1: % For Phys. Rev. appearance, change preprint to twocolumn.
2: % Choose pra, prb, prc, prd, pre, prl, prstab, or rmp for journal
3: % Add 'draft' option to mark overfull boxes with black boxes
4: % Add 'showpacs' option to make PACS codes appear
5: % Add 'showkeys' option to make keywords appear
6: %\documentclass[aps,prl,preprint,groupedaddress]{revtex4}
7: %\documentclass[aps,prl,preprint,superscriptaddress]{revtex4}
8: %\documentclass[aps,prb,twocolumn,groupedaddress]{revtex4}
9: %\documentclass[12pt,a4paper]{article}
10: %\documentclass[aps,prb,amsfonts,amssymb,twocolumn,amsmath,preprintnumbers,nofootinbib,superscriptaddress,floatfix,showpacs]{revtex4}
11: \documentclass[aps,prb,amsfonts,amssymb,twocolumn,amsmath,preprintnumbers,nofootinbib,floatfix,showpacs]{revtex4}
12: % You should use BibTeX and apsrev.bst for references
13: % Choosing a journal automatically selects the correct APS
14: % BibTeX style file (bst file), so only uncomment the line
15: % below if necessary.
16: %\bibliographystyle{apsrev}\gtrsim
17: \usepackage{amsmath,bm,amsfonts,amssymb}
18: \usepackage[dvips]{graphics}
19: \usepackage{graphicx}
20: \usepackage{bm}
21: \def\k{{\bf k}}
22: \def\r{{\bf r}}
23: \def\q{{\bf q}}
24:
25: \begin{document}
26:
27: \title{Nonequilibrium Transport in Superconductor/Ferromagnet/Superconductor Diffusive
28: Junctions: Interplay between Proximity Effect and Ferromagnetism.}
29: \author{I. V. Bobkova}
30: \affiliation{Institute of Solid State Physics, Chernogolovka,
31: Moscow reg., 142432 Russia}
32: \author{A. M. Bobkov}
33: \affiliation{Institute of Solid State Physics, Chernogolovka,
34: Moscow reg., 142432 Russia}
35:
36: \date{\today}
37:
38: \begin{abstract}
39: The theory of the I-V characteristics in diffusive
40: superconductor/weak ferromagnet/superconductor (SFS) junction is
41: developed. We show that the exchange field $h$ of the ferromagnet
42: manifests itself as an additional conductance peak at $eV \sim \Delta+h$ in
43: the phase-coherent regime, when the Thouless energy is of the
44: order of superconducting order parameter. The excess current
45: exhibits non-monotonous dependence on the exchange field and
46: non-trivial temperature behavior, which is strongly influenced by
47: the temperature dependence of the exchange field.
48: \end{abstract}
49: % insert suggested PACS numbers in braces on next line
50: \pacs{74.45.+c, 74.50.+r}
51:
52: \maketitle
53:
54: The recent progress in the experimental techniques has made
55: possible the fabrication of mesoscopic structures on the nanometer
56: scale. Hybrid structures containing superconducting and
57: ferromagnetic elements offer an opportunity to generate and
58: control coherent spin transport with otherwise conventional
59: electronics. The equilibrium transport and proximity effect in
60: such structures have been theoretically and experimentally
61: investigated recently in details as for the case of weak
62: ferromagnetic alloys so as for half-metals like $CrO_2$
63: \cite{Eschrig03,Keizer06}.In particular, Josephson current in SFS
64: junctions and $T_c$ of SF bilayers and multilayers have been
65: investigated in details for the case of weak ferromagnetic alloys
66: (see Ref.\onlinecite{buzdin} and references therein). Equilibrium
67: density of states was also studied \cite{kadigrob,Fazio99}. On the
68: other hand, to the best of our knowledge, the nonequilibrium
69: transport in SFS mesoscopic junctions has not been studied yet
70: neither from theoretical nor from experimental point of view
71: except for theoretical investigations of magnetic quantum point
72: contacts\cite{Cuevas01,Fogelstrom02,Bobkova06,Sauls06}. We address
73: our theoretical paper to the part of this problem and study the
74: phase-coherent transport in diffusive voltage-biased SFS plane
75: junctions. We consider interlayers made of a weak ferromagnetic
76: material, in which the value of the exchange field $h$ (measured
77: in the energy units) is of order of superconducting order
78: parameter $\Delta$. In ferromagnetic alloys like $CuNi$, which
79: have been intensively used by now for experimental investigation
80: of equilibrium properties of SFS heterostructures, the exchange
81: field is several times larger than $\Delta$. On the other hand, it
82: can be concluded from our analysis of the problem that for
83: studying the I-V characteristics of SFS junctions, made on the
84: basis of a weak ferromagnetic alloy, the most interesting case is
85: $h \lesssim \Delta$. As far as we know, the work on the creation
86: of appropriate alloys is in progress now, so we believe that this
87: limit can be experimentally realized in the nearest future.
88:
89: The I-V characteristics of the superconductor/normal
90: metal/superconductor (SNS) voltage-biased diffusive junction are
91: studied theoretically in details mainly in two limits. The first
92: one is the limit of short junction $d \ll \xi$, where $d$ is the
93: length of the interlayer, $\xi=\sqrt{\hbar D/\Delta}$ is the
94: superconducting coherence length and $D$ is the diffusion
95: constant. In this regime the subharmonic gap structure (SGS) in
96: the differential conductance $dI/dV$ consists of a set of
97: pronounced maxima at $eV_n=2\Delta/n$
98: \cite{Bardas97,Kupriyanov03}. The excess current, which takes
99: place at high biases $eV \gg \Delta$, behaves in dependence on
100: temperature as $\Delta(T)$ does and can be as positive so as
101: negative depending on the transparency of SN
102: interfaces\cite{Volkov93,Bardas97}. The second one is the
103: incoherent limit $d \gg \xi$. The proximity effect is negligible
104: in this regime and the transport is determined by the kinetic
105: equation for the distribution
106: function\cite{Klapwijk82,Octavio8388,Bezuglyi00}. The SGS of the
107: conductance again shows sharp features at $eV_n=2\Delta/n$. At the
108: same time in the intermediate regime $\xi \sim d$ the proximity
109: effect, which manifests itself as a minigap $\Delta_g$ in the
110: equilibrium density of states in the normal region, takes place.
111: It has been shown very recently \cite{Cuevas05} that for SNS
112: voltage-biased diffusive junction with highly-transparent NS
113: interfaces in this regime, when the interplay between proximity
114: effect and MARs takes place, the well-known subgap conductance
115: structure $eV_n=2\Delta/n$ modifies and exhibits an additional
116: maximum at roughly $eV \sim \Delta + \Delta_g$. These predictions
117: are in good agreement with the existing
118: experiments\cite{Kutchinsky97,Hoss00}. In the present work we
119: numerically calculated the I-V characteristics taking into account
120: the proximity effect and non-equilibrium distribution function in
121: the ferromagnetic interlayer. We show that in the phase-coherent
122: regime, when the proximity effect in the ferromagnetic region
123: takes place, the exchange field of a weak ferromagnet $h<\Delta$
124: explicitly manifests itself in the SGS of diffusive voltage-biased
125: SFS junction leading to the additional maximum at $eV \sim (\Delta
126: + h)$, which is the most pronounced in the case of low-transparent
127: SF boundaries. For the case $h>\Delta$ the differential
128: conductance exhibits no pronounced characteristic features. The dc
129: current at high voltage biases is also investigated. It is deficit
130: ($I_{def} = I-V/R<0$) for the case of highly-resistive interfaces
131: we consider and has a non-monotonous behavior as a function of
132: exchange field, reaching the maximum value at roughly $h \sim
133: \Delta$. This non-monotonous behavior results in the fact, that
134: the temperature dependence of the deficit current exhibits the
135: characteristic features, which are determined by the temperature
136: dependence of the exchange field and can be used for its
137: experimental mapping.
138:
139: Further the model under consideration and the method we use are
140: briefly described. We study an SFS junction, where F is a
141: diffusive weak ferromagnet of length $d$ coupled to two identical
142: superconducting (S) reservoirs. The superconductors are supposed
143: to be diffusive and have $s$-wave pairing. We assume the SF
144: interfaces to be not fully transparent and suppose that the
145: resistance of the SF boundary $R_g$ dominates the resistance of
146: the ferromagnetic interlayer $R_F$. We use the quasiclassical
147: theory of superconductivity for diffusive systems in terms of
148: time-dependent Usadel equations \cite{usadel}. The fundamental
149: quantity for diffusive transport is the momentum average of the
150: quasiclassical Green's function $\check g(x,\varepsilon, t) =
151: \langle \check g(\bm p_f, x,\varepsilon, t) \rangle_{\bm p_f}$. It
152: is a $8\times8$ matrix form in the product space of Keldysh,
153: particle-hole and spin variables. Here $x$ - is the coordinate
154: measured along the normal to the junction, $t$ stands for a time
155: variable and $\varepsilon$ is the excitation energy.
156:
157: The electric current should be calculated via Keldysh part of the
158: quasiclassical Green's function. For the plane diffusive junction
159: the corresponding expression reads as follows
160: \begin{widetext}
161: \begin{equation}
162: \frac{j^{el}}{e} = -\frac{d}{8 \pi^2 e^2 R_F} \int
163: \limits_{-\infty}^{+\infty} d \varepsilon {\rm Tr}_4
164: \left[\frac{1}{2}(\hat \tau_0 + \hat \tau_3) \hat \sigma_{0}
165: \left(\check g(x, \varepsilon, t)\otimes \frac{\partial \check
166: g(x, \varepsilon, t)}{\partial x}\right)^K \right] \label{current}
167: \enspace ,
168: \end{equation}
169: \end{widetext}
170: where $e$ is the electron charge and $\hbar = 1$ throughout the
171: paper. $\left(\check g(x, \varepsilon, t)\otimes (\partial \check
172: g(x, \varepsilon, t))/(\partial x)\right)^K$ is a $4\times4$
173: Keldysh part of the corresponding combination of full Green's
174: functions. The product $\otimes$ of two functions of energy and
175: time is defined by the noncommutative convolution $A \otimes B =
176: e^{i(\partial_\varepsilon^A\partial_t^B-\partial_t^A\partial_\varepsilon^B)}A(\varepsilon,t)B(\varepsilon,t)$.
177: $\hat \tau_i$ and $\hat \sigma_i$ are Pauli matrices in
178: particle-hole and spin spaces respectively.
179:
180: The quasiclassical Green's function $\check g(x,\varepsilon, t)$
181: satisfies the non-stationary Usadel equation. In order to solve
182: the Usadel equation it is convenient to express quasiclassical
183: Green's function $\check g$ in terms of Riccati coherence
184: functions $\hat \gamma^{R,A}$ and $\hat {\tilde \gamma}^{R,A}$ and
185: distribution functions $\hat x^K$ and $\hat {\tilde x}^K$. All
186: these functions are $2 \times 2$ matrices in spin space and depend
187: on $(x, \varepsilon, t)$. The corresponding expression for $\check
188: g$ can be found in Ref.\onlinecite{Eschrig00} and therefore is not
189: explicitly written here. Riccati coherence and distribution
190: functions obey Riccati-type transport equations\cite{Eschrig04},
191: where the appropriate self energy takes the form $\check \Sigma(x,
192: \varepsilon, t) = h \hat \sigma_3 + (1/2\pi \tau_s) \hat \sigma_3
193: \check g \hat \sigma_3$. Here $\tau_s^{-1}$ is an inverse magnetic
194: scattering time. As the F layer is supposed to be an alloy, a role
195: of magnetic scattering may be quite
196: important\cite{Sellier03,Ryazanov04}. We assume a presence of the
197: relatively strong uniaxial magnetic anisotropy which prevents
198: mixing of spin-up and spin-down Green's functions, so the magnetic
199: scattering term is a diagonal matrix in spin
200: space\cite{Ryazanov05}.
201:
202: The Riccati-type transport equations should be solved together
203: with the boundary conditions at SF interfaces. As it was mentioned
204: above we consider the case when the dimensionless conductance of
205: the boundary $G \equiv R_F/R_g \lesssim 1$, so the interface
206: transparency $T \sim G(l/d) \ll 1$ ($l$ is the mean free path).
207: Due to the smallness of the interface transparency $T$ we can use
208: Kupriyanov-Lukichev boundary conditions at SF
209: boundaries\cite{Kupriyanov}. In terms of Riccati coherence and
210: distribution functions they take the form
211: \begin{widetext}
212: \begin{equation}
213: 2 i \pi d \partial_x \hat \gamma^R_{l,r} = \pm G \left[ \hat
214: f^R_{S;l,r} + \hat \gamma^R_{l,r} \otimes \hat {\tilde
215: g}^R_{S;l,r} - (\hat g^R_{S;l,r} + \hat \gamma^R_{l,r} \otimes
216: \hat {\tilde f}^R_{S;l,r}) \otimes \hat \gamma^R_{l,r} \right]
217: \enspace , \label{boundary_cond_gamma}
218: \end{equation}
219: \begin{equation}
220: 2 i \pi d \partial_x \hat x^K_{l,r} = \pm G \left[ (1/2) \left(
221: \hat g^K_{S;l,r} + \hat \gamma^R_{l,r} \otimes \hat {\tilde
222: g}^K_{S;l,r} \otimes \hat {\tilde \gamma}^A_{l,r} \right) - \hat
223: \gamma^R_{l,r} \otimes \hat {\tilde f}^K_{S;l,r} - \left( \hat
224: g^R_{S;l,r} + \hat \gamma^R_{l,r} \otimes \hat {\tilde
225: f}^R_{S;l,r} \right) \otimes \hat x^K_{l,r} - h.c. \right]
226: \enspace . \label{boundary_cond_x}
227: \end{equation}
228: \end{widetext}
229: Here Riccati coherence and distribution functions denoted by the
230: lower case symbols $l,r$ are taken at the left and right ends of
231: the ferromagnet. The quantities denoted by the lower case symbols
232: ($S;l,r$) are corresponding Green's functions at the
233: superconducting side of the left and right SF interfaces. We
234: assume the parameter $(R_F/R_g)(\sigma_F/\sigma_s)$, where
235: $\sigma_F$ and $\sigma_s$ stand for conductivities of
236: ferromagnetic and superconducting materials respectively, to be
237: also small, what allows us to neglect the suppression of the
238: superconducting order parameter in the S leads near the interface
239: and take the Green's functions at the superconducting side of the
240: boundaries to be equal to their bulk values with the appropriate
241: shift of the quasiparticle energy due to the applied voltage
242: $V=V_r-V_l$.
243:
244: Let us now turn to the analysis of the I-V characteristics.
245:
246: \begin{figure}[!tbh]
247: \centerline{\includegraphics[clip=true,width=3in]{fig1a.eps}}
248: \centerline{\includegraphics[clip=true,width=3in]{fig1b.eps}}
249: \caption{The zero temperature differential conductance $dI/dV$,
250: normalized to its value for the normal state of the leads, as a
251: function of $eV/T_c$, taken for: (a) different exchange fields:
252: $h=0$(gray solid); $h=0.45T_c$(dotted); $h=0.9T_c$(black solid)
253: and $h=2.6T_c$(dashed); The insert to (a) shows the $R dI/dV$ as a
254: function of $eV/T_c$ for $h>\Delta$: $h=2.6T_c$(solid);
255: $h=3T_c$(dashed) and $h=3.5T_c$(dotted) are practically
256: undistinguishable; (b) different values of the Thouless energy:
257: $\varepsilon_{Th}=5T_c$(dashed); $\varepsilon_{Th}=2T_c$(black
258: solid) and $\varepsilon_{Th}=0.5T_c$(dotted). $T_c$ is the
259: superconducting critical temperature. $G=0.1$ and $1/\tau_s=0$.
260: The inset to (b) demonstrates the influence of the magnetic
261: scattering on the conductance: $1/\tau_s=0$ (solid);
262: $1/\tau_s=0.05 T_c$(dotted); $1/\tau_s=0.15 T_c$(dashed);
263: $1/\tau_s=0.3T_c$ (dashed-dotted); $1/\tau_s=0.45T_c$ (gray
264: solid); $1/\tau_s=0.9T_c$ (gray dashed). $h=0.9 T_c$.}\label{IV}
265: \end{figure}
266:
267: Figure \ref{IV}(a) shows the dependence of the zero temperature
268: differential conductance $dI/dV$ on voltage for different values
269: of the exchange field in the ferromagnet. The following main
270: features can be observed:
271:
272: (i) there is a series of peaks at $eV_n=2\Delta/n$. In fact, only
273: the peaks corresponding to $n=1,2$ are seen in the figure. The
274: other harmonics are smeared out due to the small interface
275: transparency and large enough intrinsic broadening of the
276: quasiparticle energy (modelling the inelastic scattering rate)
277: $\gamma=0.03 T_c$, which we take in the numerical calculations.
278: This value of intrinsic broadening is large, but consistent with
279: an estimate of inelastic scattering rate due to electron-phonon
280: processes $\gamma \sim \varepsilon^3/\omega_D^2$ at high enough
281: energies $\varepsilon \sim \Delta$ \cite{gantmakher}.
282:
283: (ii) for $h \lesssim \Delta$ the large additional maximum appears
284: at roughly $eV=\Delta+h$, while the curves corresponding to
285: $h>\Delta$ do not exhibit any additional pronounced characteristic
286: features, in particular at $eV=\Delta+h$ (See Insert to
287: Fig.\ref{IV}(a)). The exchange fields of order of several $\Delta$
288: only result in the suppression of the peak at $eV=\Delta$. In
289: fact, the larger the exchange field the less different from each
290: other the appropriate conductance curves. Of course, this
291: statement is related to the region of weak exchange fields we
292: study.
293:
294: Analyzing the dependence of the I-V characteristics on the
295: Thouless energy $\varepsilon_{Th}=D/d^2$, which is represented in
296: Figure \ref{IV}(b), we can conclude, that the most pronounced
297: maximum at $eV \sim \Delta+h$ exists at the region
298: $\varepsilon_{Th} \sim \Delta$. If the Thouless energy increases
299: the peak washes out and finally vanishes. At the limit of very
300: short junction $\varepsilon_{Th} \gg \Delta$ the only structure
301: $eV=2\Delta/n$ survives. On the other hand, with decreasing the
302: Thouless energy the amplitude of maximum at $eV \sim \Delta+h$
303: reduces to the zero value at the limit of very small Thouless
304: energies and again the only structure at $eV=2\Delta/n$ can be
305: observed. With the increasing of transparency of the interfaces
306: the maximum smears out and is masked by the growing peak at
307: $eV=\Delta$.
308:
309: The inset to Figure \ref{IV}(b) demonstrates the influence of the
310: magnetic scattering on the maxima discussed above. It can be seen
311: that the magnetic scattering does not influence the peak at
312: $eV=2\Delta$, while the maximum at $eV \sim \Delta + h$ is
313: broadened, but not dramatically for small compared with $T_c$
314: values of magnetic scattering rate. Of course, large enough values
315: of $1/\tau_s \sim h,T_c$ destroy the peak and we found the
316: characteristic value of $1/\tau_s$, under which the peak
317: completely washes out, to be approximately equal to $0.45T_c$. The
318: magnetic scattering is typically arises due to the magnetic
319: inhomogeneity, related above all to $Ni$-rich clusters, and can be
320: of order of the average exchange field in the ferromagnetic alloys
321: like $Cu_{1-x}Ni_x$ (see \onlinecite{Ryazanov05} and references
322: therein). However, one can believe that the magnetic inhomogeneity
323: could be weaker in the more diluted alloys with very low exchange
324: fields of order of $\Delta$ resulting in small values of magnetic
325: scattering parameter $1/\tau_s$, even compared with $T_c$.
326:
327: The appearance of this additional maximum in the intermediate
328: regime $\varepsilon_{Th} \sim \Delta$ is caused by the proximity
329: effect in the ferromagnetic region and can be qualitatively
330: understood as follows. It is well known that for a diffusive SNS
331: junction the zero-bias density of states (DOS) has a minigap in
332: the normal region. For low-transparency junctions it is $\Delta_g
333: \propto G\varepsilon_{Th}$. If the normal metal is replaced by a
334: ferromagnet, the exchange field $h$ shifts the densities of states
335: for the two spin subbands in the opposite directions, therefore
336: the minigap in the spectrum closes at $h \sim \Delta_g$
337: \cite{Fazio99}, but the sharp onsets in the DOS for the both
338: subbands survive until they merge the edge of the continuous
339: spectrum. So for small enough values of $\Delta_g < h$ and $eV
340: \sim \Delta + h \pm \Delta_g$ an electron belonging to the one
341: spin subband travels through the interlayer from one sharp onset
342: of the DOS at $-\Delta$ to another one at $h \pm \Delta_g$, while
343: an electron from the other spin subband can move from $-h \pm
344: \Delta_g$ to the edge of continuous spectrum at $+\Delta$.
345: According to this qualitative consideration the maximum should be
346: split and the splitting is roughly proportional to the value of
347: the equilibrium minigap $\Delta_g \sim G \varepsilon_{Th}$. We do
348: not observe such splitting at our curves. In addition, due to the
349: multiple Andreev reflection processes there could arise the
350: subsequent harmonics at $eV \sim (\Delta+h\pm\Delta_g)/n$. Of
351: course, the proposed qualitative explanation is very crude because
352: it deals with the equilibrium minigap instead of real
353: position-dependent pseudogap at finite bias and does not take into
354: account the behavior of the distribution function in the
355: interlayer, which is highly nonequilibrium and obviously
356: influences the I-V characteristics.
357:
358: \begin{figure}[!tbh]
359: \centerline{\includegraphics[clip=true,width=3in]{fig2a.eps}}
360: \centerline{\includegraphics[clip=true,width=3in]{fig2b.eps}}
361: \caption{(a) Possible dependencies of the weak exchange field on temperature.
362: The black solid line represents the BCS-type behavior of $\Delta(T)/T_c$. (b) The appropriate temperature
363: dependencies of the deficit current, normalized to its value at
364: $h=0$. The inset shows the deficit current $eIR/T_c$ as a function of the
365: exchange field at $T=0$, $G=0.1$, $\varepsilon_{Th} = 2 T_c$. \label{exc_cur}}
366: \end{figure}
367:
368: Let us now discuss the behavior of the high voltage $eV \gg
369: \Delta$ deficit current $I_{def} = I-V/R<0$, where $R=R_F + 2 R_g$
370: is the full resistance of the junction. The inset in Figure
371: \ref{exc_cur} represents the typical dependence of the deficit
372: current on the exchange field. The deficit current is a
373: non-monotonous function of the exchange field and reaches a
374: maximum at $h \sim \Delta$. The non-monotonous behavior of the
375: deficit current leads to very peculiar dependencies of this
376: current on temperature. The corresponding curves have the
377: characteristic features determined by the temperature dependence
378: of the exchange field. Figure \ref{exc_cur}(a) shows several
379: examples of possible behavior of exchange field on temperature in
380: comparison with the temperature dependence of $\Delta(T)$. Figure
381: \ref{exc_cur}(b) represents the corresponding dependencies of the
382: deficit current on temperature. They are normalized to the value
383: of the deficit current at zero exchange field. It is seen that (i)
384: if $h(T)<\Delta(T)$ in the whole temperature region $0<T<T_c$, the
385: curve $\tilde I_{def}(T) = I_{def}(T)/I_{def,h=0}(T)$ resembles
386: the appropriate dependence $h(T)$ up to the Curie temperature of
387: the ferromagnet $T_{Cu}$ and then follows the curve for $h=0$
388: (dotted curve), (ii) in the opposite case of $h(T)>\Delta(T)$ for
389: $0<T<T_c$ $\tilde I_{def}(T)$ monotonously increases up to $T=T_c$
390: (dashed curve), (iii) if $h(0)>\Delta(0)$ and $T_{Cu}<T_c$, then
391: $\tilde I_{def}(T)$ has a minimum approximately at the
392: intersection point of the curves $h(T)$ and $\Delta(T)$. For the
393: temperatures higher then $T_{Cu}$ it follows the curve for $h=0$
394: (dashed-dotted curve) and (iv) if $h(0)<\Delta(0)$ and
395: $T_{Cu}>T_c$, then $\tilde I_{def}(T)$ grows starting from the
396: values less then that one for $h=0$, intersects this line and
397: rises very sharply at $T \to T_c$ (gray solid curve).
398:
399: In summary, we have developed the theory of the I-V
400: characteristics of diffusive SFS junctions with highly enough
401: resistive SF interfaces. It is found that weak exchange field of
402: the ferromagnet manifests itself only in the coherent regime
403: $\varepsilon_{Th} \sim \Delta$. For $h<\Delta$ it results in the
404: appearance of the additional well-pronounced maximum in the
405: differential conductance at $eV \sim (\Delta+h)$ if the magnetic
406: scattering rate in the ferromagnetic region is small compared to
407: $T_c$. We have also studied the deficit current at high voltages
408: and found that it behaves non-monotonously as a function of the
409: exchange field exhibiting a maximum at $h \sim \Delta$. This gives
410: rise to the characteristic features in the temperature behavior of
411: the deficit current, determined by the temperature dependence of
412: the exchange field. We believe, that all the discussed features
413: can be measured in the diffusive SFS junctions, realized on the
414: basis of weak ferromagnetic alloys.
415:
416: We thank V.V. Ryazanov and A. Yu. Rusanov for discussions. The
417: support by RFBR Grants 05-02-17175 (I.V.B. and A.M.B.),
418: 05-02-17731 (A.M.B.) and the programs of Physical Science Division
419: of RAS is acknowledged. I.V.B. was also supported by the Russian
420: Science Support Foundation.
421: % Create the reference section using BibTeX:
422: %\bibliography{/users/tkm/howell/latexx/bibtexx/refs}
423:
424: \begin{thebibliography}{99}
425: %
426: \bibitem{Eschrig03}
427: M. Eschrig, J. Kopu, J.C. Cuevas, and G. Sch\"on, Phys.Rev.Lett.
428: {\bf 90}, 137003 (2003).
429: %
430: \bibitem{Keizer06}
431: R.S. Keizer, S.T.B. Goennenwein, T.M. Klapwijk, G. Miao, G. Xiao,
432: and A. Gupta, Nature {\bf 439}, 825 (2006).
433: %
434: \bibitem{buzdin}
435: A.I. Buzdin, Rev.Mod.Phys. {\bf 77}, 935 (2005).
436: %
437: \bibitem{kadigrob}
438: A. Kadigrobov, R.I. Shekhter, M. Jonson, Z.G. Ivanov, Phys.Rev.B
439: {\bf 60}, 14593 (1999); A.~Kadigrobov, R.I. Shekhter, M. Jonson,
440: cond-mat/0503643.
441: %
442: \bibitem{Fazio99}
443: R. Fazio and C. Lucheroni, Europhys. Lett. {\bf 45}, 707 (1999).
444: %
445: \bibitem{Cuevas01}
446: A. Martin-Rodero, A. Levy Yeyati, J.C. Cuevas, Physica C, {\bf
447: 352}, 117 (2001).
448: %
449: \bibitem{Fogelstrom02}
450: M. Andersson, J.C. Cuevas, M.Fogelstr\"om, Physica C, {\bf 367},
451: 117 (2002).
452: %
453: \bibitem{Bobkova06}
454: I.V. Bobkova, Phys.Rev.B {\bf 73}, 012506 (2006).
455: %
456: \bibitem{Sauls06}
457: Erhai Zhao and J.A. Sauls, cond-mat/0603610.
458: %
459: \bibitem{Bardas97}
460: A. Bardas, and D.V. Averin, Phys.Rev. B, {\bf 56}, 8518 (1997);
461: A.V. Zaitsev, and D.V. Averin, Phys. Rev. Lett. {\bf 80}, 3602
462: (1998).
463: %
464: \bibitem{Kupriyanov03}
465: A. Brinkman, A.A. Golubov, H. Rogalla, F.K. Wilheim, M.Yu.
466: Kupriyanov, Phys. Rev. B {\bf 68}, 224513 (2003).
467: %
468: \bibitem{Volkov93}
469: A.F. Volkov, A.V. Zaitsev, T.M. Klapwijk, Physica C, {\bf 210}, 21
470: (1993).
471: %
472: \bibitem{Klapwijk82}
473: T.M. Klapwijk, G.E. Blonder, and M. Tinkham, Physica B+C {\bf
474: 109-110}, 1657 (1982).
475: %
476: \bibitem{Octavio8388}
477: M. Octavio, M. Tinkham, G.E. Blonder, and T.M. Klapwijk,
478: Phys.Rev.B {\bf 27}, 6739 (1983); K. Flensberg, J.Bindslev Hansen,
479: and M. Octavio, {\it ibid.} {\bf 38}, 8707 (1988).
480: %
481: \bibitem{Bezuglyi00}
482: E.V. Bezuglyi, E.N. Bratus', V.S. Shumeiko, G. Wendin, and
483: H.Takayanagi, Phys. Rev. B, {\bf 62}, 14439, (2000); E.V.
484: Bezuglyi, E.N. Bratus', V.S. Shumeiko, G. Wendin, Phys. Rev. B,
485: {\bf 63}, 100501, (2001).
486: %
487: \bibitem{Cuevas05}
488: J.C. Cuevas, J. Hammer, J. Kopu, J.K. Viljas, and M. Eschrig,
489: Phys.Rev.B {\bf 73}, 184505 (2006).
490: %
491: \bibitem{Kutchinsky97}
492: J. Kutchinsky, R. Taboryski, T. Clausen, C. B. Sorensen, A.
493: Kristensen, P. E. Lindelof, J. Bindslev Hansen, C. Schelde
494: Jacobsen and J. L. Skov, Phys.Rev.Lett. {\bf 78}, 931 (1997); J.
495: Kutchinsky, R. Taboryski, O. Kuhn, C. B. Sorensen, P. E. Lindelof,
496: A. Kristensen, J. Bindslev Hansen, C. Schelde Jacobsen and J. L.
497: Skov, Phys.Rev.B {\bf 56}, R2932 (1997).
498: %
499: \bibitem{Hoss00}
500: T. Hoss, C. Strunk, T. Nussbaumer, R. Huber, U. Staufer and C.
501: Schonenberger, Phys.Rev.B {\bf 62}, 4079 (2000).
502: %
503: \bibitem{usadel}
504: K.D. Usadel, Phys.Rev.Lett. {\bf 25}, 507 (1970).
505: %
506: \bibitem{Sellier03}
507: H. Sellier, C. Baraduc, F. Lefloch, and R. Calemczuk, Phys.Rev. B
508: {\bf 68}, 054531 (2003).
509: %
510: \bibitem{Ryazanov04}
511: V.V. Ryazanov, V.A. Oboznov, A.S. Prokofiev, V.V. Bolginov, A.K.
512: Feofanov, Journ. Low Temp. Phys. {\bf 136}, 385 (2004).
513: %
514: \bibitem{Ryazanov05}
515: V.A. Oboznov, V.V. Bol'ginov, A.K. Feofanov, V.V. Ryazanov, and
516: A.I. Buzdin, Phys.Rev.Lett. {\bf 96}, 197003 (2006).
517: %
518: \bibitem{Eschrig00}
519: M. Eschrig, Phys. Rev. B {\bf 61}, 9061 (2000).
520: %
521: \bibitem{Eschrig04}
522: M. Eschrig, J. Kopu, A. Konstandin, J.C. Cuevas, M. Fogelstr\"om,
523: and Gerd Sch\"on, Advances in Solid State Physics, {\bf 44}, 533
524: (2004).
525: %
526: \bibitem{Kupriyanov}
527: M.Yu. Kupriyanov and V.F. Lukichev, Sov. Phys. JETP {\bf 67}, 1163
528: (1988).
529: %
530: \bibitem{gantmakher}
531: V.F. Gantmakher, Y.B. Levinson, {\it Carrier Scattering in Metals
532: and Semiconductors}, Elsevier Science Ltd., 1987.
533: %
534: \end{thebibliography}
535: \end{document}
536: %
537: % ****** End of file template.aps ******
538: