1: \documentclass[aps,twocolumn,prb,showpacs]{revtex4}

2: \usepackage{graphics,amssymb,amsmath,epsfig}

3: %\documentstyle[prb,aps,epsfig,multicol]{revtex}

4: \begin{document}

5: 

6: \title{Autonomous Stochastic Resonance in Fully Frustrated Josephson-Junction Ladders}

7: \author{Gun Sang Jeon}

8: \affiliation{Center for Strongly Correlated Materials Research,

9: Seoul National University, Seoul 151-747, Korea}

10: \author{M.Y. Choi}

11: \affiliation{Department of Physics, Seoul National University, Seoul 151-747,

12: Korea}

13: 

14: \begin{abstract}

15: We investigate autonomous stochastic resonance in fully frustrated

16: Josephson-junction ladders, which are driven by uniform constant currents.

17: At zero temperature large currents induce oscillations between the

18: two ground states, while for small currents the lattice potential forces

19: the system to remain in one of the two states.

20: At finite temperatures, on the other hand, oscillations between

21: the two states develop even below the critical current;

22: the signal-to-noise ratio

23: %decreasing monotonically with temperature for large currents,

24: is found to display array-enhanced stochastic resonance.

25: % below the critical current.

26: It is suggested that such behavior may be observed experimentally

27: through the measurement of the staggered voltage.

28: \end{abstract}

29: 

30: \pacs{74.50.+r, 74.25.Nf, 05.40.-a}

31: 

32: \maketitle

33: 

34: The stochastic resonance (SR), which describes the phenomenon

35: that a noise of finite strength optimizes the response signal to a weak

36: periodic external force in a nonlinear system, has attracted much

37: attention for the past two decades.\cite{review}

38: %The concept was originally proposed in the problem of the periodic

39: %oscillation of the ice ages~\cite{ice} and has been widely studied in

40: %a variety of fields.\cite{example}

41: Conventional SR phenomena are generally known to require an energy barrier

42: between two stable states, a weak external periodic force, and noise.

43: However, the SR behavior was also exhibited without any external periodic force,

44: in limit-cycle systems\cite{Gang} and excitable systems.\cite{CR,CR2}

45: Subsequently such coherence resonance or {\em autonomous} SR

46: has also been observed in systems with

47: delay\cite{delay} and inertia.\cite{inertia}

48: In the latter cases, SR phenomena can still be understood by the

49: time-scale matching argument since the delay or the inertia

50: provides an external time-scale.

51: In the former, on the other hand, such an external time scale

52: apparently does not exist,

53: %which makes the phenomena more interesting.

54: and different noise dependencies of the activation

55: time and the excursion time have been proposed as

56: an explanation of the coherence resonance.\cite{CR2}

57: %

58: Recently, the fully-frustrated Josephson-junction ladder has been

59: proposed to give a good physical realization of the standard two-state SR system

60: with many degrees of freedom.\cite{acSR}

61: In the presence of the external current which is periodic in time and staggered

62: in space, the SR behavior and other rich physics

63: have been demonstrated theoretically in the ladder system.

64: Furthermore, recent progress in the fabrication technique makes such

65: Josephson-junction systems available for experimental study.\cite{review2,Delsing}

66: 

67: In this work we study fully frustrated Josephson-junction

68: ladders driven by {\em uniform constant} currents,

69: paying particular attention to the possibility of the autonomous SR phenomena.

70: %

71: Note that this system is {\em bistable}, possessing two stable states;

72: this is in sharp contrast with the usual excitable system,

73: characterized by a single stable state.

74: %

75: Numerical integration of the coupled equations of motion,

76: established within the resistively shunted junction (RSJ) model,

77: shows the existence of the critical current beyond which

78: oscillations emerge between the two ground states. Below the

79: critical current, such oscillations are suppressed by the lattice

80: potential; here the addition of noise currents, relevant at finite

81: temperatures, induces again oscillations, giving rise to SR

82: behavior. In particular, the signal-to-noise ratio (SNR) is found

83: to display array-enhanced SR phenomena, which may be observed

84: experimentally through the measurement of the staggered voltage.

85: To our knowledge, such array-enhanced SR behavior has not been

86: observed in bistable system without periodic driving although

87: coherence resonance was reported in the bistable regime of the

88: FitzHugh-Nagumo model.~\cite{bistable}

89: 

90: The RSJ model, with single-junction

91: critical current $I_c$ and shunt resistance $R$,

92: is described by the set of equations of motion

93: for the phase $\phi_i$ of the superconducting order parameter

94: on grain $i$:\cite{acSR,review2}

95: \begin{equation}

96: \label{eqmotion}

97: {\sum_j}' \left[ {\hbar \over 2eR} {d \phi_{ij} \over dt}

98: + I_c \sin(\phi_{ij}-A_{ij}) + \eta_{ij}\right]

99: = I_i^{\rm ext} ,

100: \end{equation}

101: where $\phi_{ij}\equiv \phi_i - \phi_j$ is the phase difference

102: across the junction $(ij)$,

103: the thermal noise current $\eta_{ij}$ is assumed to be the white noise

104: satisfying

105: \begin{equation}

106: \langle \eta_{ij}(t{+}\tau) \eta_{kl} (t) \rangle

107: = {2 T \over R} \delta (\tau) (\delta_{ik}\delta_{jl} -

108: \delta_{il}\delta_{jk}),

109: \end{equation}

110: and the primed summation runs over the nearest neighbors of grain $i$.

111: We choose the Landau gauge and write the gauge field $A_{ij}$,

112: which describes the transverse magnetic field, in the form

113: \begin{equation}

114: A_{ij} = \left\{

115: \begin{array}{ll}

116: 0 &  \hbox{for }(ij) \hbox{ legs},\\

117: 2\pi f x &  \hbox{for }(ij) \hbox{ rungs},

118: \end{array}

119: \right.

120: \end{equation}

121: where $f \equiv \Phi / \Phi_0$ is the flux $\Phi$ per plaquette

122: in units of the flux quantum $\Phi_0 \equiv hc / 2e$

123: and $x \, (= 1,2, \ldots, L)$ denotes the position along the leg.

124: As illustrated in Fig.~\ref{fig:ladder}, the ladder, placed along the $x$ direction,

125: is driven by the uniform constant currents $I$ along the rung direction (i.e.,

126: the $y$ direction):

127: The external current $I_i^{\rm ext}$ fed into grain $i$ is thus given by

128: \begin{equation}

129: I_i^{\rm ext}(t) = (-1)^y I ,

130: \end{equation}

131: where $y \, (= 1,2)$ is the leg index of the $i$th grain.

132: Henceforth, we write the current, the temperature, and the time in units of

133: $I_c$, $\hbar I_c / 2e$, and $\hbar / 2eRI_c$, respectively.

134: \begin{figure}

135: \epsfig{file=fig1.eps,width=7.5cm}

136: \caption{Schematic diagram of a fully frustrated Josephson junction ladder

137: driven by uniform constant currents.}

138: \label{fig:ladder}

139: \end{figure}

140: 

141: \begin{figure}

142: \epsfig{file=fig2.eps,width=8cm}

143: \vspace*{0.4cm}

144: \caption{Time evolution of the staggered magnetization for various

145: currents $I = 0.2,\, 0.25$, and $0.3$ from above.

146: For clarity, the data corresponding to $I=0.2$ and $0.3$ have been shifted by $1.5$

147: upward and downward, respectively.}

148: \label{fig:zeroT}

149: \end{figure}

150: In the absence of external currents, the fully frustrated ($f=1/2$)

151: ladder is well known to have two degenerate ground states,

152: where a vortex is located on every two plaquettes.\cite{acSR}

153: (Note that the length $L$ of a fully frustrated ladder should be an even number.)

154: These states can be characterized by the {\em staggered magnetization}

155: \begin{equation}

156: m \equiv {1 \over L} \sum_X (-1)^X \left(n_X - {1\over2} \right) ,

157: \end{equation}

158: where $n_X$ is the vorticity on plaquette $X$. The vorticity is

159: given by the plaquette sum $n_X = (1/2\pi) \sum_{P_X}

160: (\phi_i{-}\phi_j{-}A_{ij} )+ 1/2$, where the gauge-invariant phase

161: difference is defined modulo $2\pi$ in the range $(-\pi,\pi]$.

162: Here uniformly applied currents in the $y$ direction drive

163: vortices to the $x$ direction, generating a flow of the vortex

164: array along the ladder. Due to the spatial periodicity of the

165: vortex array, the flow in the $x$ direction results in the

166: alternation between the two ground states, which can be manifested

167: by the oscillation of the staggered magnetization. In contrast to

168: a homogeneous thin film, the array system has a lattice potential

169: which tends to suppress the flow of a vortex array. Accordingly,

170: oscillations of the staggered magnetization should appear only

171: when the external currents exceed the critical value $I_0

172: =\sqrt{5}{-}2$.\cite{I0}

173: %

174: The time evolution of the staggered magnetization at zero temperature,

175: obtained from direct integration of Eq.~(\ref{eqmotion})

176: and presented in Fig.~\ref{fig:zeroT} for various values of $I$,

177: indeed confirms the existence of the critical value:

178: For small currents ($I=0.2 < I_0$), the staggered magnetization remains constant,

179: indicating that the ladder is pinned by the lattice potential

180: to one of the two ground states.

181: Above the critical value $I_0$,

182: there arise oscillations of the staggered magnetization.

183: Note also that the oscillation frequency increases with currents,

184: reflecting that the vortex array flows faster at larger currents.

185: 

186: To probe the possibility of autonomous SR here, we consider the

187: average rate of the change in the staggered magnetization

188: \begin{equation}

189: v(t) \equiv {\delta m \over \delta t} \equiv {m(t) - m(t{-}\delta

190: t) \over \delta t} ,

191: \end{equation}

192: %which quantifies properly the transitions between the ground states.

193: which measures the transition rate between the ground states. The

194: presence of periodic oscillations between the ground states is

195: manifested by the peak in the power spectrum $S_v(\omega)$ of

196: $v(t)$ at the corresponding frequency. We then compute the SNR:

197: \begin{equation}

198: R \equiv 10 \log_{10} \left[ {S \over N} \right],

199: \end{equation}

200: where the signal $S \equiv S_v(\omega{=}\omega_p )$ is the peak

201: value of the power spectrum (at the peak frequency $\omega_p$).

202: The background noise level $N$ is estimated by the power spectrum

203: averaged over the interval of frequency much higher than

204: $\omega_p$, where the power spectrum is flat.

205: Equation~(\ref{eqmotion}) has thus been integrated directly with

206: the time steps $\Delta t = 0.05$, from which $v(t)$ has been

207: computed. We have also reduced $\Delta t$ to $0.01$, to obtain the

208: same results within the numerical accuracy, and followed an

209: annealing schedule with the equilibration time $1000$ at each

210: temperature. The data have been averaged over $200$ to $5000$

211: independent runs, depending on the system size. For convenience,

212: $\delta t$ has been set to be $0.25$.

213: %In general reduction of $\delta t$ tends to enhance the noise level $N$

214: %and to suppress the SNR.

215: We have also considered smaller values of $\delta t$, and found

216: that the overall results remain qualitatively the same although

217: higher noise levels make it more formidable to perform precise

218: numerical analysis.  The periodic boundary conditions have been

219: imposed along the $x$ direction.

220: \begin{figure}

221: \epsfig{file=fig3.eps,width=8cm} \vspace*{0.4cm} \caption{Power

222: spectrum of the transition rate $v(t)$ (in arbitrary units) in a

223: ladder of length $L=50$ for $I=0.25$ at various temperatures $T=

224: 0.01,\, 0.02,\, 0.03,\, 0.04$, $0.06,\, 0.08,\, 0.12,\, 0.16,\,

225: 0.24$, and $0.4$ from left. } \label{fig:PSabove}

226: \end{figure}

227: \begin{figure}

228: %\vspace*{1.3cm}

229: \epsfig{file=fig4.eps,width=8cm} \caption{SNR versus temperature

230: in a ladder of length $L=50$ for $I=0.25$. Here and in other

231: figures applicable, error bars have been estimated by the standard

232: deviation, and the lines connecting data points are merely guides

233: to eyes. } \label{fig:SNRabove}

234: \end{figure}

235: 

236: We first examine the effects of noise above the critical current.

237: Figure~\ref{fig:PSabove} shows the power spectrum at

238: various temperatures in the ladder of length $L=50$

239: above the critical current ($I=0.25$).

240: The power spectrum exhibits a sharp peak even at zero temperature,

241: arising from the current-induced oscillation.

242: As the temperature is raised, however, the signal $S$ decreases monotonically,

243: reflecting that the thermal noise disturbs the

244: coherent motion of the vortex array.

245: Accordingly, the SNR displays monotonic decrease

246: with the temperature, as shown in Fig.~\ref{fig:SNRabove}.

247: \begin{figure}

248: \epsfig{file=fig5.eps,width=8cm} \vspace*{0.4cm} \caption{Power

249: spectrum of $v(t)$ (in arbitrary units)

250: %as a function of $\omega$

251: in a ladder of length $L=50$ for $I=0.2$ at various temperatures

252: $T= 0.01,\, 0.02,\, 0.03,\, 0.04,\, 0.06,\, 0.08,\, 0.12,\, 0.16,\,

253: 0.28$, and $0.4$ from below.

254: }

255: \label{fig:PSbelow}

256: \end{figure}

257: 

258: We next reduce the current $I$ to $0.2$, below $I_0$. Unlike at

259: zero temperature, there emerge oscillations between the two ground

260: states at finite temperatures, as manifested by the peak of

261: $S_v(\omega)$ in Fig.~\ref{fig:PSbelow}. The signal $S$ increases

262: as the temperature is raised from zero while further increase of

263: the temperature tends to suppress the peak, eventually forcing the

264: power spectrum to be white. The SNR plotted in

265: Fig.~\ref{fig:SNR-I} for $I=0.2$ and lower values clearly displays

266: autonomous SR, i.e., SR without periodic

267: driving.\cite{Gang,CR,CR2}

268: %

269: Here it should be stressed that the SR observed above does not

270: describe the direct response to the {\em external} dc current.

271: As shown in Fig.~\ref{fig:peakfreq}, the peak frequency, which

272: is proportional to the mean flow velocity of the vortex array,

273: increases monotonically with the temperature, exhibiting no SR behavior.

274: This is consistent with the previous study

275: which concluded that the mobility of an overdamped particle

276: in a washboard potential does not display SR behavior.\cite{washboard}

277: 

278: \begin{figure}

279: %\vspace*{1.0cm}

280: \epsfig{file=fig6.eps,width=8cm}

281: \vspace*{0.4cm}

282: \caption{SNR versus temperature in a ladder of

283: length $L=50$ for various currents below $I_0$.}

284: \label{fig:SNR-I}

285: \end{figure}

286: 

287: \begin{figure}

288: %\vspace*{0.3cm}

289: \epsfig{file=fig7.eps,width=8cm}

290: \vspace*{0.4cm}

291: \caption{Peak frequency $\omega_p$

292: as a function of the temperature in a ladder of length $L=50$ for $I=0.2$.

293: }

294: \label{fig:peakfreq}

295: \end{figure}

296: 

297: Similar SR phenomena are also observed in the behavior of the

298: staggered magnetization $m(t)$ itself.  Figure \ref{fig:PSm}

299: displays the power spectrum $S_m(\omega)$ of the staggered

300: magnetization $m(t)$ both above and below the critical current.

301: Above the critical current, indeed similarly to $S_v(\omega)$

302: shown in Fig.~\ref{fig:PSabove}, $S_m(\omega)$ exhibits a sharp

303: zero-temperature peak at a finite frequency, which gets broadened

304: gradually with the introduction of thermal noise. Below the

305: critical current, on the other hand, the zero-temperature peak

306: develops at zero frequency ($\omega=0$), indicating the absence of

307: oscillation in the staggered magnetization.  As the temperature is

308: raised from zero, the dc component $S_m(\omega{=}0)$ becomes

309: smaller; instead there emerges a peak at a finite frequency

310: $\bar{\omega}_p$, which, like the peak frequency $\omega_p$ of

311: $S_v (\omega)$ (see Fig.~\ref{fig:peakfreq}), depends on the

312: temperature.  These behaviors of $S_m (\omega)$ and those of $S_v

313: (\omega)$ are in general consistent with the relation $S_v

314: (\omega) \propto \omega^2 S_m (\omega)$, accurate for small

315: $\delta t$.  Note here that such development of a finite-frequency

316: peak at finite temperatures is an indication of the coherence

317: resonance phenomena observed in the bistable regime of the

318: Fitzhugh-Nagumo model.~\cite{bistable}

319: 

320: \begin{figure}

321: \epsfig{file=fig8a.eps,width=8cm} \centerline{(a)}

322: \epsfig{file=fig8b.eps,width=8cm} \centerline{(b)} \caption{ Power

323: spectrum of the staggered magnetization $m(t)$ (in arbitrary

324: units) at various temperatures $T= 0.01,\, 0.02,\, 0.03,\, 0.04$,

325: $0.06,\, 0.08,\, 0.12,\, 0.16,\, 0.24$, and $0.4$ (from left) in a

326: ladder of length $L=50$ for (a) $I=0.25$ and (b) $I=0.2$. }

327: \label{fig:PSm}

328: \end{figure}

329: 

330: To quantify conveniently the expected SR behavior of $m(t)$, we

331: compute the coherence measure\cite{Gang,beta}

332: \begin{equation}

333: \beta \equiv S_m(\bar{\omega}_p) \left({\Delta\omega \over

334: \bar{\omega}_p} \right)^{-1},

335: \end{equation}

336: where the half-width of the peak is given by $\Delta\omega \equiv

337: \omega_r - \bar{\omega}_p$ with $\omega_r \,(> \bar{\omega}_p)$

338: satisfying $2 S_m(\omega_r) = S_m(\bar{\omega}_p)$.  The coherence

339: measure $\beta$, which is plotted as a function of temperature in

340: Fig.~\ref{fig:beta}, also reveals the nature of the autonomous SR

341: behavior in the system.  Above the critical current it decreases

342: exponentially with the temperature [see Fig.~\ref{fig:beta}(a) for

343: $I=0.25$] while below the critical current a prominent peak

344: appears at a finite temperature, as demonstrated in

345: Fig.~\ref{fig:beta}(b) for $I=0.2$.  Thus confirmed is the

346: autonomous SR behavior of the staggered magnetization $m(t)$.

347: 

348: \begin{figure}

349: \epsfig{file=fig9a.eps,width=8cm} \centerline{(a)}

350: \epsfig{file=fig9b.eps,width=8cm} \centerline{(b)}

351: \caption{Coherence measure versus temperature in a ladder of

352: length $L=50$ for (a) $I=0.25$ and (b) $I=0.2$. } \label{fig:beta}

353: \end{figure}

354: The autonomous SR in the fully frustrated ladder is rather peculiar,

355: arising from the combination of the external uniform currents and

356: the spatial periodicity of the ground state inherent in the system.

357: %

358: %The SR behavior in the system can be understood by the

359: %following argument:

360: Although currents smaller than $I_0$ do not induce the vortex

361: array to move, they add an overall gradient to the lattice

362: potential along the $x$ direction. At finite temperatures, the

363: additional noise currents assist vortices to hop to neighboring

364: plaquettes while the potential gradient generated by the external

365: currents makes hopping in one direction more favorable than that

366: in the other. Consequently, the vortex array is encouraged to flow

367: to one direction, leading to oscillations of the transition rate

368: $v(t)$.  These oscillations grow with the temperature, which gives

369: the enhancement of the signal. At higher temperatures, however,

370: the lattice potential gradient can be neglected, and the random

371: nature of the noise disturbs the coherent flow, weakening the

372: signal. In this manner the two different roles of the noise

373: produces the autonomous SR in the system,\cite{comm} which also

374: provides a natural explanation of the current-dependence of the

375: SNR behavior in Fig.~\ref{fig:SNR-I}: At small currents the

376: lattice potential barrier between neighboring sites remain rather

377: high, and the noise required to overcome the barrier should be

378: strong. Accordingly, as the applied currents are reduced, the SR

379: temperature is expected to increase while the SNR peak diminishes

380: due to the large incoherent contributions of the noise.

381: \begin{figure}

382: %\vspace*{1.0cm}

383: \epsfig{file=fig10.eps,width=8cm}

384: \vspace*{0.4cm}

385: \caption{SNR versus temperature for various sizes and $I=0.2$.}

386: \label{fig:SNR-L}

387: \end{figure}

388: 

389: \begin{figure}

390: %\vspace*{1.0cm}

391: \epsfig{file=fig11.eps,width=8cm}

392: \caption{SNR for the staggered voltage as a function of the temperature

393: in a ladder of length $L=50$ for $I=0.2$.}

394: \label{fig:SNR-Vs}

395: \end{figure}

396: Figure~\ref{fig:SNR-L}, showing the SNR versus temperature for various lengths,

397: discloses another interesting feature of the SR behavior in the ladder system.

398: In the ladder of length $L=2$, which is the minimum length required

399: for a fully frustrated ladder, the SNR exhibits no sign of SR.

400: As the length is increased, there first appears a broad peak in the SNR, which

401: then develops into a prominent peak.

402: Such {\em array-enhanced} SR phenomena

403: reflect the stronger coherence between vortices in longer ladders.

404: However, the enhancement should not persist with the length

405: since the ladder does not evolve long-range order in the thermodynamic limit.

406: Indeed the SNR in Fig.~\ref{fig:SNR-L} eventually saturates for $L \gtrsim 30$,

407: appearing independent of the length.

408: While array-enhanced resonance was first reported in the array

409: systems exhibiting conventional SR,\cite{array}

410: recent studies have also revealed the possibility of array-enhanced

411: coherence resonance in coupled excitable systems.\cite{array2}

412: It is remarkable that similar array-enhancement of autonomous SR can also be

413: observed in the Josephson-junction ladder, which possesses two stable states

414: like the conventional SR system (but in the absence of periodic driving).

415: %

416: %This array-enhanced feature is manifested more clearly in

417: %Fig.~\ref{fig:Smax-L}, where the maximum of $S{-}N$ is plotted as a

418: %function of system size.

419: %Up to $L=20$ it increases with the system size, indicating the

420: %enhancement of the coherence between the vortices, which is the main

421: %origin of the array-enhanced SR.

422: %However, due to the absence of long-range order in the thermodynamic

423: %limit of the ladder, it indeed decreases of the form $\sim 1 / L$ for

424: %$L \gtrsim 30$.

425: %Since the noise intensity also decreases with the same form, the SNR

426: %saturates for large arrays instead of increasing indefinitely.

427: 

428: Finally, we discuss how to observe the autonomous SR in experiment.

429: As a candidate for the measurable quantity characterizing the

430: autonomous SR, we suggest the average staggered voltage:

431: \begin{equation}

432: V_s (t) \equiv {1 \over L} \sum_x (-1)^x \bar{V}_x(t),

433: \end{equation}

434: where $\bar{V}_x(t)$ is the rung voltage at position $x$ averaged

435: over the interval $\delta t$ around time $t$. The rigid motion of

436: the vortex array at low temperatures should give rise to periodic

437: oscillations of $V_s(t)$, similarly to the transition rate $v(t)$.

438: The SNR for the staggered voltage $V_s(t)$, denoted by $R_s$, is

439: plotted in Fig.~\ref{fig:SNR-Vs}, where SR behavior similar to

440: that of $v(t)$ is shown. We thus suggest that the autonomous SR

441: phenomena in the ladder should be observed through the measurement

442: of the staggered voltage.

443: 

444: In summary we have investigated fully frustrated Josephson-junction

445: ladders, driven by uniform constant currents,

446: with regard to the possibility of stochastic resonance.

447: While the system under large currents displays oscillations between the

448: two ground states, the lattice potential suppresses

449: such oscillations for small currents.

450: Still the addition of noise currents, relevant at finite temperatures,

451: induces again oscillations, giving rise to stochastic resonance behavior.

452: In particular the signal-to-noise ratio has been found to display

453: array-enhanced stochastic resonance phenomena.

454: It has also been suggested that such behavior may be observed experimentally

455: through the measurement of the staggered voltage.

456: In view of this, the fully frustrated Josephson-junction ladder

457: makes a good physical realization of the system with many degrees of freedom,

458: adequate for the study of the autonomous

459: as well as the conventional stochastic resonance phenomena.

460: 

461: We thank P.\ Minnhagen and B.J.\ Kim for useful discussions.

462: This work was supported by the Korea Science and Engineering

463: Foundation through the Center for Strongly Correlated Materials

464: Research (GSJ) and by the Ministry of Education through the BK21 Project (MYC).

465: %Grant 2000-015-DP0138 (MYC).

466: 

467: \begin{thebibliography}{99}

468: 

469: \bibitem{review} For a review, see L.\ Gammaitoni, P.\ H\"anggi,

470: P.\ Jung, and F.\ Marchesoni, Rev.\ Mod.\ Phys.\ {\bf 70}, 223 (1998).

471: %\bibitem{ice} R.\ Benzi, S.\ Sutera, and A.\ Vulpiani, J.\ Phys.\ A

472: %{\bf 14}, L453 (1981).

473: %

474: %\bibitem{example}

475: %For example,

476: %S.\ Fauve and F.\ Heslot, Phys.\ Lett.\ A {\bf 97}, 5 (1983);

477: %B.\ McNamara, K.\ Wiesenfeld, and R.\ Roy, Phys.\ Rev.\ Lett.\ {\bf

478: %60}, 2626 (1988);

479: %J.K.\ Douglass, L.\ Wilkens, E.\ Pantazelou, and F.\ Moss, Nature

480: %(London) {\bf 365}, 337 (1993).

481: 

482: \bibitem{Gang} G.\ Hu, T.\ Ditzinger, C.Z.\ Ning, and H.\ Haken, Phys.\

483: Rev.\ Lett.\ {\bf 71}, 807 (1993).

484: \bibitem{CR} A.\ Longtin, Phys.\ Rev.\ E {\bf 55}, 868 (1997);

485: A.\ Neiman, P.I.\ Saparin, and L.\ Stone, {\it ibid}. {\bf 56}, 270 (1997);

486: S.-G.\ Lee, A.\ Neiman, and S.\ Kim, {\it ibid}. {\bf 57}, 3292 (1998).

487: 

488: \bibitem{CR2} A.S.\ Pikovsky and J.\ Kurths, Phys.\ Rev.\ Lett.\ {\bf 78},

489: 775 (1997).

490: 

491: \bibitem{delay} T.\ Ohira and Y.\ Sato, Phys.\ Rev.\ Lett.\ {\bf 82},

492: 2811 (1999).

493: 

494: \bibitem{inertia} H.\ Hong and M.Y.\ Choi, Phys.\ Rev.\ E {\bf 62},

495: 6462 (2000). See also M.I.\ Dykman, R.\ Mannella, P.V.E.\ McClintock,

496: F.\ Moss, and S.M.\ Soskin, Phys.\ Rev.\ A {\bf 37}, 1303 (1988).

497: 

498: \bibitem{acSR} B.J.\ Kim, M.-S.\ Choi, P.\ Minnhagen, G.S.\ Jeon,

499: H.J.\ Kim, and M.Y.\ Choi, Phys.\ Rev.\ B {\bf 63}, 104506 (2001).

500: %

501: %\bibitem{chemical} Z.\ Hou and H. Xin, Phys.\ Rev.\ E {\bf 60}, 6329

502: %(1999).

503: %\bibitem{exp} D.E.\ Postnov, S.K.\ Han, T.G.\ Yim, and O.V.\

504: %Sosnovtseva, Phys.\ Rev.\ E {\bf 59}, 3791 (1999).

505: 

506: \bibitem{review2} For a review, see, e.g.,

507: {\it Macroscopic Quantum Phenomena and Coherence in Superconducting Networks},

508: edited by C. Giovannella and M. Tinkham (World Scientific, Singapore, 1995);

509: Physica B {\bf 222}, 253 (1996).

510: 

511: \bibitem{Delsing} H.\ Shimada and P.\ Delsing, Phys.\ Rev.\ Lett.\

512: {\bf 85}, 3253 (2000).

513: 

514: \bibitem{bistable} B.\ Lindner and L. Schimansky-Geier,

515: Phys.\ Rev.\ E {\bf 61}, 6103 (2000).

516: 

517: \bibitem{I0} B.J.\ Kim, S.\ Kim, and S.J.\ Lee, Phys.\ Rev.\ B {\bf

518: 51}, 8462 (1995).

519: 

520: \bibitem{washboard} F.\ Marchesoni, Phys.\ Lett.\ A {\bf 231}, 61 (1997).

521: 

522: \bibitem{beta} J.R.\ Pradines, G.V.\ Osipov, and J.J. Collins,

523: Phys.\ Rev.\ E {\bf 60}, 6407 (1999).

524: 

525: \bibitem{comm} Similar argument was also given as an

526: explanation of the autonomous SR in the limit-cycle system (see

527: Ref.~\onlinecite{Gang}).

528: 

529: \bibitem{array} J.F.\ Lindner, B.K.\ Meadows, W.L.\ Ditto, M.E.\

530: Inchiosa, and A.R.\ Bulsara, Phys.\ Rev.\ Lett.\ {\bf 75}, 3 (1995);

531: M.\ L\"ocher, G.A.\ Johnson, and E.R.\ Hunt, {\it ibid}.\ {\bf 77}, 4698 (1996).

532: Similar saturation behavior of the SNR in an array exhibiting conventional SR,

533: i.e., in a linear chain of coupled bistable elements under periodic driving,

534: has been explained in terms of soliton contributions.

535: See F. Marchesoni, L. Gammaitoni, and A.R. Bulsara, Phys. Rev. Lett. {\bf 76}, 2609 (1996).

536: 

537: \bibitem{array2} S.K. Han, T.G. Yim, D.E. Postnov, and D.V. Sosnovtseva,

538: Phys. Rev. Lett. {\bf 83}, 1771 (1999);

539: B.\ Hu and C.\ Zhou, Phys.\ Rev.\ E {\bf 61}, R1001 (2000);

540: Y.\ Wang, D.T.W.\ Chik, and Z.D.\ Wang, {\it ibid}. {\bf 61}, 740 (2000);

541: C.\ Zhou, J.\ Kurths, and B.\ Hu, Phys.\ Rev.\ Lett.\ {\bf 87}, 098101 (2001).

542: 

543: \end{thebibliography}

544: 

545: %\begin{figure}

546: %\vspace*{1.3cm}

547: %\epsfig{file=fig5_old.eps,width=8cm}

548: %\caption{Behavior of SNR as a function of temperature in a ladder of

549: %length $L=50$ for $I=0.2$.}

550: %\label{fig:SNRbelow}

551: %\end{figure}

552: %

553: %\begin{figure}

554: %%\vspace*{1.0cm}

555: %\epsfig{file=fig9.eps,width=8cm}

556: %\caption{Maximum of $S{-}N$ as a function of size for $I=0.2$.}

557: %\label{fig:Smax-L}

558: %\end{figure}

559: 

560: \end{document}

561: