the generalized classical and quantum dynamics we will describe.

Starting point of the Tsallis' generalization of the Boltzmann-Gibbs statistical mechanics is the introduction of a q-deformed entropy functional defined, in a phase space system, as [1]

where k_B is the Boltzmann constant, p = p(x,v) is the phase space probability distribution, dW stands for the corresponding phase space volume element and ln_qx = (x^1-q-1)/(1-q) is, for x > 0, the q-deformed logarithmic function. For the real parameter q ® 1, Eq.(1) reduces to the standard Boltzmann-Gibbs entropy functional.

In the equilibrium canonical ensemble, under the constraints imposed by the probability normalization

and the normalized q-mean expectation value of the energy [2]

the maximum entropy principle gives the probability distribution [2,3]

where

and

Let us note that, depending from the extremization procedure, in Ref.[2] the above factor b is only proportional to the Lagrange multiplier, because of the probability distribution is self-referential, while, in Ref.[3], b is actually the Lagrange multiplier associated to the energy constraint.

By using the definitions in Eqs. (4) and (6) into the relation = òdW p^q [3], the following identity holds [4]

and the normalized q-mean expectation value for a physical observable A(x,v) can be expressed as

Therefore, the probability distribution and the q-mean value of an observable have the same normalization factor Z_q, as in the extensive statistical mechanics. Such a non-trivial property does not depend on the equilibrium frame but it comes from the normalization condition and holds at any time (if we require that the transport equation conserves the probability normalization or the number of particles). This observation will play a crucial role in the following discussion.

2 From Fokker-Planck equation to nonextensive Schrödinger equation

The linear Fokker-Planck equation in the velocity space can be written as

Of course, the stationary solution of the above equation depends from the explicit expression of the drift J and the diffusion coefficient D. It is easy to see that if J and D depend on the deformation parameter q by means the following relation

where E = mv²/2, the stationary solution is the power-law Tsallis distribution function [1,5]

On the basis of the above observations and following the stochastic quantization procedure, the "nonextensive'' linear Schrödinger equation can be written as

where

_q is the q-deformed Hamiltonian

is the Hamiltonian operator and we have assumed that it contains no explicit time dependence. The right side of Eq.(12) must be seen as a power series of the operator (1-q)i

t/ for q » 1. Equation (12) does not admit factorized solutions but can be integrated to find the wave function at any time t_b from the state at any other time t_a as follows

where t_a < t_b and we have introduced the generalized q-deformed time evolution operator

In the second equivalence we have used the definition of the Tsallis q-deformed exponential

which satisfies the properties: e_q(ln_q x) = x and e_q(x)·e_q(y) = e_q[x+y+(1-q)xy].

The q-deformed time evolution satisfies the fundamental composition law. In fact, if two time translations are performed successively, the corresponding operators _q are related by

for any t_c Î (t_a,t_b). It is important to observe that the operator _q is not a unitary operator in the common sense, however if we observe that the q-deformed exponential, defined in Eq.(16), satisfies the following properties

it appears natural to define the q-adjoint of an operator _q as

On the basis of the above prescription, the q-deformed time evolution operator can be view as a q-unitary operator

and

Moreover, on the basis of the definition (20) we can see that the q-deformed Hamiltonian _q is a q-hermitian operator, in the sense that

The q-Schrödinger equation for _q can be written as

Furthermore, the q-deformed Schrödinger equation (12) conserves the probability. In fact, if we define the probability density of a single particle in a finite volume as

it is easy to show that

Finally, we want to stress that the structure of the Heisemberg's correspondence principle is invariant in the q-deformed quantum mechanics. In fact, if we define, as usual, the Heisenberg operator _qH(t) as

and

_q(t) is an arbitrary observable in the Schrödinger picture, it is easy to show that

3 From non-linear Fokker-Planck equation to non-linear and nonextensive Schrödinger equation

A class of anomalous diffusions are currently described through the non-linear Fokker-Planck equation (NLFPE)

where D and J are the diffusion and drift coefficients, respectively. Tsallis and Bukman [12] have shown that, for linear drift, the time dependence solution of the above equation is a Tsallis-like distribution with q = 1+m-n. The norm of the distribution is conserved for all times only if m = 1, therefore we will limit the discussion to the case n = 2-q.

The main characteristic of anomalous diffusion is that the mean squared displacement is not proportional to time t but rather to some power of t. If the scaling is faster than t, we say that the system is superdiffusive; if it slower than t, we say that the system is subdiffusive. Often these processes are strictly connected to memory effects and described by nonlinear evolution equations.

In this section we want to show that it is possible to introduce a different generalization of the Schrödinger equation on the basis of the above NLFPE.

Following, as before, the stochastic quantization and, for simplicity, restricting the discussion to the free case, we obtain the non-linear Schrödinger equation

The above equation appears interesting because, differently from Eq.(12), admits a factorized solution y = f(t)j(x), where f(t) obeys to the equation

and the time independent wave function j(x) can be obtained from the n-deformed eigenvalue equation

where k = . It is easy to see that the solution of the above equations can be written as

Analogously to the previous section, we can introduce the q-deformed time evolution operator by means of the following expression

where

It is also interesting to note that by using the property (18) of the q-exponential, _q(t_b,t_a) can be written in a more symmetric form as follows

From the above expressions we can see that the time evolution operator

_q(t_b,t_a) depends explicitly from the initial a and the final state b and not only from the time difference t_b-t_a. Such a behavior is motivated by the non-linear Fokker-Planck equation (30) that implies anomalous diffusion and it is related to memory effects [12]. For this reason the composition law is not satisfied, anomalous diffusion and memory effects imply that _q can not be seen as a representation of the abelian group.

Finally, we observe the properties (21) and (22) still hold in the non-linear generalization if we assume, as before, the adjoint operator rule of Eq.(20).

4 Nonextensive relativistic dynamics

In the previous sections we have introduced generalized classical and quantum evolution equation in the framework of Tsallis nonextensive conditions. Because of the growing interest to high energy physics applications of nonextensive thermostatistics [6], in this Section we want to study the formulation of the relativistic nonextensive kinetic equations. Let us start defining the basic macroscopic variables in the language of relativistic kinetic theory. Because we are going to describe a non-uniform system in the phase space, we introduce the particle four-flow as [7,8]

and the energy-momentum flow as

where we have set = c = 1, x º x^m = (t,x), p º p^m = (p⁰,p) and p⁰ = is the relativistic energy. The four-vector N^m = (n,j) contains the probability density n = n(x) (which is normalized to unity) and the probability flow j = j(x). The energy-momentum tensor contains the normalized q-mean expectation value of the energy density, as well as the energy flow, the momentum and the momentum flow per particle.

In order to derive a relativistic Boltzmann equation for a dilute system in nonextensive statistical mechanics, we consider the finite volume elements D³x and D³p in the phase space. These volume elements are large enough to contain a very large number of particles but also small enough compared to the macroscopic dimension of the system. In the Lorentz frame, the particle fraction DN(x,p) in the volume D³x D³p can be written as

where N is the total number of particles of the system. With respect to the observer's frame of reference, the above expression becomes

where d³s_m is a time-like three-surface element of a plane space-like surface s [9]. Requiring that the net flow through surface D³s of element D⁴x vanishes in absence of collisions, we have: p^m¶_mf^q(x,p) = 0. While considering collisions between particles, the Boltzmann equation becomes

C_q(x,p) is the q-deformed collision term that, under the hypothesis that only binary collisions occur in the gas, can be expressed as

In the above equation, we have set W(p,p₁|p¢,p¢₁) as the transition rate between two particle state with initial four-momentum p and p¢ and a final state with four-momenta p₁ and p¢₁; h_q[f,f₁] is the correlation function related to two particles in the same space-time position but with different four-momenta p and p₁, respectively. The factorization of h_q in two single probability distributions (uncorrelated particles at the same spatial point) is the celebrated hypothesis of molecular chaos (Boltzmann's Stosszahlansatz). Thus, the function h_q defines implicitly a generalized nonextensive molecular chaos hypothesis.

By assuming the conservation of the energy-momentum in the collisions (i.e. p^m+p¢^m = +) and requiring that the correlation function h_q is symmetric and always positive (h_q[f,f₁] = h_q[f₁,f], h_q[f,f₁] > 0), it is easy to show that collision term satisfies the following property

if

where a(x) and b(x) are arbitrary functions.

By choosing y = const. and using the Boltzmann equation (43), then Eq.(45) implies

and this is nothing else that the conservation of the probability normalization Z_q. Otherwise, by setting y = b_mp^m, we have from Eq.(45)

which implies the energy and the momentum conservation.

Let us remark that to have conservation of the probability normalization, energy and momentum, it is crucial that not only the collision term C_q be explicitly deformed by means of the function h_q, but also the streaming term p^m¶_mf^q. This matter of fact is a direct consequence of the nonextensive statistical prescription of the normalized q-mean expectation value and is not taken into account in the non-relativistic formulation of Ref.[10].

The relativistic local H-theorem states that the entropy production s_q (x) = ¶_m(x) at any space-time point is never negative.

Assuming the validity of Tsallis entropy (1), it appears natural to introduce the nonextensive four-flow entropy (x) as follows

On the basis of the above equation the entropy production can be written as

where the second identity follows from the Boltzmann equation (43) and the definition of F[y] in Eq.(45). After simple manipulations, Eq.(50) can be rewritten as

By assuming the detailed-balance property, we have that the entropy production is always an increasing function, if q > 0 and if the function h_q satisfies the general condition

in addition to be symmetric and always positive.

Because the nonextensive formalism reduces to the standard Boltzmann kinetic formulation for q ® 1, it appears natural to postulate the q-generalized Boltzmann molecular chaos hypothesis as

where e_q(x) is the Tsallis q-exponential function introduced in Eq.(16). Let us note that a similar expression for the function h_q was previously introduced in Ref.s [10] and a rigorous justification of the validity of the ansatz (53) can be found only by means of a microscopic analysis of the dynamics of correlations in nonextensive statistics.

5 Rapidity distribution in relativistic heavy-ion collisions

In this Section we want to show as nonextensive statistical effects can be very relevant also in the phenomenological interpretation of the high-energy nuclear collisions data. In fact, the quark-gluon plasma close to the critical temperature is a strongly interacting system [6]. For such a system, the color-Coulomb coupling parameter of the QGP can be defined, in analogy to the classical plasma, as

where C = 4/3 or 3 is the Casimir invariant for the quarks or gluons, respectively, a_s = g²/(4p) = 0.2 0.5 and r₀ is the mean particle distance r₀~ n^1/3~ 0.5 fm.

Near the phase transition, the interaction range is much larger than Debye screening length (small number of partons in Debey sphere). In fact, l_D = 1/m < 0.2 fm, if we use the non-perturbative estimate m = 6T. The Coulomb radius for a thermal parton with energy 3T is given by < r > = Cg²/3T = 16 fm. Therefore one obtain < r > /l_D = 530. In other words, memory effects and long-range color interactions give rise to the presence of non-Markovian processes in the kinetic equation affecting the thermalization process toward equilibrium as well as the standard equilibrium distribution.

We are going to study the evolution of the rapidity distribution from a macroscopic point of view by using the non-linear relativistic Fokker-Planck equation (30) in the space of the rapidity distribution y. We show that the observed broad rapidity shape could be a signal of non-equilibrium properties of the system. Let us note that a similar approach, within a linear Fokker-Planck equation, has been previously studied in Ref.[11] by using a linear drift in the space of the rapidity and a free parameter diffusion coefficient. With this choice, the author found a strongly violation of the fluctuation-dissipation theorem. We will see that generalizing the Brownian motion to the relativistic kinetic variables, the standard Einstein relation is satisfied and Tsallis non-extensive statistics emerges in a natural way from the non-linearity of the Fokker-Planck equation.

Basic assumption of our analysis is that the rapidity distribution y is not appreciably influenced by transverse dynamics which is considered in thermal equilibrium. Such hypothesis is well confirmed by the experimental data and adopted in many theoretical works [13,14,15]. A crucial rôle in the solution of the above NLFPE plays the choice of the diffusion and the drift coefficients. Such a choice influences the time evolution of the system and its equilibrium distribution.

Since the temperature at freeze-out exceeds 100 MeV, quantum statistical effects are negligible and the Boltzmann approximation is usually adopted, therefore the single particle equilibrium distribution is written as

where y is the rapidity, m_^ = is the transverse mass, T is the temperature. Imposing the validity of the Einstein relation for Brownian particle, we can generalize to the relativistic case the standard expressions of diffusion and drift coefficients as follows [16]

where p_|| is the longitudinal momentum and t is a common constant. It is easy to see that the above coefficients give us the Boltzmann stationary distribution (5) in the linear Fokker-Planck equation (q = n = 1). This result cannot be obtained if one assumes a linear drift coefficient as in Ref.[11]. The stationary solution of the NLFPE (30) with n = 2-q is a Tsallis-like distribution with the relativistic energy E = m_^ cosh(y). In the light of the above results, we generalize the results of Tsallis-Bukman by searching as a solution of the NLFPE (30) the following time dependent Tsallis-like distribution

The unknown functions b(t) = 1/T(t) and y_m(t) have been derived by means of numerical integration of Eq.(30) with initial d-function condition depending on the value of the experimental projectile rapidities. The rapidity distribution at fixed time is then obtained by numerical integration over the transverse mass m_^ (or transverse momentum) as follows

where c is the normalization constant fixed by the total number of the particles. Rapidity spectra calculated from (58) will ultimately depend on two parameters: the "interaction'' time t = gt and the non-extensive parameter q. Therefore, no more free parameters are used in this analysis respect to previous theoretical studies [14,15].

The obtained spectra are normalized to 164 protons and the beam rapidity is fixed to y_cm = 2.9 (in the c.m. frame) [13]. In Fig. 1, we show the shape of the rapiditity distribution at different integration time t_int of the linear Fokker-Planck equation (q = 1). In Fig. 2, we report the time evolution of the rapidity distribution corresponding to the integration of the NLFPE with q = 1.25 (the value of the q-parameter used to fit the data).

In Fig. 3, we show the calculated rapidity spectra compared with the experimental data. The full line corresponds to the NLFPE solution (58) at t_int = 0.82 and q = 1.25; the dashed line corresponds to the solution of the linear case (q = 1) at t_int = 1.2. Imposing the validity of the fluctuation-dissipation theorem, it is not possible to reproduce the experimental rapidity shape at any time. Only in the non-linear case (q¹ 1) exists a (finite) time for which the obtained rapidity spectra well reproduces the broad experimental shape. A value of q ¹ 1 implies anomalous superdiffussion in the rapidity space, i.e., [y(t)-y_M(t)]² scale like t^a with a> 1 [12].

A complete description of the applicability of nonextensive statistical effects to high-energy heavy ion collisions lies out the scope of this paper. However, we want to outline that an analysis of the transverse pion momentum spectra and the net proton rapitity distribution measured at RHIC is under investigation.

6 Conclusion

There is an increasing evidence that the generalized nonextensive thermodynamics can be considered as the more appropriate basis of a theoretical framework to deal with several physical phenomena where long-range interactions, long-range memory effects and/or fractal space-time constraints are present. A considerable variety of physical applications involve microscopic quantum and/or relativistic effects. The main effort of this paper is to study an appropriate generalization of the quantum dynamics and relativistic kinetic equation in the framework of the Tsallis nonextensive thermostatistics. We have introduced two kinds of generalized Schrödinger equations which satisfy the basic assumptions of the quantum mechanics under appropriate operator properties which depend on the deformation parameter q. Furthermore, we have studied a nonextensive generalization of the relativistic Boltzmann equation and of the nonlinear Fokker-Planck equation. The above evolution equations have been applied to the study of the nonequilibrium rapidity distribution in relativistic heavy-ion collisions obtaining a very good agreement with the experimental results.

References

[1] C. Tsallis, J. Stat. Phys. 52, 479 (1988). See also http://tsallis.cat.cbpf.br/biblio.htm for a regularly updated bibliography on the subject.

[2] C. Tsallis, R.S. Mendes, and A.R. Plastino, Physica A 261, 534 (1998).

[3] S. Martínez, F. Nicolás, F. Pennini, and A. Plastino, Physica A 286, 489 (2000).

[4] S. Martínez, F. Pennini, and A. Plastino, Phys. Lett. A 278, 47 (2000).

[5] L. Borland, Phys. Lett. A 245, 67 (1998).

[6] W.M. Alberico, A. Lavagno, and P. Quarati, Eur. Phys. J. C 12, 499 (2000); Nucl. Phys. A 680, 94c (2001).

[7] A. Lavagno, Phys. Lett. A 301, 13 (2002).

[8] A. Drago, A. Lavagno, and P. Quarati, Physica A 344, 472 (2004).

[9] S.R. Groot, W.A. van Leeuwen, and Ch. G. van Weert, Relativistic Kinetic Theory (North-Holland, Amsterdam, 1980).

[10] J.A.S. Lima, R. Silva, and A.R. Plastino, Phys. Rev. Lett. 86, 2938 (2001).

[11] G. Wolschin, Eur. Phys. J. A 5, 85 (1999); Phys. Rev. C 69, 024906 (2004).

[12] C. Tsallis, D.J. Bukman, Phys. Rev. E 54, R2197 (1996).

[13] H. Appelshäuser et al. (NA49 Collaboration), Phys. Rev. Lett. 82, 2471 (1999).

[14] I.G. Bearden et al. (NA44 Collaboration), Phys. Rev. Lett. 78, 2080 (1997).

[15] P. Braun-Munzinger et al., Phys. Lett. B 344, 43 (1995); B 356, 1 (1996).

[16] A. Lavagno, Physica A 305, 238 (2002).

Received on 20 December, 2004

Publication Dates

• Publication in this collection
  06 Sept 2005
• Date of issue
  June 2005

History

• Received
  20 Dec 2004