1: \documentclass[bibnotes,a4paper,twocolumn,prl]{revtex4}%
2: \usepackage{amsfonts}
3: \usepackage{amsmath}
4: \usepackage{amssymb}
5: \usepackage{graphicx}%
6: \setcounter{MaxMatrixCols}{30}
7: %TCIDATA{OutputFilter=latex2.dll}
8: %TCIDATA{Version=5.00.0.2606}
9: %TCIDATA{CSTFile=revtex4.cst}
10: %TCIDATA{Created=Wednesday, October 04, 2006 15:11:01}
11: %TCIDATA{LastRevised=Thursday, January 08, 2009 11:18:26}
12: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
13: %TCIDATA{<META NAME="SaveForMode" CONTENT="3">}
14: %TCIDATA{BibliographyScheme=Manual}
15: %TCIDATA{<META NAME="DocumentShell" CONTENT="Articles\SW\REVTeX 4">}
16: %TCIDATA{Language=American English}
17: \newtheorem{theorem}{Theorem}
18: \newtheorem{acknowledgement}{Remerciement}
19: \newtheorem{algorithm}{Algorithm}
20: \newtheorem{axiom}{Axiom}
21: \newtheorem{claim}{Claim}
22: \newtheorem{conclusion}{Conclusion}
23: \newtheorem{condition}{Condition}
24: \newtheorem{conjecture}{Conjecture}
25: \newtheorem{corollary}{Corollary}
26: \newtheorem{criterion}{Criterion}
27: \newtheorem{definition}{Definition}
28: \newtheorem{example}{Example}
29: \newtheorem{exercise}{Exercise}
30: \newtheorem{lemma}{Lemma}
31: \newtheorem{notation}{Notation}
32: \newtheorem{problem}{Problem}
33: \newtheorem{proposition}{Proposition}
34: \newtheorem{remark}{Remarque}
35: \newtheorem{solution}{Solution}
36: \newtheorem{summary}{Résumé}
37: \newenvironment{proof}[1][Preuve]{\noindent\textbf{#1.} }{\ \rule{0.5em}{0.5em}}
38: \graphicspath{{Z:/SUPRA/fluctuations_fflo/article_PRB/}}
39: %BeginMSIPreambleData
40: \ifx\pdfoutput\relax\let\pdfoutput=\undefined\fi
41: \newcount\msipdfoutput
42: \ifx\pdfoutput\undefined\else
43: \ifcase\pdfoutput\else
44: \msipdfoutput=1
45: \ifx\paperwidth\undefined\else
46: \ifdim\paperheight=0pt\relax\else\pdfpageheight\paperheight\fi
47: \ifdim\paperwidth=0pt\relax\else\pdfpagewidth\paperwidth\fi
48: \fi\fi\fi
49: %EndMSIPreambleData
50: \begin{document}
51: \title{Magnetic moment manipulation by a Josephson current}
52: \author{F. K\textsc{onschelle}}
53: \email{f.konschelle@cpmoh.u-bordeaux1.fr}
54: \author{A. B\textsc{uzdin}}
55: \altaffiliation{Also at \emph{Institut Universitaire de France}.\ }
56:
57: \affiliation{Condensed Matter Theory Group, CPMOH, Universit\'{e} de Bordeaux and CNRS.
58: F-33405 Talence, France}
59: \date{\today, Document published in Phys. Rev. Lett. \textbf{102}, 017001 (2009).}
60: \startpage{1}
61:
62: \begin{abstract}
63: We consider a Josephson junction where the weak-link is formed by a
64: non-centrosymmetric ferromagnet. In such a junction, the superconducting
65: current acts as a direct driving force on the magnetic moment. We show that
66: the a.c. Josephson effect generates a magnetic precession providing then a
67: feedback to the current. Magnetic dynamics result in several anomalies of
68: current-phase relations (second harmonic, dissipative current) which are
69: strongly enhanced near the ferromagnetic resonance frequency.
70:
71: \end{abstract}
72: \maketitle
73:
74:
75: \bigskip
76:
77: Many interesting phenomena have been observed recently in the field of
78: spintronics: the spin-dependent electric current and inversely the
79: current-dependent magnetization orientation (see for example
80: \cite{zutic_fabian_dassarma_RMP.76.323.2004,hauptmann_paaske_lindelof.2008}).
81: Moreover, it is well known that spin-orbit interaction may be of primary
82: importance for spintronic, namely for systems using a two-dimensional electron
83: gas \cite{b.winkler.2003}. In the superconductor/ferromagnet/superconductor
84: (S/F/S) Josephson junctions, the spin-orbit interaction in a ferromagnet
85: without inversion symmetry provides a mechanism for a direct (linear) coupling
86: between the magnetic moment and the superconducting current
87: \cite{buzdin:107005.2008}. Similar anomalous properties have been predicted
88: for Josephson junctions with spin-polarized quantum point contact in a two
89: dimensional electron gas \cite{reynoso_etal:107001.2008}. S/F/S junctions are
90: known to reveal a transition to $\pi$-phase, where the superconducting phase
91: difference $\varphi$ in the ground state is equal to $\pi$ \cite{buzdin(2005)}%
92: . However, the current-phase relation (CPR) in such a $\pi$-junction has a
93: usual sinusoidal form $I=I_{c}\sin\varphi$, where the critical current $I_{c}$
94: depends in a damped oscillatory manner on the modulus of the ferromagnet
95: exchange field. In a non-centrosymmetric ferromagnetic junction, called
96: hereafter $\varphi_{0}$-junction, the time reversal symmetry is broken and the
97: CPR becomes $I=I_{c}\sin\left( \varphi-\varphi_{0}\right) $, where the phase
98: shift $\varphi_{0}$ is proportional to the magnetic moment perpendicular to
99: the gradient of the asymmetric spin-orbit potential \cite{buzdin:107005.2008}.
100: Therefore, manipulation of the internal magnetic moment can be achieved via
101: the superconducting phase difference (\emph{i.e. }by Josephson current).
102:
103: In the present work we study theoretically the spin dynamics associated with
104: such $\varphi_{0}$-junctions. Though there is a lot of experimental progress
105: in studying the static properties of S/F/S junctions, little is known about
106: the spin-dynamics in S/F systems. Note here the pioneering work
107: \cite{bell_Milikisyants_Huber_Aarts:047002.2008} where a sharpening of the
108: ferromagnetic resonance was observed below the superconducting transition in
109: Nb/Ni$_{80}$Fe$_{20}$ system. Theoretically, the single spin dynamics
110: interplay with a Josephson effect has been studied in
111: \cite{zhu_balatsky.2003,bulaevskii_hruska_shnirman_etal:177001.2004,zhu_nussinov_shnirman_balatsky.2004,nussinov_shnirman_arovas_balatsky_zhu:214520.2005}%
112: . More recently, the dynamically induced triplet proximity effect in S/F/S
113: junctions was studied in
114: \cite{takahashi_hikino_mori:057003.2007,houzet:057009.2008}, while the
115: junctions with composite regions (including several F regions with different
116: magnetization) were discussed in
117: \cite{Waintal_Brouwer.65.054407.2002,braude:207001.2008}. Here we consider a
118: simple S/F/S $\varphi_{0}$-junction in a low frequency regime $\hslash
119: \omega_{J}\ll T_{c}$ ($\omega_{J}=2eV/\hslash$ being the Josephson angular
120: frequency \cite{b.likharev.1986}), which allows us to use the quasi-static
121: approach to treat the superconducting subsystem in contrast with the case
122: analyzed in \cite{takahashi_hikino_mori:057003.2007,houzet:057009.2008}. We
123: demonstrate that a d.c. superconducting current may produce a strong
124: orientation effect on the F layer magnetic moment. More interestingly, the
125: a.c. Josephson effect, \emph{i.e.} applying a d.c. voltage $V$ to the
126: $\varphi_{0}$-junction, would produce current oscillations and consequently
127: magnetic precession. This precession may be monitored by the appearance of
128: higher harmonics in CPR as well as a d.c. component of the current. In
129: particular regimes, a total reversal of the magnetization could be
130: observed.\ In the case of strong coupling between magnetic and superconducting
131: subsystems, complicated non-linear dynamic regimes emerge.\
132:
133: To demonstrate the unusual properties of the $\varphi_{0}$-junction, we
134: consider the case of an easy-axis magnetic anisotropy of the F material (see
135: Fig.\ref{FIG_schema}). Both the easy axis and gradient of the asymmetric
136: spin-orbit potential $\mathbf{n}$ are along the $z$-axis. Note that suitable
137: candidates for the F interlayer may be MnSi or FeGe.\ In these systems, the
138: lack of inversion center comes from the crystalline structure, but the origin
139: of broken-inversion symmetry may also be extrinsic, like in a situation near
140: the surface of a thin F film. In the following, we completely disregard the
141: magnetic induction. Indeed the magnetic induction in the $\left( xy\right) $
142: plane is negligible for the thin F layer considered in this paper, whereas the
143: demagnetization factor cancels the internal induction along the $z$-axis
144: $\left( N=1\right) $. The coupling between F and S subsystems due to the
145: orbital effect has been studied in \cite{hikino_mori.2008} and it appears to
146: be very weak, and quadratic over magnetic moment $\mathbf{M}$ for the case
147: when the flux of $\mathbf{M}$ through the F layer is small in comparison with
148: flux quantum $\Phi_{0}=h/2e$.
149:
150: The superconducting part of the energy of a $\varphi_{0}$-junction is
151: \begin{equation}
152: E_{s}\left( \varphi,\varphi_{0}\right) =E_{J}\left[ 1-\cos\left(
153: \varphi-\varphi_{0}\right) \right] ,
154: \end{equation}
155: where $E_{J}=\Phi_{0}I_{c}/2\pi$ is the Josephson energy, $I_{c}$ is the
156: critical current and $\varphi_{0}$ is proportional to the $M_{y}$ component of
157: the magnetic moment (see Fig.\ref{FIG_schema}). Therefore, when the magnetic
158: moment is oriented along the $z$-axis, we have the usual Josephson junction
159: with $\varphi_{0}=0$. Assuming the ballistic limit we may estimate the
160: characteristic Josephson energy as \cite{buzdin(2005)} $\Phi_{0}I_{c}/S\sim
161: T_{c}k_{F}^{2}\sin\ell/\ell$ with $\ell=4hL/\hslash v_{F}$, where $S,L$ and
162: $h$ are the section, the length and the exchange field of the F layer,
163: respectively. The phase shift is \
164: \begin{equation}
165: \varphi_{0}=\ell\frac{v_{\text{so}}}{v_{F}}\frac{M_{y}}{M_{0}}%
166: \end{equation}
167: where the parameter $v_{\text{so}}/v_{F}$ characterizes the relative strength
168: of the spin-orbit interaction \cite{buzdin:107005.2008}. Further on we assume
169: that $v_{\text{so}}/v_{F}\sim0.1$. If the temperature is well below the Curie
170: temperature, $M_{0}=\left\Vert \mathbf{M}\right\Vert $ can be considered as a
171: constant equal to the saturation magnetization of the F layer. The magnetic
172: energy contribution is reduced to the anisotropy energy
173: \begin{equation}
174: E_{M}=-\frac{K\mathcal{V}}{2}\left( \frac{M_{z}}{M_{0}}\right) ^{2},
175: \end{equation}
176: where $K$ is an anisotropy constant and $\mathcal{V}$ is the volume of the F
177: layer.\begin{figure}[ptb]
178: \begin{center}
179: \includegraphics[width=3.4in,angle=0]{shema_junction_PS.eps}
180: \end{center}
181: \caption{Geometry of the considered $\varphi_{0}$-junction. The ferromagnetic
182: easy-axis is directed along the $z$-axis, which is also the direction
183: $\mathbf{n}$ of the gradient of the spin-orbit potential. The magnetization
184: component $\mathbf{M}_{y}$ is coupled with Josephson current through the phase
185: shift term $\varphi_{0}\propto\mathbf{n.}\left( \mathbf{M\wedge\nabla}%
186: \Psi\right) $, where $\Psi$ is the superconducting order parameter
187: ($\mathbf{\nabla}\Psi$ is along the $x$-axis in the system considered here).}%
188: \label{FIG_schema}%
189: \end{figure}
190:
191: Naturally, we may expect that the most interesting situation corresponds to
192: the case when the magnetic anisotropy energy does not exceed too much the
193: Josephson energy. From the measurements
194: \cite{rusanov_hesselberth_aarts_buzdin:057002.2004} on permalloy with very
195: weak anisotropy, we may estimate $K\sim4.10^{-5}%
196: %TCIMACRO{\unit{K}}%
197: %BeginExpansion
198: \operatorname{K}%
199: %EndExpansion
200: .%
201: %TCIMACRO{\unit{\U{212b}}}%
202: %BeginExpansion
203: \operatorname{\text{\AA}}%
204: %EndExpansion
205: ^{-3}$. On the other hand, typical value of $L$ in S/F/S junction is $L\sim10%
206: %TCIMACRO{\unit{nm}}%
207: %BeginExpansion
208: \operatorname{nm}%
209: %EndExpansion
210: $ and $\sin\ell/\ell\sim1$.\ Then, the ratio of the Josephson over magnetic
211: energy would be $E_{J}/E_{M}\sim100$ for $T_{c}\sim10%
212: %TCIMACRO{\unit{K}}%
213: %BeginExpansion
214: \operatorname{K}%
215: %EndExpansion
216: $. Naturally, in the more realistic case of stronger anisotropy this ratio
217: would be smaller but it is plausible to expect a great variety of regimes from
218: $E_{J}/E_{M}\ll1$ to $E_{J}/E_{M}\gg1$. \ \
219:
220: Let us now \ consider the case when a constant current $I<I_{c}$ is applied to
221: the $\varphi_{0}$-junction. The total energy is (see, \emph{e.g.
222: }\cite{b.likharev.1986}):
223: \begin{equation}
224: E_{\text{tot}}=-\frac{\Phi_{0}}{2\pi}\varphi I+E_{s}\left( \varphi
225: ,\varphi_{0}\right) +E_{M}\left( \varphi_{0}\right) , \label{EQ_energy}%
226: \end{equation}
227: and both the superconducting phase shift difference $\varphi$ and the rotation
228: of the magnetic moment $M_{y}=M_{0}\sin\theta$ (where $\theta$ is the angle
229: between the $z$-axis and the direction of $\mathbf{M}$) are determined from
230: the energy minimum conditions $\partial_{\varphi}E_{\text{tot}}=\partial
231: _{\varphi_{0}}E_{\text{tot}}=0$. It results in
232: \begin{equation}
233: \sin\theta=\frac{I}{I_{c}}\Gamma\text{ \ with }\Gamma=\frac{E_{J}%
234: }{K\mathcal{V}}\ell\frac{v_{\text{so}}}{v_{F}}, \label{EQ_phi_0_current_bias}%
235: \end{equation}
236: which signifies that a superconducting current provokes the rotation of the
237: magnetic moment $M_{y}$ in the $\left( yz\right) $ plane. Therefore, for
238: small values of the rotation, $\theta\left( I\right) $ dependence is linear.
239: In principle, the parameter $\Gamma$ can be larger than one. In that case,
240: when the condition $I/I_{c}\geq1/\Gamma$ is fulfilled, the magnetic moment
241: will be oriented along the $y$-axis. Therefore, applying a d.c.
242: superconducting current switches the direction of the magnetization, whereas
243: applying an a.c. current on a $\varphi_{0}$-junction could generate the
244: precession of the magnetic moment.
245:
246: We briefly comment on the situation when the direction of the gradient of the
247: spin-orbit potential is perpendicular (along $y$) to the easy axis $z$. To
248: consider this case we simply need to take $\varphi_{0}=\ell\left(
249: v_{\text{so}}/v_{F}\right) \cos\theta$. The total energy $(\ref{EQ_energy})$
250: has two minima $\theta=\left( 0,\pi\right) $, while applying the current
251: removes the degeneracy between them. However, the energy barrier exists for
252: the switch from one minimum into another. This barrier may disappear if
253: $\Gamma>1$ and the current is large enough $I>I_{c}/\Gamma$. \ In this regime
254: the superconducting current would provoke the switching of the magnetization
255: between one stable configuration $\theta=0$ and another $\theta=\pi$. This
256: corresponds to the transitions of the junction between $\ +\varphi_{0}$ and
257: $-\varphi_{0}$ states. The read-out of the state of the $\varphi_{0}$-junction
258: may be easily performed if it is a part of some SQUID-like circuit (the
259: $\varphi_{0}$-junction induces a shift of the diffraction pattern by
260: $\varphi_{0}$).
261:
262: \begin{figure*}[ptb]
263: \begin{center}
264: \includegraphics[width=5.0in]{4figsIN1.eps}
265: \end{center}
266: \caption{Results of numerical analysis of the magnetic moment dynamics of the
267: $\varphi_{0}$-junction.\ a) Reversal of $m_{z}$ from analytical expression
268: Eq.$\left( \ref{EQ_reversal}\right) $ (dashed curve) and numerical one
269: (normal curve). The other curves are related to the $\mathbf{M}$ trajectory:
270: b) in strong damping case c) and d) in the strong coupling regime revealing
271: incomplete and complete magnetic moment reversal, respectively.}%
272: \label{FIG_numerical}%
273: \end{figure*}
274:
275: In fact, the voltage-biased Josephson junction, and thus the a.c. Josephson
276: effect provides an ideal tool to study magnetic dynamics in a $\varphi_{0}%
277: $-junction. In such a case, the superconducting phase varies with time like
278: $\varphi\left( t\right) =\omega_{J}t$ \cite{b.josephson.1968}. If
279: $\hslash\omega_{J}\ll T_{c}$, one can use the static value for the energy of
280: the junction $\left( \ref{EQ_energy}\right) $ considering $\varphi\left(
281: t\right) $ as an external potential. The magnetization dynamics are described
282: by the Landau-Lifshitz-Gilbert equation (LLG) \cite{b.landau.IX_e}%
283: \begin{equation}
284: \frac{d\mathbf{M}}{dt}=\gamma\mathbf{M\times H}_{\text{eff}}+\frac{\alpha
285: }{M_{0}}\left( \mathbf{M\times}\frac{d\mathbf{M}}{dt}\right) ,
286: \label{EQ_LLG}%
287: \end{equation}
288: where $\mathbf{H}_{\text{eff}}=-\delta F/\mathcal{V}\delta\mathbf{M}$ is the
289: effective magnetic field applied to the compound, $\gamma$ the gyromagnetic
290: ratio, and $\alpha$ a phenomenological damping constant. The corresponding
291: free energy $F=E_{s}+E_{M}$ yields
292: \begin{equation}
293: \mathbf{H}_{\text{eff}}=\frac{K}{M_{0}}\left[ \Gamma\sin\left( \omega
294: _{J}t-r\dfrac{M_{y}}{M_{0}}\right) \boldsymbol{\hat{y}}+\dfrac{M_{z}}{M_{0}%
295: }\boldsymbol{\hat{z}}\right] , \label{EQ_Heff}%
296: \end{equation}
297: where $r=\ell v_{\text{so}}/v_{F}$. Introducing $m_{i}=M_{i}/M_{0}$,
298: $\tau=\omega_{F}t$ ($\omega_{F}=\gamma K/M_{0}^{2}$ is the frequency of the
299: ferromagnetic resonance) in LLG equation $\left( \ref{EQ_LLG}\right) $ leads
300: to
301: \begin{equation}
302: \left\{
303: \begin{array}
304: [c]{l}%
305: \dot{m}_{x}=m_{z}\left( \tau\right) m_{y}\left( \tau\right) -\Gamma
306: m_{z}\left( \tau\right) \sin\left( \omega\tau-rm_{y}\right) \\
307: \dot{m}_{y}=-m_{z}\left( \tau\right) m_{x}\left( \tau\right) \\
308: \dot{m}_{z}=\Gamma m_{x}\left( \tau\right) \sin\left( \omega\tau
309: -rm_{y}\right)
310: \end{array}
311: \right. , \label{EQ_LLG_complete}%
312: \end{equation}
313: where $\omega=\omega_{J}/\omega_{F}$. The generalization of Eq.$\left(
314: \ref{EQ_LLG_complete}\right) $ for $\alpha\neq0$ is straightforward.\ One
315: considers first the "weak coupling" regime $\Gamma\ll1$ when the Josephson
316: energy $E_{J}$ is small in comparison with the magnetic energy $E_{M}$. In
317: this case, the magnetic moment precess around the $z$-axis. If the other
318: components verify $\left( m_{x},m_{y}\right) \ll1$, then the equations
319: $\left( \ref{EQ_LLG_complete}\right) $ may be linearized, and the
320: corresponding solutions are
321: \begin{equation}
322: m_{x}\left( t\right) =\dfrac{\Gamma\omega\cos\omega_{J}t}{1-\omega^{2}%
323: }~\text{and}~m_{y}\left( t\right) =-\dfrac{\Gamma\sin\omega_{J}t}%
324: {1-\omega^{2}}. \label{EQ_magnetic_resonance_without_damping}%
325: \end{equation}
326: Near the resonance $\omega_{J}\approx\omega_{F}$, the conditions of
327: linearization are violated and it is necessary to take the damping into
328: account.\ The precessing magnetic moment influences the current through the
329: $\varphi_{0}$-junction like
330: \begin{equation}
331: \frac{I}{I_{c}}=\sin\omega_{J}t+\frac{\Gamma r}{2}\dfrac{1}{\omega^{2}-1}%
332: \sin2\omega_{J}t+..., \label{EQ_current_without_damping}%
333: \end{equation}
334: \emph{i.e.}, in addition to the first harmonic oscillations, the current
335: reveals higher harmonics contributions. The amplitude of the harmonics
336: increases near the resonance and changes its sign when $\omega_{J}=\omega_{F}%
337: $. Thus, monitoring the second harmonic oscillations of the current would
338: reveal the dynamics of the magnetic system.\
339:
340: The damping plays an important role in the dynamics of the considered system.
341: It results in a d.c. contribution to the Josephson current. Indeed, the
342: corresponding expression for $m_{y}\left( t\right) $ in the presence of
343: damping becomes
344: \begin{equation}
345: m_{y}\left( t\right) =\frac{\omega_{+}-\omega_{-}}{r}\sin\omega_{J}%
346: t+\frac{\alpha_{-}-\alpha_{+}}{r}\cos\omega_{J}t,
347: \label{EQ_magnetic_resonance_with_damping}%
348: \end{equation}
349: where%
350: \begin{equation}
351: \omega_{\pm}=\dfrac{\Gamma r}{2}\dfrac{\omega\pm1}{\Omega_{\pm}}\text{ and
352: }\alpha_{\pm}=\dfrac{\Gamma r}{2}\dfrac{\alpha}{\Omega_{\pm}},
353: \end{equation}
354: with $\Omega_{\pm}=\left( \omega\pm1\right) ^{2}+\alpha^{2}$. It thus
355: exhibits a damped resonance as the Josephson frequency is tuned to the
356: ferromagnetic one $\omega\rightarrow1$. Moreover, the damping leads to the
357: appearance of out of phase oscillations of $m_{y}\left( t\right) $ (term
358: proportional to $\cos\omega_{J}t$ in Eq.$\left(
359: \ref{EQ_magnetic_resonance_with_damping}\right) $). In the result the
360: current
361: \begin{multline}
362: I\left( t\right) \approx I_{c}\sin\omega_{J}t+I_{c}\frac{\omega_{+}%
363: -\omega_{-}}{2}\sin2\omega_{J}t+\\
364: +I_{c}\frac{\alpha_{-}-\alpha_{+}}{2}\cos2\omega_{J}t+I_{0}\left(
365: \alpha\right) \label{EQ_current_with_damping}%
366: \end{multline}
367: acquires a d.c. component
368: \begin{equation}
369: I_{0}\left( \alpha\right) =\frac{\alpha\Gamma r}{4}\left( \frac{1}%
370: {\Omega_{-}}-\frac{1}{\Omega_{+}}\right) .
371: \end{equation}
372: This d.c. current in the presence of a constant voltage $V$ applied to the
373: junction means a dissipative regime which can be easily detected. In some
374: aspect, the peak of d.c. current near the resonance is reminiscent of the
375: Shapiro steps effect in Josephson junctions under external r.f. fields. Note
376: that the presence of the second harmonic in $I\left( t\right) $ Eq.$\left(
377: \ref{EQ_current_with_damping}\right) $ should also lead to half-integer
378: Shapiro steps in $\varphi_{0}$-junctions
379: \cite{sellier_baraduc_lefloch_calemczuk.2004}.
380:
381: The limit of the "strong coupling" $\Gamma\gg1$ (but $r\ll1$) can also be
382: treated analytically. In this case, $m_{y}\approx0$ and solutions of
383: Eq.$\left( \ref{EQ_LLG_complete}\right) $ yields
384: \begin{equation}
385: \left\{
386: \begin{array}
387: [c]{l}%
388: m_{x}\left( t\right) =\sin\left[ \dfrac{\Gamma}{\omega}\left( 1-\cos
389: \omega_{J}t\right) \right] \\
390: \\
391: m_{z}\left( t\right) =\cos\left[ \dfrac{\Gamma}{\omega}\left( 1-\cos
392: \omega_{J}t\right) \right]
393: \end{array}
394: \right. , \label{EQ_reversal}%
395: \end{equation}
396: which are the equations of the magnetization reversal, a complete reversal
397: being induced by $\Gamma/\omega>\pi/2$. Strictly speaking, these solutions are
398: not exact oscillatory functions in the sense that $m_{z}\left( t\right) $
399: turns around the sphere center counterclockwise before reversing its rotation,
400: and returns to the position $m_{z}\left( t=0\right) =1$ clockwise, like a
401: pendulum in a spherical potential (see Fig.\ref{FIG_numerical}.c).
402:
403: Finally, we have performed numerical studies of the non-linear LLG Eq.$\left(
404: \ref{EQ_LLG}\right) $ for some choices of the parameters when the analytical
405: approaches fail. To check the consistency of our numerical and analytical
406: approaches, we present in Fig.\ref{FIG_numerical}.a) the corresponding
407: $m_{z}\left( t\right) $ dependences for low-damping regimes. They clearly
408: demonstrate the possibility of the magnetization reversal. In
409: Figs.\ref{FIG_numerical}b-d), some trajectories of the magnetization vectors
410: are presented for general coupling regimes. These results demonstrate that the
411: magnetic dynamics of S/F/S $\varphi_{0}$-junction may be pretty complicated
412: and strongly non-harmonic.
413:
414: If the $\varphi_{0}$-junction is exposed to a microwave radiation at angular
415: frequency $\omega_{1}$, the physics that emerge are very rich. First, in
416: addition to the Shapiro steps at $\omega_{J}=n\omega_{1}$, half-integer-steps
417: will appear. Secondly, the microwave magnetic field may also generate an
418: additional magnetic precession with $\omega_{1}$ frequency. Depending on the
419: parameters of $\varphi_{0}$-junction and the amplitude of the microwave
420: radiation the main precession mechanism may be related either to the Josephson
421: current or the microwave radiation. In the last case the magnetic spin-orbit
422: coupling may substantially contribute to the amplitude of the Shapiro steps.
423: Therefore, we could expect a dramatic increase of this amplitude at
424: frequencies near the ferromagnetic resonance. When the influence of the
425: microwave radiation and Josephson current on the precession is comparable, a
426: very complicated regime may be observed.
427:
428: In the present work we considered the case of the easy-axis magnetic
429: anisotropy. If the ferromagnet presents an easy-plane anisotropy than
430: qualitatively the main conclusions of this article remain the same because the
431: coupling between magnetism and superconductivity depends only on the $M_{y}$
432: component.\ However, the detailed dynamics would be strongly affected by a
433: weak in-plane anisotropy.\
434:
435: To summarize, we have demonstrated that S/F/S $\varphi_{0}$-junctions provide
436: the possibility to generate magnetic moment precession via Josephson current.
437: In the regime of strong coupling between magnetization and current, magnetic
438: reversal may also occur. These effects have been studied analytically and
439: numerically. We believe that the discussed properties of the $\varphi_{0}%
440: $-junctions could open interesting perspectives for its applications in spintronics.\
441:
442: The authors are grateful to Z. Nussinov, J. Cayssol, M. Houzet, D. Gusakova,
443: M. Roche and D. Braithwaite for useful discussions and comments. This work was
444: supported by the French ANR Grant N$%
445: %TCIMACRO{\U{b0}}%
446: %BeginExpansion
447: {{}^\circ}%
448: %EndExpansion
449: $ ANR-07-NANO-011: ELEC-EPR.
450:
451: \bigskip
452:
453: \begin{thebibliography}{99} %
454: \providecommand{\url}[1]{\texttt{#1}} \providecommand{\urlprefix}{URL }
455:
456: \bibitem {zutic_fabian_dassarma_RMP.76.323.2004}I.~\v{Z}uti\'{c}, J.~Fabian,
457: and S.~Das~Sarma. \newblock Rev. Mod. Phys. \textbf{76}, 323 (2004).
458:
459: \bibitem {hauptmann_paaske_lindelof.2008}J.~Hauptmann, J.~Paaske, and
460: P.~Lindelof. \newblock Nature Physics \textbf{4}, 373 (2008).
461:
462: \bibitem {b.winkler.2003}R.~Winkler. \newblock \emph{Spin-orbit coupling
463: effects in two-dimensional electron and hole systems} (Springer, {N}ew {Y}ork, 2003).
464:
465: \bibitem {buzdin:107005.2008}A.~Buzdin. \newblock Phys. Rev. Lett.
466: \textbf{101}, 107005 (2008).
467:
468: \bibitem {reynoso_etal:107001.2008}A.~A. Reynoso \emph{et~al.} \newblock Phys.
469: Rev. Lett. \textbf{101}, 107001 (2008).
470:
471: \bibitem {buzdin(2005)}A.~I. Buzdin. \newblock Rev. Mod. Phys. \textbf{77},
472: 935 (2005).
473:
474: \bibitem {bell_Milikisyants_Huber_Aarts:047002.2008}C.~Bell \emph{et~al.}
475: \newblock Phys. Rev. Lett. \textbf{100}, 047002 (2008).
476:
477: \bibitem {zhu_balatsky.2003}J.-X. Zhu, and A.~V. Balatsky. \newblock Phys.
478: Rev. B \textbf{67}, 174505 (2003).
479:
480: \bibitem {bulaevskii_hruska_shnirman_etal:177001.2004}L.~Bulaevskii
481: \emph{et~al.} \newblock Phys. Rev. Lett. \textbf{92}, 177001 (2004).
482:
483: \bibitem {zhu_nussinov_shnirman_balatsky.2004}J.-X. Zhu \emph{et~al.}
484: \newblock Phys. Rev. Lett. \textbf{92}, 107001 (2004).
485:
486: \bibitem {nussinov_shnirman_arovas_balatsky_zhu:214520.2005}Z.~Nussinov
487: \emph{et~al.} \newblock Phys. Rev. B \textbf{71}, 214520 (2005).
488:
489: \bibitem {takahashi_hikino_mori:057003.2007}S.~Takahashi \emph{et~al.}
490: \newblock Phys. Rev. Lett. \textbf{99}, 057003 (2007).
491:
492: \bibitem {houzet:057009.2008}M.~Houzet. \newblock Phys. Rev. Lett.
493: \textbf{101}, 057009 (2008).
494:
495: \bibitem {Waintal_Brouwer.65.054407.2002}X.~Waintal, and P.~W. Brouwer.
496: \newblock Phys. Rev. B \textbf{65}, 054407 (2002).
497:
498: \bibitem {braude:207001.2008}V.~Braude, and Y.~M. Blanter. \newblock Phys.
499: Rev. Lett. \textbf{100}, 207001 (2008).
500:
501: \bibitem {b.likharev.1986}K.~K. Likharev. \newblock \emph{Dynamics of
502: Josephson junctions and circuits} (Gordon and Beach Science Publishers, 1986).
503:
504: \bibitem {hikino_mori.2008}S.~Hikino \emph{et~al.} \newblock J. Phys. Soc.
505: Jap. \textbf{77}, 053707 (2008).
506:
507: \bibitem {rusanov_hesselberth_aarts_buzdin:057002.2004}A.~Y. Rusanov
508: \emph{et~al.} \newblock Phys. Rev. Lett. \textbf{93}, 057002 (2004).
509:
510: \bibitem {b.josephson.1968}B.~D. Josephson. \newblock \emph{Superconductivity
511: (in two volumes), vol. 1}, chapter 9, Weakly coupled superconductors (R.D
512: Parks, Marcel Dekker, Inc., 1968).
513:
514: \bibitem {b.landau.IX_e}E.~M. Lifshitz, and L.~P. Pitaevskii.
515: \newblock \emph{Course of theoretical physics, {V}olume 9 : {T}heory of the
516: condensed state} (Butterworth Heinemann, 1991).
517:
518: \bibitem {sellier_baraduc_lefloch_calemczuk.2004}H.~Sellier \emph{et~al.}
519: \newblock Phys. Rev. Lett. \textbf{92}, 257005 (2004).
520: \end{thebibliography}
521:
522:
523: \end{document}