1: \documentclass[12pt]{article}
2: \usepackage{graphicx} % standard LaTeX graphics tool
3: \usepackage{latexsym}
4: \usepackage{amsbsy}
5: \usepackage{sectsty}
6: \usepackage{cite}
7:
8: \sectionfont{\large}
9:
10: \addtolength\textwidth{1cm}
11: \addtolength\textheight{1cm}
12:
13: \newcommand{\kB}{k_{\mathrm{B}}}
14:
15: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
16: \begin{document}
17:
18: \begin{center}
19: {\Large \bf Simulation of Magnetization Switching in Nanoparticle Systems}
20:
21: \vspace{2cm}
22:
23: {\large
24: D.\ Hinzke and U.\ Nowak\\
25: Theoretische Physik, Gerhard-Mercator-Universit\"{a}t\\
26: 47048 Duisburg, Germany}
27: \end{center}
28:
29: \vspace{2cm}
30:
31: \noindent
32: Pacs-numbers: 75.10.Hk; 75.40.Mg; 75.40.Gb
33:
34: \section*{Abstract}
35: Magnetization reversal in magnetic nanostructures is investigated
36: numerically over time-scales ranging from fast switching processes on
37: a picosecond scale to thermally activated reversal on a microsecond
38: time-scale. A simulation of the stochastic Landau-Lifshitz equation of
39: motion is used as well as a time quantified Monte Carlo method for the
40: simulation of classical spin systems modeling magnetic Co
41: nanoparticles. For field pulses larger than the Stoner-Wohlfarth limit
42: spin precession effects govern the reversal behavior of the particle
43: while for lower fields a magnetization reversal is only possible when
44: it is assisted by thermal fluctuations.
45:
46: \newpage
47:
48: \section{Introduction}
49: The miniaturization of magnetic structures plays an important role for
50: fundamental research as well as for technical applications. This
51: leads to an incremental interest in the understanding of the behavior
52: of small magnetic particles and structures down to the nanometer
53: scale. But with decreasing size of the magnetic system thermal
54: activation becomes relevant. Hence, much effort is focused now on the
55: understanding of magnetization dynamics at finite temperatures.
56: \cite{nowakARCP00}.
57:
58: In the following, we briefly describe numerical techniques for the
59: study of magnetization dynamics in nanostructures, modeled as
60: classical spin systems, where a finite temperature is taken into
61: account. First, we will focus on the underlying model and the two
62: basic methods, Langevin dynamics \cite{lyberatosJAP93} and Monte Carlo
63: simulation \cite{nowakPRL00}. Then we study the magnetization reversal
64: in Co nanoparticles. Starting with the deterministic spin dynamics on
65: short time scales which plays a crucial role in high-speed data
66: storage \cite{backPRL98,bauerJAP00,leineweberPB00}, we go on to the
67: probabilistic long-time behavior where a thermally assisted reversal
68: can occur even for magnetic fields below the coercive field
69: \cite{brownPR63,braunPRL93,wernsdorferPRL97,hinzkePRB00,hinzkeJMMM00}.
70:
71:
72: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
73:
74: \section{Classical Spin Model}
75: \label{s:model}
76:
77: The micromagnetic properties of a nanoparticle can be described using
78: a model of classical magnetic moments which are localized on a given
79: lattice. Such a spin model can be motivated following different
80: lines: on the one hand it is the classical limit of a quantum
81: mechanical, localized spin model --- the Heisenberg model
82: \cite{aharoniBOOK96}. On the other hand, a classical spin model can
83: also be interpreted as the discretized version of a micromagnetic
84: continuum model \cite{schreflEM01}, where the charge distribution for
85: a single cell of the discretized lattice is approximated by a point
86: dipole \cite{aharoniBOOK96,hubertBOOK98}.
87:
88: The interpretation as an atomic model restricts the use of computer
89: simulations to the investigation of rather small systems of only a few
90: million atoms - corresponding to particle sizes of only a few
91: nanometers. On the other hand, within a continuum model, the space
92: might be discretized on a much larger length scale, as compared to an
93: atomic distance. However, in continuum theory usually a constant
94: absolute value of the magnetization vector is assumed, an assumption
95: which fails for higher temperatures since the space averaged
96: magnetization breaks down when approaching the critical temperature.
97: Hence, one expects correct thermal properties only in the limit of
98: small cell sizes of the order of atomic distances.
99:
100: In the following, let us consider a classical three dimensional
101: Heisenberg Hamiltonian for localized spins,
102: \begin{eqnarray}
103: \label{e:ham}
104: \textstyle
105: {\cal H} &=& - J \sum_{\langle ij \rangle} {\mathbf S}_i \cdot {\mathbf
106: S}_j - \mu_s{\mathbf B}\cdot \sum_i {\mathbf S}_i -
107: d_{z} \sum_i (S_{i}^{z})^2 \nonumber \\ &&
108: - w \sum_{i<j} \frac{3({\mathbf S}_i \cdot {\mathbf e}_{ij})({\mathbf
109: e}_{ij} \cdot {\mathbf S}_j) - {\mathbf S}_i \cdot {\mathbf
110: S}_j}{r^3_{ij}},
111: \end{eqnarray}
112: where the ${\mathbf S}_i = {\boldsymbol \mu}_i/\mu_s$ are three
113: dimensional magnetic moments of unit length representing atomic
114: magnetic moments. The first sum is the ferromagnetic exchange of the
115: moments with the coupling constant $J$. The second sum is the
116: coupling of the magnetic moments to an external magnetic field $B$,
117: the third sum represents a uniaxial anisotropy, here, for $d_z > 0$
118: favoring the $z$ axis as easy axis of the system, and the last sum is
119: the dipolar interaction where $w = \mu_0 \mu_s^2 /(4 \pi a^3)$
120: describes the strength of the dipole-dipole interaction. The
121: ${\mathbf e}_{ij}$ are unit vectors pointing from lattice site $i$ to
122: $j$ and $r_{ij}$ is the distance between these lattice sites in units
123: of $a$. The dipole-dipole interaction can be computed efficiently
124: with the help of fast Fourier transformation methods
125: \cite{yuanIEEE92,berkovPSS93}. One should however note that in a
126: Monte Carlo simulation with a single-spin flip algorithm the FFT
127: method is an approximation the implementation of which was described
128: in details before \cite{hinzkeJMMM00}.
129:
130: The transformation of the above introduced atomic parameters to the
131: material parameters which are usually used in a continuum model is
132: given by $J = 2aA_{x}$ where $A_{x}$ is the exchange energy, $d_{z} =
133: Ka^{3}$ where $K$ is the anisotropy energy density, and $\mu_{s} =
134: M_{s}a^{3}$ where $M_{s}$ is the spontaneous magnetization.
135:
136:
137: \section{Landau-Lifshitz Equation and Langevin Dynamics}
138:
139: In the short time limit spin precession is important which can be
140: taken care of by studying the corresponding equation of motion. The
141: basic numerical approach which includes thermal activation, is the
142: direct numerical integration of the Langevin equation of the problem.
143: In order to obtain thermal averages one has to calculate many of these
144: trajectories starting with the same initial conditions taking an
145: average over these trajectories for the quantities of interest. This
146: method is referred to as the Langevin dynamics formalism
147: \cite{lyberatosJAP93}.
148:
149: The underlying equation of motion for a magnetic system is the
150: Landau-Lifshitz-Gilbert (LLG) equation,
151: \begin{equation}
152: \frac{\partial {\mathbf S}_i}{\partial t} =
153: - \frac{\gamma}{(1+\alpha^2)\mu_s}
154: {\mathbf S}_i \times \Big[{\mathbf H}_i(t) +
155: \alpha \big ({\mathbf S}_i \times {\mathbf H}_i(t) \big) \Big],
156: \label{e:llg}
157: \end{equation}
158: with the gyromagnetic ratio $\gamma = 1.76 \times 10^{11} ({\mathrm
159: Ts})^{-1}$, the dimensionless damping constant $\alpha$, and the
160: internal field $ {\mathbf H}_i(t) = {\boldsymbol \zeta}_i(t) -
161: \partial {\cal H} /\partial {\mathbf S}_i$. Langevin
162: dynamics is introduced here in form of the noise ${\boldsymbol \zeta}_i(t)$
163: which represents thermal fluctuations, with $\langle {\boldsymbol
164: \zeta}_i(t) \rangle = 0$ and $ \langle \zeta_i^{\eta}(t)
165: \zeta_j^{\theta}(t') \rangle = 2 \delta_{ij} \delta_{\eta \theta}
166: \delta(t-t') \alpha \kB T \mu_{s}/ \gamma$ where $i, j$ denote once
167: again lattice sites and $\eta, \theta$ Cartesian components.
168:
169: The LLG equation with Langevin dynamics is a stochastic differential
170: equation with multiplicative noise. For this kind of differential
171: equation a problem arises which is called the It\^{o}-Stratonovich
172: dilemma \cite{GreinerJSP88}. As a consequence, different time
173: discretization schemes may converge to different results with
174: decreasing time step. As was pointed out in \cite{garciaPRB98} the
175: multiplicative noise in the Langevin equation above has to be treated
176: by means of the Stratonovich interpretation. An appropriate
177: discretization scheme leading to a Stratonovich interpretation is the
178: Heun method \cite{GreinerJSP88,garciaPRB98,nowakARCP00} which is used
179: in the following.
180:
181:
182: \section{Monte Carlo Methods}
183:
184: In the long time limit only spin relaxation and thermal fluctuations
185: are relevant which can be studied very conveniently using Monte Carlo
186: methods with quantified time step \cite{nowakPRL00}. Within a Monte
187: Carlo approach \cite{binderBOOK97} trajectories in phase space are
188: calculated following a master equation \cite{reifBOOK65} for the time
189: development of the probability distribution $P_s(t)$ in phase space,
190: \begin{equation}
191: \label{e:master}
192: \frac{\mbox{d} P_s}{\mbox{d} t} = \sum_{s'} (P_{s'} w_{s' \to s} -
193: P_s w_{s \to s'}).
194: \end{equation}
195: Here, $s$ and $s'$ denote different states of the system and
196: $w_{s' \to s}$ is the transition rate for a change from a state $s'$
197: to a state $s$. These rates have to fulfill the condition
198: \cite{reifBOOK65}
199: \begin{equation}
200: \label{e:det-bal}
201: \frac{w_{s \to s'}}{w_{s' \to s}} = \exp\Big[\frac{E(S) - E(S')}{\kB
202: T}\Big].
203: \end{equation}
204: The master equation describes exclusively the coupling of the system
205: to the heat bath \cite{reifBOOK65}. Hence, only the irreversible part
206: of the dynamics of the system is considered including only the
207: relaxation and the fluctuations. A Monte Carlo simulation does not
208: include the energy conserving part of the equation of motion. Hence, no
209: precession of magnetic moments will be found.
210:
211: Monte Carlo approaches in general have no physical time associated
212: with each step of the algorithm. However, recently a time quantified
213: Monte Carlo method was proposed in \cite{nowakPRL00} and later
214: successfully applied to different model systems
215: \cite{hinzkePRB00,hinzkeJMMM00,smirnovJAP00}. Here, the
216: interpretation of a Monte Carlo step as a realistic time interval
217: $\Delta t$ was achieved by a comparison of one step of the Monte Carlo
218: process with a time interval of the LLG equation in the high damping
219: limit. We will use this algorithm in the following. The trial step of
220: this algorithm is a random movement of the magnetic moment within a
221: cone with a given size $r$ with
222: \begin{equation}
223: r^2 = \frac{20 k_B T \alpha \gamma}{(1+\alpha^2) \mu_s} \Delta t.
224: \label{e:trial}
225: \end{equation}
226: In order to achieve this efficiently one constructs a random vector with
227: constant probability distribution within a sphere of radius $r$. This
228: random vector is added to the initial moment and subsequently the
229: resulting vector is normalized \cite{nowakPRL00}.
230:
231: Using this algorithm one Monte Carlo step represents a given time
232: interval $\Delta t$ of the LLG equation in the high damping limit as
233: long as $\Delta t$ is chosen appropriately (for details see
234: \cite{nowakARCP00}).
235:
236:
237: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
238: \section{Precessional Reversal in Co Nanoparticles}
239: \label{s:results}
240:
241: In the following we consider Co nanoparticles, where the material
242: parameters are $A_x = 1.3 \cdot 10^{-11}$ J/m, $K = 6.8 \cdot 10^{5} $
243: J/m$^2$ and $M_s = 1.43 \cdot 10^6$ A/m. For simplicity we simulate a
244: simple cubic lattice with atomic distance $a = 0.25$ nm. Our
245: simulation starts with a spin configuration where all magnetic moments
246: point into the $z$ direction, aligned with the easy axis, and with the
247: $z$ component of the external magnetic field antiparallel to the
248: magnetization so that the system is in an unstable, or at least
249: metastable state.
250:
251: In sufficiently small particles the magnetic moments rotate coherently
252: during the magnetization reversal. A quantitative description of
253: coherent rotation in ellipsoidal single domain particles was developed
254: by Stoner and Wohlfarth \cite{stonerPTRS49}. Depending on the angle
255: between the applied field $\mathbf{B}$ and the $z$ (easy) axis of the
256: system, the coercive field $B_c$ varies following the so-called
257: Stoner-Wohlfarth asteroid \cite{stonerPTRS49}. Under an angle of
258: $45^{\circ}$ it is $B_c = d_{\mathrm{eff}}V /\mu_{\mathrm s}$ where
259: $V$ is the volume of the particle and $d_{\mathrm{eff}}$ is an
260: effective anisotropy constant. In our case it is $B_c \approx 0.7$ T.
261:
262: First, we are interested in fast switching processes where the applied
263: field is higher than the coercive field and the reversal is dominated
264: by spin precession. We simulate the system with Langevin dynamics as
265: described before. Figure \ref{f:mag_LD} shows the time dependence of
266: the magnetization of our Co particle in the low damping limit ($\alpha
267: = 0.1$) for a very small ellipsoidal shaped particle of length $L=4$nm
268: and diameter $L=2$nm.
269:
270: \begin{figure}
271: \includegraphics[width=.85\textwidth, bb = 100 260 450
272: 470]{nowfig1.eps}
273: \caption{Reduced magnetization vs. time for a Co nanoparticle. The
274: data are from Langevin dynamics simulations. $B = 1.13$ T and $T =
275: 16$ K.}
276: \label{f:mag_LD}
277: \end{figure}
278:
279: The $z$ and the $x$ components of the magnetization are shown as well
280: as its absolute value. The magnetic field ${\mathbf B}$ is set under
281: an angle of $45^{\circ}$ to the $z$ axis within the $yz$ plane so that
282: the response of the system to the external field sets in directly. The
283: wavering magnetization of the system clearly follows from the
284: precession of the spins.
285:
286: Note, that the precession time of our system is not simply given by
287: the precession time of a single spin in an external field ($\tau_{p} =
288: 2 \pi (1+ \alpha^2)/\gamma B \approx 32$ ps in our case). Instead, the
289: whole internal field is relevant for the spin precession, also the
290: contributions from the dipolar field, the exchange and the anisotropy.
291: This internal field is not constant in time and in general it is
292: non-homogenous within the system. However, in the very small particle
293: which we consider here, the internal field is sufficiently homogenous
294: so that the magnetization moves coherently. Hence, the absolute value
295: of the magnetization remains constant in time as is shown in the
296: figure. Note also, that even after the new stable state is reached
297: the magnetization still keeps on oscillating around its equilibrium
298: value, driven by thermal fluctuations.
299:
300: \section{Thermally Activated Reversal}
301:
302: Let us now turn to the case $B < B_c$, where the reversal process can
303: only occur when it is thermally activated. Since we are now interested
304: in the long time and high damping limit ($\alpha = 4$) where the
305: behavior of the particle is governed by thermal fluctuations we can
306: use Monte Carlo methods. Figure \ref{f:mag_MC} shows the typical time
307: dependence of the magnetization of the same Co particle as before. The
308: field $\mathbf{B}$ is set antiparallel to the initial state here, so
309: that the zero-temperature coercive field is given by $B_c =
310: 2d_{\mathrm{eff}}V /\mu_{\mathrm s}$ which in our case is $B_c \approx
311: 1.4$ T.
312:
313: \begin{figure}
314: \includegraphics[width=.85\textwidth, bb = 100 260 450
315: 470]{nowfig2.eps}
316: \caption{Reduced magnetization vs. time for the same particle as in
317: Fig. \ref{f:mag_LD}. The data are from Monte Carlo simulations. $B
318: = 1.1$T and $T = 5$ K.}
319: \label{f:mag_MC}
320: \end{figure}
321:
322: As one can see, the $z$ component of the magnetization remains nearly
323: constant for a time period which is rather long as compared to the
324: previous simulation. Then, suddenly, the magnetization drops and its
325: sign changes. As before the absolute value of the magnetization is
326: constant in time and the reversal mechanism is mainly a coherent
327: rotation. The value of the switching time is approximately 6.6 ns in
328: our simulation. However, this thermally activated switching is not a
329: deterministic process as it was the short time dynamics studied
330: before, where the switching followed mainly from the deterministic
331: part of the equation of motion. Instead, the thermal activation
332: process here is a probabilistic event. The probability distribution
333: $P(t_s)$ for switching events taking place after a time $t_s$ follows
334: an exponential law \cite{brownPR63},
335: \begin{equation}
336: P(t_s) \sim \exp{(-t_{\mathrm s}/\tau)},
337: \label{e:ver}
338: \end{equation}
339: in the limit of large time-scales. Here, $\tau$ is a characteristic
340: time scale,
341: \begin{equation}
342: \tau = \tau_{0} \exp{\big( \Delta E/ \kB T \big)},
343: \end{equation}
344: where $\tau_0$ is a prefactor and $\Delta E$ an energy barrier which
345: both are related to a certain reversal mechanism (see e.\ g.\
346: \cite{brownPR63,braunPRL93,coffeyPRL98} for analytically determined
347: prefactors and energy barriers in different systems and
348: \cite{nowakPRL00,hinzkePRB00,hinzkeJMMM00,nowakARCP00} for numerical
349: work on this subject). In general the prefactor may depend on the
350: system parameters, the temperature, the applied magnetic field and the
351: damping constant.
352:
353: For the case of a Stoner-Wohlfarth particle with the applied field
354: parallel to the easy axis the energy barrier has the form
355: \begin{equation}
356: \Delta E = d_{\mathrm{eff}} V \Big(1 - \frac{B}{B_c}\Big)^2.
357: \label{e:energy}
358: \end{equation}
359: This energy barrier as well as the prefactor have been calculated by
360: Brown \cite{brownPR63} under the assumption that all magnetic moments
361: are parallel, so that the system behaves like one single magnetic
362: moment.
363:
364: \begin{figure}
365: \includegraphics[width=.85\textwidth, bb = 100 260 450
366: 470]{nowfig3.eps}
367: \caption{Characteristic time $\tau$ vs. $1/ T$ for the same Co
368: particle as in the figures before. The slope of the solid line
369: represents the energy barrier $\Delta E$. $B = 1.1$ T.}
370: \label{f:tau}
371: \end{figure}
372:
373: For a further analysis we extract the energy barrier which governs the
374: reversal process from our numerical data. Figure \ref{f:tau} shows
375: the temperature dependence of the characteristic time, i.\ e.\ the
376: mean switching time, obtained from our simulations. The slope of the
377: solid line corresponds to the theoretical value of the energy barrier
378: for a reversal by coherent rotation obtained from Eq.~\ref{e:energy}.
379: Obviously, it is in very good agreement with our numerical data for
380: low enough temperatures.
381:
382: \section*{Acknowledgments}
383: This work was supported by the Deutsche Forschungsgemeinschaft (SFB
384: 491 and project NO290/1) and by the European Union (COST action P3,
385: working group 4).
386:
387: \newpage
388: \begin{thebibliography}{10}
389:
390: \bibitem{nowakARCP00}
391: U. Nowak, in {\em Annual Reviews of Computational Physics IX}, edited by D.
392: Stauffer (World Scientific, Singapore, 2000), p.\ 105.
393:
394: \bibitem{lyberatosJAP93}
395: A. Lyberatos and R.~W. Chantrell, J.\ Appl.\ Phys.\ {\bf 73}, 6501 (1993).
396:
397: \bibitem{nowakPRL00}
398: U. Nowak, R.~W. Chantrell, and E.~C. Kennedy, Phys.\ Rev.\ Lett. {\bf 84}, 163
399: (2000).
400:
401: \bibitem{backPRL98}
402: C.~H. Back, D. Weller, J. Heodmann, D. Mauri, D. Guarisco, E.~L. Garwin, and
403: H.~C. Siegmann, Phys.\ Rev.\ Lett. {\bf 81}, 3251 (1998).
404:
405: \bibitem{bauerJAP00}
406: M. Bauer, J. Fassbender, and B. Hillebrands, J.\ Appl.\ Phys.\ {\bf 87}, 6274
407: (2000).
408:
409: \bibitem{leineweberPB00}
410: T. Leineweber and H. Kronm{\"u}ller, Physica B {\bf 275}, 5 (2000).
411:
412: \bibitem{brownPR63}
413: W.~F. Brown, Phys.~Rev. {\bf 130}, 1677 (1963).
414:
415: \bibitem{braunPRL93}
416: H.~B. Braun, Phys.\ Rev.\ Lett. {\bf 71}, 3557 (1993).
417:
418: \bibitem{wernsdorferPRL97}
419: W. Wernsdorfer, E.~B. Orozco, K. Hasselbach, A. Benoit, B. Barbara, N. Demoncy,
420: A. Loiseau, H. Pascard, and D. Mailly, Phys.\ Rev.\ Lett. {\bf 78}, 1791
421: (1997).
422:
423: \bibitem{hinzkePRB00}
424: D. Hinzke and U. Nowak, Phys.\ Rev.\ B {\bf 61}, 6734 (2000).
425:
426: \bibitem{hinzkeJMMM00}
427: D. Hinzke and U. Nowak, J.\ Magn.\ Magn.\ Mat.\ {\bf 221}, 365 (2000).
428:
429: \bibitem{aharoniBOOK96}
430: A. Aharoni, {\em Introduction to the Theory of Ferromagnetism} (Oxford
431: University Press, Oxford, 1996).
432:
433: \bibitem{schreflEM01}
434: T. Schrefl, J. Fidler, R.~W. Chantrell, and M. Wongsam, Encyclopedia of
435: Materials: Science and Technology (2001), in press.
436:
437: \bibitem{hubertBOOK98}
438: A. Hubert and R. Sch\"{a}fer, {\em Magnetic Domains} (Springer-Verlag, Berlin,
439: 1998).
440:
441: \bibitem{yuanIEEE92}
442: S.~W. Yuan and H.~N. Bertram, IEEE Trans.\ Mag. {\bf 28}, 2031 (1992).
443:
444: \bibitem{berkovPSS93}
445: D.~V. Berkov, K.~R. Ramst\"{o}ck, and A. Hubert, Phys.\ stat.\ sol.\ (a) {\bf
446: 137}, 207 (1993).
447:
448: \bibitem{GreinerJSP88}
449: A. Greiner, W. Strittmatter, and J. Honerkamp, J.\ Stat.\ Phys. {\bf 51}, 95
450: (1988).
451:
452: \bibitem{garciaPRB98}
453: J.~L. Garc\'{\i}a-Palacios and F.~J. L\'{a}zaro, Phys.\ Rev.\ B {\bf 58},
454: 14937 (1998).
455:
456: \bibitem{binderBOOK97}
457: K. Binder and D.~W. Heermann, in {\em Monte Carlo Simulation in Statistical
458: Physics}, edited by P. Fulde (Springer-Verlag, Berlin, 1997).
459:
460: \bibitem{reifBOOK65}
461: F. Reif, {\em Fundamentals of statistical and thermal physics} (McGraw-Hill
462: Book Company, New York, 1965).
463:
464: \bibitem{smirnovJAP00}
465: R. Smirnov-Rueda, O. Chubykalo, U. Nowak, R.~W. Chantrell, and J.~M.
466: Gonz\'{a}les, J.\ Appl.\ Phys.\ {\bf 87}, 4798 (2000).
467:
468: \bibitem{stonerPTRS49}
469: E.~C. Stoner and E.~P. Wohlfarth, Philos.\ Trans.\ R.\ Soc.\ London Ser.\ A
470: {\bf 240}, 599 (1949).
471:
472: \bibitem{coffeyPRL98}
473: W.~T. Coffey, D.~S.~F. Crothers, J.~L. Dorman, Y.~P. Kalmykov, E.~C. Kennedy,
474: and W. Wernsdorfer, Phys.\ Rev.\ Lett. {\bf 80}, 5655 (1998).
475:
476: \end{thebibliography}
477:
478: \end{document}
479:
480:
481: