1: \documentstyle[aps,manuscript]{revtex}
2:
3: \begin{document}
4:
5: \draft
6:
7: \title{Fluctuations and the existence of potential in dissipative
8: semiclassical systems}
9:
10: \author{Bidhan Chandra Bag and
11: Deb Shankar Ray{\footnote{pcdsr@mahendra.iacs.res.in}}
12: }
13:
14: \address{Indian Association for the Cultivation of Science,
15: Jadavpur, Calcutta 700 032, INDIA.}
16:
17: \maketitle
18:
19: \begin{abstract}
20: We examine the weak noise limit of an overdamped dissipative
21: system within a semiclassical description
22: and show how quantization influences the growth
23: and decay of fluctuations of the thermally equilibrated systems.
24: We trace its origin
25: in a semiclassical counterpart of the generalized potential for the dissipative
26: system.
27: \end{abstract}
28:
29: \vspace{1cm}
30: \pacs{{\bf Keywords} : Dissipative quantum systems, large fluctuations,
31: generalized potential.\\
32: \\
33: PACS number(s): 05.40.-a, 05.70.Ln}
34:
35: \newpage
36:
37:
38: \section{Introduction}
39:
40: One of the important issues in nonequilibrium phenomena in the macroscopic
41: nonlinear systems is to understand the
42: interplay of nonlinearity of the system and the fluctuations of its environment.
43: The problem is fairly general in the
44: context of chemical reactions \cite{dykman2},
45: breakdown of electronic devices \cite{kautz}, phase transitions \cite{nature}
46: etc. The essential
47: description of the physical situation rests on the Fokker-Planck equations for the
48: probability distribution functions of the relevant variables of the dynamics.
49: In the weak noise limit the fluctuations have been described
50: \cite{nature,bray,graham,tel} by appropriate auxiliary
51: Hamiltonian or path integral methods.
52: The theoretical results have been
53: corroborated by remarkable experiments on fluctuations using analogue
54: electronic circuits \cite{nature} , which allow the phase space
55: trajectories of fluctuations
56: to be observed directly in a precise manner. These studies have enriched our
57: understanding in several theoretical issues, e. g. , the symmetry between
58: the growth and the decay of classical fluctuations in equilibrium
59: and its breakdown under
60: nonequilibrium conditions \cite{nature} , the existence of a nonequilibrium
61: potential of a dissipative
62: system \cite{graham,tel} etc.
63:
64: It is the purpose of this paper to extend the theory to the semiclassical
65: context. The quantization of the system itself adds a new dimension to the
66: interplay of nonlinearity and stochasticity in a dissipative
67: system. To make a fair comparison with classical theory
68: we adopt the Wigner's phase space distribution function of c-number variables.
69: The weak noise limit can then be appropriately employed to develop an auxiliary
70: Hamiltonian formulation at the semiclassical level
71: in terms of these phase space variables. This allows us to realize
72: the existence of an optimal force of purely quantum origin derivable
73: from the fluctuating field and relate it to the momentum of the auxiliary
74: Hamiltonian. The quantum correction also makes its presence felt in the growth
75: and decay of fluctuations of thermally equilibrated semiclassical
76: systems keeping the symmetry preserved.
77:
78: The outline of the paper is as follows: In Sec. II we introduce the general
79: aspects of dynamics of dissipative quantum system
80: in terms of the Wigner phase space function.
81: In Sec. III we consider the weak noise
82: and semiclassical limit under overdamped condition and take resort to the well-known auxiliary
83: Hamiltonian description. The quantum part of the optimal force is then
84: explicitly derived. The symmetry between the
85: growth and decay of fluctuations in a thermalized quantum system is
86: discussed in Sec. IV. The existence of a semiclassical contribution to the potential
87: in a dissipative system is then shown in Sec. V. The paper is concluded in
88: Sec. VI.
89:
90:
91: \section{Quantum dynamics of a dissipative system}
92:
93: We consider a dynamical system characterized by a potential $V(x)$ coupled to an
94: environment. Evolution of such an open quantum system
95: has been studied over the last several decades under a variety of reasonable
96: assumptions \cite{louisell,legg,gg1}. Specifically interesting is the semiclassical limit of an Ohmic
97: environment. The dissipative time evolution of the Wigner distribution
98: function $W(x, p, t)$ for the system with unit mass ($m = 1$)
99: under the potential $V(x)$ can be described by \cite{zurek,bag1,pat}
100:
101: \begin{equation}
102: \frac{\partial W}{\partial t} =
103: -p \frac{\partial W}{\partial x}
104: + \frac{\partial V}{\partial x} \frac{\partial W}{\partial p}
105: + \epsilon \sum_{n\geq 1}
106: \frac{ \hbar^{2n}(-1)^n}{ 2^{2n}(2n+1)! }
107: \frac{\partial^{2n+ 1}V}{\partial x^{2n+ 1} }
108: \frac{ \partial^{2n+1} W }{\partial p^{2n+1} } +
109: \gamma \frac{\partial p W}{\partial p}
110: + D \frac{\partial^2 W}{\partial p^2} \; \;,
111: \end{equation}
112:
113: \noindent
114: where $\gamma$ and $D$ are the dissipation constant and the diffusion
115: coefficient, respectively. $x$ and $p$ are c-number co-ordinate and
116: momentum variables. The drift term is a direct consequence \cite{legg} of the
117: existence of $\gamma$-dependent term in the imaginary part of the exponent
118: in the expression for the propagator for the density operator in the
119: Feynman-Veron theory and has been shown to be responsible for appearance of a
120: damping force in the classical equation of motion for the Brownian particle
121: to ensure quantum-classical correspondence. $\gamma$ and D are related by the
122: fluctuation-dissipation relation, $D = \frac{\gamma}{2} \hbar \omega
123: \coth{\frac{\hbar \omega}{2 k_b T}}$ (in the semiclassical
124: limit $D = \gamma k T$ ). $\omega$ is the renormalised linear frequency
125: of the nonlinear system.
126: The quantum correction to classical Liouville motion is contained in the
127: $\hbar$-containing terms in the sum. $\epsilon$ is a parameter (whose value is 1) which is
128: kept in the equation for bookkeeping the Wigner correction term in our further
129: analysis. We put $\epsilon=1$ at the end of calculation.
130:
131: Eq.(1) had been used earlier in several occasions. For example, Zurek and Paz
132: \cite{zurek} and others \cite{pat} have studied some interesting aspects of quantum-classical
133: correspondence in relation to decoherence and chaos. Based on this equation
134: and its variant chaotic dissipative systems has been studied.
135: \cite{bag1,mil,dit,bag2}. The equation also yields the simplest leading order
136: quantum correction term to classical Kramers' rate \cite{jr}.
137: The primary reason for choosing Eq.(3) as our starting point is
138: that it reaches the correct classical limit when $\hbar \rightarrow 0$ so that
139: $D$ becomes a thermal diffusion coefficient ($\gamma k T$) in the high
140: temperature limit and the Wigner function reduces to the corresponding classical
141: phase space distribution function
142: and we recover the Kramers' equation which describes classical Brownian
143: motion of a particle in phase space.
144:
145:
146: \section{Weak noise and semiclassical limit of quantum dissipative dynamics:}
147:
148: The weak noise limit of a dissipative system within
149: a semiclassical description can
150: be conveniently described by a ``WKB-like'' ansatz (we refer to ``WKB-like'' since we are
151: considering more than one dimension. Traditionally WKB refer to one dimension
152: only)
153: of the Eq.(1) for the Wigner function of the form
154:
155: \begin{equation}
156: W(x, p, t) = Z(x, t) \exp(-\frac{x p}{\hbar}) \exp(-\frac{s}{D_1}) \; \;.
157: \end{equation}
158:
159: where $D_1 = \frac{D} {\omega^2}$. The weak noise limit is defined \cite{nature}
160: as $D_1 \rightarrow 0$ and semiclassical limit refers to ($\hbar \rightarrow 0$).
161: $Z(x,t)$ is a prefactor and $s(x, p, t)$ is a classical action which is a
162: function of c-number variables $x$ and $p$ , satisfying the following
163: Hamilton-Jacobi equation,
164:
165: \begin{eqnarray}
166: & & \frac{\partial s}{\partial t} + p \frac{\partial s}{\partial x} -
167: V' \frac{\partial s}{\partial p} - \gamma p \frac{\partial s}
168: {\partial p}
169: + \omega^2 (\frac
170: {\partial s}{\partial p})^2 \nonumber\\
171: & & + \epsilon
172: \sum_{n\geq 1} \frac{x^{2n} (-1)^{3n+1}}{2^{2n}(2n)!} \frac{\partial^{2n+1} V}
173: {\partial x^{2n+1}} \frac{\partial s}{\partial p} = 0 \; \;.
174: \end{eqnarray}
175:
176: The derivation of Eq.(3) is based on the following consideration. Since in the
177: weak noise limit $D_1$ is the relevant small parameter one obtains with
178: ansatz (2) in leading order a term proportional to $\hbar^{2n} \left( \frac{1}
179: {D_1} \frac{\partial s}{\partial p} \right)^{2n+1}$ which is not balanced by
180: any other term of the same order ${D_1}^{-(2n+1)}$. This is because the
181: highest derivative in Eq.(1) does not have a factor scaling with the
182: corresponding power of $D_1$. The successive terms next to the leading order are
183: also $\hbar$-containing terms. All of these terms vanishes in the semiclassical
184: limit ($\hbar \rightarrow 0$). Therefore the leading order term that remains gives
185: rise to Eq.(3). It is thus obvious that an ansatz (2) with $\hbar$ finite
186: is not feasible. The semiclassical limit $\hbar \rightarrow 0$ is a necessary
187: requirement for the validity of ansatz (2).
188: The above equation can be solved by integrating the Hamiltonian equations of
189: motion,
190:
191: \begin{eqnarray}
192: \dot{x} & = & p \nonumber\\
193: \dot{X} & = & P - \gamma X \nonumber\\
194: \dot{p} & = & V' + \gamma p - 2 \omega^2 X - \epsilon
195: \sum_{n\geq 1} \frac{ (-1)^{3n+1} x^{2n} }{2^{2n}(2n)! }
196: \frac{\partial^{2n+1} V}{\partial x^{2n+1}} \nonumber\\
197: \dot{P} & = & \left[V'' - \epsilon
198: \sum_{n\geq 1} \frac{(-1)^{3n+1}}{2^{2n}(2n)!}
199: \frac{\partial}{\partial x} \left( x^{2n}
200: \frac{\partial^{2n+1} V}{\partial x^{2n+1} } \right) \right] X
201: \end{eqnarray}
202:
203: which are derived from the following effective Hamiltonian $H_{eff}$
204:
205: \begin{eqnarray}
206: H_{eff} & = & p P - V' X
207: - \gamma X p
208: + \omega^2 X^2 \nonumber\\
209: & & + \epsilon
210: \sum_{n\geq 1} \frac{x^{2n} (-1)^{3n+1}}{2^{2n}(2n)!}
211: \frac{\partial^{2n+1} V}{\partial x^{2n+1}} X \; \; .
212: \end{eqnarray}
213:
214: \noindent
215: Here we have put $\frac{\partial s}{\partial x} = P$ and
216: $\frac{\partial s}{\partial p} = X$.
217: The introduction of additional degree-of-freedom by incorporating the
218: auxiliary momentum (P) and co-ordinate (X) makes the system an effectively
219: two-degree-of-freedom system. The origin of these two variables is
220: the thermal fluctuations of the environment \cite{nature}.
221: The auxiliary Hamiltonian $H_{eff}$ is not to be
222: confused with the microscopic Hamiltonian comprising the system, the bath and
223: their coupling. Thus the phase space
224: trajectories concern fluctuations of the c-number variables.
225:
226:
227: Under overdamped condition ($\ddot{x} \ll \gamma \dot{x} ; \ddot{X}
228: \ll \gamma \dot{X}$) Eqs. (4) can be easily reduced to the following form ;
229:
230: \begin{eqnarray}
231: \dot{x} & = & K(x) + \frac{2 \omega^2 X}{\gamma} \nonumber\\
232: \dot{X} & = & - \frac{\partial K(x)}{\partial x} X
233: \end{eqnarray}
234:
235: \noindent
236: where
237: \begin{equation}
238: K(x) = \frac{1}{ \gamma} \left[-V' + \epsilon
239: \sum_{n\ge 1} \frac{(-1)^{3n+1} x^{2n}}{2^{2n}(2n)!}
240: \frac{\partial^{2n+1} V}{\partial x^{2n+1} } \right]
241: \end{equation}
242:
243: It is easy to recognize the quantity $\frac{2 \omega^2 X}{\gamma}$ as a
244: momentum and redefine it as $p_r$. Therefore Eqs(6) may
245: be rewritten as
246:
247: \begin{eqnarray}
248: \dot{x} & = & K(x) + p_r \nonumber\\
249: \dot{p_r} & = & \frac{\partial K(x)}{\partial x} p_r \; \; .
250: \end{eqnarray}
251:
252: So the overdamped motion is described by the following effective Hamiltonian,
253:
254: \begin{equation}
255: H_{od} = \frac{p^2_r}{2} + K(x) p_r \; \; .
256: \end{equation}
257:
258: The above auxiliary Hamiltonian description (8, 9) is isomorphic in form
259: to that of Luchinsky and McClintock \cite{nature}, who had considered
260: an overdamped classical Brownian motion in a force field $K(x)$, driven by a weak
261: white noise $\zeta(t)$ whose intensity $D_1 \ll 1$ as
262:
263: \begin{equation}
264: \dot{x} = K(x) + \zeta(t), \; \; \; \langle \zeta \rangle = 0 \; \; ,
265: \; \langle \zeta(t) \zeta(0) \rangle = D_1 \delta(t) \; \; \; .
266: \end{equation}
267:
268: Equivalently the corresponding Fokker-Planck equation for the probability
269: density $P(x, t)$ is
270:
271: \begin{equation}
272: \frac{\partial P(x, t)}{\partial t} = - \frac{\partial K(x)}{\partial x} P(x, t)
273: + \frac{D_1}{2} \frac{\partial^2 P(x, t)}{\partial x^2} \; \;.
274: \end{equation}
275:
276:
277: The large fluctuations of scale $\gg \sqrt {D_1}$ can therefore be treated
278: in the weak noise limit $D_1 \rightarrow 0$ \cite{nature,graham,tel}
279: by ``WKB like'' approximation of the Fokker-Planck equation (11) in the
280: form
281:
282: \begin{equation}
283: P(x, t) = y(x, t) \exp[ \frac{- \phi(x,t)}{D_1}] \; \;.
284: \end{equation}
285:
286: Here y(x, t) is the prefactor and $\phi(x, t)$ is a ``classical'' action
287: describing a Hamiltonian-Jacobi equation which can be solved by solving
288: Hamilton' s equation (8) with $p_r = \frac{\partial \phi}{\partial x}$
289: as the momentum for auxiliary system.
290:
291: The important distinguishing feature of the above description in which
292: the system is treated semiclassically is the structure of $K(x)$ which
293: is given by equation(7) and comprises of two terms ;
294:
295: \begin{equation}
296: K(x) = K_{cl} + K_{semi}(x)
297: \end{equation}
298:
299: \noindent
300: where
301:
302: \begin{equation}
303: K_{cl} = - \frac{V'(x)}{\gamma}
304: \end{equation}
305:
306: is derivable from purely classical potential $V(x)$ and $K_{semi}(x)$
307: does not
308: explicitly involve $\hbar$,
309:
310: \begin{equation}
311: K_{semi}(x) = \frac{\epsilon}{\gamma}
312: \sum_{n\geq 1} \frac{x^{2n} (-1)^{3n+1}}{2^{2n}(2n)!}
313: \frac{\partial^{2n+1} V}{\partial x^{2n+1}}
314: \end{equation}
315:
316: originates from the nonlinearity of the potential $V(x)$
317: and quantum nature of the system. The quantum
318: contribution to $K(x)$ is therefore
319: likely to influence both the fluctuational
320: and the relaxational paths of the dynamics and is also responsible for the
321: existence of a potential.
322: Our objective is now to explore these aspects in the following two
323: sections.
324:
325: \section{Large fluctuations in equilibrated semiclassical systems }
326:
327: In the thermally equilibrated systems a typical large fluctuation of the
328: variable $x$ implies a temporary departure from its stable state, $x_s$ to
329: some remote state $x_f$. This is followed by a return to $x_s$ as a result
330: of relaxation in the absence of fluctuations $p_r$. A nonzero value of $p_r$
331: which results from fluctuations due
332: to surrounding drives the system away from $x_s$ along a set of trajectories
333: which form the unstable invariant manifold and define the so called
334: fluctuational paths. On the other hand the system relaxes along the
335: relaxational return path to $x_s$ under the condition $p_r = 0$, which form
336: stable invariant manifold. The latter condition implies $\dot{x} = K(x)$.
337: In each case the trajectory represents the optimal paths along which the system
338: is expected to move with overwhelming probability. Luchinsky and McClintock \cite{nature}
339: have studied these paths in analog electronic circuits and demonstrated the
340: growth and the decay of {\it classical} fluctuations in equilibrium. We extend
341: this analysis to semiclassical domain using the same model potential,
342:
343: \begin{equation}
344: V(x) = \frac{1}{4} x^4 - \frac{1}{2} x^2 \; \;.
345: \end{equation}
346:
347: The quantum contribution to the growth and the decay of fluctuations can be understood
348: by recognizing the $K_{semi}(x)$ term in the dynamics (8). In Fig.1 we
349: compare both the deterministic fluctuational and relaxational (optimal)
350: paths for quantum and classical thermally equilibrated systems. It is
351: important to note that the maximum possible amplitude of large fluctuations
352: is almost double for the quantum system compared to that for the
353: corresponding the classical system. This is due to the addition of the
354: nonlinear force term of quantum origin, $K_{semi}(x)$ in the
355: c-number equation (8), which is shown to be derivable from the
356: fluctuating field, and is related to the momentum of the auxiliary Hamiltonian.
357:
358: Before leaving this section we make a brief remark on the thermally
359: nonequilibrated systems. Since the detailed balance is not operative here,
360: the optimal path to a given state not just the time-reversed dynamical path
361: along which the system moves from this state to the stable state in absence
362: of fluctuations $p_r$. Thus for the driven system, for example, the pattern of
363: optimal path is generically different from that for the thermally equilibrated
364: systems. It may display singularities whose topological manifestations as
365: caustics, switching line, cusps etc have been thoroughly studied
366: for classical systems. We believe that $K_{semi}(x, t)$ where t
367: signifies the driving by a periodic force in $V(x, t)$ is likely to play an
368: important role in their quantum counterparts.
369:
370: \section{Existence of a potential for dissipative semiclassical system}
371:
372: In a significant analysis Graham and coworkers \cite{graham,tel} had examined the condition for
373: existence of potential for classical dissipative systems. We now extend this analysis
374: to the present semiclassical context.
375:
376: The general criterion for a dissipative dynamical system described by
377: autonomous equations of the form
378:
379: \begin{equation}
380: \dot{x}^{\nu} = K^\nu (x)
381: \end{equation}
382:
383:
384: to have a potential $\phi(x)$ with respect to $Q^{\nu \mu}$ (positive,
385: semidefinite symmetric matrix, which are considered to be the matrix of
386: transport coefficients) if there exists a single-valued continuously
387: differentiable and globally defined function $\phi(x)$, bounded from
388: below which is stationary in the limit sets of Eq.(17) and which satisfies
389:
390: \begin{equation}
391: K^{\nu}(x) = -\frac{1}{2} Q^{\nu \mu}
392: \frac{\partial \phi(x)}{\partial x^\mu}
393: + r^{\nu} (x)
394: \end{equation}
395:
396: \noindent
397: with
398:
399: \begin{equation}
400: r^{\nu} \frac{\partial \phi (x)}{\partial x^{\nu}} = 0 \; \;.
401: \end{equation}
402:
403: Here for simplicity $Q^{\nu \mu}$ is assumed to be independent of $x$. The
404: first and the second terms of Eq.(18) correspond to irreversible and reversible
405: part, respectively.
406:
407:
408: The stochastic process $x (D_1, t)$ which involve Eq.(17) and a symmetric
409: non-negative matrix $Q^{\nu \mu}$ is governed by the Fokker-Planck equation
410: for probability distribution function $P(x, t)$
411:
412: \begin{equation}
413: \frac{\partial P(x,t)}{\partial t} = - \frac{\partial}{\partial x^{\nu}}
414: K^{\nu}(x) P + \frac{D_1}{2}
415: \frac{\partial^2}{\partial x^{\nu} \partial x^{\mu}}
416: Q^{\nu \mu} p \; \;.
417: \end{equation}
418:
419: For $D_1 = 0$ the above description reduces to (17). For $D_1 \ne 0$ the
420: steady state distribution defines the function $\phi(x, t)$ by
421:
422: \begin{equation}
423: P(x, D_1, t \rightarrow \infty) = N(D_1) \exp [ -\frac{\phi(x, D_1)}{D_1}]
424: \end{equation}
425:
426: N is the normalization constant.
427: If $\phi(x) = \lim_{D_1 \rightarrow 0}
428: \phi(x, D_1)$ is a single-valued, continuously differentiable
429: function bounded from below it satisfies
430:
431: \begin{equation}
432: K^{\nu}(x) \frac{\partial \phi(x)}{\partial x^{\nu}}
433: + \frac{1}{2} Q^{\nu \mu}
434: \frac{\partial \phi(x)}{\partial x^{\nu}}
435: \frac{\partial \phi(x)}{\partial x^\mu} = 0 \; \;
436: \end{equation}
437:
438: Eq. (22) is equivalent to Eqs. (18, 19). Interpreting Eq.(22) as usual as a
439: Hamilton-Jacobi equation by defining $\phi(x)$ as action and $\frac{\partial
440: \phi}{\partial x^{\nu}} = P_{\nu}$
441: as the momentum conjugate to $x_{\nu}$, one
442: can construct the auxiliary Hamiltonian
443:
444: \begin{equation}
445: H(x,p) = \frac{1}{2} Q^{\nu \mu} p_{\nu} p_{\mu} + K^{\nu}(x) p_{\nu} \; \;.
446: \end{equation}
447:
448: Graham and co-workers \cite{graham,tel} have argued that a potential can exist with equation (17) if the
449: above Hamiltonian is integrable for $H = 0$, because the condition implies
450: that there exist a smooth separatrix which connect smoothly the stable and unstable
451: manifolds emanating from the hyperbolic fixed points of the dynamical system.
452:
453: We now turn back to our dissipative semiclassical system described by Eqs(8)
454: and (9) where $K(x)$ is defined by Eq.(7). Recognizing Eq.(9) as Eq.(23)
455: for a one-degree-of-freedom system we identify
456:
457: \begin{eqnarray}
458: Q & = & 1 \nonumber\\
459: K(x) & = & K_{cl} + K_{semi} \; \;.
460: \end{eqnarray}
461:
462: The potential function $\phi(x)$ can therefore be calculated from Eq.(9)
463: with $H = 0$ as (since $p_r = \frac{\partial \phi(x)}{\partial x}$)
464:
465: \begin{equation}
466: \phi(x) = -2 \int K(x) dx \; \;.
467: \end{equation}
468:
469: The above expression can be made more explicit if we make use of Eq.(7) in (25).
470: We obtain
471:
472: \begin{equation}
473: \phi(x) = \phi_{cl}(x) + \phi_{semi}(x)
474: \end{equation}
475:
476: \noindent
477: where
478:
479: \begin{equation}
480: \phi_{cl} = \frac{2}{\gamma} \int V'(x) dx = \frac{2}{\gamma} V(x)
481: \end{equation}
482:
483: \noindent
484: and
485:
486: \begin{equation}
487: \phi_{semi}(x) = - \frac{2}{\gamma}
488: \sum_{n\geq 1} \int \frac{x^{2n} (-1)^{3n+1}}{2^{2n}(2n)!}
489: \frac{\partial^{2n+1} V}{\partial x^{2n+1}} dx
490: \end{equation}
491:
492: The existence of a potential for a {\it dissipative, semiclassical} dynamical system
493: is thus ascertained. The method essentially relies on a dynamical definition
494: $p_r$ as a derivative of the potential $\phi(x)$ in a system described by an
495: overdamped quantum Markov process in the weak noise and semiclassical limit.
496: As elaborated
497: earlier in Sec.III $p_r$ has a statistical origin which drives the
498: system away from its stable state $x_s$ to a preassigned remote state $x_f$
499: from which the system relaxes in absence of $p_r$. It is thus important
500: to realize that both the dynamical and statistical notions are kept
501: intact in the quantum treatment.
502:
503: \section{Conclusions}
504:
505: Keeping in view of the quantum nature of the system
506: in terms of the Wigner's phase space function we examine the
507: semiclassical dynamics of a dissipative system in an Ohmic environment.
508: The weak noise limit
509: of the stochastic process then allows us to capture the essential
510: features of the dynamics within the framework of an auxiliary
511: Hamiltonian description at the
512: semiclassical level. Our results are summarized as follows :
513:
514: \vspace{0.7cm}
515: \noindent
516: (i) The Wigner's quantum correction to classical Liouville equation gives rise
517: to an optimal force in addition
518: to usual the classical force term. This quantum optimal force is essentially
519: a result of an interplay of nonlinearity of the system and the thermal
520: fluctuations of its environment and is derivable
521: in terms of an auxiliary Hamiltonian description.
522:
523: \vspace{0.7cm}
524: \noindent
525: (ii) This term is also responsible for modification of growth and decay
526: of large fluctuations from equilibrium for the appropriately
527: thermalized quantum system (compared to its classical counterpart).
528: The symmetry of the fluctuational and the relaxational paths, signifying
529: the detailed balance, however, as expected is kept preserved.
530:
531: \vspace{0.7cm}
532: \noindent
533: (iii) The quantum correction term implies the existence
534: of a potential for the dissipative semiclassical system.
535:
536: \vspace{0.5cm}
537: Since the fluctuational and the relaxational paths have been experimentally
538: demonstrated as a part of physical reality by analogue experiments \cite{nature}
539: in the realm of large fluctuations, we believe that the essential
540: modification of the integrable, nonintegrable and the singular
541: topological features of the dynamics due to
542: semiclassical correction might be
543: relevant in several contexts. We hope to address some of these issues in a
544: future communication.
545:
546: \acknowledgments
547: B. C. Bag is indebted to the Council of Scientific and
548: Industrial Research for a fellowship.
549:
550:
551: \begin{thebibliography}{99}
552:
553: \bibitem{dykman2} M. I. Dykman, E. Mori, J. Ross and M. P. Hunt, J. Chem. Phys. {\bf 100},
554: 5737 (1994).
555:
556: \bibitem{kautz}
557: R. L. Kautz, Rep. Prog. Phys. {\bf 59}, 935 (1996).
558:
559: \bibitem{nature}
560: D. G. Luchinsky and P. V. E. McClintock, Nature {\bf 389}, 403 (1997).
561:
562: \bibitem{bray}
563: A. J. Bray and A. J. McKane, Phys. Rev. Lett. {\bf 62}, 493 (1989).
564:
565: \bibitem{graham}
566: R. Graham, in {\it Noise in Nonlinear Dynamical Systems} Vol. 1 (eds F. Moss
567: and P. V. E. McClintock) 225-278 (Cambridge Univ. Press, 1989).
568:
569: \bibitem{tel} R. Graham and T. T\'el, Phys. Rev. Lett {\bf 52}, 9 (1984).
570:
571: \bibitem{louisell}
572: W. H. Louisell, {\it Quantum Statistical Properties of Radiation} (Wiley, New
573: York, 1973); G. W. Ford, J. T. Lewis and R. J. O'Connell, Ann. Phys.(N.Y.)
574: {\bf 185}, 270 (1988).
575:
576: \bibitem{legg}
577: A. J. Leggett and A. O. Caldeira, Physica {\bf A121}, 587 (1983).
578:
579: \bibitem{gg1}
580: See, for example, the review by
581: G. Gangopadhay and D. S. Ray, in {\it Advances in Multiphoton Processes and
582: Spectroscopy} Vol 8, edited by S. H. Lin, A. A. Villayes and F. Fujimura,
583: (World Scientific, Singapore, 1993).
584:
585: \bibitem{zurek}
586: W. H. Zurek and J. P. Paz, Phys. Rev. Lett. {\bf 72}, 2508 (1994) ;
587: S. Habib, K. Shizume and W. H. Zurek, Phys. Rev. Lett. {\bf 80}, 4361 (1998).
588:
589: \bibitem{bag1}
590: B. C. Bag, S. Chaudhuri, J. Ray Chaudhuri and D. S. Ray, Physica D {\bf 125},
591: 47 (1999).
592:
593: \bibitem{pat} A. K. Pattanayak, Phys. Rev. Lett. {\bf 83}, 4526 (1999).
594:
595: \bibitem{mil} M. E. Goggin, B. Sundaram and P. W. Milonni, Phys. Rev. A {\bf 41},
596: 5705 (1990).
597:
598: \bibitem{dit} T. Dittrich and R. Graham, Ann. Phys. {\bf 200}, 363 (1990)
599:
600: \bibitem{bag2} B. C. Bag and D. S. Ray, J. Stat. Phys. {\bf 96}, 271 (1999).
601:
602: \bibitem{jr} J. Ray Chaudhuri, B. C. Bag and D. S. Ray, J. Chem. Phys.
603: {\bf 111}, 10852 (1999).
604:
605: \end{thebibliography}
606:
607:
608: \begin{figure}
609: \caption{ A plot of system co-ordinate $x$ vs time $t$ signifying the
610: fluctuational and relaxational paths according to Eq.(8) for the
611: model system with $V(x) = \frac{1}{4} x^4 - \frac{1}{2} x^2$ for $(a)$
612: semiclassical case ($\epsilon = 1$) and $(b)$ classical case ($\epsilon = 0$),
613: (units arbitrary).}
614: \end{figure}
615:
616:
617: \end{document}
618:
619: