1: \documentclass[aps,prl,twocolumn,groupedaddress]{revtex4}
2: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
3: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
4:
5: % Some other (several out of many) possibilities
6: %\documentclass[preprint,aps]{revtex4}
7: %\documentclass[preprint,aps,draft]{revtex4}
8: %\documentclass[prb]{revtex4}% Physical Review B
9:
10: \usepackage{graphicx}% Include figure files
11: \usepackage{dcolumn}% Align table columns on decimal point
12: \usepackage{bm}% bold math
13:
14:
15: \begin{document}
16:
17:
18: \title{\large Observation of McMillan-Rowell like
19: oscillations in underdoped YBCO junctions oriented along the node
20: of the d-wave order parameter}
21:
22: \author{L. Shkedy, P. Aronov, G. Koren and E. Polturak}
23: \affiliation{Physics Department, Technion - Israel Institute of Technology Haifa,
24: 32000, ISRAEL}
25:
26:
27: \email{gkoren@physics.technion.ac.il}
28: \homepage{http://physics,technion.ac.il/~gkoren}
29:
30:
31: \date{\today}
32:
33: \begin{abstract}
34: Dynamic resistance spectra of ramp junctions made of underdoped
35: $YBa_{2}Cu_{3}O_y$ electrodes and Ga-doped YBCO barrier are
36: reported. Series of equidistant peaks were observed in these
37: spectra in junctions oriented along the node direction. Junctions
38: with different barrier thickness $d_N$, showed that the distance
39: between adjacent peaks scales inversely with $d_N$. The peaks were
40: thus identified as due to McMillan-Rowell like oscillations in the
41: barrier. Analysis of the series of peaks yields an upper limit of
42: about 3.7 meV on the value of the energy gap along the node. We
43: attribute this small gap to the $is$ component of the order
44: parameter of underdoped YBCO near the interface of the junctions.
45:
46: \end{abstract}
47:
48:
49: \maketitle{\large INTRODUCTION}\\
50:
51: The fine details of the symmetry of the order parameter in the
52: high temperature superconductors (HTS) are still under debate.
53: Determining the exact symmetry could be important for
54: understanding the mechanism of the high $T_c$ superconductors.
55: Several experiments show that the dominant component of the order
56: parameter in the HTS materials has a $d_{x^2-y^2}$-wave symmetry,
57: as summarized in a review article by Tsuei and
58: Kirtley.\cite{Tsuei} Other experiments are consistent with the
59: existence of an additional sub-dominant component on the surface
60: of the HTS, of $is$ or $id_{xy}$
61: nature.\cite{Covington,Krupke,Sharoni,Koren2002,KorenJLTP}
62: Tunneling measurements of underdoped junctions show that in
63: addition to the $d_{x^2-y^2}+is$ gap, a large gap which can be
64: attributed to the pseudogap is also
65: present.\cite{DeutscherNature,Krasnov,KorenPseudogap} The
66: magnitude of the $is$ component in these
67: studies,\cite{Covington,Krupke,Sharoni,Koren2002,KorenJLTP} was
68: found to be in the range of ~1-3 meV as determined from the peak
69: to peak distance in the conductance curves. Recent self-consistent
70: calculations using Bogoliubov-DeGennes type equations, led to a
71: good fit of the data assuming a pure d-wave symmetry in the bulk,
72: and coexisting $d_{x^2-y^2}$ and $is$ order parameters near the
73: interface.\cite{Lubimova} The magnitude of the $is$ gap resulting
74: from these simulations is 2.6$\pm 0.1$ meV. In the present
75: experiment we observed series of geometrical resonances in the
76: dynamic resistance spectra of node junctions. From these series,
77: we find an upper limit of 3.7 meV on the energy of the
78: sub-dominant s-wave component of the gap, which is consistent
79: with the simulations results.\\
80:
81: \begin{figure}
82: \includegraphics[height=6cm,width=8.5cm]{Fig1.ps}
83: \caption{\label{fig:epsart} Resistivity versus temperature of a
84: Ga-doped YBCO film. The resistivity values were obtained by
85: averaging over six microbridges which were patterned in the ~80 nm
86: thick film. The inset shows a schematic cross section of a ramp
87: type junction. The tunneling current flows along the a-b plane,
88: while c-axis current is suppressed in the junction due to the
89: thick insulating STO layer.}
90: \end{figure}
91:
92: \maketitle{\large EXPERIMENTAL}\\
93:
94: The junctions used in the present study are of the same type of
95: ramp junctions used
96: before,\cite{KorenPseudogap,NesherTomasch,Koren2002,KorenJLTP,NesherSubHarmonic}
97: but instead of the Fe-doped YBCO barrier, we used a Ga-doped YBCO
98: barrier. This was done in order to check if the appearance of the
99: s-wave component depends on the nature of the barrier. In
100: particular, the possible presence of magnetic impurities in the Fe
101: doped barrier, is probably absent when a Ga doped barrier is used.
102: Therefore, observation of a signature of an $is$ gap, in both type
103: of junctions, is indicative that magnetic effects in the barrier
104: are not likely to be the source of this gap. Fig. 1 shows the
105: resistivity versus temperature of a blanket
106: $YBa_2Ga_{0.4}Cu_{2.6}O_y$ film deposited on (100) $SrTiO_3$ (STO)
107: wafer, annealled {\em in-situ} under the same annealing conditions
108: as for obtaining YBCO films with $T_c=60K$. This barrier material
109: behaves like a Mott insulator (MI) with variable range hopping
110: (VRH) in 3D with lg$(\rho) \propto T^{-1/4}$ for the whole
111: temperature range. The resistivity value at 2K is of about
112: $0.6\Omega cm$, which is more than an order of magnitude higher
113: than that of $YBa_2Fe_{0.45}Cu_{2.55}O_y$ used previously as the
114: barrier layer.\cite{NesherSubHarmonic} The geometry of the
115: junction is described schematically in the inset of Fig. 1. All
116: the YBCO and doped YBCO layers are epitaxial with the $c-axis$
117: normal to the wafer. The two superconducting YBCO electrodes are
118: coupled in the
119: $a-b$ plane via a thin barrier layer of the Ga-doped YBCO film.\\
120:
121: The multi-step junction preparation process was described
122: before.\cite{NesherTomasch} Briefly, we first prepared by laser
123: ablation deposition the base electrode which was composed of a
124: bilayer of STO on YBCO on (100) STO wafer. Patterning of the base
125: electrode was done by photolithography and Ar ion beam milling.
126: The ramps of the junctions were patterned along the node direction
127: of the YBCO film. After a thorough cleaning process, the cover
128: electrode was deposited. This included the barrier layer, a second
129: YBCO film, and an Au layer on top. The cover electrode was then
130: patterned to produce the final junction layout, as well as the
131: four gold pads for each junction. All junctions in the present
132: study had the same 90nm thick YBCO electrodes (base as well as
133: cover), the same lateral width of $5\,\mu m$, and varying barrier
134: thickness. The resistance versus temperature of the junctions was
135: measured using the standard four probe technique, and the dynamic
136: resistance was measured using a standard ac modulation
137: technique.\\
138:
139: \begin{figure}
140: \includegraphics[height=6cm,width=8.5cm]{Fig2.ps}
141: \caption{\label{fig:epsart} Resistance versus temperature of three
142: ramp junctions with 10.5, 21 and 32 nm thick Ga-doped YBCO
143: barrier, annealled in 8, 8 and 27 mTorr of oxygen pressure,
144: respectively. The three junctions were patterned along the node
145: direction at an angle of $45^{0}$ to the main axis. The junction
146: with the 10.5nm thick barrier had longer leads which led to a
147: higher normal resistance. }
148: \end{figure}
149:
150: \maketitle{\large RESULTS AND DISCUSSION}\\
151:
152: Fig. 2 shows the measured resistance as a function of temperature
153: of three node junctions with 10.5, 21 and 32 nm thick barriers.
154: The first two junction were annealled in 8 mTorr of flowing
155: oxygen, and the last one in 27 mTorr oxygen flow, respectively.
156: The corresponding normal state resistance of these three junctions
157: is typical of underdoped $YBa_{2}Cu_{3}O_y$ with $y\sim6.55$ for
158: the low oxygen annealled ones and $y\sim6.85$ for the richer
159: oxygen junction.\cite{Segawa} We note that the higher normal state
160: resistance of the junction with the thinnest barrier is due to its
161: longer leads (a different photolithographic mask pattern was
162: used). One can observe two distinct transitions in the resistance
163: of each junction. In the two oxygen deficient junctions (8 mTorr
164: annealing), the transition temperature $T_c$ of the electrodes
165: occurs at approximately 55K, while in the third oxygen rich
166: junction (27 mTorr annealing) the electrodes become
167: superconducting already at 80K. The transitions seen at 40 and 25K
168: in the first two junctions, and at 60K in the third one, result
169: presumably from an apparent proximity effect in the barrier of the
170: different junctions. At low temperatures, the junction with the
171: 10.5nm thick barrier shows a critical current of about 0.5mA at 5K
172: which yields a critical current density of $\sim 1.1\times
173: 10^4\,A/cm^{2}$. The other two junctions with the 21 and 32nm
174: thick barriers are resistive at low temperatures and have
175: resistance values at low bias of about 200 and 2$\Omega$,
176: respectively. In these two cases, the barriers exhibit insulating
177: behavior with the resistance either increasing slightly with
178: decreasing temperature or staying almost
179: constant.\\
180:
181: \begin{figure}
182: \includegraphics[height=6cm,width=8.5cm]{Fig3.ps}
183: \caption{\label{fig:epsart} Dynamic resistance spectra of the
184: junction in Fig. 2 with the 21 nm thick barrier and 8 mTorr oxygen
185: annealing pressure. }
186: \end{figure}
187:
188: \begin{figure}
189: \includegraphics[height=6cm,width=8.5cm]{Fig4.ps}
190: \caption{\label{fig:epsart} Dynamic resistance spectra of the
191: junction in Fig. 2 with the 10.5 nm thick barrier and 8 mTorr
192: oxygen annealing pressure (three traces, main panel, peak no. -3
193: is missing). In the inset, the results of another node junction on
194: the same wafer are shown (two traces, peak no. 1 is missing). }
195: \end{figure}
196:
197: The dynamic resistance spectra of the two oxygen deficient node
198: junctions with 21 and 10.5nm thick barriers are shown in Figs. 3
199: and 4, respectively. The different behavior at low bias, namely
200: tunneling like in Fig. 3, and critical current and zero bias
201: conductance peak (ZBCP) in Fig. 4, is due to the much higher
202: normal resistance of the first junction ($R_N(3K)\sim 200\Omega$).
203: We prepared another node junction with a 32nm thick barrier, but
204: annealled it in a higher oxygen pressure of 27mTorr. This was done
205: in order to avoid a very high $R_N(5K)$, and yielded $R_N(5K)\sim
206: 2\Omega$ (see Fig. 1). The dynamic resistance spectra of this
207: junction (not shown here) was similar to that of Fig. 3, but with
208: a ZBCP and a more closely spaced series of peaks. In all three
209: junctions, the position of the series of peaks on the voltage axis
210: in the dynamic resistance curves seem to be almost independent of
211: the oxygen content, or the presence of a ZBCP. It was however
212: strongly dependent on the thickness of the barriers. In order to
213: determine the origin of the series of peaks, we plotted the peak
214: voltages versus peak number for each of the three junctions. It is
215: generally nontrivial to associate a peak number to each peak
216: because not all of them are present or have the same intensity.
217: Some of the peaks are missing, some are enhanced while others can
218: also overlap and interfere with one another. Nevertheless, by
219: comparison with data of other node junctions, we could determine
220: the peak numbers properly, and in Fig. 5 we plot the peak voltages
221: of four series versus the peak number including the data of Figs.
222: 3 and 4. Fig. 5 shows that each series of peaks appears with a
223: constant voltage difference between adjacent peaks. Linear fits of
224: the four series yield $V_n=5.1+10.1n$ and $V_n=6+12n$ for the two
225: junctions in Fig. 4 with the thinnest barrier, $V_n=1.9+6.5n$ for
226: the junction in Fig. 3, and $V_n=1.9+4.5n$ for the fourth junction
227: with the 32nm thick barrier. The intersects at n=0 (5.1, 6, 1.9
228: and 1.9 mV) are of the same order of magnitude as the bias
229: voltages of the corresponding first peaks in the series (5, 6, 3.4
230: and 2.4 mV). Fig. 5 also shows that the adjacent peaks spacing is
231: sensitive to the barrier thickness. Its ratio in the three types
232: of junctions is 10.1-12:6.5:4.5 which is approximately equal to
233: the inverse ratio of the barrier thickness (1/10.5):(1/21):(1/32).
234: Since the superconducting electrodes in all our junctions have the
235: same thickness (90nm), the above result indicates that the series
236: of peaks in the dynamic resistance spectra originate in
237: geometrical resonances in the barrier layer. Next we discussed
238: this result in the context of the nature of the present S/MI/S
239: junctions (MI is Mott insulator with VRH).\\
240:
241:
242: \begin{figure}
243: \includegraphics[height=6cm,width=8.5cm]{Fig5.ps}
244: \caption{\label{fig:epsart} Peak voltages of the series in Figs. 3
245: and 4 versus the peak number, together with the series of peaks of
246: the oxygen rich junction in Fig. 2 with the 32 nm thick barrier.
247: The straight lines are linear fits to the data. }
248: \end{figure}
249:
250: It has already been demonstrated in the past that a-axis
251: YBCO/$PrBa_{2}Cu_{3}O_{7-\delta}$/YBCO junctions which are
252: basically S/MI/S junctions as we have here, carry significant
253: critical currents $I_c$ at low temperatures even when the barriers
254: are up to 100nm thick.\cite{Yoshida} Surprisingly for these kind
255: of junctions, it was found that $I_c(T)$ behaves as
256: exp$[-aT^{0.5}]$ where $a$ is a constant, which is exactly the
257: expected behavior for SNS type junctions with a normal metal
258: barrier in the dirty limit. There are many more reports on
259: observations of a long range proximity effect and Andreev
260: reflections in similar type of junctions with insulating VRH
261: barriers.\cite{Kabasawa,Adrian,Frydman} The puzzling question is
262: why such barrier materials with a resistivity of 0.1-1 $\Omega cm$
263: which is two to three orders of magnitude higher than $\sim 1
264: m\Omega cm$ the maximum resistivity a metal can have, when in
265: contact with a superconductor should behave like normal metals?
266: This is a complicated problem to deal with theoretically, and only
267: sketchy reports on this issue exist.\cite{Tanaka} We shall not
268: attempt to speculate here what is the reason for this normal metal
269: like behavior of the VRH barriers in the S/MI/S junctions, but
270: simply take it as given, and use as a first approximation the
271: formulas derived for SNS junctions. We note however, that unlike
272: the previous reports where the HTS electrodes of the S/MI/S
273: junctions were close to optimal
274: doping,\cite{Yoshida,Kabasawa,Adrian} in our case the underdoped
275: 60K YBCO based junctions yield critical current only for the
276: thinnest 10.5nm barrier. This can perhaps be related to the fact
277: that the present node junctions have a weaker superconductivity
278: coupling and shorter proximity penetration lengths, but detailed
279: investigation of this effect is outside of the scope of the
280: present study. We thus proceed with the treatment of our
281: S/MI/S junctions, along the line used with SNS junctions.\\
282:
283:
284:
285: We now discuss four possible types of geometrical resonances in
286: SNS junctions that can lead to similar series of peaks in the
287: dynamic resistance spectra like we observed here. Two are due to
288: sub-gap scattering processes of quasiparticles, and two to above
289: gap processes. The first two involve sub-harmonic resonances,
290: which are caused by multiple Andreev reflections, or
291: DeGennes-Saint James bound
292: states.\cite{NesherSubHarmonic,Andreev,DGSJboundStates} Both
293: result in series of peaks which are not equally spaced, and
294: therefore are not similar to the presently observed series.
295: Moreover, the junctions in the present study are oriented along
296: the node direction where the d-gap vanishes, and the s-gap is too
297: small to allow us to attribute the peaks at large voltage to
298: sub-gap structures. The second group of scattering processes that
299: leads to above gap series involves Tomasch and McMillan-Rowell
300: oscillations.\cite{Tomasch,MMR} The Tomasch oscillations are due
301: to resonances in the superconducting electrode, and their peak
302: energies are given by
303:
304: \begin{equation}\label{1}
305: eV_{n}=\sqrt{(2\Delta)^{2}+[\frac{nhv_{FS}}{2d_{S}}]^{2}},
306: \end{equation}
307:
308: \noindent where $\Delta$ is the superconducting gap, $v_{FS}$ is
309: the Fermi velocity in the electrodes, $d_{S}$ is the
310: superconducting electrode thickness, and n is the serial number of
311: the peak. Thus the resonances are not equally spaced, but for
312: small energy gap values the deviation from a constant voltage
313: difference between adjacent peaks is quite small, and generally
314: cannot be observed due to experimental error. McMillan-Rowell
315: oscillations (MRO),\cite{MMR} are also seen as series of
316: equidistant peaks in the dynamic resistance spectra above the gap,
317: and are caused by geometrical resonances of quasiparticles in the
318: barrier. The voltage difference between adjacent peaks in this
319: series is given by
320:
321: \begin{equation}\label{1}
322: \Delta V=\frac{hv_{FN}}{4ed_{N}},
323: \end{equation}
324:
325: \noindent where $v_{FN}$ is the Fermi velocity in the barrier, and
326: $d_{N}$ is the barrier thickness. We thus find that both the
327: Tomasch like and McMillan-Rowell like oscillations can yield the
328: linear behavior of the peak voltage versus peak number seen in
329: Fig. 5, provided the node gap is small compared to the adjacent
330: peak spacing. The voltage difference between adjacent peaks
331: however, depends on either the thickness of the superconducting
332: electrode in the Tomasch scenario as seen in Eq. (1), or on the
333: barrier thickness in the MRO case as seen in Eq. (2). Since the
334: thickness of the superconducting base and cover electrodes is the
335: same for all our junctions, and the observed series of peaks
336: depend on the thickness of the barrier as seen in Fig. 5, it seems
337: that these series are due to McMillan-Rowell like oscillations.
338: Compared to the original study of MRO,\cite{MMR} the presently
339: observed conductance peaks are much sharper, especially those in
340: Fig. 4. As mentioned above, the linear fits in Fig. 5 show that
341: the ratio of adjacent peaks spacing is approximately equal to the
342: ratio of the inverse thickness of the barriers. It thus follows
343: that scaling with the barrier thickness $d_N$ as depicted by Eq.
344: (2), is found here, and the observed resonances are due to
345: McMillan-Rowell like oscillations. We stress that this result is
346: independent of either the barrier strength or the different oxygen
347: content of the junctions. For a normal metal barrier, Eq. (2)
348: would have allow us to determine also the Fermi velocity of
349: quasiparticles in the barrier. In the present case of S/MI/S
350: junctions with SNS like behavior, analysis yields an effective
351: velocity $v_{FN} = 1.2\pm0.2 \times 10^{7}$cm/sec, which compares
352: well with a previous result of $1.5\times 10^{7}$ cm/sec measured
353: in the same kind of junction with a Fe-doped YBCO barrier. The
354: later has a much lower resistivity value at low temperatures, of
355: the order of $10-20 m\Omega cm$, thus being much closer to a
356: normal metal than the present Ga doped YBCO. A-priori the Fermi
357: velocity is not well defined here since there is no Fermi surface
358: at all in isolated VRH materials. When the thin VRH layer however,
359: is in contact with a superconductor like in the present junctions,
360: it is possible that the Fermi surface is recovered, and the Fermi
361: velocity is thus well defined. The fact that similar numbers were
362: obtained for $v_{FN}$ of the Fe and Ga doped YBCO barriers which
363: have very different resistivity values, further supports the
364: notion that the S/MI/S junctions have N like features. These
365: $v_{FN}$ values of course are only approximate ones. They are
366: quite smaller than the value $v_{F}\approx 2.5\times 10^7$
367: obtained by ARPES along the node direction in differently doped
368: LSCO crystals.\cite{Shen} This ARPES study also shows that there
369: is no direct link between the measured Fermi velocity in the
370: cuprates and the size of the resistivity.\\
371:
372: Since the McMillan-Rowell like oscillations occur at energies
373: above the gap, the voltage of the first peak in each series
374: constitutes an upper limit on the magnitude of the gap energy.
375: Previously, Nesher and Koren measured McMillan-Rowell oscillations
376: in junctions that were directed along a main crystallographic axis
377: of YBCO where the d-gap is at it's
378: maximum.\cite{NesherSubHarmonic} The first peak in their series
379: appears at 16 mV, which is an upper limit on the value of the
380: dominant d-wave component of the gap in the 55K phase of
381: underdoped YBCO. In the present study, on the contrary, the
382: junctions are aligned along the node direction where the dominant
383: d-wave gap vanishes, and we can thus measure an upper limit on the
384: gap energy of the sub-dominant component. The first peak (knee) in
385: Fig. 3 is found at 3.4 mV, and its voltage in Fig. 4 is 5-6 mV.
386: This yields an upper limit on the value of the s-gap in the range
387: of 3.4-6 meV, which is in reasonable agreement with the measured
388: gap values of 2.5$\pm$0.5 meV found
389: previously.\cite{Koren2002,KorenJLTP} We note however, that if we
390: take the average value of the intercepts of the four straight
391: lines in Fig. 5, we obtain a value of 3.7 mV. This represents a
392: four series average of the voltage of the first peak, which is
393: thus a more reliable upper limit on the energy of the s-gap
394: near the interface.\\
395:
396: \maketitle{\large CONCLUSIONS}\\
397:
398: The present study shows that the experimental properties of S/MI/S
399: junctions made of underdoped YBCO have several common features
400: with normal SNS junctions. Measurements of dynamic resistance
401: spectra in underdoped YBCO junctions along the node direction,
402: show geometrical resonances in the barrier which behave like
403: McMillan-Rowell oscillations above the gap. From these we find an
404: upper limit of 3.7 meV on the magnitude of the
405: $is$ component of the gap near the interface.\\
406:
407:
408: {\em Acknowledgments:} This research was supported in part by the
409: Israel Science Foundation, the Heinrich Hertz Minerva Center for
410: HTSC, the Karl Stoll Chair in advanced materials, and by the Fund
411: for the Promotion of Research at the Technion.\\
412:
413:
414:
415: \bibliography{AndDepBib.bib}
416: \begin{thebibliography}{99}
417: \label{Bib}
418:
419: \bibitem{Tsuei} C. C. Tsuei and J. R. Kirtley, Rev. Mod. Phys.
420: \textbf{72}, No 4, 969 (2000).
421: \bibitem{Covington} M. Covington, M. Aprili, E. Paraoanu,
422: L. H. Greene, F. Xu, J. Zhu and C. A. Mirkin, Phys. Rev. Lett.
423: {\bf 79}, 277 (1997).
424: \bibitem{Krupke} R. Krupke and G. Deutscher, Phys. Rev. Lett.
425: {\bf 86}, 4634 (1999).
426: \bibitem{Sharoni} Sharoni A., Koren G. and Millo O., {\em Europ. Phys.
427: Lett.}, {\bf 54} (2001) 675.
428: \bibitem{Koren2002} G. Koren and N. Levy, Europhys. Lett.
429: \textbf{59}, 121 (2002).
430: \bibitem{KorenJLTP} G. Koren, N. Levy and E. Polturak,
431: J. of Low Temp. Phys. {\bf 131}, 849 (2003).
432: \bibitem{DeutscherNature} G. Deutscher, Nature {\bf 397}, 410
433: (1999).
434: \bibitem{Krasnov} V. M. Krasnov, A. Yurgens, D. Winkler, P. Delsing
435: and T. Claeson, Physica C {\bf 352}, 89 (2001), and V. M. Krasnov,
436: A. E. Kovalev, A. Yurgens and D. Winkler, Phys. Rev. Lett. {\bf
437: 86}, 2657 (2001).
438: \bibitem{KorenPseudogap} G. Koren, L.Shkedy and E. Polturak,
439: Physica C, in press (2003). Cond-mat/0306594.
440: \bibitem{Lubimova} I. Lubimova and G.Koren, Phys. Rev. B, in press
441: (BE9094). Cond-mat/0306030 (2003).
442: \bibitem{NesherTomasch} O. Nesher and G. Koren, Appl. Phys. Lett. \textbf{74},
443: 3392 (1999).
444: \bibitem{NesherSubHarmonic} O.Nesher and G.Koren, Phys. Rev. B \textbf{60}, 9287
445: (1999).
446: \bibitem{Segawa} Y. Ando and K. Segawa, Phys. Rev.
447: Lett. {\bf 88}, 167005 (2002).
448: \bibitem{Yoshida} T. Hashimoto, M. Sagoi, Y. Mizutani, J. Yoshida
449: and K. Mizushima, Appl. Phys. Lett. {\bf 60}, 1756 (1992).
450: \bibitem{Kabasawa} U. Kabasawa, Y. Tarutani, M. Okamoto, T.
451: Fukazawa, A. Tsukamoto, M. Hiratani and K. Takagi, Phys. Rev.
452: Lett. {\bf 70}, 1700 (1993).
453: \bibitem{Adrian} C. Stozel, M. Siegel, G. Adrian, C. Krimmer,
454: J. Sollner, W. Wilkens, G. Schulz and H. Adrian, Appl. Phys. Lett.
455: {\bf 63}, 2970 (1993).
456: \bibitem{Frydman} A. Frydman and Z. Ovadyahu, Europhys. Lett. {\bf
457: 33}, 217 (1996).
458: \bibitem{Tanaka} Y. Tanaka, {\it Coherence in high temperature
459: superconductors}, editted by G. Deutscher and A. Revcolevschi,
460: World Scientific, Singapore (1995), pp 393-411.
461: %\bibitem{Gallaghere} W. Gallagher, Phys. Rev. B \textbf{22}, 1233
462: %(1980).
463: \bibitem{Andreev} A. F. Andreev, Sov. Phys. JETP {\bf 19}, 1228 (1964).
464: \bibitem{DGSJboundStates} P. G. de Gennes and D. Saint-James,
465: Phys. Lett. \textbf{4}, 151 (1963).
466: %\bibitem{KBT} T. M. Klapwijk, G. E. Blonder and M. Tinkham, Physica
467: %{\bf 109} and {\bf 110}B, 1657 (1982).
468: \bibitem{Tomasch} W. J. Tomasch, Phys. Rev. Lett. \textbf{15}, 672
469: (1965); {\em ibid} \textbf{16}, 16 (1966).
470: \bibitem{MMR} J. M. Rowell and W. L. McMillan, Phys. Rev. Lett.
471: \textbf{16}, 453 (1966); W. L. McMillan and P. W. Anderson, Phys.
472: Rev. Lett. \textbf{16}, 85 (1966).
473: \bibitem{Shen} Z. X. Shen, Cond-mat/0305576 (2003).
474:
475:
476:
477: \end{thebibliography}
478:
479:
480: \bibliography{apssamp}% Produces the bibliography via BibTeX.
481:
482: \end{document}
483: