1: \documentstyle[aps,prl]{revtex}
2: %\documentstyle[aps,prl]{revtex}
3: %\documentstyle[aps,prl,preprint]{revtex}
4: \input epsf
5: \input{epsf.sty}
6: \begin{document}
7: \draft
8: % for two column activate the line below...
9: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname@twocolumnfalse\endcsname
10:
11: %
12: \title {Abrupt Change of
13: Josephson Plasma Frequency at the Phase Boundary of\\
14: the Bragg Glass in Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$}
15: \author{M.B. Gaifullin$^1$, Yuji Matsuda$^{1,\dag}$, N. Chikumoto$^2$, J.
16: Shimoyama$^3$ and K. Kishio$^3$} \address{$^1$Institute for Solid State Physics,
17: University of Tokyo, Roppongi 7-22-1, Minato-ku, Tokyo 106, Japan} \address{$^2$Superconductivity
18: Research Laboratory, ISTEC, Shibaura 1-16-25 , Minato-ku, Tokyo 105, Japan}
19: \address{$^3$Department of Superconductivity, University of Tokyo, Bunkyo-ku,
20: Tokyo 113, Japan}
21:
22:
23: \date{received 16 December 1999}
24: \maketitle
25: \begin{abstract}
26:
27: We report the first detailed and quantitative study of the Josephson coupling
28: energy in the vortex liquid, Bragg glass and vortex glass phases of
29: Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ by the Josephson plasma resonance. The
30: measurements revealed distinct features in the $T$- and $H$-dependencies of the
31: plasma frequency $\omega_{pl}$ for each of these three vortex phases. When
32: going across either the Bragg-to-vortex glass or the Bragg-to-liquid transition
33: line, $\omega_{pl}$ shows a dramatic change. We provide a quantitative
34: discussion on the properties of these phase transitions, including the first
35: order nature of the Bragg-to-vortex glass transition.
36:
37: \end{abstract}
38: \pacs{74.25.Nf, 74.50+r, 74.60.Ec, 74.72.Hs}
39:
40: ]
41:
42:
43: \narrowtext
44:
45: The vortex matter in
46: high-$T_c$ superconductors exhibits a fascinatingly rich phase diagram with a
47: variety of phase transitions. There, thermal fluctuation and disorder alter
48: dramatically the vortex phase diagram which has been observed in the
49: conventional superconductors. At high temperature, the strong thermal
50: fluctuation melts a vortex lattice into a vortex liquid well below the upper
51: critical field. On the other hand, at low temperature or low field where the
52: vortex liquid freezes into a solid phase, disorder plays an important role. The
53: disorder is known to destroy the long-range order of the Abrikosov lattice
54: \cite{lo}. Recent investigations have revealed that the vortex solid phase is
55: comprised of two distinct phases; a highly disordered phase at high field and a
56: rather ordered phase at low field \cite{chiku,cubt,khay}. The former phase is
57: the vortex glass or entangled solid phase which is characterized by divergent
58: barriers for vortex motion \cite{ffh}. The latter phase is the Bragg glass or
59: quasilattice phase in which no dislocation exists and quasi-long-range
60: translational order is preserved \cite{giam}. In very clean single crystals,
61: thermodynamical measurements have revealed that the Bragg glass undergoes a
62: first order transition (FOT) to the vortex liquid \cite{zeld,dods,hu}. The
63: transition from the Bragg glass to the vortex glass, on the other hand, is
64: characterized by the second magnetization peak at which the critical current
65: shows a sharp increase \cite{khay}. It was proposed that the crossover from the
66: FOT to the second peak regime is governed by a critical point $T_{cp}$ in the
67: phase diagram, which in Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ is located near 40~K.
68: While the nature of the vortex liquid has been extensively studied, the
69: properties of the Bragg glass and the nature of the thermally induced FOT from
70: the Bragg glass to the vortex liquid are still not quite understood. Moreover,
71: the phase transition from the Bragg glass to the vortex glass at lower
72: temperatures has been a longstanding issue, though this transition is proposed
73: to be disorder driven, caused by competition between the elastic and pinning
74: energies \cite{khay,nels,ryu,vino,horo,kosh2}. A major obstacle has been that
75: most of the previous experiments had to been performed under a strongly
76: nonequilibrium condition because most part of the Bragg and vortex glasses are
77: located deep inside the irreversibility line $T_{irr}$.
78:
79:
80: The most direct way to clarify the nature of these phases and the phase
81: transitions among them is to measure the interlayer phase coherence for each
82: vortex phases, because the CuO$_2$ layers are connected by the Josephson effect.
83: One of the most powerful probes for the interlayer phase coherence is the
84: Josephson plasma resonance (JPR) which provides a direct measurement of the
85: Josephson plasma frequency $\omega_{pl}$ related to the maximum Josephson
86: critical current $j_{J}=\varepsilon_0\Phi_0\omega_{pl}^2/8\pi^2cd$ and the
87: Josephson coupling energy $U_J=\Phi_0j_J/2 \pi c$, where $\varepsilon_0$ and $d$
88: are the dielectric constant and interlayer spacing, respectively.
89: \cite{mats1,mats2,shib,bula,kosh1,hwan,gaif,kosh3}. Especially in
90: Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ with large anisotropy, a very precise
91: determination of $\omega_{pl}$ is possible because $\omega_{pl}$ falls within
92: the microwave window.
93:
94:
95: All of the JPR measurements of Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ up to now have
96: been carried out in the cavity resonator by reducing $\omega_{pl}$ by $H$
97: \cite{mats1,mats2,shib}. Unfortunately, sweeping $H$ below $T_{irr}$ drives the
98: vortex system into a strongly nonequilibrium state due to the Bean critical
99: current induced by the field gradient inside the crystal, as was demonstrated in
100: Refs.\cite{mats2} and \cite{rodr}. Therefore, in order to investigate the Bragg
101: and vortex glass phases, it is crucial to measure the JPR as a function of the
102: microwave frequency $\omega_{pl}$ while holding $H$ at a constant value. In
103: this Letter, we report the first detailed and quantitative study of the
104: Josephson coupling energy in the Bragg glass, the vortex glass and the vortex
105: liquid phases and the phase transitions among them by the JPR which has been
106: preformed by sweeping $\omega$ continuously. The measurements revealed distinct
107: features in the $T$- and $H$-dependencies of $\omega_{pl}$ for each of the three
108: different vortex phases. When going across either the Bragg-to-vortex glass
109:
110: \begin{figure}
111: \centerline{\epsfxsize 7.5cm \epsfbox{figure1.eps}}
112: \vspace{-1cm}
113: \caption{ Vortex phase diagram determined by
114: the magnetization and the JPR. The filled circles represent $H_{m}$ determined
115: by the magnetization step (see inset). The filled triangles represent $H_{sp}$.
116: The filled diamonds represent the field at which $\omega_{pl}$ shows an abrupt
117: change. The dashed line is the irreversibility line. Lower inset:
118: Magnetization measured in the field cooling (FC) and zero field cooling (ZFC)
119: conditions. Upper inset: Bolometric detection of the microwave absorption. The
120: crystal is supported inside the waveguide by a thermally isolated diamond plank
121: ({\bf H}$\parallel c$). The JPR is caused by the microwave electric field {\bf
122: E}$_{ac}\parallel c$.}
123:
124: \end{figure}
125: \noindent or
126: the Bragg-to-liquid transition line, $\omega_{pl}$ shows a dramatic change. We
127: provide a quantitative discussion on the nature of these phase transitions in
128: the light of these results.
129:
130:
131: All experiments were performed on a slightly underdoped
132: Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ single crystals ($T_c$=82.5~K) with
133: dimensions $1.2\times0.5\times0.03$mm$^3$ grown by the traveling floating zone
134: method. The inset of Fig.1 shows a typical magnetization step measured by SQUID
135: magnetometer which can be attributed to the FOT of the vortex lattice. This FOT
136: terminates at $\approx40$~K and the step is followed by the second magnetization
137: peak located at $\sim$230~Oe. Figure 1 shows the phase diagram obtained by the
138: magnetization measurements. The JPR is measured by sweeping $\omega$
139: continuously from 20~GHz to 150~GHz \cite{gaif}. The sample was placed at the
140: center of the broad wall of the waveguide in the traveling wave TE$_{01}$ mode.
141: We used a bolometric technique to detect very small microwave absorption by the
142: sample and employed a leveling loop technique to ckeep the microwave power
143: constant when sweeping frequency. For this crystal $\omega_{pl}$=125~GHz at
144: $T$=0, corresponding to the anisotropy parameter
145: $\gamma=\lambda_c/\lambda_{ab}\approx 550$, where $\lambda_{ab}$ and $\lambda_c$
146: are the in-plane and out-of-plane penetration lengths, respectively. Here we
147: used $\lambda_{ab}\approx 200$~nm and
148: $\lambda_c=c/\omega_{pl}\sqrt{\varepsilon_0}\approx$110~$\mu$m. We determined
149: $\varepsilon_0=11.5\pm1$ from the dispersion of the transverse plasma mode. All
150:
151: \begin{figure}
152: \centerline{\epsfxsize 7.5cm \epsfbox{figure2.eps}}
153: \caption{The JPR as a function of frequency
154: when crossing (a) the second peak (6.4~K) and (b) the FOT (50~K). We picked up
155: only the longitudinal plasma mode which is sample size independent. The arrows
156: indicate the peak position. }
157: \end{figure}
158:
159: \noindent JPR measurements were performed in {\bf H}$\parallel c$ {\em under the field
160: cooling condition} (FCC) where the field is very uniform. In this condition,
161: the system is in equilibrium or at worst is trapped in a metastable state which
162: we expect should be much closer to equilibrium compared to the state obtained in
163: the field sweeping condition (FSC). In fact, while the resonance frequency
164: below $T_{irr}$ did not change at all with time for more than 48 hours in the
165: FCC, it increases gradually with time in the FSC. We also confirmed that the
166: resonance curves are exactly the same in different cooling cycles.
167:
168:
169: Figures 2(a) and (b) depict the resonant absorption as a function of
170: $\omega$ when crossing the second peak field $H_{sp}$ and the FOT field
171: $H_m$, respectively. When $\omega$ coincides with $\omega_{pl}$, the
172: resonant absorption of the microwave occurs. These are the JPR measured
173: in the Bragg and vortex glass phases under the FCC for the first time.
174: In the magnetic field, $\omega_{pl}$ can be written as \cite{bula}
175: \begin{equation}
176: \omega_{pl}^2(B,T)=\omega_{pl}^2(0,T)\langle\cos\phi_{n,n+1}\rangle.
177: \end{equation}
178: Here $\langle\cos\phi_{n,n+1}\rangle$ represents the thermal and disorder
179: average of the cosine of the gauge invariant phase difference between layer $n$
180: and $n+1$. If the vortex forms a straight line along the $c$-axis,
181: $\langle\cos\phi_{n,n+1}\rangle$ is unity. The reduction of
182: $\langle\cos\phi_{n,n+1}\rangle$ from unity is caused by the Josephson strings
183: that are created by the deviation from the straight alignment of the pancake
184: vortices along the $c$-axis. Thus $\omega_{pl}$ gives a direct information on
185: the vortex alignment and therefore the phase transition of the vortex matter.
186: After gradual decrease with $H$ at lower $H$, $\omega_{pl}$ shows a sharp
187: decrease in the field range between 215~Oe and 220~Oe at 6.5~K and between
188: 140~0e and
189:
190: \begin{figure}
191: \centerline{\epsfxsize 7.5cm \epsfbox{figure3.eps}}
192: \caption{Inset: $H$-dependence of
193: $\langle\cos\phi_{n,n+1}\rangle$ when going across the FOT at high temperatures.
194: Solid squares, diamonds, circles, and triangles show the data at 40~K, 50~K,
195: 60~K, and 70~K respectively. Main panel: Same data plotted as a function of
196: $H/H_m$. Open squares show $\langle\cos\phi_{n,n+1}\rangle$ as a function of
197: $H/H_{sp}$ at 30~K when crossing the second peak. The dashed line is the result
198: of Eq.2. The solid line is the fit to Eq.3. }
199: \end{figure}
200: \noindent 160~Oe at 50~K. At 217.5~Oe in Fig.2(a) and at 150~Oe in Fig.2(b),
201: the resonance lines become broader, indicating a very rapid change of
202: $\omega_{pl}$ with $H$. At higher $H$, $\omega_{pl}$ again decreases gradually.
203: In Fig.1, we plot the fields at which $\omega_{pl}$ shows an abrupt change.
204: These fields coincide well with the second peak and FOT fields determined by
205: magnetization measurements.
206:
207:
208: We first discuss the resonance when going across the FOT. The inset of Fig.3
209: depicts the $H$-dependence of $\langle\cos\phi_{n,n+1}\rangle$ obtained from
210: $\omega_{pl}^2(B,T)/\omega_{pl}^2(0,T)$. Although similar results have been
211: reported \cite{mats2,shib}, quantitative analysis was very difficult because the
212: JPR measurements in the Bragg glass had been done under the strongly
213: nonequilibrium condition, as we have already mentioned. Figure 3 depicts
214: $\langle\cos\phi_{n,n+1}\rangle$ as a function of $H$ normalized by $H_m$.
215: Interestingly, $\langle\cos\phi_{n,n+1}\rangle$ exhibits very similar
216: $H/H_m$-dependence at all temperatures. Obviously, the $H$-dependence of
217: $\langle\cos\phi_{n,n+1}\rangle$ above FOT is very different from that below
218: FOT; the curvature changes from negative to positive. We found that
219: $\langle\cos\phi_{n,n+1}\rangle$ in the Bragg glass phase can be fitted as,
220: \begin{equation}
221: \langle\cos\phi_{n,n+1}\rangle=
222: 1-A_1\frac{H}{H_m}-A_2\left(\frac{H}{H_m}\right)^2,
223: \end{equation}
224: with $A_1$=0.16 and $A_2$=0.19 above 40~K as shown in the dashed line in Fig.3.
225: On the other hand, according to high temperature expansion theory \cite{kosh1},
226: $\langle\cos\phi_{n,n+1}\rangle$ in the liquid phase above FOT can be written
227: as,
228: \begin{equation}
229: \langle\cos\phi_{n,n+1}\rangle=\frac{U_J\Phi_0}{2k_BTH},
230: \end{equation}
231:
232: \begin{figure}
233: \centerline{\epsfxsize 7.5cm \epsfbox{figure4.eps}}
234: \caption{$H$-dependence of
235: $\langle\cos\phi_{n,n+1}\rangle$ when going across the Bragg-to-vortex glass
236: transition at low temperatures. }
237: \end{figure}
238:
239: \noindent when the Josephson energy is negligible compared with the energy of thermal
240: fluctuations, {\it i.e.}$U_J \ll k_BTH/\Phi_0$. It has been shown
241: experimentally that $\langle\cos\phi_{n,n+1}\rangle$ is inversely proportional
242: to $H$ in the liquid phase \cite{mats1,mats2,shib}. The present results provide
243: a further rigorous test to Eq.3, because we now have no ambiguous fitting
244: parameter and also have the data of the very detailed $H$-dependence of
245: $\langle\cos\phi_{n,n+1}\rangle$ obtained by sweeping $\omega$. The solid line
246: in Fig.3 shows the result of the calculation. In the calculation, we used
247: $\varepsilon_0$=12.0. The fit to the data is excellent
248: in the whole $H$-range at $H>1.2H_m$, indicating that the vortex
249: liquid is decoupled on the scale of the interlayer distance. Small deviation
250: from Eq.3 is observed at $H\leq1.2H_m$. This suggests that the vortex-vortex
251: correlation effect in the $ab$-plane which gives rise to the deviation from
252: $1/H$-dependence of $\langle\cos\phi_{n,n+1}\rangle$ in the liquid phase is
253: important just above the FOT \cite{kosh3}.
254:
255:
256: At $H_m$, $\langle\cos\phi_{n,n+1}\rangle$ is reduced to $\approx$0.7 at all
257: temperatures, showing {\em an occurrence of large vortex wandering in the Bragg
258: glass.} Near $T_c$, we note that $\omega_{pl}$ at $H$=0 is already suppressed by
259: the phase fluctuations. If this effect is taken into account, it is expected
260: that $\langle\cos\phi_{n,n+1}\rangle$ at $H_m$ slowly increases with $T$,
261: indicating that the melting becomes more linelike at higher $T$. The values of
262: $\langle\cos\phi_{n,n+1}\rangle$ at $H_m$ are close to the recent results of
263: computer simulations for systems with small anisotropies\cite{hu}.
264:
265:
266: The internal energy $U$ experiences a jump $\Delta U$ at the FOT. This latent
267: heat $\Delta U$ can be represented as a sum of the jumps in the in-plane energy,
268: in the electromagnetic coupling energy, and in the Josephson energy $\Delta U_J$
269: \cite{hu}. To understand the nature of the FOT in detail, it is important to
270: establish the relative jump in Josephson energy $\Delta U_J/\Delta U$. At 60~K,
271: $\Delta U_J/T$ can be estimated to be $\approx 0.21k_B$ from
272: $\langle\cos\phi_{n,n+1}\rangle$ which drops approximately from 0.70 to 0.45.
273: On the other hand, $\Delta U/T$ at 60~K obtained from the magnetization step
274: $\Delta M$ using the Clausius-Clapeyron relation,
275: \begin{equation}
276: \Delta U/T_m=\Delta S=-d\Phi_0\frac{\Delta M}{B_m}\frac{dB_m}{dT},
277: \end{equation}
278: is $\approx1.34k_B$. Here $\Delta S$ is the entropy jump at the FOT point
279: ($T_m$, $B_m$). Thus we find that $\Delta U_J$ constitutes approximately 16\%
280: of the latent heat, showing that $\Delta U_J$ occupies a substantial part of the
281: latent heat at the FOT even in Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ with very
282: large anisotropy.
283:
284: We now move on to the subject of the Josephson coupling at low temperatures when
285: going across the transition from the Bragg glass to the vortex glass. Figure 4
286: shows the $H$-dependence of $\langle\cos\phi_{n,n+1}\rangle$ below 35~K. Below
287: 100~Oe, $\langle\cos\phi_{n,n+1}\rangle$ shows a hump structure which may be
288: related with the lower critical field. Above 100~Oe, the $H$-dependence of
289: $\langle\cos\phi_{n,n+1}\rangle$ is very similar to that at high $T$ when
290: crossing the FOT. At all temperatures, $\langle\cos\phi_{n,n+1}\rangle$ shows an
291: abrupt change at the second peak field $H_{sp}\approx$220~Oe. In similarity to
292: the high temperature behavior, the $H$-dependence of
293: $\langle\cos\phi_{n,n+1}\rangle$ below and above $H_{sp}$ are very different,
294: showing clearly that $H_{sp}$ separates two distinct vortex phases. In a very
295: narrow field interval less than 5~Oe at $H_{sp}$,
296: $\langle\cos\phi_{n,n+1}\rangle$ drops from approximately 0.7 to 0.5 (see also
297: Fig.1(a)), corresponding to a nearly 20\%-reduction of $U_J$. This strong
298: reduction of $\langle\cos\phi_{n,n+1}\rangle$ provides a direct evidence of the
299: decoupling nature of the Bragg-to-vortex glass transition \cite{horo}. At
300: $H_{sp}$, $\langle\cos\phi_{n,n+1}\rangle$ is reduced to $\approx$0.7 from the
301: zero field value similar to that below FOT. Although we do not show here,
302: $\langle\cos\phi_{n,n+1}\rangle$ in the vortex glass phase deviates from the
303: $1/H$ dependence in the whole $B$-regime above $H_{sp}$, which is to be
304: contrasted to the behavior in the vortex liquid phase.
305:
306: We finally discuss the phase transition from the Bragg glass to the vortex glass
307: inferred from the JPR. The first question is the order of the transition. The
308: abrupt change of $\langle\cos\phi_{n,n+1}\rangle$ shown in Figs.3 and 4 provides
309: a direct evidence of the abrupt changes of the $c$-axis correlation length of
310: the pancakes and of $U_{J}$ which composes a substantial part in the free
311: energy. In Fig.3 we plot the change of $\langle\cos\phi_{n,n+1}\rangle$ at
312: $H_{sp}$ ($T$=30~K), for the comparison with the change of the same quantity at
313: the FOT. {\em The change of $\langle\cos\phi_{n,n+1}\rangle$ at $H_{sp}$ is
314: comparable or even sharper than that at the FOT}. This fact strongly indicates
315: the first order nature of the phase transition from the Bragg glass to the
316: vortex glass. The second issue is the critical point $T_{cp}$ of the FOT which
317: has been proposed to terminate at $\approx$40~K \cite{khay,zeld}. This proposal
318: was made from the observation that $\Delta S$ becomes extremely small which can
319: be seen from the $T$-independence of FOT line below $T_{cp}$. However, the
320: vanishing of $\Delta S$ does not immediately imply the termination of the FOT,
321: which suggests that the issue of the termination is nontrivial. As seen in
322: Figs.3 and 4, there is no discernible difference in the $H$-dependence of
323: $\langle\cos\phi_{n,n+1}\rangle$ as we go through the Bragg-to-liquid transition
324: regime, into the Bragg-to-vortex glass transition regime, except for a gradual
325: decrease of the change of $\langle\cos\phi_{n,n+1}\rangle$ at the transition.
326: These results imply that the FOT does not terminate at $\approx$40~K, but that
327: there is no critical point or the FOT persists at least below 6.4~K. We note
328: that a similar conclusion has been reached very recently using the
329: magneto-optical imaging technique \cite{beek}.
330:
331: In summary, we have performed the JPR experiments in the Bragg glass, vortex
332: glass, and vortex liquid phases in the FCC. We found an abrupt change in the
333: Josephson coupling energy when going across either the FOT line or the second
334: magnetization peak line. We showed that this change occupies a substantial part
335: of the latent heat at the FOT. The results suggest that the Bragg-to-vortex
336: glass transition is first order in nature and that the critical point of the FOT
337: does not terminate at $\approx$40~K.
338:
339: We thank B. Horovitz, X. Hu, Y. Kato, P.H. Kes, T. Onogi, A. Sudb\o, and A.
340: Tanaka for discussions. We are indebted to L.N. Bulaevskii
341: for several valuable comments. We also thank A.E.Koshelev for the critical
342: reading of the manuscript.
343:
344: \begin{thebibliography}{99}
345:
346: \bibitem[\dag].Corresponding author, \\E-mail Address :ym@issp.u-tokyo.ac.jp
347: \bibitem{lo}A.I. Larkin, and Y.N. Ovchinnikov, JETP {\bf 38}, 854 (1974).
348: \bibitem{chiku}N. Chikumoto {\it et al.}, Phys. Rev. Lett. {\bf 69}, 1260
349: (1992)
350: \bibitem{cubt}R. Cubtt {\it et al.}, Nature (London), {\bf 365}, 407
351: (1993).
352: \bibitem{khay}B. Khaykovich {\it et al.}, Phys. Rev. Lett. {\bf 76},
353: 2555 (1996); Phys. Rev. B {\bf 56}, R517 (1997).
354: \bibitem{ffh}D.S. Fisher
355: {\it et al.}, Phys. Rev. B {\bf 43}, 130 (1991).
356: \bibitem{giam}T. Giamarchi,
357: and P.Le Doussal, Phys. Rev. B {\bf 55}, 6577 (1997). \bibitem{zeld} E.
358: Zeldov {\it et al.}, Nature (London), {\bf 375}, 373 (1995), A. Shilling {\it et
359: al.}, Nature (London) {\bf 382}, 791 (1996).
360: \bibitem{dods}M.J.W. Dodgson {\it
361: et al.}, Phys. Rev. Lett. {\bf 80}, 837 (1998).
362: \bibitem{hu}X. Hu {\em et
363: al.}, Phys. Rev. Lett. {\bf 79}, 3498 (1997), A.E. Koshelev, Phys. Rev. B
364: {\bf 56}, 11201 (1997), A.K. Nguyen and Sudb\o, Phys. Rev. B {\bf 58}, 2802
365: (1998), Y. Nonomura, and X. Hu, Physica B in press.
366: \bibitem{nels}D.Ertas and D.R. Nelson, Physica C {\bf
367: 272}, 79 (1996).
368: \bibitem{ryu}S. Ryu {\it et al.}Phys. Rev. Lett. {\bf 77},
369: 2300 (1996).
370: \bibitem{vino}V. Vinokur {\it et al.}, Physica C {\bf 295}, 209
371: (1998).
372: \bibitem{horo}B. Horovitz and T.R. Goldin, Phys. Rev. Lett. {\bf
373: 80}, 1734 (1998), B. Horovitz, Phys. Rev. B {\bf 60}, R9939 (1999).
374: \bibitem{kosh2}A.E. Koshelev and V.M. Vinokur, Phys. Rev. B {\bf 57}, 8026
375: (1998).
376: \bibitem{mats1}Y. Matsuda {\em et al.}, Phys. Rev. Lett., {\bf 75},
377: 4512 (1995), O.K.C. Tsui {\em et al.} Phys. Rev. Lett. {\bf 76}, 819 (1996)
378: \bibitem{mats2}Y. Matsuda {\it et al.} Phys. Rev. Lett. {\bf 78}, 1972
379: (1997).
380: \bibitem{shib}T. Shibauchi {\it et al.}, Phys. Rev. Lett. {\bf 83},
381: 1010 (1999).
382: \bibitem{bula}L.N. Bulaevskii {\em et al.}, Phys. Rev. Lett.
383: {\bf 74}, 801 (1995).
384: \bibitem{kosh1}A.E.Koshelev, Phys. Rev. Lett. {\bf
385: 77}, 3901 (1996).
386: \bibitem{hwan}I.J. Hwang and D. Stroud, Phys. Rev. B {\bf
387: 59}, 3896 (1999).
388: \bibitem{gaif}M.B. Gaifullin {\it et al.} Phys. Rev. Lett.
389: {\bf 83}, 3928 (1999).
390: \bibitem{kosh3}A.E. Koshelev,L.N. Bulaevskii, and
391: M.P.Maley, Phys. Rev. Lett. {\bf 81}, 902 (1998).
392: \bibitem{rodr}E.Rodrigues
393: {\it et al.}, Phys. Rev. Lett. {\bf 71}, 3375 (1993).
394: \bibitem{beek}C.J. van der Beek {\it et al.}, cond-mat/9912276.
395:
396:
397: \end{thebibliography}
398:
399:
400:
401:
402:
403:
404:
405:
406:
407: \end{document}
408:
409:
410:
411:
412: