1: \documentstyle[preprint,epsfig,aps]{revtex}
2: %\documentstyle[prl,epsfig,aps,multicol]{revtex}
3: \begin{document}
4: \preprint{SNUTP 99-028}
5: \title{Boundary Effects on Dynamic Behavior of
6: Josephson-Junction Arrays}
7: \author{M.Y. Choi, Gun Sang Jeon, and Mina Yoon~\cite{MY}}
8: \address{Department of Physics and Center for Theoretical Physics\\
9: Seoul National University\\
10: Seoul 151-742, Korea}
11: \maketitle
12: \draft
13:
14: \begin{abstract}
15: The boundary effects on the current-voltage characteristics
16: in two-dimensional arrays of resistively shunted Josephson junctions
17: are examined.
18: In particular, we consider both the conventional boundary
19: conditions (CBC) and the fluctuating twist boundary conditions (FTBC),
20: and make comparison of the obtained results.
21: It is observed that the CBC, which have been widely adopted in
22: existing simulations,
23: may give a problem in scaling, arising from rather large boundary effects;
24: the FTBC in general turn out to be effective in reducing the finite-size
25: effects, yielding results with good scaling behavior.
26: To resolve the discrepancy between the two boundary conditions,
27: we propose that the proper scaling in the CBC should be performed
28: with the boundary data discarded: This is shown
29: to give results which indeed scale well and are the same as those
30: from the FTBC.
31:
32: \end{abstract}
33: \bigskip
34: \thispagestyle{empty}
35:
36: %{\tiny \noindent
37: %74.50.+r Proximity effects, weak links, tunneling phenomena, and Josephson effects\\[-\baselineskip]
38: %74.25.Fy Transport properties (electric and thermal conductivity, thermoelectric effects, etc.)\\[-\baselineskip]
39: %74.40.+k Fluctuations (noise, chaos, nonequilibrium superconductivity, localization, etc.)\\[-\baselineskip]
40: %64.60.Ht Dynamic critical phenomena }
41: \pacs{PACS number: 74.50.+r, 74.25.Fy, 74.40.+k, 64.60.Ht}
42:
43: %\begin{multicols}{2}
44: %\narrowtext
45:
46: %\section{Introduction}
47: During the past decades superconducting arrays have been extensively
48: studied both theoretically and experimentally~\cite{proceeding}.
49: The superconducting arrays in equilibrium are well described by
50: the $XY$ model, which exhibits the
51: Berezinskii-Kosterlitz-Thouless (BKT) transition
52: at a nonzero temperature in two dimensions~\cite{BKT}.
53: Theoretically this transition is manifested by the characteristic behavior of static
54: quantities such as the universal jump of the helicity modulus.
55: On the other hand, the standard experimental probe of the transition
56: is to measure the dynamical response to the external current
57: driving~\cite{Exp}.
58: The need for the analysis of the experimental data
59: together with the interest in the dynamic properties themselves
60: has motivated extensive theoretical examinations of the dynamics
61: of the system, where numerical simulations
62: have been of great help.
63:
64: It has been recognized that the
65: boundary conditions employed in those simulations, performed in finite systems,
66: may affect the results in a crucial way, which makes it necessary to
67: examine their effects.
68: Most existing dynamic simulations of the system in the absence of driving currents have been performed
69: with periodic boundary conditions along both directions;
70: under driving currents, the boundary conditions along the current-injected direction are changed
71: to be free
72: whereas periodic boundary conditions are retained along the perpendicular direction.
73: Such conventional boundary conditions (CBC), however, are known to produce rather large
74: edge effects, and the system is expected to approach the
75: thermodynamic limit rather slowly,
76: compared with that in the busbar boundary conditions~\cite{Simkin}.
77: Of much interest here are the recent results on the dynamical exponent.
78: Under the periodic boundary conditions, remarkably different values of the dynamical exponent
79: have been obtained in the resistively shunted Josephson junction (RSJ) dynamics and
80: in the time-dependent Ginzburg-Landau (TDGL) dynamics~\cite{Jose,Kim}, suggesting
81: that the two kinds of dynamics may be inherently different.
82: In particular the dynamical exponent in the RSJ dynamics turns out to be
83: much smaller than expected and apparently inapplicable to the flux-noise
84: experiment~\cite{Shaw}.
85: On the other hand, no difference has been observed
86: when the fluctuating twist boundary conditions (FTBC),
87: forcing the vortex interaction to be periodic~\cite{Ol}, is employed~\cite{Kim};
88: this implies that the difference in the dynamical exponent stems from the
89: employed boundary conditions rather than from the intrinsic property of
90: the dynamics.
91:
92: This paper is to clarify the boundary effects and thus to resolve the discrepancy
93: between the two types of the boundary conditions.
94: For this purpose, we investigate in detail the current-voltage $(IV)$
95: characteristics and the corresponding dynamic critical behaviors
96: of the system in the CBC and in the FTBC.
97: The finite-size-scaling analysis is performed to reveal that the
98: two types of the boundary conditions yield different values of the dynamical exponent.
99: In particular, the data obtained in the CBC are found
100: not to display proper scaling behavior.
101: We show that the discrepancy has its origin in the rather large
102: boundary effects in the CBC
103: and give a simple proposal that the proper scaling analysis
104: in the CBC should be performed with the boundary data discarded.
105: It turns out that the resulting $IV$ characteristics indeed exhibit
106: good scaling behavior and are consistent with those from the FTBC,
107: confirming that the poor scaling behavior originates from the peculiar
108: boundary effects in the CBC and that correct description of the
109: dynamic behavior can be achieved by appropriate extraction of the
110: bulk properties from the data.
111:
112: %This paper is organized as follows:
113: %In Sec.~\ref{set:IV}, we compute the $IV$ characteristics
114: %of the superconducting arrays under the two boundary conditions,
115: %CBC and FTBC, and perform the finite-size scaling analysis
116: %of the $IV$ data.
117: %It is demonstrated that the data from the CBC do not collapse within
118: %the one-parameter scaling curve while those from the FTBC display
119: %nice scaling behavior.
120: %In Sec.~\ref{set:BE} the failure of the scaling is attributed to
121: %the large boundary effects produced in the CBC. Indeed good
122: %scaling behavior is shown to be restored when the boundary data are discarded
123: %in the analysis, thus confirming the crucial role of boundary conditions
124: %in the finite-size scaling analysis.
125: %Finally, Sec.~\ref{set:CON} summarizes the main results of this
126: %paper.
127: %==================================================
128:
129: %\section{Finite-Size Scaling Analysis of Current-Voltage Characteristics}
130: %\label{set:IV}
131: %==================================================
132: %In this section, the $IV$ characteristics of an
133: %$N{\times}N$ RSJ array are computed.
134: %We consider two different boundary conditions:
135: %the CBC, consisting of free boundary conditions
136: %along the current-injected direction and
137: %periodic boundary conditions along the perpendicular direction,
138: %and the FTBC, where new variables $\Delta_x$ and $\Delta_y$
139: %are introduced with the periodic
140: %boundary conditions imposed in both directions.
141:
142: We begin with the set of equations of motion describing
143: the dynamics of an RSJ array with
144: critical current $I_c$ and shunt resistance $R_s$:
145: \begin{equation}
146: \label{eq:dyn1}
147: {\sum_j}' \left[ {\hbar \over 2eR_s} {d \phi_{ij} \over dt}
148: + I_c \sin\phi_{ij} + L_{ij}\right]
149: = I_i^{\rm ext} .
150: \end{equation}
151: Here $\phi_{ij}\equiv \phi_i - \phi_j$ is the phase difference
152: across a junction with
153: $\phi_i$ being the phase of the superconducting order parameter
154: on grain $i$ (at position ${\bf r}_i$),
155: $I_i^{\rm ext}$ is the external current fed into grain $i$,
156: and the primed summation runs over the nearest neighbors of grain $i$.
157: The thermal noise current $L_{ij}$ is assumed to be a white noise
158: satisfying
159: \begin{equation}
160: \langle L_{ij}(t+\tau) L_{kl} (t) \rangle
161: = {2 k_B T \over R_s} \delta (\tau) (\delta_{ik}\delta_{jl} -
162: \delta_{il}\delta_{jk})
163: \end{equation}
164: with the angular bracket denoting an ensemble average and $T$
165: being the temperature.
166: In the CBC, uniform current $I$ is fed into each grain at
167: one edge ($x=0$) and extracted from each grain at the other edge ($x=N$),
168: leading to $I_i^{\rm ext} = I (\delta_{x_i0} - \delta_{x_iN})$~\cite{Choi}.
169: The boundary conditions for the phase variables are taken to be
170: free in the $x$ direction and periodic in the $y$ direction.
171: The average voltage across each junction is given by the relation
172: \begin{equation} \label{eq:V1}
173: V = {\hbar \over 2eN^2} \sum_{y=1}^N \left(
174: \left\langle {d \phi_{(0,y)} \over dt} \right\rangle_t -
175: \left\langle {d\phi_{(N,y)} \over dt} \right\rangle_t \right),
176: \end{equation}
177: where $\langle\ldots\rangle_t$ stands for the time average.
178: %
179: In the FTBC, on the other hand,
180: periodic boundary conditions are imposed on $\{\phi_i\}$
181: in both directions, yielding
182: the equations of motion in the form~\cite{Kim}
183: \begin{equation} \label{eq:dyn2}
184: {\sum_j}' \Big[ {\hbar \over 2eR_s}
185: {d \phi_{ij} \over dt}
186: + I_c \sin (\phi_{ij} - {\bf r}_{ij} \cdot {\bf \Delta})
187: + L_{ij}\Big]=0
188: \end{equation}
189: with ${\bf r}_{ij} \equiv {\bf r}_i - {\bf r}_j$.
190: The dynamics of the twist variables
191: ${\bf \Delta} \equiv (\Delta_x,\Delta_y)$
192: is governed by
193: \begin{eqnarray}
194: %\label{eq:dyn3}
195: \frac{\hbar}{2eR_s}\frac{d\Delta_{x}}{dt} &=& \frac{I_c}{N^2}
196: \sum_{\langle ij \rangle_{x}}
197: \sin(\phi_{ij} -\Delta_{x}) - {I \over N} + L_{\Delta_{x}} , \nonumber \\
198: \label{eq:dyn4}
199: \frac{\hbar}{2eR_s}\frac{d\Delta_{y}}{dt} &=& \frac{I_c}{N^2}
200: \sum_{\langle ij \rangle_{y}}
201: \sin(\phi_{ij} -\Delta_{y}) + L_{\Delta_{y}} ,
202: \end{eqnarray}
203: where $\sum_{\langle ij \rangle_{x(y)}}$ denotes the summation over
204: all nearest-neighboring pairs in the $x(y)$ direction and
205: the thermal noise current $L_{\Delta_{x(y)}}$ satisfies
206: \begin{equation}
207: \langle L_{\Delta_{x(y)}}(t+\tau) L_{\Delta_{x(y)}} (t) \rangle =
208: {2 k_B T \over N^2 R_s} \delta (\tau).
209: \end{equation}
210: The average voltage can then be obtained from the time evolution of the twist
211: variable $\Delta_x$:
212: \begin{equation}
213: V = - {\hbar \over 2e}
214: \left\langle {d \Delta_x \over dt} \right\rangle_t.
215: \end{equation}
216:
217: To obtain the $IV$ characteristics in the two boundary conditions, we
218: integrate directly the coupled equations of motion given by Eq.~(\ref{eq:dyn1})
219: for the CBC and Eqs.~(\ref{eq:dyn2}) and (\ref{eq:dyn4}) for the FTBC, with the time
220: step $\Delta t = 0.05$ (in units of $\hbar /2 e R_s I_c$).
221: The data are averaged over 30 independent runs while
222: in each run the averaging time is chosen in such a way that at least 25
223: vortices are created.
224: We
225: %consider several system sizes ranging from $N=4$ to $N=64$
226: %to clarify size effects, and
227: concentrate on the temperature $T=0.84$ (in units of
228: $\hbar I_c / 2e$), which is below the BKT transition temperature $T_c
229: \approx 0.89$~\cite{Ol2}, and
230: henceforth write the current and the voltage in units of
231: $I_c$ and $I_c R_s$, respectively.
232: %
233:
234: Figure~\ref{fig:IV} presents the resulting $IV$ characteristics
235: in the CBC and in the FTBC.
236: We have considered several system sizes ranging from $N=4$ to $N=64$,
237: to clarify size effects.
238: At high currents the average voltage drop
239: is found to be almost independent of the size in both boundary conditions.
240: As the current is decreased, however, the system undergoes a crossover from the
241: nonlinear-resistance regime to the linear-resistance regime, exhibiting
242: size-dependence.
243: The crossover current apparently decreases with the increase of the size, suggesting that
244: the linear resistance at low currents reflects the finite-size
245: effects~\cite{Simkin,Wallin} rather than the lattice effects~\cite{Choi}.
246: Here one can observe two striking differences between the
247: $IV$ characteristics in the CBC and those in the FTBC:
248: First, at low currents the voltage drop in
249: the CBC is much higher than that in the FTBC.
250: Furthermore, the ratio of the former to the latter grows larger as
251: the system size is increased.
252: Secondly, the decrease of the crossover current with the system size is
253: faster in the FTBC than in the CBC; this implies that the FTBC is more effective in
254: reducing the finite-size effects.
255: %
256: %The resulting $IV$ characteristics are not shown here because
257: %similar figures can be found in the existing studies~\cite{Simkin,Kim,Choi}.
258:
259: %==================================================
260: %\section{Finite-Size scaling}
261: %\label{set:FSS}
262: We now analyze the $IV$ data by means of the finite-size scaling
263: method, and investigate the dynamic critical behavior both in the CBC and
264: in the FTBC.
265: We assume the dynamic scaling form~\cite{Kim,Wallin}
266: \begin{equation}
267: N R^{1/z} = f(N I),
268: \label{eq:FSS}
269: \end{equation}
270: where $R$ is the resistance of the system and $z$ is the dynamical exponent.
271: The dimensionless scaling function $f(x)$ is expected to possess the asymptotic behavior:
272: $f(x) \propto {\rm constant}$ for small $x$ and $f(x) \propto x$ for large $x$.
273: %\begin{equation} \label{eq:asym}
274: %f(x) \propto \left\{
275: %\begin{array} {cl}
276: %\hbox{const.}& \hbox{ for small $x$},\\
277: %x & \hbox{ for large $x$}.
278: %\end{array}
279: %\right.
280: %\end{equation}
281: The small-$x$ behavior follows from the behavior of the linear resistance for $I \rightarrow 0$:
282: \begin{equation}
283: R_L \sim N^{-z}
284: \label{RL}
285: \end{equation}
286: whereas the finiteness of the resistance in the limit $N\rightarrow \infty$
287: elucidates the behavior of $f(x)$ for large $x$.
288:
289: We first fit the linear resistance $R_L$ in the CBC to the form
290: in Eq.~(\ref{RL}), and display in the inset of Fig.~\ref{fig:FSS}(a)
291: the log-log plot of the linear resistance versus the system size,
292: which shows that the linear resistance is well described by Eq.~(\ref{RL})
293: with the dynamical exponent $z=0.91 \pm 0.07$.
294: Note, however, that this is presumably deceptive since
295: the value of $z$ less than two is likely to be erroneous,
296: especially, below the transition temperature:
297: In a two-dimensional Josephson-junction array, the value is
298: generally believed to be $z=2$ at the transition temperature
299: and to grow larger as the temperature is lowered~\cite{Wallin}.
300: Indeed the scaling plot with the value $z=0.91$, which is presented
301: in Fig.~\ref{fig:FSS}(a), manifests that the system in the CBC does not
302: follow the one-parameter dynamic scaling given by Eq.~(\ref{eq:FSS}).
303:
304: We next examine the $IV$ data in the FTBC,
305: and fit the corresponding linear resistance $R_L$.
306: This yields the dynamical exponent $z=2.73 \pm 0.02$,
307: as shown in the inset of Fig.~\ref{fig:FSS}(b),
308: which is quite different from the value in the CBC
309: and indeed consistent with the results of previous studies~\cite{Kim,Wallin}.
310: In particular, unlike the result of the CBC,
311: the scaling plot shown in Fig.~\ref{fig:FSS}(b) reveals that
312: the scaling form in Eq.~(\ref{eq:FSS}) is observed well
313: in the FTBC. Namely, the $IV$ data in the
314: nonlinear-resistance regime as well as in the linear-resistance regime
315: collapse nicely into a single curve and the scaling function follows the expected
316: asymptotic behaviors.
317:
318: %In the next section we will investigate in detail
319: %the origin of the discrepancy and propose its possible resolution.
320:
321: %==================================================
322: %\section{Boundary Effects}
323: %\label{set:BE}
324: %==================================================
325: Both the values of the dynamical exponent and the resulting scaling plots
326: indicate that the CBC are not appropriate for describing
327: the dynamical behavior of the system whereas the FTBC give a correct
328: description.
329: This discrepancy is likely to have its origin in the
330: effects of the boundary, which is formed for the purpose of the
331: current injection in the CBC but is absent in the FTBC.
332: To find a clue to the boundary effects,
333: we examine the distribution of the voltage drop along the $x$
334: direction, which is, according to
335: Eq.~(\ref{eq:V1}), proportional to the gradient of the phase profile.
336: The phase profile
337: \begin{equation} \label{eq:profile}
338: \Delta \phi(x) \equiv
339: {1 \over N} \sum_y [\phi_{(x,y)} (\tau)-\phi_{(x,y)}(0)],
340: \end{equation}
341: where $\tau$ is the average time,
342: has been obtained for several system sizes and currents,
343: %is displayed in Fig.~\ref{fig:profile} for the system of size $N=32$ at current $I=0.04$.
344: %Figure~\ref{fig:profile}
345: which turns out to be essentially the same as the profile
346: obtained in Ref.~\cite{Simkin}.
347: It demonstrates that the voltage drop across a few
348: columns near the boundary is comparable to, or even larger than, that
349: across the rest of the system.
350: Such tendency becomes prominent at lower currents,
351: and leads to an overestimation of the voltage drop in the system,
352: especially at low currents; this may
353: have significant effects on the dynamic behavior of a system
354: of finite size~\cite{Simkin}.
355:
356: To make clear that the boundary effects are the main origin
357: of the failure of the dynamic finite-size scaling in the CBC,
358: we consider the $(N{+}10)\times N$ array, consisting of
359: $N{+}10$ columns and $N$ rows under the CBC,
360: and calculate the voltage drop between the grains at
361: $x=5$ and $x=N{+}5$.
362: Namely, we propose that the proper scaling in the CBC should be performed
363: with the boundary data discarded,
364: which helps to estimate correctly the voltage drop in the bulk~\cite{comment}.
365: Here the number five has been chosen since the phase profile reveals
366: that the boundary effects are severe, i.e., the anomalous voltage drop
367: is substantial, across the five columns at each edge,
368: apparently regardless of the system size.
369: The $IV$ characteristics for several sizes, obtained in this scheme,
370: are plotted in Fig.~\ref{fig:IVMCBC},
371: which shows marked difference from Fig.~\ref{fig:IV}(a):
372: %that computed in the original CBC.
373: The voltage drop is significantly reduced at small currents,
374: resulting in the $IV$ characteristics
375: which look almost the same as those obtained from the FTBC.
376:
377: Indeed the finite-size scaling analysis, shown in the inset of Fig.~\ref{fig:FSSMCBC},
378: leads to the dynamical exponent $z=2.84 \pm 0.06$, which is consistent with
379: that obtained from the FTBC.
380: Further, with this value of the dynamical exponent,
381: the $IV$ data in Fig.~\ref{fig:IVMCBC}
382: collapse remarkably well into a single curve,
383: as illustrated in Fig.~\ref{fig:FSSMCBC}.
384: >From these results we thus reach the conclusion that the
385: CBC give a correct
386: description of the dynamical behavior of the Josephson-junction
387: array only when the rather large voltage drop
388: near the boundary is discarded in the analysis.
389: Such boundary effects should be irrelevant in the thermodynamic limit, but
390: can affect significantly the properties of a finite system,
391: generating substantial errors in the scaling analysis.
392: %
393: %It is of interest to note that under periodic boundary conditions
394: %the values of $z$ far smaller than expected were also
395: %obtained from the scaling analysis of the relaxation time
396: %in the RSJ model both above the BKT transition temperature $T_c$~\cite{Jose}
397: %and below $T_c$~\cite{Kim}.
398: %In contrast the same analysis of the TDGL model was found to
399: %give $z \approx 2$ at the same temperature~\cite{Jose};
400: %this led to the conclusion that the TDGL model describes the flux-noise
401: %experiment~\cite{Shaw} better than the RSJ model.
402: %In a subsequent study of the low-temperature regime ($T\lesssim T_c$)
403: %via the FTBC, however,
404: %both models were shown to possess the same value of
405: %$z\, (\gtrsim 2)$~\cite{Kim}.
406: %Accordingly, the erroneous value of $z$ obtained in the RSJ model
407: %was proposed to be an artifact of the periodic boundary conditions
408: %rather than the inherent property of the RSJ model.
409: Taking into account that the periodic boundary conditions are naturally
410: generalized to the CBC in the presence of driving currents, we can infer
411: that the erroneous value of $z$ obtained
412: in Ref.~\cite{Jose} also has its origin in the
413: boundary conditions rather than in the RSJ dynamics itself.
414:
415: %==================================================
416: %\section{Summary}
417: %\label{set:CON}
418: %==================================================
419: In summary, we have studied the boundary effects on the
420: dynamical behavior of a two-dimensional RSJ array, both
421: in the CBC and in the FTBC.
422: %In particular, we have investigated in detail the current-voltage
423: %characteristics and the corresponding dynamic critical behaviors
424: %of the system in the conventional boundary conditions and
425: %in the fluctuating twist boundary conditions.
426: The finite-size-scaling analysis has been performed to reveal
427: %that the
428: %two types of the boundary conditions yield different
429: not only the discrepancy in the value of the dynamical exponent
430: but also the lack of scaling in the CBC,
431: %obtained in the the conventional boundary conditions
432: %are found not to exhibit proper scaling behavior
433: %while the fluctuating twist boundary conditions
434: %yielded results displaying good scaling behavior.
435: %It has been demonstrated that the discrepancy arises
436: steming from the rather large boundary effects.
437: %in the conventional boundary conditions.
438: We have thus given a simple proposal that in the CBC
439: the proper scaling should be performed with the boundary data discarded,
440: and shown that such removal of the boundary effects indeed restores
441: the correct scaling behavior.
442: %in the CBC~\cite{comment}
443: This confirms the crucial role of boundary conditions
444: in the finite-size scaling analysis and
445: resolves the discrepancy between the two types of the boundary conditions.
446:
447: %\acknowledgements
448: We thank B.J. Kim for useful discussions and acknowledge the partial support
449: from the Seoul National University Research Fund and from the
450: Ministry of Education through the BK21 Program.
451: G.S.J. was also supported by the Korea Science and Engineering
452: Foundation postdoctoral fellowship.
453:
454: %===========================================================================
455: \begin{references}
456: %===========================================================================
457: %\bibitem{intro}
458: \bibitem[*]{MY} Present address: Department of Physics and Astronomy,
459: Michigan State University, East Lansing, MI 48824.
460:
461: \bibitem{proceeding} For a list of references, see, e.g., {\it
462: Proceedings of the 2nd CTP workshop on Statistical Physics: KT
463: Transition and Superconducting Arrays}, edited by D.\ Kim, J.S.\ Chung,
464: and M.Y.\ Choi (Min-Eum Sa, Seoul, 1993); Physica B {\bf 222} 253--406
465: (1996).
466:
467: \bibitem{BKT} V.L.\ Berezinskii, Zh.\ Eksp.\ Teor.\ Fiz.\ {\bf 59},
468: 907 (1970) [Sov.\ Phys.\ JETP {\bf 32}, 493 (1971)];
469: J.M.\ Kosterlitz and D.J.\ Thouless, J.\ Phys.\ C {\bf 6}, 1181 (1973);
470: J.M.\ Kosterlitz, {\it ibid.} {\bf 7}, 1046 (1974).
471:
472: \bibitem{Exp}
473: D.J.\ Resnick {\it et al.}, Phys.\ Rev.\ Lett.\ {\bf 47}, 1542 (1981);
474: D.W.\ Abraham {\it et al.}, Phys.\ Rev.\ B {\bf 26}, 5268 (1982);
475: H.S.J.\ van der Zant, H.A.\ Rijken, and J.E.\ Mooij, J.\ Low.\
476: Temp.\ Phys.\ {\bf 79}, 289 (1990).
477: %\bibitem{Exp}
478: %D.J.\ Resnick, J.C.\ Garland, J.T.\ Boyd, S.\ Shoemaker, and R.S.\
479: %Newrock, Phys.\ Rev.\ Lett.\ {\bf 47}, 1542 (1981);
480: %D.W.\ Abraham, C.J.\ Lobb, M.\ Tinkham, and T.M.\ Klapwijk, Phys.\
481: %Rev.\ B {\bf 26}, 5268 (1982);
482: %H.S.J.\ van der Zant, H.A.\ Rijken, and J.E.\ Mooij, J.\ Low.\
483: %Temp.\ Phys.\ {\bf 79}, 289 (1990).
484:
485: \bibitem{Simkin}
486: M.V.\ Simkin and J.M.\ Kosterlitz, Phys.\ Rev.\ B {\bf 55}, 11646 (1997).
487:
488: \bibitem{Jose}
489: P.H.E.\ Tiesinga {\it et al.}, Phys.\ Rev.\ Lett.\ {\bf 78}, 519 (1997).
490: See also I.J.\ Hwang and D. Stroud, Phys. Rev. B {\bf 57}, 6036 (1998);
491: B.J.\ Kim and P. Minnhagen, {\it ibid}. {\bf 60}, 6834 (1999).
492: %\bibitem{Jose}
493: %P.H.E.\ Tiesinga, T.J.\ Hagenaars, J.E.\ van Himbergen, and J.V.\
494: %Jos\'e, Phys.\ Rev.\ Lett.\ {\bf 78}, 519 (1997).
495: %See also I.J.\ Hwang and D. Stroud, Phys. Rev. B {\bf 57}, 6036 (1998);
496: %B.J.\ Kim and P. Minnhagen, {\it ibid}. {\bf 60}, 6834 (1999).
497:
498: \bibitem{Kim}
499: B.J.\ Kim, P.\ Minnhagen, and P.\ Olsson, Phys.\ Rev.\ B {\bf 59},
500: 11506 (1999).
501:
502: \bibitem{Shaw}
503: T.J.\ Shaw {\it et al.}, Phys.\ Rev.\ Lett.\ {\bf 76}, 2551 (1996).
504: %\bibitem{Shaw}
505: %T.J.\ Shaw, M.J.\ Ferrari, L.L.\ Sohn, D.-H.\ Lee, M.\ Tinkham, and
506: %J.\ Clarke, Phys.\ Rev.\ Lett.\ {\bf 76}, 2551 (1996).
507:
508: \bibitem{Ol}
509: P.\ Olsson, Phys.\ Rev.\ B {\bf 46}, 14598 (1992).
510:
511: \bibitem{Choi}
512: S.\ Kim and M.Y.\ Choi, Phys.\ Rev.\ B {\bf 48}, 322 (1993).
513: M. Yoon, M.Y. Choi, and B.J. Kim, {\it ibid.} {\bf 61}, 3263 (2000).
514:
515: \bibitem{Ol2}
516: P.\ Olsson, Phys.\ Rev.\ Lett.\ {\bf 73}, 3339 (1994);
517: Phys.\ Rev.\ B {\bf 52}, 4526 (1995).
518:
519: \bibitem{Wallin}
520: M.\ Wallin and H.\ Weber, Phys.\ Rev.\ B {\bf 51}, 6163 (1995);
521: H.\ Weber, M.\ Wallin, and H.J.\ Jensen, {\it ibid}. {\bf 53}, 8566 (1996).
522:
523: \bibitem{comment}
524: %Whereas the currents fed or extracted make the grains at the edges
525: %$x=0$ and $x=N$ different from those inside, no such distinction is
526: %present along the $y$ direction. Accordingly,
527: Here the boundary conditions
528: along the transverse direction are not expected to have significant
529: effects on the voltage drop along the current-injected direction
530: in Eq. (3), which is averaged over the transverse direction.
531: Thus the periodic boundary conditions
532: may be maintained along the transverse direction, without affecting
533: qualitatively the scaling behavior, as manifesteded by Fig.~\ref{fig:FSSMCBC}.
534:
535: \end{references}
536:
537: \begin{center}
538:
539:
540: \begin{figure}
541: \vspace{0.5cm}
542: \centerline{\epsfig{width=10cm,file=fig2a.eps}}
543: \centerline{\large (a)}
544: \vspace{1cm}
545: \centerline{\epsfig{width=10cm,file=fig2b.eps}}
546: \vspace{-0.3cm}
547: \centerline{\large (b)}
548: \vspace{0.5cm}
549: \caption
550: {$IV$ characteristics of $N{\times}N$ square Josephson-junction arrays
551: at $T =0.84$ for several sizes (a) in the CBC and (b) in the FTBC.
552: Lines are merely guides to the eye.}
553: \label{fig:IV}
554: \end{figure}
555:
556: \newpage
557: \begin{figure}
558: \vspace*{0.5cm}
559: \centerline{\epsfig{width=12cm,file=fig3a.eps}}
560: \centerline{\large (a)}
561: \vspace*{0.5cm}
562: \centerline{\epsfig{width=12cm,file=fig3b.eps}}
563: %\vspace{-0.3cm}
564: \centerline{\large (b)}
565: \vspace{0.5cm}
566: \caption
567: {Scaling plots of the $IV$ data
568: (a) in the CBC with $z=0.91$ and (b) in the FTBC with $z=2.73$.
569: Each inset displays the log-log plot of the linear resistance versus
570: the size and its power-law fitting curve, from which the value of the dynamical
571: exponent $z$ is estimated.
572: }
573: \label{fig:FSS}
574: \end{figure}
575:
576: \newpage
577: \begin{figure}
578: \centerline{\epsfig{width=10cm,file=fig5.eps}}
579: \vspace{0.5cm}
580: \caption
581: {$IV$ characteristics of $(N{+}10){\times}N$ square Josephson-junction arrays
582: at $T =0.84$ for several sizes in the CBC.
583: The voltage is measured between the grains at $x=5$ and at $x=N{+}5$.
584: Lines are merely guides to the eye.}
585: \label{fig:IVMCBC}
586: \end{figure}
587:
588: \begin{figure}
589: \centerline{\epsfig{width=12cm,file=fig6.eps}}
590: \vspace{0.5cm}
591: \caption
592: {Scaling plot of the $IV$ data presented in Fig.~\ref{fig:IVMCBC},
593: with $z=2.84$.
594: Inset: the log-log plot of the linear resistance versus
595: the size and its power-law fitting curve.
596: }
597: \label{fig:FSSMCBC}
598: \end{figure}
599:
600: %\begin{figure}
601: %\vspace{1cm}
602: %\centerline{\epsfig{width=6cm,file=fig1.eps}}
603: %%
604: %%\vspace{1cm}
605: %
606: %\caption
607: %{Schematic diagram of an $N{\times}N$ square Josephson-junction array
608: %(corresponding to the case with $N=4$) in the CBC,
609: %which consist of the periodic and free boundary conditions
610: %for phases in the directions perpendicular and parallel to
611: %the current flow, respectively.}
612: %\label{fig:array}
613: %\end{figure}
614:
615: %\newpage
616:
617:
618:
619: %\vspace{1cm}
620:
621: %\begin{figure}
622: %\centerline{\epsfig{width=11mc,file=fig4.eps}}
623: %\vspace{0.5cm}
624: %\caption
625: %{Phase profile, which is defined by Eq.~(\ref{eq:profile}),
626: %as a function of the position along the direction of the current
627: %in the 32{$\times$}32 array, driven by the current $I = 0.04$
628: %under the CBC.
629: %}
630: %\label{fig:profile}
631: %\end{figure}
632: %
633:
634:
635: %\vspace{1cm}
636:
637:
638: \end{center}
639: %\end{multicols}
640:
641: \end{document}
642: