1: \documentclass{jetpl}
2:
3: \hoffset=-20mm % for arXiv stamp
4: \voffset=-30mm
5:
6: \usepackage{graphicx}
7: %\usepackage{hyperref}
8:
9: \twocolumn \lat
10:
11: \DeclareMathOperator{\sgn}{sgn}
12: \DeclareMathOperator{\Tr}{Tr}
13:
14:
15: \title{Conductance of a junction between a normal metal\\ and a Berezinskii superconductor}
16: \rtitle{Conductance of a junction between a normal metal and a Berezinskii superconductor}
17: \sodtitle{Conductance of a junction between a normal metal and a Berezinskii superconductor}
18:
19: \author{Ya.\,V.~Fominov\/\thanks{e-mail: fominov@landau.ac.ru}}
20: \rauthor{Ya.\,V.~Fominov}
21: \sodauthor{Fominov}
22:
23: \address{L.\,D.~Landau Institute for Theoretical Physics RAS, 119334 Moscow, Russia}
24:
25: \dates{15 March 2008}{*}
26:
27: \abstract{The conductance of a junction between a normal metal and a superconductor having the symmetry proposed by
28: Berezinskii is studied theoretically. The main feature of this symmetry is the odd frequency dependence of the anomalous
29: Green function, which makes possible the \textit{s}-wave triplet superconducting state (the Berezinskii superconductor).
30: The Andreev reflection (which links positive and negative energies) is sensitive to the energetic symmetry; as a result,
31: the conductance of the junction involving the Berezinskii superconductor is qualitatively different from the case of a
32: conventional superconductor. Experimentally, the obtained results can be employed to test the possibility of the
33: Berezinskii superconductivity proposed for Na$_x$CoO$_2$ and to identify the odd-$\omega$ component predicted for
34: superconductor-ferromagnet systems.\\ \centerline{\textbf{Published as: JETP Letters \textbf{86}, 732 (2007) [Pis'ma v
35: ZhETF \textbf{86}, 842 (2007)]}}}
36:
37: \PACS{74.45.+c, 74.20.Rp, 74.50.+r, 74.25.Fy}
38:
39: % 74.45.+c Proximity effects; Andreev effect; SN and SNS junctions
40:
41: % 74.20.Rp Pairing symmetries (other than s-wave)
42:
43: % 74.50.+r Tunneling phenomena; point contacts, weak links, Josephson effects (for SQUIDs, see 85.25.Dq; for Josephson
44: % devices, see 85.25.Cp; for Josephson junction arrays, see 74.81.Fa)
45:
46: % 74.25.Fy Transport properties (electric and thermal conductivity, thermoelectric effects, etc.)
47:
48:
49: \begin{document}
50: \maketitle
51:
52:
53:
54: The superconducting pairing can be described by the anomalous Green function $F(1;2)$ (also referred to as the pair wave
55: function) in the Matsubara technique. The Pauli principle requires antisymmetry under permutation of two electrons in a
56: Cooper pair, $F(2;1) = -F(1;2)$, which leads to the standard classification of superconducting phases \cite{Mineev}: if
57: the coordinate dependence of $F$ is even (\textit{s}-wave, \textit{d}-wave, etc.), then the spin dependence must be odd
58: (singlet) and vice versa. This assumes that $F$ is an even function of the imaginary time $\tau = \tau_1-\tau_2$ or of
59: the frequency $\omega$ in the Fourier representation.
60:
61: In 1974, Berezinskii suggested \cite{Berez} that the \textit{s}-wave triplet phase (not listed above) is also possible
62: if $F$ is an odd function of $\omega$. The Pauli principle is satisfied and there are no symmetry restrictions on the
63: existence of such a phase. Of course, the question remains whether an interaction necessary for the realization of the
64: odd-$\omega$ phase exists in real materials. Berezinskii discussed the odd-$\omega$ phase as a possible phase of $^3$He
65: and argued (without a full microscopic derivation) that it could be formed due to the retarded paramagnon exchange.
66: However, up to now there are no indications for such a state in $^3$He. Moreover, there is no microscopic theory
67: producing the Berezinskii state in a bulk material. At the same time, properties of hypothetic odd-$\omega$
68: superconductors were studied in a number of papers, e.g. \cite{AB}.
69:
70: Surprisingly, in 2001 the ``long-range'' Berezinskii superconductivity was theoretically discovered by Bergeret, Volkov,
71: and Efetov \cite{BVElong} (see also \cite{Kadigrobov}) in a SF system consisting of a conventional (\textit{s}-wave
72: singlet) superconductor and a ferromagnet. They demonstrated that the superconducting component with the symmetry
73: proposed by Berezinskii arises in the case of inhomogeneous magnetization of the ferromagnet due to the proximity effect
74: and penetrates the ferromagnet over a much longer distance compared to the singlet component, since the exchange field
75: does not suppress the triplet superconducting correlations with projections $S_z = \pm 1$. This feature enables to
76: spatially single out the Berezinskii component. Many unusual properties of this behavior have been recently investigated
77: \cite{BVE_review}. At the same time, experimental verification of the $S_z=\pm 1$ component in SF structures is still
78: under debate. A number of experiments observed a long-range proximity effect \cite{long-range} and a Josephson coupling
79: through a half-metallic ferromagnet \cite{half-metal}, which can be interpreted in terms of the long-range component.
80: However, the data on the structure of the magnetic inhomogeneity in experimental samples is lacking.
81:
82: At the same time, as argued in \cite{Mazin}, the Berezinskii superconductivity can be realized in Na$_x$CoO$_2$. This
83: conjecture is based on the band-structure calculations and on the experimental results which indicate the triplet
84: superconductivity (from the Knight-shift data) and the \textit{s}-wave symmetry (from insensitivity to impurities).
85:
86: The question arises: can we suggest an experiment which is sensitive to the most nontrivial feature of the Berezinskii
87: superconductivity, the odd-$\omega$ dependence? If yes, then this experiment could be a good test for such a state,
88: similarly to the famous experiments sensitive to the nontrivial spatial symmetry of anisotropic superconductivity (the
89: idea proposed in \cite{Geshkenbein} was employed to experimentally verify the \textit{d}-wave symmetry in YBaCuO, see
90: review \cite{VanHarlingen}). It seems natural that the odd-$\omega$ superconductivity should lead to unusual features of
91: the Andreev reflection \cite{Andreev} from such a superconductor, since this process links an electron with positive
92: energy to a hole with negative energy (with respect to the Fermi level).
93:
94: In this Letter, the differential conductance of the normal-metal---superconductor (NS) junction shown in
95: Fig.~\ref{fig:junction} is studied at zero temperature. I consider three possibilities for the superconducting
96: reservoir: S~--- a conventional superconductor with a gap (in this case previous results are reproduced, see reviews
97: \cite{LR-PC} and references therein), S$_0$~--- a conventional superconductor without a gap (e.g., due to paramagnetic
98: impurities \cite{AG}), and S$_B$~--- a Berezinskii superconductor. S$_B$ has two main features: odd-$\omega$ symmetry
99: and gapless spectrum. The case of S$_0$ is considered in order to reveal the features of S$_B$ which are due to the
100: unusual symmetry and not simply due to its gapless spectrum. Physical examples of the S$_B$ state are the bulk
101: Berezinskii superconductor and the SF system (in the latter case, the normal wire should be attached to a region where
102: only the long-range triplet component survives, while the short-range ones are negligible).
103:
104: The differential conductance of the junction is
105: \begin{equation}
106: G_{NS} (V) = \frac{dI(V)}{dV} ;
107: \end{equation}
108: its normal-state value is $G_0 = \left( G_N^{-1} + G_T^{-1} \right)^{-1}$. The
109: diffusive normal wire of length $L$ is characterized by the Thouless energy $E_\mathrm{Th} = D/L^2$, where $D$ is the
110: diffusion constant. At low voltages, the Andreev reflection plays an important role in the transport. I assume $eV$,
111: $E_\mathrm{Th} \ll E_0$, where $E_0$ is the energy scale on which the Green function of the superconductor varies (in a
112: conventional superconductor $E_0$ coincides with the static order parameter $\Delta$). At the same time, the relation
113: between $eV$ and $E_\mathrm{Th}$ is arbitrary.
114:
115: \begin{figure}
116: \centerline{\includegraphics[width=4cm]{Junction.eps}}
117: \caption{Fig.\ref{fig:junction}. NS junction. The conductances of the normal wire and the NS interface are $G_N$ and
118: $G_T$, respectively. The voltage $V$ is applied to the normal reservoir N$_\mathrm{res}$.}
119: \label{fig:junction}
120: \end{figure}
121:
122: I consider the dirty limit and employ the Usadel equation \cite{Usadel,Kopnin} for the Green function $g$ which is a
123: $8\times 8$ matrix in the Keldysh $\otimes$ Nambu-Gor'kov $\otimes$ spin space:
124: \begin{equation}
125: g = \begin{pmatrix} \check g^R & \check g^K \\ 0 & \check g^A \end{pmatrix} .
126: \end{equation}
127: The Usadel equation for $g$ is
128: \begin{equation}
129: D \nabla ( g \nabla g) + \left[ i E \hat\tau_3 \hat\sigma_0 \hat 1, g \right] =0,
130: \end{equation}
131: where $\hat\tau$ and $\hat\sigma$ denote the Pauli matrices in the Nambu-Gor'kov and spin spaces, respectively, while
132: $\hat 1$ is the unit matrix in the Keldysh space. At the interface with the normal reservoir ($x=-L$), $g$ must be equal
133: to the bulk normal-metallic function, while the Kupriyanov-Lukichev boundary condition \cite{Kupriyanov} at the NS
134: interface ($x=0$) reads
135: \begin{equation} \label{KL}
136: g \nabla g = \frac{G_T}{2 G_N L} \left[ g, g_S \right],
137: \end{equation}
138: where $g_S$ is the Green function in the superconductor.
139:
140: The current is $I = \int j dE$ with the spectral current
141: \begin{equation} \label{j}
142: j(E,x) = - \frac{G_N L}{16e} \Tr\left[ \hat\tau_3 \hat\sigma_0 ( g \nabla g )^K \right].
143: \end{equation}
144:
145: Due to the normalization condition $g^2 =1$, we can parametrize $\check g^K$ as $\check g^K = \check g^R \check f -
146: \check f \check g^A$, then the general relation $\check g^A = - \hat\tau_3 \left( \check g^R \right)^\dagger \hat\tau_3$
147: allows us to consider only $\check g^R$ and $\check f$ as independent functions.
148:
149: In S and S$_0$, we can write
150: \begin{equation} \label{g_S}
151: \check g^R = \hat\tau_3 \hat\sigma_0 G^R + \hat\tau_1 \hat\sigma_0 F^R,
152: \end{equation}
153: while S$_B$ is a spin-triplet superconductor and the anomalous part is a vector in the spin space. Choosing its
154: direction as $z$, we write \cite{z}
155: \begin{equation} \label{g_SB}
156: \check g^R = \hat\tau_3 \hat\sigma_0 G^R + \hat\tau_1 \hat\sigma_3 F^R.
157: \end{equation}
158: Physically, this is a triplet superconducting state with zero projection of the Cooper pair's spin on the $z$ axis,
159: while the $1$ and $-1$ projections on any axis in the $xy$ plane are equiprobable.
160:
161: The Berezinskii superconductivity is odd in the Matsubara frequency, $F(-\omega) = -F(\omega)$, therefore in the
162: real-energy representation we obtain $F^R(-E) = - F^A (E)$. Together with the general relation $F^A (E) = \left( F^R (E)
163: \right)^*$ this yields \cite{TG} $F^R(-E) = -\left( F^R (E) \right)^*$. We need to know the low-energy behavior of
164: $F^R(E)$, which depends on the choice of a model. The question of a microscopic model for a bulk Berezinskii
165: superconductor is not clear at present. At the same time, we know that such a state is realized in inhomogeneous SF
166: structures due to the proximity effect. Taking the low-energy behavior of the odd-$\omega$ triplet component from, e.g.,
167: \cite{FVE}, we obtain $F(\omega) = iA \sgn\omega$, hence $F^R(E) = iA$ at $E\to 0$, with real constant $A$. This
168: behavior also follows from the standard relation $F^R = \Delta/\sqrt{\Delta^2-E^2}$ in the case of the linear low-energy
169: behavior $\Delta (E) = E / (1+A^{-2})$.
170:
171: The constraint $(\check g^R)^2 =1$ allows us to parametrize the normal and anomalous Green functions by a (complex)
172: angle $\theta$ as $G^R = \cos\theta$ and $F^R = \sin\theta$ in the cases of both the conventional and Berezinskii
173: superconductor, Eqs.\ (\ref{g_S}) and (\ref{g_SB}). The equation and the boundary conditions for $\theta$ are
174: \begin{gather}
175: \frac D2 \theta'' + i E \sin\theta =0, \label{Usadel} \\
176: \theta = 0 \Bigr|_{x=-L},\quad \theta' = \frac{G_T}{G_N L} \sin (\theta_S -\theta) \Bigr|_{x=0}. \label{bc0}
177: \end{gather}
178: The superconductor is described by $\theta_S$. In order to consider low voltages, we need to find the low-energy
179: solution of the Usadel equation. If $E \ll E_0$, then $\theta_S = \pi/2$ in~S, $0< \theta_S <\pi/2$ in~S$_0$, and
180: $\theta_S = i \vartheta_S$ with real $\vartheta_S$ in~S$_B$. Thus the type of the superconductor enters our
181: consideration only through the low-energy value $\theta_S$.
182:
183: In the case of conventional superconductors, S or S$_0$, components of the distribution function can be chosen as
184: \begin{equation}
185: \check f = \hat\tau_0 \hat\sigma_0 f_0 + \hat\tau_3 \hat\sigma_0 f_3,
186: \end{equation}
187: while in the case of S$_B$ the number of components is doubled in the general case:
188: \begin{equation}
189: \check f = \hat\tau_0 \hat\sigma_0 f_0 + \hat\tau_3 \hat\sigma_0 f_3 + \hat\tau_0 \hat\sigma_3 \bar f_0 + \hat\tau_3
190: \hat\sigma_3 \bar f_3.
191: \end{equation}
192:
193: The spectral current (\ref{j}) at $x=0$ can be rewritten with the help of the boundary condition (\ref{KL}) as
194: \begin{multline} \label{I}
195: j(E) = (G_T / 8 e) f_3 \times \\
196: \times \left[ (G^R-G^A) (G_S^R-G_S^A) + (F^R+F^A) (F_S^R+F_S^A) \right].
197: \end{multline}
198: The two terms in the square brackets have clear physical meaning. Since $(G^R-G^A)/2$ is the single-particle density of
199: states (DOS), the first term in the square brackets describes the quasiparticle contribution to the current. At the same
200: time, the second term (of $FF$ type) describes the supercurrent due to the Andreev reflection.
201:
202: The subsequent derivation is similar to the conventional case \cite{VZK,LR-PC}. At zero temperature the integral over
203: energies, which determines $G_{NS}(V)$, reduces to the sum of the two terms with $E=\pm eV$. Finally, using the symmetry
204: $\theta (-E) = \theta^*(E)$ for the conventional even-$\omega$ superconductor or $\theta(-E) = - \theta^*(E)$ for the
205: odd-$\omega$ one \cite{TG}, we obtain
206: \begin{multline} \label{GNS}
207: \frac 1{G_{NS} (V)} = \frac 1{G_N L} \int_{-L}^0 \frac{dx}{\cosh^2 \theta_2 (x)} + \\
208: + \frac 1{G_T \cos(\theta_{S1}-\theta_1) \cosh\theta_{S2} \cosh\theta_2} ,
209: \end{multline}
210: where the right-hand side (r.h.s.) is taken at $E=eV$ and we have separated $\theta$ into the real and imaginary parts,
211: $\theta = \theta_1 +i\theta_2$. The angles $\theta$ and $\theta_S$ in the second term are taken at the NS interface. The
212: problem now reduces to solving Eqs.\ (\ref{Usadel})-(\ref{bc0}) and calculating the r.h.s. of Eq.\ (\ref{GNS}).
213:
214:
215:
216: We start from the simplest case of zero bias $V=0$. At $E=0$, Eqs.\ (\ref{Usadel}) and (\ref{bc0}) are solved by a
217: linear function. In the case of S or S$_0$ we obtain
218: \begin{equation}
219: G_{NS}^{-1}(0) = G_N^{-1} + \frac{G_T^{-1}}{\cos(\theta_S-\theta_0)},
220: \end{equation}
221: where $\theta_0$ must be determined from the equation $\theta_0 = (G_T/G_N) \sin(\theta_S-\theta_0)$. In the case of
222: S$_B$:
223: \begin{equation}
224: G_{NS}^{-1}(0) = \frac{G_N^{-1} \tanh\vartheta_0}{\vartheta_0} + \frac{G_T^{-1}}{\cosh\vartheta_0 \cosh\vartheta_S} ,
225: \end{equation}
226: where $\vartheta_0$ must be determined from the equation $\vartheta_0 = (G_T/G_N) \sinh (\vartheta_S-\vartheta_0)$.
227:
228: An immediate consequence of these results is that $G_{NS}(0) < G_0$ in the cases of S and S$_0$, while $G_{NS}(0) > G_0$
229: for S$_B$. In the S case, where the low-energy DOS is zero, the current is entirely due to the Andreev contribution. In
230: the S$_0$ case, a finite DOS appears and the current has both the Andreev and quasiparticle contributions.
231: Interestingly, in the S$_B$ case, only the quasiparticle processes contribute to $G_{NS}(0)$; this fact can be
232: interpreted as the absence of the Andreev reflection from S$_B$ at $E\to 0$.
233:
234: At $eV \ll E_\mathrm{Th}$, we can develop a perturbation theory finding corrections to the zero-bias solution of the
235: Usadel equation. Two orders in $eV / E_\mathrm{Th}$ yield a quadratic low-bias behavior: $G_{NS}(V) = G_{NS}(0) + a
236: V^2$. The explicit form of $a$ is cumbersome and I only present its most important features. The sign of $a$ depends on
237: the ratio $G_N / G_T$ and on the type of the superconductor: $a > 0$ for S and S$_0$ if $G_N / G_T < g_c$ and $a < 0$ if
238: $G_N / G_T > g_c$ where $g_c$ is of the order of unity and weakly depends on $\theta_S$ (as $\theta_S$ changes from
239: $\pi/2$ to $0$, the critical value $g_c$ stays between $0.8$ and $0.9$), while $a<0$ for S$_B$ at arbitrary $G_N/G_T$.
240:
241: Now let us consider in more detail the limiting cases of large and small ratio $G_N / G_T$.
242:
243: \textbf{Tunneling limit:} $G_N \gg G_T$. In this case, we can retain only the second term in the r.h.s.\ of Eq.\
244: (\ref{GNS}). The proximity effect is weak, $|\theta| \ll 1$.
245:
246: In the case of S$_0$ and S$_B$ we can set $\theta=0$, then
247: \begin{equation} \label{ts}
248: \frac{G_{NS} (V)}{G_0} = \nu_S(eV) \approx \nu_S(0) ,
249: \end{equation}
250: where $\nu_S (E)$ is the DOS in the superconductor, normalized to the normal-metallic value; this is nearly constant at
251: $E\ll E_0$ (the more accurate analysis presented above gives the small bias-dependent correction to this constant). The
252: physical meaning of Eq.\ (\ref{ts}) is the tunneling spectroscopy of the superconductor with the normal probe. An
253: essential difference between S$_0$ and S$_B$ is that $\nu_S(0) = \cos\theta_S < 1$ for S$_0$, while $\nu_S(0) =
254: \cosh\vartheta_S > 1$ for S$_B$ (this fact for S$_B$ was pointed out in \cite{TG}).
255:
256: In the case of S, $\nu_S(eV)=0$ below the gap, therefore we cannot neglect the proximity effect. Linearizing Eqs.\
257: (\ref{Usadel}) and (\ref{bc0}) over $\theta$, we find the solution and finally obtain
258: \begin{equation} \label{G_tun}
259: \frac{G_{NS} (V)}{G_0} = \frac{G_T}{G_N} \frac{\sinh (2\sqrt\varepsilon) + \sin (2\sqrt\varepsilon)}{4\sqrt\varepsilon
260: [\sinh^2 (\sqrt\varepsilon) + \cos^2 (\sqrt\varepsilon)]},\quad \varepsilon = \frac{eV}{E_\mathrm{Th}} ,
261: \end{equation}
262: hence $G_{NS} (V) \ll G_0$ [Eq.\ (\ref{G_tun}) also follows from a more general result obtained in \cite{VZK}].
263:
264: The results for the tunneling limit are summarized in Fig.~\ref{fig:G_NS_V_tun}. Note that the physics related to the
265: Andreev reflection is not essential for S$_B$ in this limit, since the transport is due to the quasiparticle
266: contribution. At the same time, the Andreev reflection plays a crucial role in the transparent limit considered below.
267:
268: \begin{figure}
269: \centerline{\includegraphics[width=80mm]{GV_tun.eps}}
270: \caption{Fig.\ref{fig:G_NS_V_tun}. Differential conductance at $eV\ll E_0$ in the tunneling limit ($G_N \gg G_T$). In
271: the S case, $G_{NS}(V)$ demonstrates the zero-bias anomaly \cite{ZBA}: it has quadratic and $1/\sqrt V$ dependences at
272: $eV \ll E_\mathrm{Th}$ and $eV \gg E_\mathrm{Th}$, respectively. In the S$_0$ and S$_B$ cases, $G_{NS}(V)$ is nearly
273: constant, smaller than $G_0$ for S$_0$ and larger than $G_0$ for S$_B$.}
274: \label{fig:G_NS_V_tun}
275: \end{figure}
276:
277: \begin{figure}
278: \centerline{\includegraphics[width=80mm]{GV_transp.eps}}
279: \caption{Fig.\ref{fig:G_NS_V}. Differential conductance at $eV\ll E_0$ in the transparent limit ($G_N \ll G_T$). In all
280: the three cases, $G_{NS}(V)$ is quadratic at $eV \ll E_\mathrm{Th}$ and approaches unity as $1/\sqrt V$ at $eV \gg
281: E_\mathrm{Th}$. In the S and S$_0$ cases, the behavior is reentrant (see \cite{reentrance} for the S case). The maximum
282: value of $G_{NS}(V)/G_0$ (achieved at $eV$ of the order of several $E_\mathrm{Th}$) approximately equals 1.15 for the S
283: case, while for S$_0$ the curve is closer to unity. In the S$_B$ case, $G_{NS}(V)$ monotonically decreases.}
284: \label{fig:G_NS_V}
285: \end{figure}
286:
287:
288:
289: \textbf{Transparent limit:} $G_N \ll G_T$. In this case, we can retain only the first term in the r.h.s.\ of Eq.\
290: (\ref{GNS}).
291:
292: At $eV \ll E_\mathrm{Th}$ we calculate a small correction to the zero-bias conductance. For S and S$_0$ we obtain
293: \begin{equation}
294: \frac{G_{NS} (V)}{G_0} = 1+ A \left( \frac{eV}{E_\mathrm{Th}} \right)^2
295: \end{equation}
296: with the positive coefficient
297: \begin{equation}
298: A = \frac 4{\theta_S^4} \left( \frac 12 + \frac{\sin^2 \theta_S}3 + \frac{3 \sin 2\theta_S}{4\theta_S} - \frac{2\sin^2
299: \theta_S}{\theta_S^2} \right).
300: \end{equation}
301: $A(\theta_S)$ monotonically grows from zero at $\theta_S = 0$ to approximately $0.015$ at $\theta_S =\pi/2$.
302:
303: The low-bias conductance for S$_B$ is
304: \begin{equation}
305: \frac{G_{NS} (V)}{G_0} = \frac{\vartheta_S}{\tanh\vartheta_S} -B \left( \frac{eV}{E_\mathrm{Th}} \right)^2
306: \end{equation}
307: with the positive coefficient
308: \begin{equation}
309: B = \frac 2{\vartheta_S^2 \tanh^2 \vartheta_S} + \frac 3{\vartheta_S^3 \tanh\vartheta_S} + \frac{3 \cosh^2
310: \vartheta_S}{\vartheta_S^4} - \frac{4 \sinh 2\vartheta_S}{\vartheta_S^5}.
311: \end{equation}
312: $B(\vartheta_S)$ is a monotonically growing function starting from zero at $\vartheta_S = 0$. Although $\vartheta_S$ is
313: an unknown parameter, it can in principle be determined from measurements in the tunneling limit [see Eq.\ (\ref{ts})
314: with $\nu_S(0) = \cosh\vartheta_S$].
315:
316: At $eV \gg E_\mathrm{Th}$ all the three cases (S, S$_0$, and S$_B$) are treated in a similar manner. Since $G_N \ll
317: G_T$, the boundary condition (\ref{bc0}) at $x=0$ reduces to $\theta = \theta_S$. The well-known solution of the
318: sine-Gordon equation (\ref{Usadel}) with fixed surface value is
319: \begin{equation} \label{4atan}
320: \theta(x) = 4 \arctan \biggl\{ \tan\left( \frac{\theta_S}4 \right) \exp\biggl( -(1-i) |x| \sqrt{\frac ED} \biggr)
321: \biggr\}.
322: \end{equation}
323: This solution satisfies the boundary condition $\theta(-L) = 0$ with good accuracy, because $\theta(-L)$ is
324: exponentially small at $E \gg E_\mathrm{Th}$. Substituting Eq.\ (\ref{4atan}) into the first term in the r.h.s.\ of Eq.\
325: (\ref{GNS}), we obtain
326: \begin{equation}
327: \frac{G_{NS}(V)}{G_0} = 1 + \int_{-\infty}^0 \frac{dx}L \tanh^2 \theta_2 (x) = 1 + C \sqrt{\frac{E_\mathrm{Th}}{eV}},
328: \end{equation}
329: where the positive coefficient $C$ depends only on $\theta_S$, i.e., on the type of the superconductor. In the S case,
330: $C\approx 0.3$.
331:
332: The results for the transparent limit are summarized in Fig.~\ref{fig:G_NS_V}. Note that in \cite{BVElong} and
333: \cite{BVE_review} a different, so-called cross geometry was considered under assumption of weak proximity effect. Then a
334: small correction to the normal-state conductance due to the S$_B$ component was numerically shown to monotonically
335: decrease as a function of temperature at zero bias.
336:
337: In conclusion, the conductance of the junction between a normal metal and a Berezinskii superconductor (odd-$\omega$
338: spin-triplet \textit{s}-wave state) has been studied. The main differences from the case of a conventional
339: superconductor are: (i)~in the tunneling limit, $G_{NS}(V)$ is larger than the normal-state conductance
340: (Fig.~\ref{fig:G_NS_V_tun}), and (ii)~in the transparent limit, $G_{NS}(V)$ monotonically decreases
341: (Fig.~\ref{fig:G_NS_V}). These features can be used as an experimental test for the Berezinskii superconductivity in
342: bulk samples (e.g., Na$_x$CoO$_2$) or in superconductor-ferromagnet proximity systems.
343:
344:
345:
346:
347:
348:
349:
350:
351:
352: I am grateful to M.\,V.\ Feigel'man and I.\,I.\ Mazin for drawing my attention to this problem and for helpful
353: discussions. I am also grateful to A.\,A.\ Golubov, Y.\ Tanaka, and A.\,F.\ Volkov for useful discussions of the
354: results. The research was supported by the Dynasty Foundation, CRDF, the Russian Ministry of Education, RF Presidential
355: Grant No.\ MK-4421.2007.2, RFBR Grants Nos.\ 07-02-01300 and 07-02-00310, and the program ``Quantum Macrophysics'' of
356: the RAS.
357:
358:
359:
360:
361: \begin{thebibliography}{99}
362:
363: % experimental ZBA \cite{Kastalsky}; explanation by impurities \cite{vanWees}; $G_T^2/G_N$ \cite{HN}
364:
365: % experimental reentrance \cite{Charlat}
366:
367: \bibitem{Mineev}
368: V.\,P.\ Mineev and K.\,V.\ Samokhin, \textit{Introduction to Unconventional Superconductivity} (Gordon and Breach,
369: London, 1999).
370:
371: \bibitem{Berez}
372: V.\,L.\ Berezinskii, Pis'ma Zh.\ Eksp.\ Teor.\ Fiz.\ \textbf{20}, 628 (1974) [JETP Lett.\ \textbf{20}, 287 (1974)].
373:
374: \bibitem{AB}
375: E.\ Abrahams, A.\ Balatsky, D.\,J.\ Scalapino, and J.\,R.\ Schrieffer, Phys.\ Rev.~B \textbf{52}, 1271 (1995).
376:
377: \bibitem{BVElong}
378: F.\,S.\ Bergeret, A.\,F.\ Volkov, and K.\,B.\ Efetov, Phys.\ Rev.\ Lett.\ \textbf{86}, 4096 (2001).
379:
380: \bibitem{Kadigrobov}
381: A.\ Kadigrobov, R.\,I.\ Shekhter, and M.\ Jonson, Europhys.\ Lett.\ \textbf{54}, 394 (2001).
382:
383: \bibitem{BVE_review}
384: F.\,S.\ Bergeret, A.\,F.\ Volkov, and K.\,B.\ Efetov, Rev.\ Mod.\ Phys.\ \textbf{77}, 1321 (2005).
385:
386: \bibitem{long-range}
387: V.\,T.\ Petrashov, V.\,N.\ Antonov, S.\,V.\ Maksimov, and R.\,Sh.\ Shaikhaidarov, Pis'ma Zh.\ Eksp.\ Teor.\ Fiz.\
388: \textbf{59}, 523 (1994) [JETP Lett.\ \textbf{59}, 551 (1994)]; M.\,D.\ Lawrence and N.\ Giordano, J.~Phys.: Condens.\
389: Matter \textbf{8}, L563 (1996); M.\ Giroud, H.\ Courtois, K.\ Hasselbach et al., Phys.\ Rev.~B \textbf{58}, R11872
390: (1998).
391: %M.\ Giroud, H.\ Courtois, K.\ Hasselbach, D.\ Mailly, and B.\ Pannetier, Phys.\ Rev.~B \textbf{58}, R11872 (1998).
392:
393: \bibitem{half-metal}
394: V.\ Pe$\mathrm{\tilde n}$a, Z.\ Sefrioui, D.\ Arias et al., Phys.\ Rev.~B \textbf{69}, 224502 (2004);
395: %V.\ Pe\~{n}a, Z.\ Sefrioui, D.\ Arias, C.\ Leon, J.\ Santamaria, M.\ Varela, S.~J.\ Pennycook, and J.~L.\ Martinez,
396: %Phys.\ Rev.~B \textbf{69}, 224502 (2004).
397: R.\,S.\ Keizer, S.\,T.\,B.\ Goennenwein, T.\,M.\ Klapwijk et al., Nature (London) \textbf{439}, 825 (2006).
398: %R.~S.\ Keizer, S.~T.~B.\ Goennenwein, T.~M.\ Klapwijk, G.\ Miao, G.\ Xiao, and A.\ Gupta, Nature (London) \textbf{439},
399: %825 (2006).
400:
401: \bibitem{Mazin}
402: M.\,D.\ Johannes, I.\,I.\ Mazin, D.\,J.\ Singh, and D.\,A.\ Papaconstantopoulos, Phys.\ Rev.\ Lett.\ \textbf{93}, 097005
403: (2004).
404:
405: \bibitem{Geshkenbein}
406: V.\,B.\ Geshkenbein, A.\,I.\ Larkin, and A.\ Barone, Phys.\ Rev.~B \textbf{36}, 235 (1987).
407:
408: \bibitem{VanHarlingen}
409: D.\,J.\ Van Harlingen, Rev.\ Mod.\ Phys.\ \textbf{67}, 515 (1995).
410:
411: \bibitem{Andreev}
412: A.\,F.\ Andreev, Zh.\ Eksp.\ Teor.\ Fiz.\ \textbf{46}, 1823 (1964) [Sov.\ Phys.\ JETP \textbf{19}, 1228 (1964)].
413:
414: \bibitem{LR-PC}
415: C.\,J.\ Lambert and R.\ Raimondi, J.\ Phys.: Condens.\ Matter \textbf{10}, 901 (1998); B.\ Pannetier and H.\ Courtois,
416: J.\ Low Temp.\ Phys.\ \textbf{118}, 599 (2000).
417:
418: \bibitem{AG}
419: A.\,A.\ Abrikosov and L.\,P.\ Gor'kov, Zh.\ Eksp.\ Teor.\ Fiz.\ \textbf{39}, 1781 (1960) [Sov.\ Phys.\ JETP \textbf{12},
420: 1243 (1961)].
421:
422: \bibitem{Usadel}
423: K.\,D.\ Usadel, Phys.\ Rev.\ Lett.\ \textbf{25}, 507 (1970).
424:
425: \bibitem{Kopnin}
426: N.\,B.\ Kopnin, \textit{Theory of Nonequilibrium Superconductivity} (Oxford University Press, Oxford, 2001).
427:
428: \bibitem{Kupriyanov}
429: M.\,Yu.\ Kuprianov and V.\,F.\ Lukichev, Zh.\ Eksp.\ Teor.\ Fiz.\ \textbf{94}, 139 (1988) [Sov.\ Phys.\ JETP
430: \textbf{67}, 1163 (1988)].
431:
432: \bibitem{z}
433: Of course, the direction in the spin space can be chosen arbitrarily, hence the difference from papers considering the
434: long-range triplet component (LRTC) in SF structures (where the $z$ axis is usually chosen as the direction of the
435: exchange field in ferromagnetic regions with constant magnetization, while the LRTC corresponds to $\hat\sigma_1$ or
436: $\hat\sigma_2$ \cite{BVE_review,FVE}) is only in notations.
437:
438: \bibitem{TG}
439: Y.\ Tanaka and A.\,A.\ Golubov, Phys.\ Rev.\ Lett.\ \textbf{98}, 037003 (2007); Y.\ Tanaka, Y.\ Asano, A.\,A.\ Golubov,
440: and S.\ Kashiwaya, Phys.\ Rev.~B \textbf{72}, 140503(R) (2005).
441:
442: \bibitem{FVE}
443: Ya.\,V.\ Fominov, A.\,F.\ Volkov, and K.\,B.\ Efetov, Phys.\ Rev.~B \textbf{75}, 104509 (2007).
444:
445: \bibitem{VZK}
446: A.\,F.\ Volkov, A.\,V.\ Zaitsev, and T.\,M.\ Klapwijk, Physica~C \textbf{210}, 21 (1993).
447:
448: \bibitem{ZBA}
449: A.\ Kastalsky, A.\,W.\ Kleinsasser, L.\,H.\ Greene et al., Phys.\ Rev.\ Lett.\ \textbf{67}, 3026 (1991); B.\,J.\
450: van~Wees, P.\ de~Vries, P.\ Magn\'ee, and T.\,M.\ Klapwijk, Phys.\ Rev.\ Lett.\ \textbf{69}, 510 (1992); F.\,W.\,J.\
451: Hekking and Yu.\,V.\ Nazarov, Phys.\ Rev.~B \textbf{49}, 6847 (1994).
452:
453: \bibitem{reentrance}
454: A.\ Volkov, N.\ Allsopp, and C.\,J.\ Lambert, J.\ Phys.: Condens.\ Matter \textbf{8}, L45 (1996); P.\ Charlat, H.\
455: Courtois, Ph.\ Gandit et al., Phys.\ Rev.\ Lett.\ \textbf{77}, 4950 (1996).
456:
457: \end{thebibliography}
458:
459: \end{document}
460: