1: \documentstyle[aps,prl,epsfig,floats]{revtex}
2: \begin{document}
3: \draft
4: \wideabs{
5: \title {Critical currents in Josephson junctions, with unconventional pairing symmetry: $d_{x^2-y^2}+is$ versus $d_{x^2-y^2}+id_{xy}$}
6: \author{ N. Stefanakis, N. Flytzanis}
7: \address{Department of Physics , University of Crete,
8: P.O. Box 2208, GR-71003, Heraklion, Crete, Greece}
9:
10: \date{\today}
11:
12: \maketitle
13:
14: \begin{abstract}
15: Phenomenological Ginzburg-Landau theory is used to
16: calculate the possible spontaneous vortex states that may exist
17: at corner junctions of $d_{x^2-y^2}+ix$-wave, (where $x=s$ or $x=d_{xy}$)
18: and $s$-wave superconductors. We study
19: the magnetic flux and the critical current modulation
20: with the junction orientation angle $\theta$, the magnitude
21: of the order parameter, and the magnetic field $H$.
22: It is seen that the critical current $I_c$
23: versus the magnetic flux $\Phi$ relation is symmetric / asymmetric
24: for $x=d_{xy}/s$ when the orientation is exactly
25: such that the lobes of
26: the dominant $d_{x^2-y^2}$-wave order parameter points towards the two
27: junctions, which are at right angles for the corner junction.
28: The conclusion is that a measurement of the $I_c(\Phi)$ relation
29: may distinguish which symmetry ($d_{x^2-y^2}+is$ or
30: $d_{x^2-y^2}+id_{xy}$) the order parameter has.
31: \end{abstract}
32:
33: \pacs{}
34: }
35:
36: \section{Introduction}
37: One of the main questions in the research
38: activity on high-$T_c$ superconductors nowadays is the identification
39: of the order parameter symmetry and its underlying mechanism\cite{scalapino,vanh}.
40: The most possible scenario is that the
41: bulk pairing state has a $d_{x^2-y^2}$-wave character.
42: Theoretical calculations, suggest
43: that an imaginary $s$-wave component which breaks the time reversal
44: symmetry is induced in some cases, wherever the $d_{x^2-y^2}$-wave
45: order parameter varies spatially such as near a vortex,
46: or near the surface
47: \cite{shiba}.
48: Also the observation of fractional vortices on a triangular grain boundary in
49: YBa$_2$Cu$_3$O$_7$ by Kirtley $et$ $al.$ \cite{kirtley2},
50: may indicate a possible violation
51: of the time-reversal symmetry near grain boundary.
52: Theoretical explanation of this experiment is given by Bailey et. al. in Ref. \cite{bailey}
53: where they study a triangular grain boundary in $d$-wave
54: superconductors. They conclude that under the assumption of
55: $d$-wave symmetry, the flux at the edges of this triangle
56: can take the values $\pm \Phi_0/2$, which does not agree with the
57: experiment.
58: However under the assumption of $d_{x^2-y^2}+is$-wave symmetry
59: an intrinsic phase shift $\phi_c(x)$ exists in each
60: triangle edge. In turn the phase $\phi(x)$
61: must change in order to connect the different values of
62: $\phi_c$ in each segment. This arrangement leads to
63: fractional vortices or antivortices at each three corners,
64: in agreement with the experiment.
65:
66: Another pairing state which breaks the time reversal symmetry is
67: the $d_{x^2-y^2}+id_{xy}$-wave.
68: Patches of complex $d_{xy}$ components are induced around
69: magnetic impurities at low temperatures in a $d_{x^2-y^2}$-wave superconductor
70: forming a phase coherent state as a result of tunneling between
71: different patches \cite{balatsky1}.
72: Violation of parity and time reversal symmetry occurs in this state.
73: Also on the high field region, $H\le H_{c2}$ the $d_{x^2-y^2}$-wave
74: state can be perturbed by the external filed, producing a
75: $d_{x^2-y^2}+{\it i}d_{xy}$ state in the bulk \cite{balatsky2}.
76:
77: The observation of the splitting of the zero energy
78: peak in the conductance spectra at low temperatures indicates
79: that a secondary component is
80: induced which violates locally the time reversal symmetry
81: \cite{covington}. Theoretical
82: explanation based on surface-induced Andreev states,
83: has been proposed \cite{fogelstrom}.
84: Recently the field dependence of this splitting has been
85: observed in the tunneling spectra of YBCO \cite{aprili,krupke}.
86: This observation
87: is consistent with a $d_{x^2-y^2}+is$ surface order parameter
88: as well as a $d_{x^2-y^2}+id_{xy}$ bulk order parameter.
89: Another question which can be asked is to what extend, the
90: observation of a symmetric magnetic interference pattern in
91: the corner junction experiments \cite{vanh} is an identification
92: of $d_{x^2-y^2}$-wave symmetry, or could also imply a
93: $d_{x^2-y^2}+id_{xy}$ pairing state also?
94: In this work we propose a phase sensitive experiment
95: based on the Josephson effect, which may be used to
96: distinguish which symmetry ($d_{x^2-y^2}+is$ or $d_{x^2-y^2}+id_{xy}$)
97: the order parameter has near
98: the surface.
99: We study the static properties of a frustrated
100: junction which is made of two one-dimensional junctions,
101: of $d_{x^2-y^2}+ix$-wave, (where $x=d_{xy}$
102: or $x=s$) and $s$-wave superconductors.
103: By introducing an extra relative phase in one part of this
104: junction, the above junction can be mapped into the corner junctions
105: experiments\cite{vanh,yanoff}.
106: We examine the spontaneous flux and the critical current modulation
107: of the vortex states with the
108: junction orientation angle $\theta$, the magnitude
109: of the secondary component $n_s$, and the magnetic field $H$.
110: In each case we derive simple arguments which are useful to
111: discriminate between the time reversal symmetry broken states.
112: For example, when the orientation is exactly
113: such that the lobes of
114: the dominant $d_{x^2-y^2}$-wave order parameter points towards
115: the junction interface the magnetic interference pattern
116: is symmetric (asymmetric) when the secondary order parameter
117: is $x=d_{xy}(s)$.
118: This is verified for small junctions
119: as well as in the long junction limit, and can be used to
120: distinguish between broken time reversal symmetry states.
121:
122: The rest of the paper is organized as follows. In Sec. II we
123: discuss the Josephson effect between a superconductor with
124: broken time reversal symmetry and an $s$-wave superconductor.
125: In Sec. III the geometry of the corner junction is
126: discussed. In Sec. IV
127: we present the results for the magnetic flux of the spontaneous
128: vortex states in corner junctions with some intrinsic magnetic flux.
129: In Sec. V the parameters which can modulate the spontaneous
130: flux and the critical currents are considered.
131: In Sec. VI a connection with the experiment is made.
132: Finally, a summary and discussion are presented in the last section.
133:
134: \section{Josephson effect between two superconductors with mixed wave symmetry}
135: \begin{figure}
136: \centerline{\psfig{figure=fig1.eps,width=6.5cm,angle=0}}
137: \caption{
138: (a) A single Josephson junction between superconductors $A$ and
139: $B$ with a two component order parameter. The angle between the
140: crystalline $a$ axis of $A$ and the junction interface is
141: $\theta$. (b) The geometry of the corner junction between
142: a mixed symmetry superconductor, and an $s$-wave superconductor.}
143: \label{fig1.fig}
144: \end{figure}
145:
146: We consider the junction shown in Fig. \ref{fig1.fig}(a), where two superconductors
147: ($A$ in the region $z>t$ and $B$ in the region $z<0$),
148: are separated by
149: an intermediate layer. We assume that each superconductor has a
150: two component order parameter. The order parameter for each
151: component $k (k=1,2)$ in the superconductors, can be written as
152: \begin{equation}
153: n_k = \left\{
154: \begin{array}{ll}
155: \widetilde{n}_k^Ae^{{\it i} \phi_k^A}
156: , & z > t \\
157: \widetilde{n}_k^Be^{{\it i} \phi_k^B}
158: , & z < 0
159: \end{array}.~~~\label{ni}
160: \right.
161: \end{equation}
162: Here $\phi_k^{A(B)}$ is the phase of the order
163: parameter $n_k$ in superconductor $A(B)$. Then phenomenological Ginzburg-Landau theory is used to
164: calculate the supercurrent density given by
165: \cite{zhu}
166: \begin{equation}
167: J=\sum_{k,l=1}^2 J_{ckl}\sin(\phi_k^B-\phi_l^A),~~~\label{j}
168: \end{equation}
169: where
170: \begin{equation}
171: \begin{array}{ll}
172: J_{c11} = & (2 e \hbar / m_a^{*} d) \widetilde{n}_1^A \widetilde{n}_1^B\\
173: J_{c21} = & (2 e \hbar / m_{\nu}^{*} d) \widetilde{n}_1^A \widetilde{n}_2^B\\
174: J_{c12} = & (2 e \hbar / m_{\nu}^{*} d) \widetilde{n}_2^A \widetilde{n}_1^B\\
175: J_{c22} = & (2 e \hbar / m_b^{*} d) \widetilde{n}_2^A \widetilde{n}_2^B
176: \end{array},~~~\label{jkl}
177: \end{equation}
178: $m_a^{*}, m_{\nu}^{*}, m_b^{*}$ are the effective masses that enter into the
179: Ginzburg-Landau equations. In the following these masses are taken equal to
180: an effective mass $m^{*}$.
181:
182: We restrict to the case where $B$ is $s$-wave. In this case
183: $\widetilde{n}_1^B=0$, and
184: $\widetilde{n}_2^B$=constant.
185: We define
186: $\phi=\phi_2^B-\phi_1^A$, as the relative phase difference
187: between the two superconductors. We consider the case where the intrinsic
188: phase difference within superconductor $A$ is $\phi_2^A-\phi_1^A=
189: \pi/2$. Then the order parameter in $A$ is complex and breaks the
190: time reversal symmetry. The supercurrent density can be written as:
191:
192: \begin{equation}
193: J(\phi)=\widetilde J_c \sin(\phi+ \phi_c),~~~\label{jphidis}
194: \end{equation}
195: with
196: \begin{equation}
197: \widetilde
198: J_c=\sqrt{J_1^2+J_2^2},~~~\label{jcdis}
199: \end{equation}
200:
201: \begin{equation}
202: \phi_c = \left\{
203: \begin{array}{ll}
204: \tan^{-1}\frac{J_2}{J_1}, & J_1 > 0 \\
205: \pi + \tan^{-1}\frac{J_2}{J_1}, & J_1 < 0
206: \end{array},~~~\label{phic}
207: \right.
208: \end{equation}
209: where $J_1=J_{c21}$, $J_2=-J_{c22}$. The Josephson critical current
210: density $\widetilde J_c$ is scaled in units of
211: $J_{c0}=\frac{e \hbar}{m^{*} d}$.
212: Two special cases are the following:
213:
214: i) For $d_{x^2-y^2}+{\it i}s$ wave case
215: the magnitude of the
216: $d_{x^2-y^2}$-wave component in (\ref{ni}) is $\widetilde{n}_1^{A}=
217: n_{10}\cos(2\theta)$
218: , where $\theta$ is the angle of the crystalline $a$-axis
219: of superconductor $A$ with the junction interface.
220: The magnitude of the secondary order parameter in superconductor
221: $A$ is $\widetilde{n}_2^{A}=
222: n_{20}=0.1n_{10}$.
223:
224: ii) For $d_{x^2-y^2}+{\it i}d_{xy}$ wave case,
225: the magnitude of the $d_{x^2-y^2}$-wave component
226: in (\ref{ni}) is given by $\widetilde{n}_1^{A}=
227: n_{10}\cos(2\theta)$, while the $d_{xy}$ wave component is
228: $\widetilde{n}_2^{A}=n_{20}\sin(2\theta)$, where $n_{20}=0.1n_{10}$.
229: This order parameter
230: can occur in the following way: The order parameter magnitude for the
231: $d$-wave state $\Delta_0(\theta)=\Delta_0\cos(2\theta)$
232: is an equal admixture of pairs with orbital moment
233: $L_z=\pm2$, and can be written as $\Delta_0(\theta)=
234: (\Delta_0/2)[\exp(2{\it i}\theta)+\exp(-2{\it i}\theta)]$.
235: In the presence of perturbation such as (ferromagnetically) ordered impurity
236: spins $S_z$ the coefficients of $L_z=\pm2$ components will shift
237: linearly in $S_z$ with opposite signs. The final state will be
238: $\Delta_0(\theta)\rightarrow
239: \Delta_0(\theta)+{\it i}S_z\Delta_1(\theta)$, where
240: $\Delta_1(\theta)=\sin(2\theta)$. The strength of the secondary
241: component is proportional to the perturbation $S_z$.
242:
243: \section{The corner junction geometry}
244:
245: We consider the corner junction shown in Fig. \ref{fig1.fig}(b), between
246: a superconductor with broken time reversal symmetry at the surface
247: and an s-wave superconductor.
248: If the angle of $a$-axis with the interface in the $x$-direction
249: is $\theta$, then the corresponding angle in the $z$-direction
250: will be $\pi/2-\theta$. We map the two segments each of length $L/2$
251: where $L=10\lambda_J$ of this junction
252: into a one-dimensional axis. In this case the two dimensional junction can
253: be considered as being made of two one dimensional junctions described
254: in Sec. II connected in parallel. Their characteristic phases
255: $\phi_{c1}$ and $\phi_{c2}$ depend upon the angle $\theta$.
256: We call this junction frustrated since the two segments have
257: different characteristic phases $\phi_{c1}, \phi_{c2}$.
258: The fabrication details of
259: corner junctions or superconducting quantum interference device (SQUID),
260: between sample faces at different angles
261: can be found in Ref. \cite{vanh,yanoff}.
262:
263: The superconducting phase difference $\phi$ across the junction is
264: then the solution of the sine-Gordon (s-G) equation
265: \begin{equation}
266: \frac{d^2 { \phi}(x)}{dx^2} = \widetilde J_c \sin[{
267: \phi(x)+\phi_c(x)}] - I^{ov},~~~\label{eq01}
268: \end{equation}
269: with the boundary conditions
270: \begin{equation}
271: \frac{d \phi}{dx}|_{x=0,L} = H
272: .~~~\label{eqbc}
273: \end{equation}
274: The length $x$ is scaled in units of the
275: the Josephson penetration depth given by
276: \[
277: \lambda_J=\sqrt{\frac{\hbar c^2}{8\pi e d J_{c0}}},
278: \]
279: where $d$ is the sum of the $s$-wave, and mixed wave $\lambda_{ab}$
280: penetration depths plus the thickness of the
281: insulator layer.
282: The relative phase
283: $\phi_c(x)$ is $\phi_{c1} (\phi_{c2})$ in the left (right) part of the
284: junction.
285: The external magnetic field $H$, scaled in units of $H_c=\frac{\hbar c}{2 e d \lambda_J}$
286: is applied in the $y$ direction, which
287: is considered small compared to $\lambda_J$.
288: The bias current per unit length $I^{ov}$ in the overlap geometry
289: is scaled in units of $\frac{c}{4\pi}H_c$,
290: and is uniformly distributed
291: along the entire $x$ axis of the junction.
292:
293: We can classify the different solutions obtained from
294: Eq. (\ref{eq01}) with their magnetic flux content
295: \begin{equation}
296: \Phi = \frac{1}{2 \pi} (\phi_R-\phi_L) ,~~~\label{phi}
297: \end{equation}
298: where $\phi_{R(L)}$ is the value of $\phi$ at the right(left) edge of
299: the junction, in units of the flux quantum
300: $\Phi_0= \frac{h c}{2 e}$.
301:
302: \section{Spontaneous vortex states}
303:
304: \begin{figure}
305: \centerline{\psfig{figure=fig2.eps,width=6.5cm,angle=0}}
306: \caption{
307: The stable solutions $\phi = - \phi_{c1}+2n_1\pi$
308: ($\phi = - \phi_{c2}+2n_2 \pi$), for $n_i=0,-1$, $i=1,2$, that exist in the
309: left(right) junction,
310: of $d_{x^2-y^2}+is$-wave and $s$-wave superconductors, when considered
311: uncoupled, at zero current, versus the orientation angle $\theta$.
312: Each junction has length $L/2$, where $L=10\lambda_J$, and
313: $\phi_{c1}, (\phi_{c2})$
314: is the extra phase difference in the left (right) junction due to the
315: different orientations.
316: The arrows denote the variation of the phase $\phi$ in order to
317: connect these stable solutions in the frustrated junction geometry.
318: We present three possible solutions i.e.
319: $n=0,-1,1$, and
320: down(up) arrow denotes
321: negative(positive) magnetic flux.
322: }
323: \label{fig2.fig}
324: \end{figure}
325:
326: Firstly let us consider the case where the two one-dimensional
327: junctions of $d_{x^2-y^2}+ix$-wave where $x=s$ or $x=d_{xy}$,
328: and $s$-wave supeconductors,
329: each of length $L/2$, described in Sec. II are uncoupled.
330: Then for $0<x<L/2$ the stable solutions for the s-G equation are
331: $\phi(x)=-\phi_{c1}+2n_1 \pi$, where $n_1=0,\pm 1,\pm 2,...$,
332: while for
333: $L/2<x<L$ the stable solutions for the s-G equation are
334: $\phi(x)=-\phi_{c2}+2n_2 \pi$, where $n_2=0,\pm1,\pm2,...$,
335: where $\phi_{c1}$, $\phi_{c2}$, are the relative phases in
336: each part of the junction due to different orientations.
337: These solutions are plotted in Fig. \ref{fig2.fig}, for $n_i=0, -1$, $i=1,2$
338: as a function of the orientation angle $\theta$.
339: When the frustrated junction is formed, and we consider the
340: above junctions in parallel, the phase $\phi$ is forced to
341: change around $x=L/2$, to connect these stable solutions.
342: This variation of the phase $\phi$, along the junction
343: describes the Josephson vortices.
344: The flux content of these states (in units of $\Phi_0$) is \cite{sigrist}
345: \begin{equation}
346: \Phi=[\phi(L)-\phi(0)]/2\pi =
347: (-\phi_{c2}+\phi_{c1}+2n\pi)/2\pi,~~~\label{con}
348: \end{equation}
349: where the $n$-value ($n=n_1-n_2=0,\pm 1,\pm 2,...$)
350: distinguishes between solutions with
351: different flux content. We will
352: concentrate to solutions called modes with the minimum flux content i.e.,
353: $n=0, 1, -1$.
354: Their magnetic flux in terms of $\phi_{c1}, \phi_{c2}$
355: is shown in table \ref{n=01_1}.
356: Generally the flux content is fractional i.e. is neither integer
357: nor half-integer, as a consequence of the broken time reversal symmetry
358: of the problem.
359:
360: \begin{table}
361: \caption{
362: The magnetic flux ($\Phi$) in terms of $\phi_{c1}$, $\phi_{c2}$
363: for the spontaneous solutions that exist in the corner junction
364: geometry between a superconductor with time reversal broken symmetry
365: and an $s$-wave superconductor ($\phi_{c1}, \phi_{c2}$
366: is the extra phase difference in the two edges of the corner junction due to the
367: different orientations, of the $a$-axis of the dominant $d_{x^2-y^2}$-wave
368: superconductor). We present only the minimum
369: flux states $n=0,-1,1$.
370: }
371: \begin{tabular}{cc}
372: Vortex state $n$ & Magnetic flux ($\Phi$)\\ \tableline
373: $0$ & $(-\phi_{c2}+\phi_{c1})/2\pi$\\
374: $1$ & $(-\phi_{c2}+\phi_{c1}+2\pi)/2\pi$\\
375: $-1$ & $(-\phi_{c2}+\phi_{c1}-2\pi)/2\pi$\\
376: \end{tabular}
377: ~~~\label{n=01_1}
378: \end{table}
379:
380: In the actual numerical simulations, the stable solutions of
381: the sine-Gordon equation in the left(right) part of the
382: junction are taken as the initial conditions for the iteration
383: procedure. For example for the $n=0$
384: solution the phase $\phi(x)$ is taken $\phi(x)=-\phi_{c1}$
385: $(-\phi_{c2})$ in the left (right) part of the junction, as an
386: initial condition and then is iterated until convergence.
387: Besides if we take as initial condition
388: , $\phi(x)=-\phi_{c1}$, in the left side, and
389: $\phi(x)=-2\pi-\phi_{c2}$ in the right side,
390: the final state of the system, after the
391: iteration procedure, is the solution which we call $n=-1$,
392: with negative magnetic flux, and not exactly opposite
393: to $n=0$. We comment here that the solutions after the
394: iteration procedure have smooth variation as a function
395: of the position, as opposed to the step function variation
396: of the initial conditions.
397:
398: For the $0-0$ junction $\phi_{c1}=\phi_{c2}=0$, and the flux becomes
399: $\Phi=n$, so we say that the flux is quantized in integer units
400: of $\Phi_0$.
401: In this case, there exist solutions with flux
402: $\Phi=...,-1,0,1,...$ \cite{caputo}. These solutions,
403: when $n \neq 0$ are stabilized
404: by the application of an external magnetic field.
405: In the case of a junction with some spontaneous flux,
406: at least for the modes with lower flux content,
407: the external field is not necessary
408: since the spontaneous magnetization state is stable.
409:
410: In the case of $0-\pi$ junction, where the intrinsic phase in the
411: right (left)
412: part of the junction is $\phi_{c2}=-\pi$ ($\phi_{c1}=0$), the stable
413: solutions of the s-G equation are $\phi(x)=2n\pi$ for the left part, while
414: $\phi(x)=\pi(2n+1)$ for the right part of the junction. In this case
415: a $0-\pi$ junction is formed. The corresponding flux
416: becomes
417: $\Phi=(n+1/2)\pi$, and the particular values of $n=0$, $n=-1$
418: give the half vortex and antivortex solutions, with opposite fluxon
419: content, $\Phi=0.5$ and $\Phi=-0.5$ respectively.
420:
421: \section{Magnetic flux and critical current modulation}
422:
423: In the following we will describe three parameters which
424: can alter the spontaneous flux and the critical currents
425: of the vortex states described
426: in the previous section, in a corner junction between a superconductor
427: with time reversal broken symmetry and an $s$-wave superconductor.
428: These include the orientation angle $\theta$,
429: the magnitude of the secondary order parameter $n_s$,
430: and the magnetic field $H$. In each parameter separately
431: we will point out the differences between the
432: $d_{x^2-y^2}+is$-wave, and $d_{x^2-y^2}+id_{xy}$-wave.
433: \subsection{Junction orientation}
434: For the $d_{x^2-y^2}+{\it i}s$-wave case, we consider first the situation where $\theta$
435: is varied from $0$ to $\pi/2$.
436: In Fig. \ref{fig3.fig} we plot the spontaneous magnetic flux versus the angle
437: ($\theta$) for the different modes $n=0,-1,1$
438: in the corner junction geometry.
439: As we can see the magnetic flux changes with orientation.
440: For angle $\theta$ close to $0$ or $\pi/2$ the spontaneous modes
441: existing at $H=0$ are separated by an integer value of the
442: magnetic flux. This is also the case in the pure $s$-wave
443: superconductor junction
444: problem. The difference is that the modes are found displaced
445: to fractional values of magnetic flux, contrary to the
446: $s$-wave case where the magnetic flux takes on integer
447: values at $H=0$.
448: In particular the vortex solution in the $n=0$ mode (solid line) contains
449: less that half a fluxon for $\theta=0$, and as we increase the
450: angle $\theta$ towards $\pi /4$ it continuously reduces its flux,
451: i.e. it becomes flat exactly at $\theta=\pi /4$ and then it
452: reverses its sign and becomes an antivortex with exactly opposite
453: flux content at $\theta=\pi /2$ from that at $\theta=0$.
454: In addition we have plotted in Fig. \ref{fig4.fig}a
455: the phase distributions for the mode $n=0$ in different
456: orientations $\theta=0, \pi /4, \pi /2$. The transition form the vortex
457: to the antivortex mode as the orientation changes is clearly
458: seen in this figure.
459: Note that the solutions in this mode remain stable for all
460: junction orientations. This is seen in Fig. \ref{fig5.fig} where we plot
461: the lowest eigenvalue ($\lambda_1$) of the linearized eigenvalue
462: problem as a function of the angle $\theta$ \cite{caputo}.
463: We see that $\lambda_1>0$, denoting stability for all values
464: of the angle $\theta$ in this mode.
465:
466: \begin{figure}
467: \centerline{\psfig{figure=fig3.eps,width=6.5cm,angle=0}}
468: \caption{
469: The magnetic flux $\Phi$ as a function of the
470: angle $\theta$, for the various vortex states,
471: $n=0,-1,1$, that exist spontaneously
472: in a corner junction between a
473: $d_{x^2-y^2}+is$-wave and an $s$-wave superconductor, with length
474: $L=10\lambda_J$. The flux for $\theta=0$
475: is fractional.
476: }
477: \label{fig3.fig}
478: \end{figure}
479:
480: \begin{figure}
481: \centerline{\psfig{figure=fig4.eps,width=6.5cm,angle=0}}
482: \caption{The phase distribution of the vortex solutions a) $n=0$,
483: at $\theta= 0$, $\pi /4$, $\pi /2$;
484: b) $n=-1$, at $\theta= 0$, $0.242 \pi$, where the instability sets in,
485: and $\pi /2$;
486: c) $n=1$, at $\theta=0$, $0.258 \pi$, at the point where the
487: instability occurs, and $\pi /2$,
488: for a corner junction of $d_{x^2-y^2}+is$-wave and $s$-wave
489: superconductors, with length
490: $L=10\lambda_J$, and zero overlap external current $I^{ov}=0$.}
491: \label{fig4.fig}
492: \end{figure}
493:
494: Let as now examine the solution in the $n=-1$ mode, (dotted line
495: in Fig. \ref{fig3.fig}). We see that at
496: $\theta=0$ it has negative flux, which in absolute value is
497: more than $\Phi_0 /2$ and as we increase the angle $\theta$
498: it decreases its flux to a full antifluxon when the
499: orientation is slightly greater than $\pi /4$ and than to flux
500: greater than $\Phi_0$ when $\theta$ reaches $\pi /2$.
501: As seen in Fig. \ref{fig5.fig} this solution becomes unstable at a point
502: to the left of $\theta= \pi /4$ (point $\iota$) due to the abrupt change of
503: flux at this angle. More strictly the instability sets in
504: due to the competition between the slope of the phase at
505: the edges of the junction and at the junction center as the
506: angle $\theta$ approaches the value $\pi /4$. At this point
507: the slope competition makes
508: the antivortex unstable.
509: This is seen in Fig. \ref{fig4.fig}b) (dotted line)
510: where the phase distribution for the
511: $n=-1$ mode solution is plotted
512: at the point where the instability
513: starts i.e. $\theta=0.242 \pi$.
514:
515:
516: \begin{figure}
517: \centerline{\psfig{figure=fig5.eps,width=6.5cm,angle=0}}
518: \caption{
519: The lowest eigenvalue $\lambda_1$ of the linearized eigenvalue
520: problem
521: as a function of angle $\theta$, for the $n=0,-1,1$
522: solutions. In the range where $\theta$ is close to zero, the eigenvalues
523: for both
524: $n=0$, and $-1$ are positive and correspond to stable solutions.
525: }
526: \label{fig5.fig}
527: \end{figure}
528:
529: Finally the solution in the $n=1$ mode contains more than one
530: fluxon at $\theta=0$ and is clearly unstable. It becomes stable
531: at an angle slightly on the right of $\theta=\pi /4$, (point $\nu$ in
532: Fig. \ref{fig5.fig})
533: where
534: the flux varies more smoothly, see
535: $\theta=0.258 \pi$ in Fig. \ref{fig4.fig}c.
536: At $\theta=\pi /2$ it contains more than $\Phi_0 /2$ in flux.
537: We expect a time reversal broken symmetry state like
538: $d_{x^2-y^2}+is$ to be characterized by either
539: the solution in the fractional vortex or antivortex mode,
540: because due to the different character of these solutions
541: a change from one variant to the other would demand the application
542: of an external current or magnetic field and in this sense it
543: would cost additional energy. So since these states are stable in external
544: perturbations, once the system is prepared in one of
545: these it will remain to that state.
546:
547: \begin{figure}
548: \centerline{\psfig{figure=fig6.eps,width=6.5cm,angle=0}}
549: \caption{
550: Overlap critical current $I_c^{ov}$ per unit length versus the angle $\theta$
551: for a corner junction of $d_{x^2-y^2}+is$-wave and $s$-wave superconductors,
552: with length
553: $L=10\lambda_J$,
554: for the vortex solutions $n=0,-1,1$
555: that exist spontaneously in the junction.
556: }
557: \label{fig6.fig}
558: \end{figure}
559:
560: In general we see that for each value of $\theta$ there exist in
561: the junction a pair of stable solutions which when
562: applying an external bias current will lead
563: to observable critical currents.
564: In Fig. \ref{fig6.fig} we plot the overlap critical current per unit
565: length $I_c^{ov}$ as a function
566: of $\theta$, at $H=0$, for the $n=0,-1,1$-mode solutions, in
567: the $d_{x^2-y^2}+is$-wave case.
568: In the overlap geometry the current is distributed in the entire $x$-axis.
569: In the calculations we have taken into account that the Josephson critical
570: current density $\widetilde J_c$ has a characteristic variation with
571: the orientation.
572: We find that
573: for a given orientation
574: it is possible for the junction current density to vary in the
575: way that several modes with different critical currents
576: can exist.
577: In Fig. \ref{fig7.fig} we plot
578: the current density when the total current is maximum, for different
579: modes, and orientations,
580: which will give us information about the actual shapes
581: of the vortices.
582: Let us consider the situation where the
583: junction contains a solution in the mode $n=0$, at $\theta=0$,
584: when the net current is maximum.
585: The spatial variation of $\phi$ is described by a fractional vortex
586: which is displaced around the value $\phi= \pi$, from the
587: corresponding distribution at zero current which is around $\pi /2$
588: (see Fig. \ref{fig4.fig}a).
589: The current density distribution as seen in Fig. \ref{fig7.fig}a (solid line)
590: at the maximum current is flat
591: above unit with a small variation around the junction center giving
592: rise to the large value on the net current, seen in Fig. \ref{fig6.fig}.
593: Also at $\theta= \pi/4$ the flat phase distribution corresponding
594: to the $n=0$ solution at zero current is displaced towards the
595: value $\phi= \pi$ when applying an external current.
596: The corresponding current distribution seen in Fig. \ref{fig7.fig}a, (dotted line)
597: is straight line and the net current is small
598: for this orientation.
599: For the $n=-1$ solution at the point where the instability sets in
600: i.e. $\theta=0.242 \pi$,
601: the current density distribution is symmetric around zero as
602: seen in Fig. \ref{fig7.fig}b (dotted line) and carries
603: zero net current at this point. Thus the instability occurs just
604: before the angle where a full antifluxon enters the junction. A
605: slightly different situation occurs in the magnetic interference
606: pattern of a pure $s$-wave superconductor junction \cite{owen}
607: where, the net current is zero at the
608: magnetic field where a full fluxon or antiluxon enters the junction,
609: in the no flux $0$-mode.
610: At the point $\theta= \pi/2$, of the $n=-1$-mode the junction
611: contains more than
612: one fluxon causing the characteristic oscillations in the current
613: density around the junction center as seen in Fig. \ref{fig7.fig}b
614: (dashed line).
615: This reduces the critical current
616: for this orientation.
617:
618: \begin{figure}
619: \centerline{\psfig{figure=fig7.eps,width=6.5cm,angle=0}}
620: \caption{The current density distribution $J(x)$
621: of the vortex solutions a) $n=0$,
622: at $\theta= 0$, $\pi /4$, $\pi /2$,
623: b) $n=-1$, at $\theta= 0$, $0.242 \pi$, where the instability sets in,
624: and $\pi /2$,
625: c) $n=1$, at $\theta=0$, $0.258 \pi$, at the point where the
626: instability occurs, and $\pi /2$,
627: for a corner junction of $d_{x^2-y^2}+is$-wave and $s$-wave
628: superconductors, with length
629: $L=10\lambda_J$, and maximum external overlap current $I_c^{ov}$.}
630: \label{fig7.fig}
631: \end{figure}
632:
633: For the $d_{x^2-y^2}+{\it i}d_{xy}$ pairing symmetry state, we plot in
634: Fig. \ref{fig8.fig}a) the flux content for the $n=0,-1,1$, versus the
635: angle $\theta$.
636: Note the half integer or multiplies value of $\Phi$ at $\theta$
637: close to $0$ or $\pi/2$.
638: For this grain orientation
639: the magnetic flux is only sensitive to the real part of the
640: order parameter, which has a sign change but does not break
641: time-reversal symmetry.
642: In the $d_{x^2-y^2}+is$-wave state the order
643: parameter is complex for all junction orientations and
644: breaks the time-reversal symmetry. Close to $0$ or $\pi/2$
645: the flux is fractional.
646: The flux quantization at $\theta=0$ can be used to
647: discriminate between
648: these states.
649:
650: In Fig. \ref{fig8.fig}b) we plot the critical current per unit length
651: evolution
652: with the grain angle $\theta$ in the
653: $d_{x^2-y^2}+{\it i}d_{xy}$-wave state. Close to $\theta=0$ we see that
654: the $I_c^{ov}$ for the $n=0,-1$ solutions, coincide.
655: This
656: happens also at $\theta=\pi/2$ for the $n=0,1$ solutions.
657: In these orientations the order parameter becomes pure real
658: and does not break the time-reversal symmetry. As a result
659: the critical current at these angles is the same as in a junction with
660: pure $d$-wave symmetry.
661: At $\theta=\pi /4$ the order parameter is pure imaginary
662: and has the same magnitude
663: for both pairing states.
664: As a consequence
665: for $\theta=\pi/4$, the critical currents for both junctions
666: coincide.
667: Also the unstable part of the $n=1$ branch, in the $I_c$ vs $\theta$
668: is almost the same for the two symmetry states, due to the
669: small difference in the flux, compared with the
670: large flux content of the solutions in this region.
671:
672: \begin{figure}
673: \centerline{\psfig{figure=fig8a.eps,width=6.5cm,angle=0}}
674: \centerline{\psfig{figure=fig8b.eps,width=6.5cm,angle=0}}
675: \caption{
676: a) The spontaneous magnetic flux $\Phi$ as a function of the
677: angle $\theta$, for the various vortex states,
678: $n=0,-1,1$, for a corner junction between a
679: $d_{x^2-y^2}+id_{xy}$-wave and an $s$-wave superconductor, with length
680: $L=10\lambda_J$. The flux for $\theta=0$
681: is integer multiply of $\Phi_0 /2$.
682: b) The corresponding critical current $I_c^{ov}$ per unit length.
683: }
684: \label{fig8.fig}
685: \end{figure}
686:
687:
688: \subsection{Magnitude of the secondary order parameter}
689: In the above calculations the magnitude of the secondary order
690: parameter is small compared to the dominant (i.e. $n_{20}=0.1n_{10}$).
691: However the maximum fraction of the secondary component that has
692: been observed in phase coherent experiments employing different
693: materials, geometries, and techniques is up to $25\%$ of the
694: dominant \cite{vanh}. This triggered our interest to study the magnetic
695: flux and also the critical currents as a function of the
696: strength ($n_s$) of the secondary order parameter,
697: where the magnitude of the dominant order parameter $n_d$ is also
698: varied in a way that $n_s+n_d=1$. When $n_s=0$ only the $d_{x^2-y^2}$-wave
699: order parameter is present, while when $n_s=1$ only the $s$-wave
700: order parameter appears. This situation can be realized
701: for example near the surface where the $d_{x^2-y^2}$-wave order parameter
702: is suppressed and the $s$-wave order parameter is enhanced.
703: The result is presented in Fig. \ref{fig9.fig}a) and \ref{fig9.fig}b)
704: for the $d_{x^2-y^2}+is$-wave
705: case at $\theta=0$. We see that when the secondary component
706: is absent (i.e. $n_s=0$) the picture of the
707: $d_{x^2-y^2}$-wave state is reproduced. The same picture also
708: holds for the $d_{x^2-y^2}+id_{xy}$-wave state at $\theta=0$, since
709: the order parameter for the $d_{x^2-y^2}+id_{xy}$-wave state at
710: $\theta=0$ is real not breaking the time-reversal symmetry.
711: So
712: for $\theta=0$ the magnetic flux and the critical current
713: for the $d_{x^2-y^2}+id_{xy}$-wave state would not show
714: any change with the variation of the secondary order parameter
715: $d_{xy}$.
716: As $n_s$ is increasing the modes $n=0$ and $n=-1$ are
717: no more degenerate, in the sense that their flux deviates
718: from the value $\Phi_0/2$ and $-\Phi_0/2$ respectively
719: and also their critical currents are no
720: longer equal. The mode $n=0$ has larger critical current
721: because it has smaller flux content in absolute value.
722: For values of $n_s$ close to unity, the different modes
723: contain integer magnetic flux, as in the junction
724: between $s$-wave superconductors, and also their
725: critical currents have the same values
726: as in the perfect junction problem.
727: The conclusion is that the larger the secondary component
728: is in a sample the easier is to be detected in a flux measurement
729: experiment.
730:
731: \begin{figure}
732: \centerline{\psfig{figure=fig9a.eps,width=6.5cm,angle=0}}
733: \centerline{\psfig{figure=fig9b.eps,width=6.5cm,angle=0}}
734: \caption{
735: a) The spontaneous magnetic flux $\Phi$
736: and b) the critical current $I_c^{ov}$ per unit length
737: versus the strength $n_s$ of the
738: secondary $s$-wave component
739: for a corner junction of $d_{x^2-y^2}+is$-wave and
740: $s$-wave superconductors,
741: with length
742: $L=10\lambda_J$,
743: for the vortex solutions $n=0,-1,1$
744: that exist spontaneously in the junction. The magnitude $n_d$ of the
745: $d_{x^2-y^2}$-wave order parameter is given by the relation
746: $n_s+n_d=1$.
747: }
748: \label{fig9.fig}
749: \end{figure}
750:
751: \subsection{Magnetic field}
752: We now examine the influence of the magnetic field on the spontaneous
753: vortices for broken time reversal symmetry pairing states. In Fig. 10
754: we plot the magnetic flux at zero current versus the magnetic field $H$
755: for the $d_{x^2-y^2}+id_{xy}$-wave state at $\theta=0$. In the pure
756: $s$-wave superconductor junction
757: there is no overlap between different modes in the magnetic flux, and
758: each mode has magnetic flux which is more than $n\Phi_0$ and less
759: than $(n+1)\Phi_0$. In this problem due to spontaneous magnetization
760: the range of the modes is different and in some cases overlapping,
761: and the labeling
762: is with a single index $n$, corresponding to the
763: pure $s$-wave superconductor
764: junction (n,n+1) mode \cite{owen}. Moreover the range in magnetic flux
765: of each mode is displaced compared to the pure $s$-wave
766: superconductor junction
767: problem by an amount which corresponds to the intrinsic flux.
768: Also we have the existence of
769: stable vortex states i.e. $n=0,-1$, together with the unstable ones
770: i.e. $n=1$, $-2$ in a large interval of the magnetic field, which
771: is almost the same.
772: The $n=-2$ mode extends to zero magnetic field, and the reason we
773: didn't examined this mode in Sec. IV is because the stability analysis
774: shows negative eigenvalues for all the range of junction orientations,
775: at $H=0$.
776: In the long $s$-wave junction the extremum of the mode $(0,1)$ in $H$
777: is the critical field for one fluxon (antifluxon) penetration from the edges,
778: [denoted by $H_{cr}$ $(H_{cl})$, for the right (left) edge], and
779: is equal to $2 (-2)$. The solution for the phase at these extremum values
780: of the field becomes unstable because the value of the phase at
781: the junction edges reaches a critical value.
782: In the problem of a junction with some spontaneous flux, we
783: consider here, the range of the corresponding mode $0$ in $H$ is
784: significantly broadened and also the instability at the boundaries
785: sets in due to different reasons. In particular the instability
786: occurs due to the interaction of the flux entering from the
787: junction edges, when the magnetic field reaches the critical
788: value $H_{cr}(H_{cl})$, with the spontaneous flux at the center.
789: Similar features are encountered in the problem of flux pinning
790: from a macroscopic defect in a conventional $s$-wave junction. \cite{defect}
791:
792: \begin{figure}
793: \centerline{\psfig{figure=fig10.eps,width=6.5cm,angle=0}}
794: \caption{ Magnetic flux $\Phi / \Phi_0$ at zero external current versus the
795: magnetic field $H$
796: for a corner junction of $d_{x^2-y^2}+id_{xy}$-wave and $s$-wave
797: superconductors, with length
798: $L=10\lambda_J$, for angle $\theta=0^{\circ}$.
799: $H_{cl}(H_{cr})$ denotes the critical values of the magnetic field
800: where the mode $n=0$, terminates.
801: }
802: \label{fig10.fig}
803: \end{figure}
804:
805:
806: We now examine the magnetic-interference pattern for the two symmetries
807: where the bias current enters in the overlap geometry.
808: In the $d_{x^2-y^2}+id_{xy}$-wave case, where $\theta=0$,
809: this pattern has a symmetric form as we can see
810: from Fig. \ref{fig11.fig}(a).
811: This is because this result is only sensitive to the real
812: part of the order parameter, which has a sign change but
813: does not break time-reversal symmetry.
814: For the angle $\theta=0.5$ where the order parameter has a
815: finite imaginary part and breaks the time-reversal symmetry this
816: pattern becomes asymmetric and the ''dip'' appears to a value of
817: flux slightly different than zero. Note that the asymmetry refers mainly to the
818: modes $n=0$, and $n=-1$. The other modes are not influenced much
819: due to their higher flux content.
820: Also the critical current is suppressed
821: compared to the case where $\theta=0$ as can be seen in Fig.
822: \ref{fig11.fig}(b), due to a drop in $J_c$.
823:
824: \begin{figure}
825: \centerline{\psfig{figure=fig11a.eps,width=6.5cm,angle=0}}
826: \centerline{\psfig{figure=fig11b.eps,width=6.5cm,angle=0}}
827: \centerline{\psfig{figure=fig11c.eps,width=6.5cm,angle=0}}
828: \centerline{\psfig{figure=fig11d.eps,width=6.5cm,angle=0}}
829: \caption{(a) Overlap critical current $I_c^{ov}$ per unit length versus the
830: magnetic flux $\Phi$ in units of $\Phi_0$,
831: for a corner junction of $d_{x^2-y^2}+id_{xy}$-wave and $s$-wave
832: superconductors, with length
833: $L=10\lambda_J$, for angle $\theta=0^{\circ}$.
834: (b) The same as in a) but for $\theta=0.5$.
835: (c) The same as in a) but for $d_{x^2-y^2}+is$-wave and $s$-wave
836: superconductors for angle $\theta=0^{\circ}$.
837: (d) The same as in (c) but for $\theta=0.5$.
838: }
839: \label{fig11.fig}
840: \end{figure}
841:
842:
843: In the $d_{x^2-y^2}+is$-wave symmetry,
844: in the limit where
845: $\theta\rightarrow 0$, the order parameter is complex and
846: the pattern is asymmetric
847: as can be seen in Fig. \ref{fig11.fig}c, for the
848: angle $\theta=0$.
849: This is in
850: agreement with our previous work for the inline current input
851: for a junction with $d_{x^2-y^2}+is$ symmetry
852: \cite{stefan}.
853: There it was found that the pattern is asymmetric for lengths as long
854: as $L=10\lambda_J$.
855: For angles close to $\pi/4$,
856: the magnetic interference
857: pattern is similar with the $d_{x^2-y^2}+id_{xy}$-state.
858: This is because the $\sin(2\theta)$ dependence of the $d_{xy}$
859: component is almost unity.
860: This is seen in Fig. \ref{fig11.fig}d where we present
861: the variation of the critical current per unit length versus the enclosed flux for
862: $\theta=0.5$, and the symmetry state is $d_{x^2-y^2}+is$.
863:
864: In the short junction limit $L<\lambda_J$ the same argument can
865: be applied without any explicit reference to fractional
866: vortex and antivortex solutions. However as we found in our
867: previous work \cite{stefan}, both $n=0$ and $n=-1$ (there $f_{va}, f_a$)
868: exist, with
869: reduced flux content, in
870: this limit as a continuation of the corresponding solutions
871: in the large junction limit. In this case the external applied
872: magnetic field becomes equal to the self field, and the maximum
873: current can be calculated analytically \cite{zhu},
874: \begin{equation}
875: \frac{I_m(\Phi)}{I_{m0}}=\left|\frac{\sin(\pi \Phi /2\Phi_0) \cos[\pi \Phi/2\Phi_0 +
876: (\phi_{c2}-\phi_{c1})/2]}{\pi\Phi/2\Phi_0}\right|
877: .~~~\label{short}
878: \end{equation}
879: As we see at $\theta=0$ for the $d_{x^2-y^2}+id_{xy}$-wave case, the relation
880: $\phi_{c2}-\phi_{c1}=n\pi$ holds, and the magnetic interference
881: pattern becomes symmetric, while for the $d_{x^2-y^2}+is$,
882: this difference is a fraction of $\pi$ and the pattern is asymmetric.
883: However as we increase the junction length, we expect this symmetric
884: pattern for the $d$-wave order parameter to be continued.
885: This symmetry in the large junction limit, is described more effectively
886: by the assumption of the $n=0,-1$ solutions which give a
887: symmetric magnetic interference pattern as we presented.
888: Also the $n=-1$ solution extends to values for the
889: magnetic flux, where the
890: $n=0$ solution is absent. Eliminating one of them will break the
891: symmetry of the diagram.
892:
893: \section{Experimental relevance}
894: The symmetric pattern with a minimum at zero applied field
895: observed in corner junction experiments between YBCO and Pb
896: at $\theta=0$ has been interpreted as an indication of
897: $d_{x^2-y^2}$-wave symmetry. \cite{wollman,wollman1}
898: This result refers to short junctions
899: where the junction size is much smaller than the Josephson penetration
900: depth. However as we found here these experimental data are also
901: consistent with an order parameter with $d_{x^2-y^2}+id_{xy}$
902: pairing symmetry at $\theta=0$.
903:
904: Also the critical current $I_c$
905: versus the magnetic flux $\Phi$ of a SQUID, consisting of two planar
906: Josephson junctions on the faces of YBCO superconducting crystal,
907: connected by a loop of a second superconductor,
908: for $\theta=0$ or $\theta=\pi /2$
909: is found shifted by
910: $\Phi=0.5\Phi_0$ and has a minimum at $\Phi=0$
911: (instead of a maximum as in a SQUID involving conventional
912: $s$-wave superconductors or the edge SQUID in which both junctions are
913: on the same crystal face) but is still symmetric.
914: This result has been attributed to an order parameter
915: with $d_{x^2-y^2}$-wave symmetry. However the
916: theoretical analysis done by Beasley et al. \cite{beasley}
917: shows that it is also consistent with an order parameter
918: with $d_{x^2-y^2}+id_{xy}$-pairing symmetry at $\theta=0$.
919:
920: In both cases of SQUID and corner junction the symmetric pattern
921: observed at $\theta=0$ rules out the $d_{x^2-y^2}+is$-wave
922: pairing state where the order parameter is complex everywhere
923: resulting in an asymmetric $I_c$ versus $\Phi$ pattern for all
924: angles $\theta$.
925: However the small asymmetry (less than $2\%$) observed at $\theta=0$
926: in some experiments can be attributed to various complicating
927: factors e.g. fluxon trapping as will be discussed latter in
928: this section.
929:
930: The experiment proposed here to
931: resolve ambiguity between $d_{x^2-y^2}+id_{xy}$ and
932: $d_{x^2-y^2}$ at $\theta=0$, is to execute the
933: same experiments using SQUID or corner junction
934: at an angle between sample faces $\theta$ between $0$ and $\pi /2$.
935: Our theory predicts symmetric(asymmetric) pattern for the
936: $d_{x^2-y^2}$-wave ($d_{x^2-y^2}+id_{xy}$)-wave pairing state for
937: the corner junction case.
938: This kind of experiment has already been done in the case of
939: SQUID geometry. \cite{yanoff}
940: The tunneling directions are defined lithographically and
941: patterned by ion milling of a $c$-axis oriented film.
942: A YBCO thin film is patterned into a circle with a series of
943: Nb-Au-YBCO edge junctions at orientations spaces every $7.5^\circ$.
944: The measurement of the $I_c$ vs $\theta$, which probes mainly the
945: magnitude of the order parameter has an angular anisotropy, indicating
946: an anisotropic order parameter. Also the execution of this
947: experiments is not easy due to the difficulty in cleaning, polishing
948: a crystal at angle $\theta$, between $0$ and $\pi /2$.
949:
950:
951: Also in an experiment analogous to the corner junction
952: Miller, Ying et. al. \cite{miller} used frustrated thin-film tricrystal
953: samples to probe the pairing symmetry of YBCO. They found a minimum
954: in the $I_c$ vs the externally applied flux
955: $\Phi_e$ diagram at $\Phi_e=0$ in the short junction limit
956: and a maximum at $\Phi_e=0$ for a wide junction where the junction
957: length is much larger than the $\lambda_J$. However
958: for a wide junction the correct quantity to be compared should
959: be the total flux $\Phi$ which involves contribution both
960: from the externally applied flux and the intrinsic flux.
961: Also in the tricrystal magnetometry experiments on half-flux
962: quantum Josephson vorticies one can only observe spontaneous
963: magnetization of $\Phi_0/2$ in a frustrated geometry only in the
964: large junction length limit \cite{kirtley}.
965:
966: There is a number of complicating factors in the interpretation of the
967: experiments involving corner junctions that could lead to
968: an asymmetric ($I_c$ vs $\Phi$) pattern even for $\theta=0$. These are
969: the asymmetry of the junction (meaning that the
970: critical current of the two junction faces are not equal).
971: This will only cause the dip
972: to be shallower and will maintain the symmetry of the $I_c$
973: vs $\Phi$ diagram.
974: Also these experiments are influenced by the sample geometry and
975: the effect of flux
976: trapping i.e. there can be vortices trapped between the planes
977: of the cuprate superconductors that could affect the $I_c$ vs $\Phi$ diagram.
978: In the corner junction case, it creates an asymmetry in the
979: flux modulation curves.
980: However these flux trapping effects are not sufficiently large
981: to change the qualitative interpretation of these experiments.
982:
983:
984: \section{Conclusions}
985: We studied numerically the possible spontaneous vortex states
986: that may exist in a corner junction between a superconductor with
987: time reversal symmetry broken, (i.e. $d_{x^2-y^2}+id_{xy}$
988: or $d_{x^2-y^2}+is$), and an $s$-wave superconductor,
989: in
990: the long junction limit.
991: We studied separately three parameters which can be used
992: to modulate the spontaneous flux. These are the magnetic field
993: $H$, the interface orientation $\theta$, and the magnitude of
994: the subdominant order parameter $n_s$. We pointed out the differences
995: between time reversal broken states under these modulation
996: parameters.
997:
998: We found that in flux modulation experiments involving superconductors
999: with some spontaneous flux the range in
1000: magnetic flux of each mode is displaced compared to the
1001: case of a pure $s$-wave superconductor junction by an amount which corresponds
1002: to the intrinsic flux.
1003: In particular when the magnetic field $H$ is considered as the
1004: modulation parameter, the range in $H$ of the lower fluxon modes
1005: is significantly broadened compared to the $s$-wave case, and
1006: the instability at the boundary values of the field sets in
1007: due to the interaction of the flux entering from the
1008: junction edges with the intrinsic flux.
1009: In any case, for each value of the parameter
1010: which changes the flux,
1011: the modes are
1012: separated by a single flux quantum.
1013:
1014: We also derived some simple arguments to discriminate between
1015: the different pairing states that break the time reversal symmetry.
1016: For the $d_{x^2-y^2}+id_{xy}$-wave pairing state, the
1017: junction orientation where $\theta=0$ i.e. the lobes of the
1018: dominant $d_{x^2-y^2}$-wave order parameter are at right angles
1019: for the corner junction, give flux quantization condition
1020: $\Phi=n\Phi_0/2$ as in
1021: the $d_{x^2-y^2}$-wave state, which is different from the
1022: corresponding flux quantization for the $d_{x^2-y^2}+is$-wave
1023: pairing state, at $\theta=0$, which is $\Phi=(n/2+f)\Phi_0$,
1024: where $f$ is a small quantity.
1025: These different conditions provide a way experimentally to
1026: distinguish between time reversal broken symmetry states.
1027: Note that since the magnitude of the secondary order
1028: parameter is small compared to the dominant, the detection of
1029: time reversal broken states requires a very precise measurement
1030: of the spontaneous magnetic flux.
1031:
1032: Also we showed that the magnetic interference pattern at $\theta=0$
1033: is symmetric
1034: (asymmetric) for the $d_{x^2-y^2}+id_{xy}$ ($d_{x^2-y^2}+is$), and
1035: this also can be used to probe which symmetry the order
1036: parameter has, at least where the junctions are formed.
1037: We expect our findings, for the magnetic field dependence of the
1038: critical current, to hold even in the short junction limit, where
1039: the most experiments on corner junctions have been performed
1040: \cite{vanh,yanoff}.
1041:
1042: \section{Acknowledgments}
1043: One of us N.S. is grateful to A.V. Balatsky, J. Betouras for useful discussions
1044: that led to this article. Also N.S. would like to acknowledge
1045: the ESF/FERLIN programme for partial support to participate
1046: to conferences.
1047:
1048: \begin{references}
1049: \bibitem{scalapino} D.J. Scalapino, Phys. Rep. {\bf 250}, 329 (1995).
1050:
1051: \bibitem{vanh} D.J. van Harlingen, Rev. Mod. Phys. {\bf 67}, 515 (1995).
1052:
1053: \bibitem{shiba} M. Matsumoto, and H. Shiba, J. Phys. Soc. Jpn.
1054: {\bf 64}, 1703 (1995).
1055:
1056: \bibitem{kirtley2} J.R. Kirtley, P. Chaudhari, M.B. Ketchen, N. Khare,
1057: S.Y. Lin, and T. Shaw, Phys. Rev. B {\bf 51}, 12 057 (1995).
1058:
1059: \bibitem{bailey} D.B. Bailey, M. Sigrist, and R.B. Laughlin, Phys. Rev.
1060: B {\bf 55}, 15 239 (1997).
1061:
1062: \bibitem{balatsky1} A.V. Balatsky, Phys. Rev. Lett. {\bf 80}, 1972 (1998).
1063:
1064: \bibitem{balatsky2} A.V. Balatsky, Phys. Rev. B {\bf 61}, 6940 (2000).
1065:
1066: \bibitem{covington} M. Covington, M. Aprili, E. Paraoanu, L.H. Greene,
1067: F. Xu, J. Zhu, and C.A. Mirkin, Phys. Rev. Lett. {\bf 79}, 277 (1997).
1068:
1069: \bibitem{fogelstrom} M. Fogelstrom, D. Rainer, and J.A. Sauls,
1070: Phys. Rev. Lett. {\bf 79}, 281 (1997).
1071:
1072: \bibitem{aprili} M. Aprili, E. Badica, and L.H. Greene,
1073: Phys. Rev. Lett. {\bf 83}, 4630 (1999).
1074:
1075: \bibitem{krupke} R. Krupke, and G. Deutscher,
1076: Phys. Rev. Lett. {\bf 83}, 4634 (1999).
1077:
1078: \bibitem{yanoff} D.J. van Harlingen, J.E. Hilliard, B.L.T. Plourde and
1079: B.D. Yanoff, Physica C {\bf 317-318}, 410 (1999).
1080:
1081: \bibitem{zhu} J.-X. Zhu, W. Kim, and C.S. Ting,
1082: Phys. Rev. B {\bf 58}, 6455 (1998).
1083:
1084: \bibitem{sigrist} M. Sigrist, Prog. Theor. Phys.
1085: {\bf 99}, 899 (1998).
1086:
1087: \bibitem{caputo}
1088: J.-G. Caputo, N. Flytzanis, Y. Gaididei, N. Stefanakis, and
1089: E. Vavalis, Supercond. Sci. Technol. {\bf 13}, 423 (2000).
1090:
1091: \bibitem{owen} C.S. Owen, and D.J. Scalapino,
1092: Phys. Rev. {\bf 164}, 538 (1967).
1093:
1094: \bibitem{defect} N. Stefanakis, and N. Flytzanis,
1095: Supercond. Sci. Technol. {\bf 14}, 16 (2001).
1096:
1097: \bibitem{stefan} N. Stefanakis, and N. Flytzanis, Phys. Rev.
1098: B {\bf 61}, 4270 (2000).
1099:
1100: \bibitem{wollman} D.A. Wollman, D.J. van Harlingen, W.C. Lee,
1101: D.M. Ginsberg, and A.J. Leggett, Phys. Rev. Lett.
1102: {\bf 71}, 2134 (1993).
1103:
1104: \bibitem{wollman1} D.A. Wollman, D.J. van Harlingen, J. Giapintzakis,
1105: and D.M. Ginsberg, Phys. Rev. Lett.
1106: {\bf 74}, 797 (1995).
1107:
1108: \bibitem{beasley} M.R. Beasley, D. Lew, and R.B. Laughlin, Phys. Rev.
1109: B {\bf 49}, 12 330 (1994).
1110:
1111: \bibitem{miller} J.H. Miller Jr., Q.Y. Ying, Z.G. Zou, N.Q. Fan,
1112: J.H. Xu, M.F. Davis, and J.C. Wolfe,
1113: Phys. Rev. Lett.
1114: {\bf 74}, 2347 (1995).
1115:
1116: \bibitem{kirtley} J.R. Kirtley, C.C. Tsuei, J.Z. Sun, L.S. Yu-Jahnes,
1117: A. Gupta, M.B. Ketchen, K.A. Moler, and M. Bhushan,
1118: Phys. Rev. Lett. {\bf 76}, 1336 (1996).
1119:
1120: \end{references}
1121:
1122: \end{document}
1123:
1124: