1: %\documentclass[pre,aps,preprint,eqsecnum]{revtex4}
2: %\documentclass[aps,prl,eqsecnum,twocolumn]{revtex4}
3: \documentclass[pre,aps,eqsecnum,twocolumn]{revtex4}
4: \usepackage[centertags]{amsmath}
5: \usepackage{amssymb}
6: \usepackage{amsthm}
7:
8: \begin{document}
9: \title{Quantum escape kinetics over a fluctuating barrier}
10:
11: \author{Pulak Kumar Ghosh$^2$, Debashis Barik$^2$, Bidhan Chandra Bag$^1$ and
12: Deb~Shankar~Ray$^2$ {\footnote {e-mail address:
13: pcdsr@mahendra.iacs.res.in}}}
14:
15:
16:
17: \affiliation{Indian Association for the Cultivation of Science,
18: Jadavpur, Kolkata 700 032, India}
19:
20:
21:
22:
23:
24: \begin{abstract}
25: The escape rate of a particle over a fluctuating barrier in a double
26: well potential exhibits resonance at an optimum value of correlation
27: time of fluctuation. This has been shown to be important in several
28: variants of kinetic model of chemical reactions . We extend the
29: analysis of this phenomenon of resonant activation to quantum domain
30: to show how quantization significantly enhances resonant activation
31: at low temperature due to tunneling.
32: \end{abstract}
33:
34: \maketitle
35:
36:
37: \section{Introduction}
38: The enhancement of small periodic signal by noise in a nonlinear
39: system has been a theme\cite{ben} of topical interest over more than
40: two decades. Ever since the observation of this phenomenon of
41: stochastic resonance in varied theoretical and experimental
42: contexts, a number of noise-induced resonance effects have been
43: reported \cite{mcn,roy,luc1,jul,rei,lin}. A prototypical effect of
44: this kind,
45: known as resonant
46: activation discovered by Doering and Godua \cite{dor}, concerns a
47: resonance effect in the escape rate of a particle over a fluctuating
48: barrier in a bistable potential, where the resonance
49: \cite{Bier,van,brey,shep,rein,gam,han,ros,mar}
50: can be achieved by
51: observing the variation of mean first passage time as function of
52: flipping rate of fluctuation of the barrier height. In the simplest
53: possible term the phenomenon can be realized in a model with a
54: linear barrier with a slope which fluctuates between two values. The
55: phenomenon has triggered a lot of theoretical activity around the
56: Markovian and non-Markovian variants of kinetic models for chemical
57: reactions \cite{Bier,van,brey,rein} and
58: several related issues and been experimentally observed by
59: Mantegna and Spagnolo \cite{ros}. This fluctuation of potential is
60: also important in various problems of chemical physics, notably in
61: molecular dissociation dynamics \cite{mad}, protein folding
62: \cite{wan} among others. The purpose of this paper is to extend the
63: analysis of resonant activation to quantum domain. Since at low
64: temperature thermal activation is accompanied by tunneling, the
65: question that naturally arises is how noise-induced resonance
66: effects manifest themselves in a quantum system. Specifically our
67: object here is two-fold: First, we intend to understand the
68: counterintuitive role of external noise in bringing out the
69: resonance behaviour in a bistable quantum system where the mean
70: escape time is varied as a function of correlation time of noise.
71: Second, it is worthwhile to examine the nature of this resonance in
72: presence of generic quantum effects like tunneling at low
73: temperature and allow ourselves a fair comparison with the classical
74: results. In what follows we carry out a theoretical and a numerical
75: study of quantum stochastic dynamics of a bistable system with a
76: barrier height fluctuating due to an external Ornstein-Uhlenbeck
77: noise and show how the mean escape time is profoundly influenced by
78: the statistical properties of the noise particularly the correlation
79: time in exhibiting the quantum resonance activation.
80:
81:
82: \section{Quantum stochastic dynamics}
83:
84: \subsection{General aspects}
85: To derive quantum Langevin equation from a microscopic picture
86: consider the well-known standard system-reservoir model with
87: following form of the Hamiltonian\cite{zwa}
88:
89: \begin{equation}\label{1.1}
90: \hat{H}=\frac{\hat{p}^2}{2 m}+V(\hat{x},t)+\sum_{j=1}^N \left\{
91: \frac{\hat{p}^2_j}{2}+\frac{1}{2} \kappa_j (\hat{q}_j-\hat{x})^2
92: \right\}
93: \end{equation}
94:
95: Here $\hat{x}$ and $\hat{p}$ are the coordinate and momentum
96: operators of the particle and $\{\hat{q}_j, \hat{p}_j\}$ are the
97: set of coordinate and momentum operators for the reservoir
98: oscillators coupled linearly through the coupling constants
99: ${\kappa}_j (j=1,2,...)$. The potential $V(\hat{x},t)$ is due to
100: the external force field for the Brownian particle. The coordinate
101: and momentum operators follow the usual commutation rules
102: $\{\hat{x}, \hat{p}\}=i\hbar$ and $\{\hat{q}_i,
103: \hat{p}_j\}=i\hbar{\delta}_{ij}$. Eliminating the bath degrees of
104: freedom in the usual way we obtain the operator Langevin equation
105: for the particle
106:
107: \begin{equation}\label{2.2}
108: m \ddot{\hat{x}}+\int^{t}_0
109: dt'\gamma(t-t')\dot{\hat{x}}(t')+V'(\hat{x},t) = \hat{F}(t)
110: \end{equation}
111:
112: where the noise operator $\hat{F}(t)$
113: and the memory kernel $\gamma(t)$ are given by
114:
115: \begin{equation}\label{2.3}
116: \hat{F}(t) =
117: \sum_j\left[\{\hat{q}_j(0)-\hat{x}(0)\}\kappa_j\cos\omega_jt
118: +\kappa_j^{1/2}\hat{p}_j(0) \sin\omega_jt\right]
119: \end{equation}
120:
121: and
122:
123: \begin{equation}\label{2.4}
124: \gamma(t) =\sum_j\kappa_j\cos\omega_jt
125: \end{equation}
126:
127: respectively with $\kappa_j=\omega_j^2$. On the basis of quantum
128: mechanical average $ \langle...\rangle $ over the bath modes with
129: coherent states and over the system mode with an arbitrary state
130: Eq.(2.2) can be cast into the form of the generalized quantum
131: Langevin equation \cite{db1,dbb1,dbb2,db2,db3,db4,d41,db5}.
132:
133: \begin{equation}\label{2.5}
134: m \ddot{x} + \int_0^{t}dt'\gamma(t-t') \dot{x}(t')+V'(x,t)= f(t) +
135: Q(x ,\langle\delta\hat{x}^n\rangle)
136: \end{equation}
137:
138: where the quantum mechanical mean value of the position operator
139: $\langle\hat{x}\rangle =x$. Here the quantum dispersion term $Q$
140: to the potential, is given by
141:
142: \begin{equation}\label{2.6}
143: Q(x ,\langle\delta\hat{x}^n\rangle) = V'(x,t) - \langle
144: V'(\hat{x},t)\rangle
145: \end{equation}
146:
147: which by expressing $ \hat{x}(t) = x(t) + \delta\hat{x}(t)$ in
148: $V(\hat{x},t)$ and using a Taylor series expansion around $x$ may
149: be rewritten as
150:
151: \begin{equation}\label{2.7}
152: Q(x ,\langle\delta\hat{x}^n\rangle) =-\sum_{n\geq
153: 2}\frac{1}{n!}V^{(n+1)}(x,t)\langle\delta\hat{x}^n\rangle
154: \end{equation}
155: Here $V^{(n+1)}$ the $(n+1)$th derivative with respect to $x$. The
156: calculation of $Q$ rests on the quantum correction terms
157: $\langle{\delta\hat{x}^n}\rangle$ which can be calculated order by
158: order by solving a set of quantum correction equations(as
159: discussed in the later part of this section). Furthermore the
160: quantum mechanical mean Langevin force is given by
161:
162: \begin{equation}\label{2.8}
163: f(t) =
164: \sum_j\left[\langle\hat{q}_j(0)\rangle-\langle\hat{x}(0)\rangle
165: \kappa_j\cos\omega_jt +\kappa_j^{1/2}\hat{p}_j(0)
166: \sin\omega_jt\right]
167: \end{equation}
168:
169: which must satisfy noise characteristics of the bath at
170: equilibrium ,
171:
172: \begin{eqnarray}
173: \langle{f(t)} \rangle_S & = &
174: 0\label{2.9}\\
175: \langle f(t) f(t') \rangle_S &=& \frac{1}{2} \sum_j \kappa_j\; \hbar
176: \omega_j \left( \coth \frac{\hbar \omega_j}{2 k T} \right) \cos
177: \omega_j (t-t')\nonumber\\\label{2.10}
178: \end{eqnarray}
179:
180: Eq.(\ref{2.10}) expresses the quantum fluctuation-dissipation
181: relation. The above conditions Eq.(\ref{2.9})-Eq.(\ref {2.10}) can
182: be fulfilled provided the initial shifted co-ordinates
183: $\{\langle\hat{q}_j(0)\rangle-\langle\hat{x}(0)\rangle\}$ and
184: momenta $\langle{\hat{p}_j}(0)\rangle$ of the bath oscillators are
185: distributed according to the canonical thermal Wigner distribution
186: \cite{wig,hil} of the form
187:
188: \begin{eqnarray}
189: &&P_j([\langle\hat{q}_j(0)\rangle-\langle\hat{x}(0)\rangle],
190: \langle\hat{p}_j(0)\rangle) \nonumber
191: \\
192: &=& N \exp\left\{-\;\frac{\frac {1}{2}\langle\hat{p}_j(0)\rangle^2 +
193: \frac
194: {1}{2}\kappa_j[\langle\hat{q}_j(0)\rangle-\langle\hat{x}(0)\rangle]^2}
195: {\hbar\omega_j[\overline{n}(\omega_j) +
196: \frac{1}{2}]}\right\}\nonumber \\\label{2.11}
197: \end{eqnarray}
198:
199: so that the statistical averages $\langle...\rangle_s $ over the
200: quantum mechanical mean value $O_j$ of the bath variables are
201: defined as
202:
203: \begin{equation}\label{2.12}
204: \langle O_j \rangle_s=\int O_j\; P_j\; d\langle
205: \hat{p}_j(0)\rangle\;
206: d\{\langle\hat{q}_j(0)\rangle-\langle\hat{x}(0)\rangle\}
207: \end{equation}
208:
209: Here $\overline{n}(\omega)$ is given by Bose-Einstein
210: distributions $(e^{\frac{\hbar\omega}{kT}}-1)^{-1}$. $P_j$ is the
211: exact solution of Wigner equation for harmonic oscillator
212: \cite{wig,hil} and forms the basis for description of the quantum
213: noise characteristics of the bath kept in thermal equilibrium at
214: temperature $T$. $N$ is the normalization constant. In the
215: continuum limit the fluctuation-dissipation relation (\ref{2.10})
216: can be written as
217:
218: \begin{eqnarray}
219: &&\langle {f(t)f(t')}\rangle_s\nonumber \\
220: &=&\frac{1}{2}\;\int_0^\infty d\omega\;
221: \kappa(\omega)\;\rho(\omega)\; \hbar\omega\;
222: \coth({\frac{\hbar\omega}{2kT}})\;\cos{\omega(t-t')}\nonumber
223: \\\label{2.13}
224: \end{eqnarray}
225:
226: where we have introduced the density of the modes $\rho(\omega)$.
227: Since we are interested in the Markovian limit in the present
228: context, we assume $\kappa(\omega)\rho(\omega)
229: =\frac{2}{\pi}\gamma$, and Eq.(\ref{2.13}) then yields
230:
231: \begin{equation}\label{2.14}
232: \langle f(t)f(t')\rangle_s =2 D_q\delta(t-t')
233: \end{equation}
234:
235: with
236:
237: \begin{equation}\label{2.15}
238: D_q
239: =\frac{1}{2}\gamma\hbar\omega_0\coth{\frac{\hbar\omega_0}{2kT}}
240: \end{equation}
241: (The passage from Eq.(\ref{2.13}) to Eq.(\ref{2.14}) is given in the
242: appendix A)
243:
244:
245: $\omega_0$ refers to the static frequency limit.
246: Furthermore from Eq.(\ref{2.4}) in the continuum limit we have
247:
248: \begin{equation}\label{2.16}
249: \gamma(t-t') = \gamma\;\delta(t-t')
250: \end{equation}
251:
252: $\gamma$ is the dissipation constant in the Markovian limit. In
253: this limit Eq.(\ref{2.5}) therefore reduces to
254:
255: \begin{equation}\label{2.17}
256: m \ddot{x} + \gamma \dot{x}+V'(x,t)= f(t ) + Q(x
257: ,\langle\delta\hat{x}^n\rangle)
258: \end{equation}
259: In order to consider the activated processes in a bistable
260: potential with a fluctuating barrier height we consider the
261: potential of the form
262: \begin{equation}\label{2.18}
263: V(x,t) = U(x)+g(x)\xi(t),
264: \end{equation}
265: where $U(x)$ is a bistable potential
266: ($-\frac{a}{2}x^2+\frac{b}{4}x^4$, $a$, $b$ being constants ) with
267: a barrier at the metastable point $x=0$ and two stable points at
268: $x=\pm (\frac{a}{b})^{\frac{1}{2}}$. The fluctuations in the
269: potential are driven by an Ornstein-Uhlenbeck noise process
270:
271: \begin{equation}\label{2.19}
272: \dot\xi(t)=-\frac{\xi(t)}{\tau}+\frac{\sqrt{2\sigma^2}}{\tau}\eta(t)
273: \end{equation}
274: where $\eta(t)$ is zero mean $\delta$-correlated Gaussian noise. The
275: stochastic process $\xi(t)$, can be characterized by the following
276: set of equations. The probability function $\bar{\rho}(\xi)$, the
277: variance $\sigma^2$ and the correlation function of the noise are
278: given
279: \begin{equation}\label{2.20}
280: \bar{\rho}(\xi)=\sqrt{2\pi\sigma^2}\exp(-\frac{\xi^2}{2\sigma^2})\;,
281: \end{equation}
282: where
283: \begin{equation}\label{2.21}
284: \sigma^2=\int_{-\infty}^{+\infty}\xi^2\bar{\rho}(\xi) d\xi\;,
285: \end{equation}
286: and
287: \begin{equation}\label{2.22}
288: \langle \xi(t)\xi(0)\rangle=\sigma^2\exp(-\frac{|t|}{\tau})\;,
289: \end{equation}
290: respectively. The barrier can also be subjected to dichotomic
291: fluctuation \cite{luc}, $\xi(t)=\{-\alpha,\beta\}$ that flips
292: between two values with flipping rate $\mu_{\alpha}$ and
293: $\mu_{\beta}$ respectively. This process can also be taylored as
294: zero mean valued, exponentially correlated and with a Gaussian
295: distribution.
296:
297: \subsection{Overdamped limit}
298: To proceed further we now confine ourselves to overdamped
299: condition so that the quantum Langevin equation in one variable
300: takes the following form:
301: \begin{equation}\label{2.23} \gamma \dot{x}=ax-bx^3+g'(x)
302: \xi(t)+f(t)+Q(x ,\langle\delta\hat{x}^n\rangle)
303: \end{equation}
304: The above classical-like stochastic differential equation contains
305: the essential quantum features, through the terms $f(t)$ which
306: represent the quantum noise of the heat bath and the another term
307: $Q(x ,\langle\delta\hat{x}^n\rangle)$ which essentially arises due
308: to nonlinear part of the system potential. The nonlinearity and the
309: quantum effects are entangled in the latter quantity modifying the
310: classical part of the potential. Thus the classical potential force
311: $-V'(x,t)$ is modified by the quantum dispersion. In absence of
312: quantum dispersion term $Q(x ,\langle\delta\hat{x}^n\rangle)$ and
313: with $D_q\rightarrow \gamma kT $ as one approaches the classical
314: limit ($kT\gg\hbar\omega_0$), the quantum Langevin equation reduces
315: to classical one. A probabilistic description of the system Eq.(\ref
316: {2.23}) is provided by the time-dependent probability distribution
317: function $\bar{P}(x,\xi,t)$. For a finite correlation time of the
318: noise driving the barrier to fast fluctuations it is
319: possible\cite{han} to go over to an approximate (exact in the limit
320: $\tau\rightarrow 0$) description which allows one to transform
321: externally driven Langevin equation Eq.(\ref {2.23}) into an
322: autonomous Langevin equation for the position variable $x$. The
323: corresponding equation for probability density function $P(x,t)$ is
324: given by
325:
326: \begin{eqnarray}\label{2.24}
327: \frac{\partial P(x,t)}{\partial t}&=&-\frac{\partial}{\partial x}
328: \left[
329: \kappa(x,\tau)f(x)\right]P(x,t)\nonumber
330: \\
331: &+&\sigma^2\frac{\partial}{\partial
332: x} \left[ \kappa(x,\tau)g'(x)\frac{\partial}{\partial x}
333: \kappa(x,\tau)g'(x)\right]P(x,t)\nonumber
334: \\
335: &+&D_q\frac{\partial}{\partial x}
336: \left[ \kappa(x,\tau)\frac{\partial}{\partial x} \kappa(x,\tau)\right]P(x,t)
337: \end{eqnarray}
338:
339: where $\gamma$ is assumed to unity, $\kappa(x,\tau)=\left[1-\tau
340: g'(x) \left(\frac{f(x)}{g'(x)} \right)'\right]^{-1}$ and
341: $f(x)=-U'(x)+Q(x ,\langle\delta\hat{x}^n\rangle)$. The quantum
342: nature of the above Fokker-Planck equation governing evolution of
343: $P(x,t)$ is manifested through the quantum correction to the
344: potential term and the quantum diffusion coefficient
345: characterizing the thermal bath.
346:
347: The quantity of special interest here is the mean thermally
348: activated escape time of the particle from the well whose barrier is
349: subjected to fluctuation. The particle is governed by the stochastic
350: dynamics Eq.(\ref {2.24}). It is important to emphasize that the
351: typical mean escape time should be much larger than the time scale
352: of the deterministic quantum dynamics $\dot{x}=-U'(x)+Q(x
353: ,\langle\delta\hat{x}^n\rangle)$ for nonzero fluctuation of the
354: potential well, and it is further required that the strengths of
355: thermal noise and barrier fluctuation must be small in comparison
356: with the barrier height. The solution of Eq.(\ref {2.24}) in the
357: stationary state reads as
358:
359: \begin{eqnarray}\label{2.25}
360: &&P_s(x,\tau)= \frac{\left[1-\tau g'(x) \left(\frac{f(x)}{g'(x)}
361: \right)'\right]}{{D_q^{\frac{1}{2}}\left(1+\frac{\sigma^2}{D_q}g'(x)^2\right)}
362: ^{\frac{1}{2}}}\nonumber
363: \\
364: &\times& \exp\left(\int_0^x f(y) \frac{\left[1-\tau g'(y)
365: \left(\frac{f(y)}{g'(y)}
366: \right)'\right]}{D_q\left(1+\frac{\sigma^2}{D_q}g'(y)^2\right)}dy
367: \right)
368: \end{eqnarray}
369: The stationary probability distribution function $P_s(x,\tau)$ of the stochastic
370: process $x(t)$, which is the quantum mechanical mean position, is bimodal
371: in nature with nonzero probability current at the barrier top even at zero
372: temperature characterizing the zero point contribution of the thermal bath
373: (as $T\rightarrow 0$, $D_q \rightarrow \hbar\omega_0$). $P(x,\tau)$ is additionally
374: modified over its classical nature by a quantum contribution characterizing
375: a correction due to anharmonic part of the system potential.
376:
377:
378:
379:
380: To proceed further it is necessary to find out the quantum
381: correction term $Q(x ,\langle\delta\hat{x}^n\rangle)$ more
382: explicitly. To this end we return to the overdamped operator
383: equation (\ref{2.2}) and use $\hat{x}( t)=x( t)+
384: \delta\hat{x}(t)$, where $x( t)(=\langle\hat{x}(t)\rangle)$ is
385: the quantum mechanical mean value of the operator $\hat{x}$. By
386: construction $[\delta\hat{x},\delta\hat{p}]=i\hbar$ and
387: $\langle\delta \hat{x}\rangle=0$. We then obtain the quantum
388: correction equation in the overdamped limit after quantum
389: mechanical averaging with the coherent states over the bath
390: operators as
391:
392: \begin{equation}\label{2.26}
393: \gamma\delta\dot{\hat{x}} +V''(x,t)\delta\hat{x}+ \sum_{n\geq
394: 2}\frac{1}{n!}{V}^{n+1}(x,t)(\delta\hat{x}^n- \langle
395: \delta\hat{x}^n\rangle)=0
396: \end{equation}
397:
398: With the help of operator equation (\ref{2.26}) we obtain the
399: equations for $\langle\delta\hat{x}^n\rangle$
400:
401: \begin{equation}\label{2.27}
402: \frac{d}{dt}\langle\delta\hat{x}^2\rangle=-\frac{1}{\gamma}\left[2V''(x,t)\langle\delta\hat{x}^2\rangle
403: +V'''(x,t)\langle\delta\hat{x}^3\rangle\right]
404: \end{equation}
405:
406: \begin{eqnarray}\label{2.28}
407: \frac{d}{dt}\langle\delta\hat{x}^3\rangle&=&-\frac{1}{\gamma}\left[
408: 3V''(x,t)\langle\delta\hat{x}^3\rangle
409: +\frac{3}{2}V'''(x,t)\langle\delta\hat{x}^4\rangle\right]\nonumber
410: \\
411: &&+\left[\frac{3}{2 \gamma}V'''(x,t)
412: \langle\delta\hat{x}^2\rangle^2\right]
413: \end{eqnarray}
414: and so on. To take into account of the leading order contribution
415: $\langle\delta\hat{x}^2\rangle$ explicitly we may write
416: \begin{equation}\label{2.29}
417: d\langle\delta\hat{x}^2\rangle =
418: -\frac{2}{\gamma}V''(x,t)\langle\delta\hat{x}^2\rangle dt
419: \end{equation}
420: The overdamped deterministic motion on the other hand gives $\gamma
421: dx = -V'(x,t)dt$ which when used in Eq.(\ref {2.29}) yields after
422: integration
423: \begin{equation}\label{2.30}
424: \langle\delta\hat{x}^2\rangle=\Delta_q \left[ V'(x,t)\right]^2
425: \end{equation}
426: $\Delta_q$ is the quantum correction parameter, as given by
427: $\Delta_q=\frac{\langle\delta\hat{x}^2\rangle_{x_c}}{\left[
428: V'(x_c,t)\right]^2}$, $x_c$ being a given quantum mechanical mean
429: position.
430:
431:
432:
433:
434: \section{Resonant activation in the quantum system}
435:
436: We are now in a position to analyze the quantum resonant activation.
437: In our present problem the fluctuating part of the potential
438: associated with harmonic term is assumed to be of the form
439: $g(x)=\frac{x^2}{2}$ and $x= \pm \sqrt{\frac{a}{b}}$ and $x=0$ are
440: the absolute minima and maximum of the potential $U(x)$
441: respectively. Having known the stationary distribution
442: Eq.(\ref{2.25}) along with the quantum correction in $f(x)$ the
443: calculation of stochastic dynamics is straightforward. This can be
444: obtained using the standard result on mean first passage time
445: \cite{str} as
446: \begin{equation}\label{3.1}
447: \langle T\rangle=\int^{0}_{-\sqrt{\frac{a}{b}}}
448: \frac{dx}{D(x,\tau)P_s(x)}\int_{-\infty}^x P_s(x)dy
449: \end{equation}
450: where $D_{eff}(x,\tau)=D_q\left(1+Rg'(x)^2\right)\left[1-\tau g'(x)
451: \left(\frac{f(x)}{g'(x)} \right)'\right]^{-2}$
452: to give
453: \begin{equation}\label{3.2}
454: \langle T\rangle=\frac{2\pi \sqrt{1+2\tau(a+\Delta_1)}}
455: {a\sqrt{2(1-\Delta_2)}}\exp\left[\frac{\Delta V_{eff}(R,\tau)}{D_q}\right]
456: \end{equation}
457:
458: Here $\Delta_1$ and $\Delta_2$ both are the leading order quantum correction terms as given by
459: \begin{equation}\label{3.3}
460: \Delta_1=\left[-g'(x)\left(\frac{Q(x
461: ,\langle\delta\hat{x}^n\rangle)}{g'(x)}\right)\right]_{x=-\sqrt{\frac{a}{b}}}
462: \end{equation}
463:
464: \begin{equation}\label{3.4}
465: \Delta_2=\left[Q'(x
466: ,\langle\delta\hat{x}^n\rangle)\right]_{x=-\sqrt{\frac{a}{b}}}
467: \end{equation}
468: and $R=\frac{\sigma^2}{D_q}$ ; $\Delta V_{eff}(R,\tau)$ is the
469: effective barrier height of the following form
470:
471: \begin{eqnarray}
472: &&\Delta V_{eff}(R,\tau)\nonumber \\
473: &=& -\int^{0}_{-\sqrt{\frac{a}{b}}}\frac{\left( -U'(x)+Q\right)
474: \left[ 1+\tau
475: g'(x)\left(\frac{-U'(x)+Q}{g'(x)}\right)\right]dx}{\left(1+R
476: g'(x)^2\right)}\nonumber\\\label{3.5}
477: \end{eqnarray}
478: Eq.(\ref {3.2}) is the central analytical result of this paper. To
479: analyse the theoretical results on the essential features of the
480: activated escape following quantum stochastic dynamics over a
481: fluctuating potential well we now resort to numerical simulation of
482: the Eqs.(\ref {2.23}),(\ref{2.19}) and the quantum correction
483: equations (\ref {2.27}) and (\ref {2.28}) simultaneously using
484: standard Heun's algorithom.
485: A very small time step ($\triangle t$) $0.001$ for
486: numerical integration has been used. In our simulation we follow the
487: dynamics of each particle starting in the left well at
488: $x=-\sqrt{\frac{a}{b}}$ till it arrives at the barrier top at $x=0$.
489: The first passage time being a statistical quantity due to the
490: random force we calculate the statistical average of the first
491: passage time over $1,000$ trajectories. We present the numerical
492: results in Fig.1 to Fig.4 for different parameters such as
493: temperatures, barrier height of the potential and noise strength of
494: the Ornstein-Unhenbeck noise process. All the curves exhibit minimum
495: at optimal $\tau$ values. Increasing strength of thermal or
496: non-thermal noise results in enhancement of escape rate and the
497: minima are shifted towards the origin. Physically this implies that
498: with increase in $D_q$ or $\sigma^2$, the escape time over the
499: barrier changes as a result of which the switching time of the
500: barrier fluctuations matches the order of escape time at lower
501: $\tau$ values. In order to examine the influence of the barrier
502: height on resonant activation we plot in Fig.3 the variation of mean
503: escape time as function of $\tau$ for several values of barrier
504: heights. Increase in the barrier height results in enhancement of
505: escape time and the minima are shifted to higher $\tau$ values.
506:
507:
508:
509:
510: We find that the barrier height Eq.(\ref {3.5}) monotonically
511: decreases to a limiting value
512: with increasing $\tau$.
513: Due to the presence of system
514: nonlinearity the barrier height
515: is slightly modified by an added slope. It is thus expected that
516: during the temporary stay of the particle close to a minimum it
517: makes unsuccessful attempts to escape and once the escape takes
518: place it should occur at a slower rate in quantum case due to
519: added slope to the effective potential than the corresponding
520: classical case.
521: However, our theoretical and numerical calculations reveal that the
522: mean escape time is lower in magnitude for the quantum particle at low
523: temperature and the difference becomes insignificant at higher
524: temperature. This is shown in the Fig.4. This behaviour can be
525: interpreted in terms of an interplay between the quantum
526: diffusion coefficient $D_q$ and the quantum correction due to
527: system nonlinearity appearing in $V_{eff}(R,\tau)$.
528: When the temperature of the system is very low, i. e., in the vacuum
529: limit or in the deep tunneling region the anharmonic terms in the
530: potential do not contribute significantly. On the other hand as
531: temperature of the system increases significantly, $D_q$ increases
532: resulting in decrease of the effective potential and hence $D_q$ and
533: $Q$ compete to cancel the effect of each other at higher
534: temperature. Finally in Fig.5 we make a comparison between the
535: numerical simulation and the corresponding theoretical result. The
536: theoretical result agrees fairly well with our numerical result.
537:
538: In addition to the barrier height the prefactor is also affected by
539: the quantum correction term. The approximate theoretical mean escape
540: time Eq.(\ref {3.2}) implies that both the barrier height and the
541: frequency factor have an important role to play with the activated
542: escape process. We are now in a position to point out that the
543: thermally activated resonance phenomenon is controlled by three
544: different time scales as emphasized earlier \cite{rein}. The three
545: relavent time scales are the corresponding time of barrier
546: fluctuation ($0\ll\tau\ll \infty$), typical escape time $\hat T$,
547: which is much larger than the time scale describing deterministic
548: motion $\dot x=-U'(x)+Q'(x ,\langle\delta\hat{x}^n\rangle)$, the
549: third time scale being $T_a$ of the escape attempt ($T_a\ll \hat
550: T$).
551:
552:
553: \section{conclusion}
554: Based on the theoretical study and numerical simulation of quantum
555: stochastic dynamics in a double-well system with a fluctuating
556: barrier under influence of a Gaussian color noise, we have examined
557: the phenomenon of resonant activation. The governing equations are
558: classical looking in form but quantum mechanical in their content. A
559: key point of the present analysis is to describe the thermal bath in
560: terms of a canonical thermal Wigner distribution of harmonic
561: oscillators. This distribution remain positive definite even at
562: absolute zero signifying a pure state and allows us to look for the
563: external noise-induced resonance when the generic quantum effects in
564: the system make their presence felt. We have shown that quantization
565: significantly enhances resonant activation at low temperature due to
566: tunneling. For higher temperature as well as for stronger noise
567: strength the resonance effects get more pronounced. Since tunneling
568: accompanies activation, it is expected that the resonant activation
569: can be observed even at absolute zero, where the resonance effect
570: essentially is due to the vacuum field.
571:
572:
573: \begin{appendix}
574:
575: \section{\textbf{(The passage from Eq.(\ref {2.13}) to Eq. (\ref {2.14}))}}
576: We start from basic
577: definition(Louisell p-426) \cite{loui}
578: \begin{equation}\label{A1}
579: 2D_q=\frac{1}{2\Delta{t}}\int_t^{t+\Delta{t}}dt\int_t^{t+\Delta{t}}dt'\;\;
580: \langle f(t)f(t')\rangle_s
581: \end{equation}
582: Using Eq.(\ref {2.13}) in Eq.(\ref{A1}) yields
583: \begin{eqnarray}\label{A2}
584: 2D_q&=&\frac{1}{2\Delta{t}}\int_0^\infty
585: d\omega\;\kappa(\omega)\;\rho(\omega)\hbar\omega\;\coth\left(\frac{\hbar\omega}{2kT}\right)
586: \nonumber
587: \\
588: &&\times\int_t^{t+\Delta{t}}dt\int_t^{t+\Delta{t}}dt'\;\cos\omega(t-t')\nonumber\\\label{A2}
589: \end{eqnarray}
590: Explicit integration over time gives
591: \begin{equation}\label{A3}
592: 2D_q=\frac{1}{2\Delta{t}}\int_0^\infty
593: d\omega\;\kappa(\omega)\;\rho(\omega)\hbar\omega\;\coth\left(\frac{\hbar\omega}{2kT}
594: \right)\;\;\;I(\omega,\Delta{t} )
595: \end{equation}
596: where
597: \begin{equation}\label{A4}
598: I(\omega,\;\Delta{t})=\frac{4}{w^2}\;\;\sin^2\frac{\omega\Delta{t}}{2}
599: \end{equation}
600: Putting $\kappa(\omega)\;\rho(\omega)=\frac{2}{\pi}\;\gamma$, we
601: obtain
602: \begin{equation}\label{A5}
603: 2D_q=\frac{\gamma}{\Delta{t}\;\pi}\int_0^\infty
604: d\omega\;\hbar\omega\;\coth\left(\frac{\hbar\omega}{2kT}\right)\;\;
605: \frac{\sin^2\frac{\omega\Delta{t}}{2}}{(\frac{\omega}{2})^2}
606: \end{equation}
607:
608: Following Louisell(p $-426$) \cite{loui} we have under Markovian
609: condition, the correlation time $\tau_c\;\ll\;\Delta{t}$, the
610: coarse-grain time(over which the probability distribution function
611: evolves)
612:
613: Thus as $\Delta{t}\rightarrow \infty$(in scale of $\tau_c$ which
614: goes to zero) the function
615: $\frac{\sin^2\frac{\omega\Delta{t}}{2}}{(\frac{\omega}{2})^2}$
616: oscillates violently so that one takes the slowly varying quantity
617: $[\hbar\omega\;\coth\frac{\hbar\omega}{2kT}]$ out of the integration
618: over frequency with an average value
619: $\hbar\omega_0\coth\frac{\hbar\omega_0}{2kT}$, $\omega_0$ be an
620: average static frequency. Since the integral
621: $\int_\infty^\infty\frac{\sin^2x\Delta{t}}{x^2}dx=\pi\Delta{t}$ it
622: follows immediately from Eq.(\ref{A5})
623:
624: \begin{equation}\label{A6}
625: 2 D_q=\gamma\; \hbar \omega_0 \coth\left(\frac{\hbar \omega_0}{2 k
626: T}\right)
627: \end{equation}
628:
629: as given in Eq.(\ref{2.15}).
630:
631: Again starting from Eq.(\ref{2.13}), we use the same argument as
632: before to have
633:
634: \begin{eqnarray}\nonumber
635: &&\langle f(t)f(t')\rangle_s
636: \nonumber \\
637: &&= \frac{1}{2}\int_0^\infty d\omega\; \kappa(\omega)\rho(\omega)
638: \hbar \omega \coth\left(\frac{\hbar \omega}{2 k T}\right)
639: \cos\omega(t-t')\nonumber
640: \end{eqnarray}
641:
642: and we use
643:
644: \begin{eqnarray}
645: \int_0^\infty d\omega \cos \omega \tau = \pi\; \delta(\tau)\nonumber
646: \end{eqnarray}
647:
648: to obtain
649:
650: \begin{eqnarray}
651: &&\langle f(t)f(t') \rangle_s\nonumber \\
652: &&= \frac{1}{2}
653: \int_0^\infty d\omega \left[\frac{2}{\pi} \gamma \right] \hbar
654: \omega \coth\left(\frac{\hbar \omega}{2 k
655: T}\right) \cos\omega(t-t')\nonumber\\
656: &&= \frac{\gamma\;\hbar \omega_0}{\pi}\coth\left(\frac{\hbar
657: \omega_0}{2 k T}\right) \pi \delta(t-t')\nonumber\\
658: &&=\gamma\; \hbar \omega_0 \coth\left(\frac{\hbar \omega_0}{2 k
659: T}\right) \delta(t-t')\label{A7}
660: \end{eqnarray}
661:
662: Therefore from Eq.(\ref{A6}) and Eq.(\ref{A7}) we have
663:
664: \begin{eqnarray}
665: \langle f(t)f(t') \rangle_s &=& 2D_q \delta(t-t')\nonumber
666: \end{eqnarray}
667:
668: which is Eq.(\ref{2.14}).
669:
670: Thus the derivation within Markovian approximation clearly depends
671: on the time scale separation. The results are valid even at absolute
672: zero as emphasized by Louisell.
673:
674:
675: \end{appendix}
676:
677:
678: \acknowledgments Thanks are due to the Council of Scientific and
679: industrial research, Govt. of India, for partial financial
680: support.
681:
682:
683: \begin{thebibliography}{99}
684: \bibitem{ben} R. Benzi, G. Parisi, A. Sutera, and A. Vulpiani,
685: Tellus {\bf 34}, 10 (1982); R. Benzi, A. Sutera, and A. Vulpiani, J.
686: Phys. A {\bf 14} L 453 (1981)
687:
688: \bibitem{mcn} B. McNamara and K. Wiesenfeld, Phys. Rev. A, {\bf
689: 39}, 4854 (1988); L. Gammaitoni, F. Marchesoni, E. Menichella-Saetta
690: and S. Santucci, Phys. Rev. Lett. 62, 349 (1989).
691:
692: \bibitem{roy} B. McNamara, K. Wiesenfeld and R. Roy, Phys. Rev.
693: Lett. 60, 2626 (1988)
694:
695: \bibitem{luc1} L. Gammaitoni, P. H\"{a}nggi, P. Jung and
696: F.Marchesoni, Rev. Mod. Phys., {\bf 70}, 223(1998)
697:
698: \bibitem{jul} F. Julicher, A. Ajdari and J. Prost, Rev. Mod. Phys.
699: {\bf 69}, 1269 (1997).
700:
701: \bibitem{rei} M. O. Magnasco,
702: Phys. Rev. Lett. {\bf 71}, 1477 (1993), C. R. Doering, W. Horsthemke
703: and J. Riordan, Phys. Rev. Lett. {\bf 72}, 2984 (1994).
704:
705:
706:
707: \bibitem{lin} H. Linke, T. E. Humphrey, A. Löfgren, A. O. Sushkov,
708: R. Newbury, R. P. Taylor, and P. Omling, Science {\bf 286}, 2314
709: (1999); L. Q. Zhou, X. Jia and Q. Ouyang, Phys. Rev. Lett.
710: \textbf{88}, 138301(2002)
711:
712: \bibitem{j2} J. Garcia-Ojalvo, A. Hernandez-Machado and J. M.
713: Sancho, Phys. Rev. Lett. \textbf{71}, 1542 (1993)
714:
715: \bibitem{a} A. Becker and L. Kramer, Phys. Rev. Lett.
716: \textbf{73}, 955 (1994), A. Becker and L. Kramer, Physica D
717: \textbf{90}, 408(1995)
718:
719: \bibitem{p} J. M. R. Parrondo, C. Van den Broeck, J. Buceta and
720: F. J. de la Rubia, Physica A \textbf{224}, 153(1996)
721:
722: \bibitem{aa}A. A. Zaikin and L. Schimansky-Geier, Phys. Rev.
723: E. \textbf{58}, 4355(1998)
724:
725: \bibitem{b}J. Buceta, M. Ibanes, J. M. Sancho and K.
726: Lindenberg, Phys. Rev. E \textbf{67}, 021113(2003)
727:
728: \bibitem{sd}S. Dutta, S. S. Riaz and D. S. Ray, Phys. Rev. E
729: {\bf 71}, 036216 (2005)
730:
731: \bibitem{dor} C. R. Doering and J. C. Gadoua, Phys. Rev. Lett. {\bf 69},
732: 2318 (1992)
733:
734: \bibitem{Bier} M. Bier and R. D. Astumian, Phys. Rev. Lett. {\bf 71},
735: 1649 (1993); U. Z\"{u}rcher and C. R. Doering, Phys. Rev. E {\bf 47},
736: 3862 (1993)
737: \bibitem{van} C. Van den Broeck, Phys. Rev. E {\bf 47},
738: 4579 (1994)
739: \bibitem{brey} J. J. Brey and J. Casado-Pascual, Phys. Rev. E. {\bf 50},
740: 116 (1994); O. Flomenbom and J. Klafter, Phys. Rev. E. {\bf 69}, 051109 (2004)
741:
742: \bibitem{shep} T. D. Shepherd and R. Hernandez J. Chem. Phys. {\bf 115},
743: 2430 (2001); A. M. Berezhkovskii, A. Szabo, G. H. Weiss, and H. Zhou
744: J. Chem. Phys. 111, 9952 (1999)
745:
746: \bibitem{rein} M. Boqu\"{n}\'{a}, J. M. Porr\.{a}, J. Masoliver and K. Lindenberg,
747: Phys. Rev. E {\bf 57}, 003990 (1998)
748:
749: \bibitem{gam} M. Marchi, F. Marchesoni, L. Gammaitoni, E. Menichella -Saetta,
750: and S. Santucci
751: Phys. Rev. E {\bf 54}, 3479 (1995)
752: \bibitem{han} P. H\"{a}nggi, Chem. Phys., {\bf 180} 157 (1994); R.
753: F. Fox, Phys. Rev. A. {\bf 34}, R4525 (1986); Y. Jia and J. R. Li
754: Phys. Rev. E. {\bf 53}, 5786 (1996); X. Luo and S. Zhu, Phys. Rev. E
755: {\bf 67}, 021104 (2003)
756:
757: \bibitem{ros} R. N. Mantegna and B. Spagnolo, Phys. Rev. Lett. {\bf 84},
758: 3025 (2000)
759: \bibitem{mad} J. Maddox , Nature(London) {\bf 359}, 771 (1992).
760: \bibitem{wan} J. Wang and P. Wolynes, Comput. Phys. {\bf 180} 141 (1994)
761: \bibitem{mar} J. M. Martinis, M. H. Devoret and J. Clarke, Phys. Rev. Lett.
762: {\bf 55}, 1543 (1985)
763:
764: \bibitem{str} R. L. Straonovich, \textit{Topics in the Theory of Random
765: Noise}, Vol. 1 (Gordon and Breach, New York, 1963); K. Lindenberg
766: and B.J. West, J. Stat. Phys. {\bf 42}, 201 (1986); J. Masoliver,
767: B.J. West, and K. Lindenberg, Phys. Rev. A {\bf 35}, 3086 (1987); D.
768: Mei, G. Xie, Li Cao, and D. Wu, Phys. Rev. E {\bf 59}, 3880 (1999)
769:
770:
771: \bibitem{db1} D. Banerjee, B. C. Bag, S. K. Banik and D. S. Ray,
772: Phys. Rev. E {\bf 65}, 021109 (2002).
773: \bibitem{dbb1} D. Barik and D. S. Ray, J. Chem. Phys. {\bf 119},
774: 12973 (2003).
775: \bibitem{dbb2} D. Barik, S. K. Banik and D. S. Ray, J. Chem. Phys. {\bf 119},
776: 680 (2003).
777:
778:
779: \bibitem{db2} D. Banerjee, B. C. Bag, S. K. Banik and D. S. Ray,
780: J. Chem. Phys. {\bf 120}, 8960 (2004).
781:
782: \bibitem{db3} D. Barik and D. S. Ray, J. Chem. Phys. {\bf 121},
783: 1681 (2004).
784:
785: \bibitem{db4} P. K. Ghosh, D. Barik and D. S. Ray,
786: Phys. Rev. E {\bf 71}, 041107 (2005)
787: \bibitem{d41}P. K. Ghosh, D.
788: Barik and D. S. Ray, Physics Letters A {\bf 342} 12 (2005).
789:
790: \bibitem{db5} D. Barik and D. S. Ray, J. Stat. Phys. (2005)
791: To be published.
792:
793: \bibitem{zwa} R. Zwanzig, J. Stat. Phys.{\bf 9}, 215 (1973).
794:
795: \bibitem{luc} A. Mielke, Ann. Phys. {\bf 4}, 476 (1995); J. M. Sancho, Phys. Rev. A
796: {\bf 31}, R3523 (1985); J. M. Gutiérrez, A. Iglesias,
797: and M. A. Rodríguez, Phys. Rev. E. {\bf 48}, 2507 (1993)
798:
799:
800: \bibitem{wig} E. P. Wigner, Phys. Rev. {\bf 40}, 749 (1932).
801:
802: \bibitem{hil} M. Hillery, R. F. O'Connell, M. O. Scully and E. P.
803: Wigner, Phys. Reps. {\bf 106}, 121 (1984).
804: \bibitem{loui} W. H. Louisell, \textit{Quantum Statistical
805: Properties of Radiation}, (John Wiley and Sons, New York, 1990);
806: \end{thebibliography}
807:
808: \begin{center}
809: {\bf Figure Captions}
810: \end{center}
811:
812: Fig.1. A plot of mean escape time vs $\tau$ using Gaussian colour
813: noise for different temperatures [(i) $T=0.1$ (dotted line), (ii)
814: $T=1.0$ (dashed line), (iii) $T=2.0$ (solid line)] and the other
815: parameter set $a=0.5$,
816: $b=0.005$ and $\sigma^2=1.0$.
817:
818: Fig.2. A plot of mean escape time vs $\tau$ using Gaussian colour
819: noise for different $\sigma^2$ [(i)$\sigma^2=0.5$ (dashed line),
820: (ii) $\sigma^2=1.0$ (dotted line), (iii) $\sigma^2=1.5$ (solid
821: line)] and the other parameter
822: set $a=0.5$, $b=0.005$ and $T=1.0$.
823:
824:
825: Fig.3. A plot of mean escape time vs $\tau$ using Gaussian colour
826: noise for different barrier heights [(i) $a=0.3$ (dashed line), (ii)
827: $a=0.5$ (dotted line), (iii) $a=0.7$ (solid line)]
828: and the other parameter set $b=0.005$, $\sigma^2=1.0$ and
829: $T=1.0$.
830:
831: Fig.4. A comparison between classical(dotted line) and quantum
832: (solid line) mean escape time describing resonant activation
833: phenomenon using Gaussian colour noise for the parameter set
834: $a=0.5$, $b=0.005$ and $\sigma^2=1.0$ at two different temperatures
835: $T=1.5$ and $T=0.1$.
836:
837:
838: Fig.5. A comparison between numerical(dashed dot dot line and dashed
839: line) and analytical (solid and dotted line) results using Gaussian
840: colour noise for different $\sigma^2$ [(i) $\sigma^2=0.5)$ (solid
841: and dashed dot dot line), (ii) $\sigma^2=1.5$ (dotted and dashed
842: line)] and the other parameter set $a=0.5$, $b=0.005$ and $T=1.5$.
843:
844:
845: \end{document}
846: