cond-mat0012238/icm.tex
1: \documentstyle[12pt]{article}
2: \input epsfig.sty      
3: \tolerance = 10000
4: \textheight=22cm
5: \textwidth=17cm
6: \hoffset=-1.5cm
7: \voffset=-2cm
8:  
9: 
10: \begin{document}
11: 
12: %\baselineskip=24pt
13: 
14: %\title{    { \tt  \large Abstract Designation: 5E-03; \qquad  Date: 04/27/2000 } \\
15: \title { {\tt \it \small Proceedings of International Conference on Magnetism 2000, 
16: to appear in J. Magn. Magn. Mater. } \\
17: { \ } \\
18: Phase-coherence transition in granular superconductors with $\pi$ junctions}
19: 
20: \author{Enzo Granato \\
21: Laborat\'orio Associado de Sensores e Materiais, \\
22: Instituto Nacional de Pesquisas Espaciais, \\
23: 12201-190 S\~ao  Jos\'e dos Campos, SP Brasil }
24: 
25: \date{}
26: \maketitle
27: 
28: \begin{abstract}                       
29: We study the three-dimensional XY-spin glass as a model for the resistive behavior of 
30: granular superconductors containing  a random distribution of $\pi$ junctions, as in 
31: high-$T_c$ superconducting materials with d-wave symmetry. The $\pi$ junctions leads to 
32: quenched in circulating currents (chiralities) and to a  chiral-glass state at low 
33: temperatures, even in the absence of an external magnetic field.   
34: Dynamical simulations in the phase representation are used to determine the 
35: nonlinear current-voltage characteristics as a function of temperature. 
36: Based on  dynamic scaling analysis, we find a phase-coherence transition at finite 
37: temperature below which the linear resistivity should vanish and determine the corresponding 
38: critical exponents. The results suggest that the phase and chiralities may order simultaneously 
39: for decreasing  temperatures into a superconducting chiral-glass state. 
40: \end{abstract}
41: 
42: \medskip
43: Keywords:  Superconductors-high-$T_c$ , Granular systems, Resistivity-scaling, Spin glass 
44: 
45: \medskip
46: %Corresponding author:
47: 
48: %Enzo Granato
49: 
50: %Laborat\'orio Associado de Sensores e Materiais, 
51: 
52: %Instituto Nacional de Pesquisas Espaciais, 
53: 
54: %12201-190 S\~ao Jos\'e dos Campos, SP Brasil 
55: 
56: %Tel: (12) 3456708; \qquad Fax: (12) 3456717;  \qquad   e-mail: enzo@las.inpe.br 
57: 
58: \newpage
59: 
60: Granular superconductors, including the high $T_c$ materials, 
61: are often described by arrays of superconducting grains coupled together
62: by  Josephson junctions. In a system of  conventional junctions ($s$-wave pairing)
63: the phases of the superconducting order parameter of neighboring grains  
64: tend to be locked with zero phase shift and a phase-coherence transition 
65: occurs for decreasing temperature into a state  
66: with long-range phase coherence and vanishing linear resistivity \cite{fisher}. 
67: On the other hand, there is increasing evidence for $d$-wave pairing symmetry 
68: in high $T_c$ materials with the remarkable consequence of the appearance of 
69: "$\pi$" junctions (negative coupling) characterized by a phase 
70: shift \cite{sigrist} of  $\pi$. 
71: This leads to frustration effects in granular samples, even in zero external 
72: magnetic field, since a closed loop containing an odd number of  $\pi$ junctions 
73: gives rise to an spontaneous circulating current (orbital magnetic moment) and can 
74: be a possible explanation of the  paramagnetic Meissner 
75: effect \cite{sigrist,kusmat,kawamura}. There is a close connection between this system and 
76: the XY-spin glass model \cite{kawamura} where the two-component spins, 
77: $s=(\cos(\theta),\sin(\theta))$,  correspond to superconducting grains and the 
78: random ferro or antiferromagnetic interactions to the Josephson couplings.
79: A chiral order parameter can be defined in the XY-spin glass model measuring the
80: direction of circulating current (vortex) in the loops for the superconducting array. 
81: Based on this analogy and numerical results for the XY-spin glass in 
82: three dimensions \cite{kawamura}, it has been argued that the equilibrium low-temperature
83: state is a chiral glass but with no long-range phase coherence and therefore 
84: not a true superconductor. However, recent results for XY-spin glass suggest that 
85: a spin glass (phase coherence) transition is possible \cite{grempel} in addition
86: to the  chiral glass transition. These results however rely on simulations for 
87: the ground state defect energy and spin-glass order parameter which are inaccessible
88: experimentally. Experiments often measure transport properties and it would be of interest
89: to study directly the current-voltage behavior from dynamical simulations and examine 
90: the issue of the finite-temperature resistive transition.  
91: 
92: To study the current-voltage characteristics we assume a resistively shunted 
93: Josephson-junction  model for the current flow between grains represented 
94: by the phases $\theta_i$ on a cubic lattice and  simulate the nonequilibrium behavior 
95: using the  Langevin equations \cite{eg}
96: \begin{equation}
97: C_o \frac{d^2 \theta_i}{dt^2} + \frac{1}{R_o}\sum_j \frac{d(\theta_i-\theta_j)}{dt} =
98: -J_o \sum_j \sin(\theta_i-\theta_j-A_{ij})+\sum_j \eta_{ij} 
99: \end{equation}
100: where $R_o$ is a shunt resistance, $C_o$ is the capacitance to the ground, 
101: $J_o>0$ is the Josephson coupling, $\eta_{ij}$ represents 
102: uncorrelated thermal noise with $<\eta_{ij}(t)^2>=2k_B T/R_o$ to ensure thermal 
103: equilibrium, and  $A_{ij}=0$ or $\pi $, represents the phase
104: shift across the junction, with equal probability of $0$ and $\pi$ corresponding to 
105: the standard XY spin glass \cite{kawamura,grempel,eg,wengel}. 
106: We use units where $\hbar/2e=1$, $R_o=1$, $I_o=1$ 
107: and set  $J_oR_o^2C_o=0.5$, corresponding to the overdamped regime. 
108: The equations were integrated numerically \cite{eg} for a system of linear size $L$
109: and the voltage $V$ (electric field $E/L$)  computed as
110: a function of the driving current $I$ (current density $J=I/L^2$) averaging 
111: over $5-10$ different realizations of the $A_{ij}$  distribution. 
112: 
113: The nonlinear resistivity $E/J$ as a function of temperature $T$ 
114: is shown in Fig. 1a, for $L=12$. The behavior is consistent with a linear resistivity 
115: $\rho_L=\lim_{J\rightarrow 0} E/J$ which is finite above a critical temperature
116: $T_c$. At lower temperatures, it appear to extrapolate to zero, indicating  
117: a superconducting transition with $T_c$ in the range $0.3 - 0.5$. 
118: This is confirmed by a scaling analysis of the nonlinear resistivity which
119: assumes the existence of a continuous superconducting transition at a finite temperature
120: \cite{fisher}. Near the transition, measurable quantities scale with the diverging correlation 
121: length $\xi\propto |T-T_c|^{-\nu}$ and relaxation time $\tau \propto \xi^z$, 
122: where $\nu$ and $z$ are the correlation-length and  dynamical critical exponents, 
123: respectively.  The nonlinear resistivity should then satisfy the scaling form \cite{fisher}
124: $ T E \xi^{z-d+2}/J= g_\pm(J\xi^{d-1}/T) $
125: where $d$ is the system dimension and $g(x)$ is a scaling function. The $+$ and $-$ 
126: signs correspond to $T>T_c$ and $T<T_c$, respectively. A scaling plot according to this
127: equation can be used to verify the scaling arguments 
128: and the assumption of an equilibrium transition at $J=0$ and also provide an estimate 
129: of the critical temperature and critical exponents. This is shown in Fig. 1b, 
130: obtained by adjusting 
131: the unknown parameters so that the best data collapse is obtained. We estimate
132: $T_c=0.41(3)$, $z=4.6(4)$ and $\nu = 1.2(4)$. The value of  $\nu$ 
133: agrees  with the result obtained from calculations of $\rho_L$ using the 
134: fluctuation-dissipation relation (zero current bias) in the vortex 
135: representation \cite{wengel}. However, $z$ is significantly larger 
136: but this could be a result of  different dynamics. 
137: We also note that $T_c$ and the critical exponents agree with the estimates for the 
138: gauge-glass model \cite{fisher} suggesting a common universality glass, although in 
139: the latter case $A_{ij}$ has a continuous distribution in the interval $[0,2\pi]$. 
140: Similar agreement has been found in the vortex representation \cite{wengel}. 
141: Since, in the phase model, 
142: the superconducting transition should correspond to the critical temperature 
143: where long-range  phase coherence sets in \cite{eg}, our results suggest a 
144: phase coherence transition at finite temperatures. 
145: This is in contrast with other numerical results indicating a phase coherence transition 
146: only  at zero temperature \cite{kawamura}. However, it is consistent with more
147: recent calculations  showing evidence of a  lower-critical dimension 
148: below or equal $3$ for the   XY-spin glass model \cite{grempel} and therefore a 
149: phase-coherence transition at finite temperature is possible. 
150: On the other hand, evidence of a chiral-glass transition has been found 
151: at $T_c\sim 0.32$ from Monte Carlo simulations \cite{kawamura}, 
152: below our estimate for the phase-coherence transition, which could suggest an 
153: intermediate phase. However, this 
154: requires phase coherence in presence of uncorrelated chiral (vortex) disorder which is 
155: unlikely. Thus, the finite-temperature transition found in the present  calculations 
156: should correspond to a single transition in the system where chirality and phase coherence
157: order simultaneously into a low temperature superconducting chiral glass phase. Although
158: our present conclusion is based on the scaling analysis of the nonlinear current-voltage 
159: characteristics which is a nonequilibrium property of the system, the good agreement with 
160: the previous results in the vortex representation  obtained from calculations of 
161: the (equilibrium) linear resistivity \cite{wengel}, suggests that this transition 
162: corresponds to the underlying equilibrium behavior. Additional calculations for the linear
163: resistivity will help to settle this interesting issue.
164: 
165: \medskip
166: This work was supported by FAPESP (grant: 99/02532). 
167: 
168: 
169: \begin{thebibliography}{99}
170: 
171: \bibitem{fisher} D.S. Fisher {\it et al.}
172: %, M.P.A. Fisher, and D.A. Huse, 
173: Phys. Rev. B {\bf 43} (1991)130 ; 
174: J.D. Reger {\it et al.}
175: %, T.A. Tokuyasu, A.P. Young, and M.P.A. Fisher, Phys. Rev.  
176: B {\bf 44} (1991) 7147. 
177: 
178: \bibitem{sigrist} M. Sigrist and T.M. Rice, Rev. Mod. Phys. {\bf 67} (1995) 503.
179: 
180: \bibitem{kusmat}  F.V. Kusmartsev, Phys. Rev. Lett. {\bf 69} (1992) 2268.
181: 
182: \bibitem{kawamura} H. Kawamura, J. Phys. Soc. Jpn {\bf 64} (1995) 711;
183: H. Kawamura and M.S. Li, Phys. Rev. Lett. {\bf 78} (1997) 1556.
184: 
185: \bibitem{grempel}  J. Maucourt and D.R. Grempel, Phys. Rev. Lett. {\bf 88} (1998) 770.
186: 
187: \bibitem{eg} E. Granato, Phys. Rev. B{\bf 58} (1998) 11161; Phys. Rev.  
188: B{\bf 61}, (2000) 391.
189: 
190: \bibitem{wengel} C. Wengel and A.P. Young, Phys. Rev. B {\bf 56} (1997) 5918.
191: 
192: \end{thebibliography}
193: 
194: \newpage
195: 
196: %\centerline {Figure Caption}
197: 
198: \bigskip
199: \begin{figure}
200: \centering\epsfig{file=fig1.eps,bbllx=1cm,bblly=0.5cm,bburx=20cm,bbury=26cm,width=8.5cm}
201: \caption{
202: Fig. 1. (a) Nonlinear resistivity $E/J$ as a function of temperature. (b) Scaling plot
203: near the transition obtained by adjusting $T_c$, $z$ and $\nu$ 
204: where $\xi\propto|T/T_c-1|^{-\nu}$. }
205: \end{figure}
206: 
207: \end{document}