1: % Author: Minchul Lee
2: % $Date: 2008/07/07 08:26:37 $
3: % $Revision: 1.8 $
4:
5: \documentclass[aps,prl,twocolumn,showpacs,floatfix,superscriptaddress]{revtex4}
6:
7: \usepackage{amsmath,amssymb}
8: \usepackage[dvips]{graphics}
9: \usepackage{epsfig}
10:
11: % definitions of new commands
12:
13: \newcommand{\varA}{{\mathcal{A}}}
14: \newcommand{\varE}{{\mathcal{E}}}
15: \newcommand{\varH}{{\mathcal{H}}}
16:
17: \newcommand{\bfk}{{\mathbf{k}}}
18: \newcommand{\bfS}{{\mathbf{S}}}
19: \newcommand{\bfsigma}{{\boldsymbol{\sigma}}}
20:
21: \newcommand{\E}{{\rm e}}
22:
23: \newcommand{\up}{{\uparrow}}
24: \newcommand{\down}{{\downarrow}}
25:
26: \newcommand{\ode}[3][]{\frac{d^{#1}{#2}}{d{#3}^{#1}}}
27:
28: \newcommand{\eqnref}[1]{Eq.~(\ref{#1})}
29: \newcommand{\Eqnref}[1]{Equation~(\ref{#1})}
30: \newcommand{\figref}[1]{Fig.~\ref{#1}}
31: \newcommand{\Figref}[1]{Figure~\ref{#1}}
32: \newcommand{\Figsref}[1]{Figures~\ref{#1}}
33:
34: \begin{document}
35: \title{Josephson Effect through an Isotropic Magnetic Molecule}
36: % comment, where does it say that there is a spin ? there is no indication that
37: % the metal inside the carbon cage is magnetic.
38: \author{Minchul Lee}
39: \affiliation{Centre de Physique Th\'eorique, UMR6207, Case 907, Luminy, 13288 Marseille Cedex 9, France}
40: \author{Thibaut Jonckheere}
41: \affiliation{Centre de Physique Th\'eorique, UMR6207, Case 907, Luminy, 13288 Marseille Cedex 9, France}
42: \author{Thierry Martin}
43: \affiliation{Centre de Physique Th\'eorique, UMR6207, Case 907, Luminy, 13288 Marseille Cedex 9, France}
44: \affiliation{Universit\'e de la M\'editerran\'ee, 13288 Marseille Cedex 9, France}
45:
46: \begin{abstract}
47: We investigate the Josephson effect through a molecular quantum dot magnet
48: connected to superconducting leads. The molecule contains a magnetic atom,
49: whose spin is assumed to be isotropic. It is coupled to the electron spin on
50: the dot via exchange coupling. Using the numerical renormalization group
51: method we calculate the Andreev levels and the supercurrent and examine
52: intertwined effect of the exchange coupling, Kondo correlation, and
53: superconductivity on the current. Exchange coupling typically suppresses the
54: Kondo correlation so that the system undergoes a phase transition from 0 to
55: $\pi$ state as the modulus of exchange coupling increases. Antiferromagnetic
56: coupling is found to drive exotic transitions: the reentrance to the $\pi$
57: state for a small superconducting gap and the restoration of 0 state for
58: large antiferromagnetic exchange coupling. We suggest that the asymmetric
59: dependence of supercurrent on the exchange coupling could be used as to
60: detect its sign in experiments.
61: \end{abstract}
62:
63: \pacs{
64: 73.63.-b, % Condensed Matter: Electronic Structure, Electrical, Magnetic, ...
65: % - Electronic structure and electrical properties of ...
66: % - Electronic transport in nanoscale materials and structure
67: 74.50.+r, % Condensed Matter: Electronic Structure, Electrical, Magnetic, ...
68: % - Superconductivity
69: % - Tunneling phenomena: point contacts, weak links, Josephson effects
70: 72.15.Qm, % - Scattering mechanisms and Kondo effect
71: 73.63.Kv % Condensed Matter: Electronic Structure, Electrical, Magnetic, ...
72: % - Electronic structure and electrical properties of ...
73: % - Quantum dots
74: }
75: \maketitle
76:
77: Molecular spintronics \cite{Rocha05} aims at exploring spin-dependent
78: electronic transport through molecules with intrinsic degrees of freedom such
79: as spin, connected to leads of various
80: nature. % (normal metal, ferromagnetic or superconducting).
81: On the theoretical and experimental side, recent advances have concerned both
82: coherent \cite{Romeike1} and incoherent \cite{Elste,Romeike2,Heersche06}
83: transport through these molecular quantum dot magnets (MQDM). They consist of
84: a magnetic molecule with either a large \cite{Sessoli93} or a small anisotropy,
85: as is the case for a endofullerene molecule \cite{Grose}.
86: % (a carbon shell with one or more magnetic atoms inside it).
87:
88: Here, we provide a nonperturbative computation of the low temperature transport
89: properties of a MQDM connected to superconducting leads using a numeral
90: renormalization group (NRG) approach. The Josephson current allows a diagnosis
91: of the interaction between the intrinsic spin of the molecule, its itinerant
92: electron spin, and the polarization of the leads. It has been known for some
93: time \cite{Shiba69,Glazman89,Spivak91,Rozhkov}, and recently analyzed in
94: experiments \cite{vanDam06}, that a quantum dot sandwiched between
95: superconducting leads can show a $\pi$ junction behavior \cite{ExpPiJunction}.
96: At the same time, a quantum dot connected to leads at low enough temperatures
97: exhibits the Kondo effect \cite{golhaber_gordon}. It was shown
98: \cite{Glazman89,Choi04,Siano04} that with superconducting leads, at low
99: temperature the 0 junction state of the Josephson current is restored when the
100: Kondo temperature exceeds the superconducting gap. The stability of this Kondo
101: phase is put in question in the presence of additional spin degrees for freedom
102: \cite{Bergeret06} which may compete with Kondo screening. Here the Josephson
103: current flows through an isotropic MQDM which can describe a endofullerene
104: molecule \cite{Kasumov05}. The electron spin in the quantum dot and the
105: magnetic ion inside it interact via an exchange coupling \cite{Elste}. We
106: calculate the Andreev level (AL) spectrum and the supercurrent and determine
107: the spin of the ground state. We find that the exchange coupling typically
108: suppresses the Kondo effect and drives a transition from 0 to $\pi$
109: state. Moreover, antiferromagnetic coupling is found to drive exotic
110: transitions: the reentrance to $\pi$ state for small superconducting gap and
111: the restoration of 0 state for large $J$.
112:
113: The MQDM connected to two $s$-wave superconducting leads (inset of
114: \figref{fig:1}) is modeled by a single-impurity Anderson model: $\varH =
115: \varH_{\rm M} + \varH_{\rm L} + \varH_{\rm T}$, where
116: \begin{align}
117: \varH_{\rm M}
118: & =
119: \epsilon_0 n + U n_\up n_\down + J \bfS\cdot\bfS_\E
120: \\
121: \varH_{\rm L}
122: & =
123: \sum_{\ell\bfk}
124: \left[
125: \epsilon_{\bfk} n_{\ell\bfk}
126: -
127: \left(
128: \Delta\, e^{i\phi_\ell} c_{\ell\bfk\up}^\dag c_{\ell-\bfk\down}^\dag
129: + (h.c.)
130: \right)
131: \right]
132: \\
133: \varH_{\rm T}
134: & =
135: \sum_{\ell\bfk\mu}
136: \left[t\, d_\mu^\dag c_{\ell\bfk\mu} + (h.c.)\right].
137: \end{align}
138: Here $c_{\ell\bfk\mu}$ ($d_\mu$) destroys an electron with energy
139: $\epsilon_{\bfk}$,
140: %momentum $\bfk$,
141: and spin $\mu$ on lead $\ell=L,R$ (on the carbon cell); $n_{\ell\bfk}$
142: %\equiv \sum_\mu c_{\ell\bfk\mu}^\dag c_{\ell\bfk\mu}$
143: and $n$
144: % \equiv \sum_\mu d_\mu^\dag d_\mu$
145: are occupation operators for the leads and the cell. The single-particle
146: energy $\epsilon_0$ can be tuned by gate voltages. $J$ denotes the exchange
147: energy between the ion spin $\bfS$ and the electron spin $\bfS_\E = \frac12
148: \sum_{\mu\mu'} d_\mu^\dag \bfsigma_{\mu\mu'} d_{\mu'}$.
149: %, where $\bfsigma =(\sigma_x,\sigma_y,\sigma_z)$ are Pauli matrices.
150: $\Delta$ is the superconducting gap. Except for the finite phase difference
151: $\phi=\phi_L-\phi_R$, the leads are identical and their coupling to the MQDM is
152: symmetric. The hybridization between the molecule and the leads is well
153: characterized by a tunneling rate $\Gamma = \pi \rho_0 |t|^2$, where $\rho_0$
154: is the density of states of the leads at the Fermi energy. As we are interested
155: in the low temperature behavior, we concentrate for the most part on the Kondo
156: regime with a localized level $-\epsilon_0 \gg \Gamma$ with large charging
157: energy $U\gg|\epsilon_0|$. Specifically, we choose $\epsilon_0 = -0.1 D$ (the
158: band width $D$ is taken as a unit of energy), $\Gamma = 0.01 D$, and $U=\infty$
159: and introduce the bare Kondo temperature $T_K^0 = \sqrt{D\Gamma/2} \exp
160: \left[\frac{\pi\epsilon_0}{2\Gamma}\left(1 + \frac{\epsilon_0}{U}\right)
161: \right]$ (at $J=\Delta=0$). The energy spectrum is found with the NRG method
162: \cite{NRG} extended to superconducting leads \cite{Yoshioka00,Choi04}. Within
163: the NRG method, the supercurrent is directly obtained by evaluating the
164: expectation value of the current operator \cite{Choi04}.
165:
166: \begin{figure}[!t]
167: \centering
168: \includegraphics[width=6.5cm]{Fig1}
169: \caption{(color online) Schematic phase diagram of a MQDM superconducting
170: junction system [see the upper inset] indicating the 0, 0' (blue), $\pi'$
171: (green), and $\pi$ regions. Each region is divided into two subregions
172: according to the ground-state spin: $S$ and $S{-}1/2$ for $0_1^{(\prime)}$
173: and $0_2$ regions and $S{-}1/2$ and $S{+}1/2$ for $\pi_1^{(\prime)}$ and
174: $\pi_2^{(\prime)}$ regions, respectively. Note that the $0_1$ state exists
175: only along the line $J{=}0$ [see the lower inset]. For larger molecular
176: spin $S' {>} S$ (see the dotted lines), the phase boundaries between $0_1$
177: and $\pi_{1/2}$ are shifted toward smaller $|J|$, and one between $0_2$ and
178: $\pi_1$ moves toward larger $J$.}
179: \label{fig:1}
180: \end{figure}
181:
182: \figref{fig:1} shows the phase diagram of our system, which constitutes the
183: main result. The junction property switches between 0 and $\pi$ state,
184: depending on the strengths of $J$ and $\Delta$ with respect to $T_K^0$. For
185: $J=0$, the system undergoes the Kondo-driven phase transition
186: \cite{Glazman89,Choi04,Siano04}: The ground-state wave function is of spin
187: singlet kind for $\Delta < \Delta_c \approx 1.84\, T_K^0$ and of spin doublet
188: for $\Delta > \Delta_c$. In the strong coupling limit ($\Delta < \Delta_c$)
189: Kondo correlations screen out the localized spin and Cooper pairs tunnel
190: through the Kondo resonance state, resulting in a 0-junction \cite{%Rozhkov,
191: Choi04,Siano04}. In the weak coupling limit ($\Delta > \Delta_c$), strong
192: superconductivity in the leads leaves the local spin unscreened and the
193: tunneling of Cooper pairs subject to strong Coulomb interaction acquires an
194: additional phase $\pi$, making a $\pi$-junction
195: \cite{Shiba69,Glazman89,Spivak91,Choi04,Siano04}. It is also found
196: \cite{Choi04} that the transition is $\phi$-dependent so that a narrow region
197: of the intermediate states $0'$ and $\pi'$ exists; see the enlarged view in
198: \figref{fig:1}.
199:
200: \begin{figure}[!t]
201: \centering
202: \includegraphics[width=8cm]{Fig2}
203: \caption{(color online) (a) ALs in units of $\Delta$ and (b) supercurrents
204: $I$ in units of $I_c^s \equiv e\Delta/\hbar$ as functions of $\phi$ in the
205: strong coupling limit ($\Delta/T_K^0 = 0.1$) for various values of
206: $J/T_K^0$: see the line $aa'$ in \figref{fig:1}. Here the ion spin $S$ is
207: set to $1/2$.}
208: \label{fig:2}
209: \end{figure}
210:
211: Finite exchange coupling between electron spins and the ion spin introduces
212: another electronic correlation and affects Cooper pair
213: transport. \figref{fig:2} shows typical variations of ALs and supercurrents
214: with $J$ along the line $aa'$ (see \figref{fig:1}) in the strong coupling limit
215: ($\Delta/T_K^0 = 0.1$). Any finite $J$ clearly induces a splitting in subgap
216: excitations and consequently causes a crossing between the ground state and the
217: lowest excitation at $\phi\ne\pi$ (at least for $|J/T_K^0| \lesssim O(1)$); the
218: level crossing otherwise takes place only at $\phi=\pi$. Across the crossing,
219: the ground state spin is changed from $S$ to $S{\mp}1/2$ for
220: $J\gtrless0$. Similarly, the ALs defined as the one-electron/hole subgap
221: excitations (identified as the poles of the dot Green's functions
222: \cite{Andreev}) exhibit discontinuities like kinks in the spectra; for
223: $J\gtrless0$ two outmost ALs with spin $S{\pm}1/2$ with respect to the spin-$S$
224: ground state cannot remain as (spin-1/2) one-electron excitations with respect
225: to the ground state with spin $S{\mp}1/2$ at the transition and are replaced by
226: new ALs with spin $S{\mp}1$. In parallel with an abrupt change in ALs, the
227: supercurrent-phase relation (SPR) shows a discontinuous sign change (note that
228: $I \propto - \partial E_A/\partial\phi$, as the continuum-excitation
229: contribution is negligible \cite{Andreev}), culminating in a transition from 0
230: to $\pi$ state: two $\pi^{(\prime)}$ states labeled as $\pi_{1,2}^{(\prime)}$
231: are identified according to the ground-state spin $S\mp1/2$, respectively. The
232: intermediate states $0'_1$ and $\pi'_{1/2}$ are defined as in Ref.
233: \cite{Rozhkov}. The full 0 state exists only at $J=0$ because any small $J$
234: drives the system to the $\pi$ state at $\phi=\pi$; see \figref{fig:2}. The
235: curve of $I(\phi)$ then has three distinct segments \cite{Choi04}. The central
236: segment resembles that of a short ballistic junction, while the two surrounding
237: segments are parts of $\pi$-junction curve. As $J$ grows in magnitude the
238: central segment shrinks and eventually vanishes. The SPR then becomes
239: sinusoidal like in a tunnel junction. It should be noted that the 0-$\pi$
240: transition is asymmetric with respect to the sign of $J$: the transition for
241: $J>0$ takes place at $\delta E_{\rm S}\sim T_K^0$, where $\delta E_{\rm S} =
242: \frac{J}{2}(2S{+}1)$ is the exchange-coupling energy gap, while the 0 state
243: survives much larger ferromagnetic coupling ($J<0$). Once the $\pi$-junction is
244: fully established, stronger ferromagnetic coupling does not lead to any
245: qualitative change in the SPR, while a second transition back to 0 state is
246: observed for large antiferromagnetic coupling $(J\gg\Delta)$. The NRG results
247: distinguish the second 0 state ($0_2$) from the former one ($0_1$) in three
248: points: (1) the ground state has spin $S-1/2$ like the $\pi_1$ phase, (2) the
249: SPR is that of a tunneling junction, and (3) the $\pi_1$-$0_2$ transition has
250: no intermediate state. \Figref{fig:3}~(c) shows that the critical current has
251: its maximum at $J=0$ and decreases with increasing $|J|$ rapidly across the
252: phase boundary for $J>0$ or rather gradually for $J<0$. The critical current
253: totally vanishes at the $\pi_1$-$0_2$ boundary and increases again slowly with
254: $J$ in $0_2$ phase (see the curve for $\Delta/T_K^0=0.01$).
255:
256: The 0-$\pi$ transitions ($0_1$-$\pi_1$ and $0_1$-$\pi_2$) can be attributed to
257: the competition between superconducting and Kondo correlations as in the
258: absence of exchange coupling. The relevant parameters are then the Kondo
259: temperature $T_K$ and the superconducting gap $\Delta$, and the 0-$\pi$ phase
260: transition occurs when they are comparable to each other: In our choice of
261: parameters the transition happens at $\Delta_c/T_K \approx 1.84$. The exchange
262: coupling manifests itself by renormalizing the Kondo temperature $T_K(J)$. To
263: see this, we applied the poor man's scaling theory to a corresponding Kondo
264: Hamiltonian with no superconductivity and $S=1/2$: $\varH_{\rm KM} = \sum_\bfk
265: \epsilon_\bfk n_\bfk + J \bfS\cdot\bfS_\E + (J_{\rm K} \bfS_\E + J_{\rm M}
266: \bfS) \cdot\bfS_{\rm L}$, where $\bfS_{\rm L}$ is the spin operator for the
267: lead electrons at molecule site. The last term $\bfS\cdot\bfS_{\rm L}$
268: describing direct coupling between spins of the ion and the lead electrons
269: arises during the scaling process. The renormalization group analysis leads to
270: the following scaling equations: together with $J \approx J(\Lambda=D)$,
271: \begin{equation}
272: \label{eq:se}
273: \ode{J_{\rm K/M}}{\ln\Lambda}
274: \approx
275: - \rho_0 J_{\rm K/M}^2 + \frac{J}{4D} (2 J_{\rm K} J_{\rm M} - J_{\rm M/K}^2)~.
276: \end{equation}
277: %\begin{subequations}
278: % \begin{align}
279: % \label{eq:se}
280: % \ode{J_{\rm K}}{\ln\Lambda}
281: % & \approx
282: % - \rho_0 J_{\rm K}^2 + \frac{J}{4D} (2 J_{\rm K} J_{\rm M} - J_{\rm M}^2)
283: % \\
284: % \ode{J_{\rm M}}{\ln\Lambda}
285: % & \approx
286: % - \rho_0 J_{\rm M}^2 + \frac{J}{4D} (2 J_{\rm K} J_{\rm M} - J_{\rm K}^2).
287: % \end{align}
288: %\end{subequations}
289: As the band width $\Lambda$ is decreased from $D$ to $T_K$, the coefficient
290: $J_{\rm K}$, responsible for the Kondo correlation, diverges and the scaling
291: breaks down. In the presence of finite exchange coupling, however, since $J
292: J_{\rm K} J_{\rm M} > 0$ with $J_{\rm M}(\Lambda=D) = 0$ and $|J_{\rm M}| \ll
293: J_{\rm K}$, the term proportional to $J$ in \eqnref{eq:se} turns out to slow
294: down the flow of $J_{\rm K}$ and accordingly lowers the Kondo temperature. This
295: point is confirmed by NRG calculations applied in the absence of
296: superconductivity. As can be seen in \figref{fig:3}~(a) and (b), the width of
297: the spectral density for dot electrons, identified as the Kondo temperature
298: $T_K(J)$, decreases with increasing $|J|$ (for $J<0$ this decrease, being
299: marginal, is not clearly shown with the logarithmic scale). We find out that
300: for the ferromagnetic case the ratio $T_K(J)/T_K^0$ coincides with
301: $\Delta_c(J)/\Delta_c(J=0)$. For the antiferromagnetic case, the Kondo
302: correlation is observed to be suppressed not only by the Kondo peak narrowing
303: but by lowering the peak height.
304:
305: \begin{figure}[!t]
306: \centering
307: \includegraphics[width=4.25cm]{Fig3a}%
308: \includegraphics[width=4.25cm]{Fig3b}\\
309: \includegraphics[width=4.5cm]{Fig3c}
310: \caption{(color online) Spectral weights $\varA(\omega)$ for dot
311: electrons coupled to normal leads with antiferromagnetic [(a)] and
312: ferromagnetic [(b)] exchange coupling to ion spin for various values of
313: $J/T_K^0$ (as annotated). (c) Critical currents as functions of $J/T_K^0$
314: for different values of $\Delta/T_K^0$ (see the annotations). The arrows
315: locate transition points corresponding to data with the same color. Here we
316: have used $S=1/2$.}
317: \label{fig:3}
318: \end{figure}
319:
320: Antiferromagnetic exchange coupling can, on the other hand, exert a more
321: profound effect than simply renormalizing the Kondo temperature: it gives rise
322: to a reentrant transition to the $\pi$ state at small $\Delta$ and restoration
323: of the 0 state for large $J$. It is known that small antiferromagnetic exchange
324: coupling $(J\lesssim T_K^0)$, studied in the context of coupled impurities
325: \cite{Vojta02} and side-coupled quantum dot systems \cite{SCQD} and observed in
326: experiments \cite{Roch08}, can produce a two-stage Kondo effect. After the
327: magnetic moment of the dot is screened by conduction electrons below
328: $T_K$, at a much lower energy scale (denoted as $T_K^J$) the ion spin is
329: screened by the local Fermi liquid that is formed on the dot. $T_K^J$ is then
330: the Kondo temperature of a magnetic moment screened by electrons of a bandwidth
331: $\sim T_K$ and density of states $\sim 1/(\pi T_K)$ \cite{SCQD}: $T_K^J \sim
332: T_K \exp\left[-\frac{\pi T_K}{J}\right]$. The second Kondo effect leads to a
333: Fano resonance and makes a dip in the dot electron density of states as shown
334: in \figref{fig:3}~(a). The dip becomes widened with $J$ and overrides the Kondo
335: peak when $T_K^J \approx T_K$ so that the Kondo effect is completely
336: overridden. As long as $\Delta > T_K^J$, the second Kondo effect does not
337: appear since the superconducting gap blocks any quasi-particle excitation with
338: energy less than $\Delta$. For $\Delta\lesssim T_K^J$, however, Cooper pairs
339: notice the suppression of the Kondo resonance level, and their tunneling is
340: governed by cotunneling under strong Coulomb interaction, forming a
341: $\pi$-junction again. Since $T_K^J$ decreases with decreasing $J$, $\Delta_c$
342: decreases to zero as $J\to0$. Note that the extremely small $T_K^J\ll T_K$
343: (unless $\delta E_{\rm S}\sim T_K^0$) might make it hard to detect the
344: reentrance even under rather weak thermal fluctuations with $T_K > T > T_K^J$.
345:
346: The revival of the 0-state for strong antiferromagnetic coupling can be
347: explained in the picture of cotunneling of Cooper pairs \cite{Spivak91}. In
348: weak coupling limit, the fourth-order perturbation theory leads to the
349: supercurrent:% given by
350: \begin{align}
351: \label{eq:Ip}
352: I
353: & =
354: \frac{4e}{\hbar} \sin\phi
355: \sum_{\bfk\bfk'}
356: t_{\rm L}^2 t_{\rm R}^2
357: \frac{u_\bfk u_{\bfk'} v_\bfk v_{\bfk'}}{\varE_{\bfk} \varE_{\bfk'}}
358: \\
359: \nonumber
360: & \qquad\mbox{}
361: \times
362: \frac{1}{2S{+}1}
363: \left(
364: \frac{1}{E_\bfk {+} E_{\bfk'}}
365: -
366: \frac{2S{+}2}{\delta E_{\rm S} {+} E_\bfk {+} E_{\bfk'}}
367: \right),
368: \end{align}
369: where $E_\bfk = \sqrt{\Delta^2 + \epsilon_\bfk^2}$, $u_\bfk = \sqrt{(1 +
370: \epsilon_\bfk/E_\bfk)/2}$, $v_\bfk = \sqrt{(1 - \epsilon_\bfk/E_\bfk)/2}$,
371: and $\varE_\bfk = - \epsilon_d - \frac{J}{2}(S+1) - E_\bfk < 0$. For
372: antiferromagnetic coupling, the ground state for the uncoupled system has spin
373: $S-1/2$. After one electron tunnels through the molecule the system can be in
374: spin eigenstate of either $S-1/2$ and $S+1/2$. The latter virtual process,
375: costing more energy by the gap $\delta E_{\rm S}$, turns out to acquire a $\pi$
376: phase, contributing to a negative supercurrent. The larger amplitude of this
377: process by a factor $2S+2$ (degeneracy of the spin state $S+1/2)$ dominates
378: over spin-preserving process as long as the gap $\delta E_{\rm S}$ is
379: small. For a large gap $\delta E_{\rm S}$, however, this process becomes
380: negligible and the sign of the supercurrent is reversed. Note that according to
381: \eqnref{eq:Ip} the SPR is always sinusoidal and the current should vanish at
382: the transition, which is also confirmed in our NRG calculations.
383:
384: \begin{figure}[!t]
385: \centering
386: \includegraphics[width=7.8cm]{Fig4}
387: \caption{(color online) (a) ALs in units of $\Delta$ and (b) supercurrents
388: $I$ in units of $I_c^s$ as functions of $\phi$ with $J/T_K^0 = 10$ and
389: $\Delta/T_K^0 = 0.02$ (at $\epsilon_d = -0.1$) while the gate voltage
390: $\epsilon_d$ is tuned from $-0.1$ to $-0.04$. See the line $bb'$ in
391: \figref{fig:1}.}
392: \label{fig:4}
393: \end{figure}
394:
395: The physical arguments for the 0-$\pi$ transitions discussed so far are valid
396: for arbitrary values of the ion spin $S$, while the phase boundaries are
397: shifted with changing $S$ as shown in \figref{fig:1}. The exchange-coupling
398: energy gap $\delta E_{\rm S}$ that is supposed to compete with $T_K$ increases
399: with $S$ so that for larger $S$ the transitions can occur at smaller $J$. On
400: the other hand, we have observed that the $\pi_1$-$0_2$ transition takes place
401: at slightly larger $J$ for larger $S$. This is because the increase in the
402: degeneracy factor $2S+2$ overwhelms the decrease in matrix elements due to a
403: larger energy cost by $\delta E_{\rm S}$ [see \eqnref{eq:Ip}].
404:
405: Finally, we present potential experimental manifestations of
406: exchange-coupling-driven 0-$\pi$ transition. While the direct control of
407: exchange coupling in molecules is difficult to achieve, the relative strength
408: $J/T_K^0$ can be controlled by the gate voltage which can tune the Kondo
409: temperature. \figref{fig:4} proposes a possibility to observe a double
410: transition (along the line $bb'$ in \figref{fig:1}) as the gate voltage is
411: swept. Note that the double transition is an evidence of strong exchange
412: coupling $(J\gg T_K^0\gg\Delta)$: for examples, with $T_K^0\sim3{\rm K}$
413: measured in a recent C$_{60}$ single-molecular transistor \cite{Roch08}, one
414: estimates $J\sim30{\rm K}$. Asymmetry of the phase diagram enables the sign
415: and possibly the amplitude of $J$ to be determined without ambiguity by
416: observing the evolution of the SPR or the critical current.
417:
418: The authors thank W. Wernsdorfer, F. Balestro, and Mahn-Soo Choi for helpful
419: discussions. This work is supported by ANR-PNANO Contract MolSpintronics
420: No. ANR-06-NANO-27.
421:
422: \begin{thebibliography}{99}
423:
424: \bibitem{Rocha05}
425: A. R. Rocha {\it et al.}, % V. M. Garcia-Suarez, S. W. Bailey, C. J. Lambert, J. Ferrer, and S. Sanvito,
426: Nature Materials \textbf{4}, 335 (2005).
427:
428: \bibitem{Romeike1}
429: C. Romeike {\it et al.}, % M. R. Wegewijs, W. Hofstetter, and H. Schoeller,
430: Phys. Rev. Lett. \textbf{96}, 196601 (2006);
431: {\it ibid}. \textbf{97}, 206601 (2006).
432:
433: \bibitem{Romeike2}
434: C. Romeike {\it et al.}, % M. R. Wegewijs, W. Hofstetter, and H. Schoeller,
435: Phys. Rev. Lett. \textbf{96}, 196805 (2006).
436:
437: \bibitem{Elste}
438: F. Elste, and C. Timm, Phys. Rev. B \textbf{71}, 155403 (2005);
439: {\it ibid.} \textbf{73}, 235304 (2006);
440: {\it ibid.} \textbf{73}, 235305 (2006).
441:
442: \bibitem{Heersche06}
443: H. B. Heersche {\it et al.}, % Z. de Groot, J. A. Folk, H. S. J. van der Zant, C. Romeike, M. R. Wegewijs, L. Zobbi, D. Barreca, E. Tondello, and A. Cornia
444: Phys. Rev. Lett. \textbf{96}, 206801 (2006).
445:
446: \bibitem{Sessoli93}
447: R. Sessoli {\it et al.}, % D. Gatteschi, A. Caneschi, and M. A. Novak,
448: Nature (London) \textbf{365}, 141 (1993).
449:
450: \bibitem{Grose}
451: J. E. Grose {\it et al.}, arXiv:0805.2585v1.
452:
453: \bibitem{Shiba69}
454: H. Shiba and T. Soda, Prog. Theor. Phys. \textbf{41}, 25 (1969).
455:
456: \bibitem{Glazman89}
457: L. I. Glazman and K. A. Matveev, Pis'ma Zh. Teor. Fiz. \textbf{49}, 570 (1989) [JETP Lett. \textbf{49}, 659 (1989)].
458:
459: \bibitem{Spivak91}
460: B. I. Spivak and S. A. Kivelson, Phys. Rev. B \textbf{43}, 3740 (1991).
461:
462: \bibitem{Rozhkov}
463: A. V. Rozhkov and Daniel P. Arovas, Phys. Rev. Lett. \textbf{82}, 2788 (1999);
464: A. V. Rozhkov and Daniel P. Arovas, Phys. Rev. B \textbf{62}, 6687 (2000);
465: A. V. Rozhkov {\it et al.}, Phys. Rev. B \textbf{64}, 233301 (2001).
466:
467: \bibitem{vanDam06}
468: J. A. van Dam {\it et al.}, % Y. V. Nazarov, E.P.A.M. Bakkers, S. De Franceschi, and L. K. Kouwenhoven,
469: Nature (London) \textbf{442}, 667 (2006).
470:
471: \bibitem{ExpPiJunction}
472: V. V. Ryazanov {\it et al.}, % V. A. Oboznov, A. Yu. Rusanov, A. V. Veretennikov, A. A. Golubov, and J. Aarts,
473: Phys. Rev. Lett. \textbf{86}, 2427 (2001);
474: T. Kontos {\it et al.}, % M. Aprili, J. Lesueur, F. Gen\^et, B. Stephanidis, and R. Boursier,
475: Phys. Rev. Lett. \textbf{89}, 137007 (2002).
476:
477: \bibitem{golhaber_gordon}
478: D. Goldhaber-Gordon {\it et al.}, % Hadas Shtrikman, D. Mahalu, David Abusch-Magder, U. Meirav, and M.A. Kastner,
479: Nature (London) \textbf{391}, 156 (1998);
480: S. M. Cronenwett {\it et al.}, % T. H. Oosterkamp, and L. P. Kouwenhoven,
481: Science \textbf{281}, 540 (1998).
482:
483: \bibitem{Choi04}
484: M.-S. Choi {\it et al.}, % M. Lee, K. Kang, and W. Belzig,
485: Phys. Rev. B \textbf{70}, 020502 (2004).
486:
487: \bibitem{Siano04}
488: F. Siano and R. Egger, Phys. Rev. Lett. \textbf{93}, 047002 (2004).
489:
490: \bibitem{Bergeret06}
491: F. S. Bergeret {\it et al.}, % A. Levy-Yeyati, and A. Martin-Rodero,
492: Phys. Rev. B \textbf{74}, 132505 (2006).
493:
494: \bibitem{Kasumov05}
495: A. Yu. Kasumov {\it et al.}, % K. Tsukagoshi, M. Kawamura, T. Kobayashi, Y. Aoyagi, K. Senba, T. Kodama, H. Nishikawa, I. Ikemoto, K. Kikuchi, V. T. Volkov, Yu. A. Kasumov, R. Deblock, S. Guéron, and H. Bouchiat,
496: Phys. Rev. B \textbf{72}, 033414 (2005).
497:
498: \bibitem{NRG}
499: K. G. Wilson, Rev. Mod. Phys. \textbf{47}, 773 (1975);
500: H. R. Krishnamurthy {\it et al.}, % J. W. Wilkins, and K. G. Wilson,
501: Phys. Rev. B \textbf{21}, 1003 (1980);
502: {\it ibid.}, \textbf{21}, 1044 (1980).
503:
504: \bibitem{Yoshioka00}
505: T. Yoshioka and Y. Ohashi, J. Phys. Soc. Jpn. \textbf{69}, 1812 (2000).
506:
507: \bibitem{Andreev}
508: E. Vecino {\it et al.}, % A. Martin-Rodero, and A. Levy Yeyati,
509: Phys. Rev. B \textbf{68}, 035105 (2003);
510: Phys. Rev. Lett. \textbf{91}, 266802 (2003);
511: R. L\'opez {\it et al.}, % M.-S. Choi, and R. Aguado,
512: Phys. Rev. B \textbf{75}, 045132 (2007).
513:
514: \bibitem{Vojta02}
515: M. Vojta {\it et al.}, % R. Bulla, and W. Hofstetter
516: Phys. Rev. B \textbf{65}, 140405(R) (2002).
517:
518: \bibitem{SCQD}
519: P. S. Cornaglia and D. R. Grempel, Phys. Rev. B \textbf{71}, 75305 (2005);
520: R. Zitko and J. Bonca, Phys. Rev. B \textbf{73}, 35332 (2006).
521:
522: \bibitem{Roch08}
523: N. Roch {\it et al.}, % Serge Florens, Vincent Bouchiat, Wolfgang Wernsdorfer, and Franck Balestro
524: Nature \textbf{453}, 633 (2008).
525: \end{thebibliography}
526:
527: \end{document}