1: \documentclass[aps,prl,onecolumn,showpacs,preprintnumbers,superscriptaddress]{revtex4}
2: \usepackage{bm,amsmath,amssymb,latexsym,mathrsfs,enumerate}
3: \usepackage[mathcal]{euscript}
4: \usepackage[hypertex]{hyperref}
5: \usepackage{graphicx}
6:
7: \begin{document}
8:
9: \title{STATIONARY PRECESSION TOPOLOGICAL SOLITONS WITH NONZERO HOPF INVARIANT IN A UNIAXIAL FERROMAGNET}
10:
11: \author{A. B. Borisov}
12: \email{Borisov@imp.uran.ru}
13: \affiliation{Institute of Metal Physics, Ural Division, Russian Academy of Sciences, Yekaterinburg 620041, Russia}
14:
15: \author{F. N. Rybakov}
16: \email[Corresponding author. Electronic address: ]{F.N.Rybakov@gmail.com}
17: \affiliation{Institute of Metal Physics, Ural Division, Russian Academy of Sciences, Yekaterinburg 620041, Russia}
18:
19: \date{30 June, 2008}
20: \begin{abstract}
21: Three-dimensional stationary precession solitons with nonzero Hopf indices are found numerically by solving
22: the Landau–Lifshitz equation. The structure and existence domain of the solitons are found.
23: \end{abstract}
24:
25: \pacs{03.50.-k, 11.27.+d, 47.32.Cc, 75.10.Hk, 75.60.Ch, 94.05.Fg}
26:
27: \maketitle
28:
29: Topological three-dimensional solitons attract considerable
30: interest in many fields of physics including
31: hydrodynamics, particle physics, cosmology, and condensed
32: matter physics. In models with the three-component
33: unit vector field ${\bf n}=(n_{1}, n_{2}, n_{3})$, where ${\bf n}^{2}=1$, localized structures exist if the field
34: $n$ asymptotically approaches the vector ${\bf n}_{0}$ as $|{\bf r}|\to\infty$. Such fields map
35: the $R^3\cup \{\infty \}$ space to the two-dimensional sphere $S^2$ êand are classified by the homotopy classes $\pi_3(S^2)=Z$ and characterized by the Hopf invariant $Í$~\cite{bib:BW} given by
36: the expression:
37: \begin{equation}
38: H=-\frac{1}{(8\pi)^2}\int{\bf F}\cdot{\bf A} d{\bf r},\label{eq:Hdefault}
39: \end{equation}
40: where $F_{i}=\epsilon_{ijk}{\bf n}\cdot(\bigtriangledown_j{\bf n}\times \bigtriangledown_k{\bf n})$ and $curl{\bf A}=2{\bf F}$. The invariant $H$
41: admits the simple geometric interpretation
42: as the linking number of two preimage closed curves
43: corresponding to an arbitrary pair of points on the
44: $S^2$ sphere.
45:
46: \begin{figure}
47: \includegraphics[width=0.8\columnwidth, viewport=80 505 370 720, clip]{fig1.eps}
48: \caption{Distribution of magnetization in a three-dimensional topological soliton with $H=3$. The thin and thick curves are
49: the preimages of the points $(\Theta_1, \Phi_1) = (0.5\pi, 1)$ and $(\Theta_2, \Phi_2) = (0.2\pi, 3)$ on the $S^2$ sphere, respectively.}
50: \label{F:fig1}
51: \end{figure}
52:
53: Stable three-dimensional solitons with $H\neq0$ (the
54: so-called knotted solitons) were studied numerically in
55: the Faddeev-Niemi model~\cite{bib:LD,bib:LN,bib:LN1,bib:BS,bib:Glad}, i.e., the nonlinear
56: $\sigma$-model including terms with fourth-order derivatives.
57: Defects with a nonzero Hopf invariant have been discussed
58: in condensed matter physics since the pioneering
59: works by Volovik and Mineev for superfluid
60: $^3He$~\cite{bib:VM} and by Dzyaloshinskii and Ivanov~\cite{bib:DI} for a uniaxial ferromagnet.
61: Simple reasoning based on the Derrick theorem~\cite{bib:Derrick} indicates the absence of nontrivial static threedimensional
62: solitons with finite energy in the aforementioned
63: media. However, dynamical structures of
64: this kind, which are stabilized by the precession of the
65: magnetization, can exist~\cite{bib:KBK,bib:PT}. Studies of threedimensional
66: magnetic structures are of both academic
67: and engineering interest in view of the development of new memory elements based on topological solitons in
68: uniaxial ferromagnets~\cite{bib:OBS}.
69:
70: Three-dimensional magnetic structures have been
71: poorly studied until recently. A few original theoretical
72: and experimental works on magnetic structures in
73: superfluid $^3He$ were published in recent years~\cite{bib:Volovik1,bib:Volovik2,bib:Volovik3,bib:Volovik4}.
74: Spiral and cnoidal hedgehogs with ñ $H=0$ were analytically
75: described in ~\cite{bib:Borisov} in the Heisenberg model for an
76: isotropic ferromagnet. Constant-velocity precession solitons with
77: $H=1$ were numerically found in the same
78: model~\cite{bib:Cooper}. Both stationary (magnon drops) ~\cite{bib:IvKos1} and
79: uniformly moving~\cite{bib:Sut2001} precession radially symmetric
80: nontopological solitons in a uniaxial ferromagnet were
81: analyzed. In addition, the moving topological soliton
82: was recently analyzed~\cite{bib:Sut2007}, but stable configurations
83: with $H\neq0$ were not found.
84:
85: In this work, we find dynamical structures, namely,
86: stationary precession three-dimensional solitons with
87: nonzero Hopf indices, in a uniaxial ferromagnet. Hereafter,
88: the existence domain of such solitons is determined
89: and their fine structure and main features are
90: studied.
91:
92: The dynamics of the magnetization vector is
93: described by the Landau-Lifshitz equation. In the case
94: of negligible relaxation, this equation has the form
95: \begin{equation}
96: \frac {\partial {\bf M}} {\partial t} = - \frac {2 \mu _ 0} {\hbar}\left[\frac {\delta { E}} {\delta {\bf M}}\times {\bf M}\right],\label{eq:LLeq}
97: \end{equation}
98: where the ferromagnet energy $E=E_{exch.}+E_{anis.}$ is the
99: sum of the exchange energy $E_{exch.}$ and the magnetic anisotropy
100: energy $E_{anis.}$ with the parameter $\beta$:
101: \begin{align}
102: E_{exch.}&=\frac{\alpha }{2}\int \left(\frac {\partial {\bf M}} {\partial {r _ i}}\right)^2 d{\bf r},\label{eq:ExchEnergy}\\
103: E_{anis.}&=\frac{\beta }{2}\int \left({M_ x}^2+ {M_ y}^2\right) d{\bf r}.\label{eq:AnizEnergy}
104: \end{align}
105: The energy of the magnetic-dipole interaction is negligible
106: for large $\beta$ values typical for many magnetic media.
107:
108: Using the parameterization
109: \begin{equation}
110: \begin{cases}
111: {\bf M}=M_{0} {\bf n},\\
112: {\bf n}=(sin \Theta cos\Phi, sin\Theta sin\Phi, cos\Theta),
113: \end{cases}\label{eq:Mparam}
114: \end{equation}
115: we seek solutions describing stationary precession solitons of the form
116: \begin{equation}
117: \begin{cases}
118: \Phi=\omega t + Q \varphi + \phi(r,z),\\
119: \Theta=\theta(r,z),
120: \end{cases}\label{eq:Qvortex}
121: \end{equation}
122: where $\varphi$ is the polar angle of the cylindrical coordinate
123: system $(r,\varphi,z)$ and $\phi$ and $\theta$ are the odd and even functions
124: of $z$, respectively, and the $z$ axis is the symmetry
125: axis. It is known~\cite{bib:Tjon,bib:PT} that studying the dynamics of
126: such structures is equivalent to solving a variational
127: problem of minimizing the energy functional $E$ for a
128: fixed invariant of motion, $N$, equal to the total number
129: of spin deviations:
130: \begin{equation}
131: N=\frac {M_0} {2 \mu _ 0}\int \left(1- n_z\right) d{\bf r}=const.\label{eq:Nintegral}
132: \end{equation}
133:
134: The precession frequency is determined by the formula
135: \begin{equation}
136: \omega=\frac {1} {\hbar}\frac {\partial E} {\partial N}.\label{eq:omega1}
137: \end{equation}
138: The Derrick method~\cite{bib:Derrick} applied to finite-energy field
139: configurations yields the convenient formula for the
140: precession frequency:
141: \begin{equation}
142: \omega=\frac {1} {\hbar}\frac {E_{exch.} +3 E_{anis.}} {3 N}\label{eq:omega2}
143: \end{equation}
144:
145: In view of Eqs.~(\ref{eq:Qvortex}), the energy functional takes the form
146: \begin{equation}
147: E=\int_{-\infty}^{\infty} \int_{0}^{\infty} w_E d{r}d{z},\label{eq:4}
148: \end{equation}
149: Here,
150: \begin{equation}
151: w_E= \alpha \pi M_0^2 \left[ r { \left( \frac {\partial {\bf n^*}} {\partial r} \right)}^2 + r {\left(\frac {\partial {\bf n^*}} {\partial z} \right)}^2 +\left( \frac {Q^2} {r}+{\frac \beta \alpha} r \right) ({n_x^*}^2+{n_y^*}^2) \right],\label{eq:wE}
152: \end{equation}
153: where
154: \begin{equation}
155: {\bf n^*}=(sin \theta cos \phi, sin \theta sin \phi, cos \theta ) .\label{eq:nParam}
156: \end{equation}
157:
158: \begin{figure}
159: \includegraphics[width=0.6\columnwidth, viewport=80 460 370 720, clip]{fig2.eps}
160: \caption{Contours of the angles parameterizing the unit vector
161: ${\bf n^*}$ for $H = 3$ and $\omega/\omega_0=0.5$. The solid and dashed curves
162: correspond to $\theta=const$ and $ctg(\phi)=const$, respectively.}
163: \label{F:fig2}
164: \includegraphics[width=0.7\columnwidth, viewport=15 610 310 760, clip]{fig3.eps}
165: \caption{Distance dependence of the normalized energy density
166: for solitons with the same precession frequency for various $H$ values. The solid, dashed, and dash–dotted curves correspond to $H = 2$, $3$, and $4$, respectively.}
167: \label{F:fig3}
168: \end{figure}
169: \begin{figure}
170: \includegraphics[width=0.6\columnwidth, viewport=20 310 300 765, clip]{fig4.eps}
171: \caption{Distance dependence of the normalized energy density
172: of a soliton for several values of the precession frequency and $H = 3$.}
173: \label{F:fig4}
174: \end{figure}
175:
176: \begin{figure}
177: \includegraphics[width=0.6\columnwidth, viewport=20 495 300 690, clip]{fig5.eps}
178: \caption{Energy versus the precession frequency for nontopological
179: ($H = 0$) and topological ($H = 3$) solitons. The points show the numerical results.}
180: \label{F:fig5}
181: \end{figure}
182:
183: The formula describing the Hopf index simplifies
184: drastically for fields specified by Eqs.~(\ref{eq:Qvortex})~\cite{bib:KUR,bib:Glad}
185: \begin{equation}
186: H=Q T,\label{eq:HopfIndex}
187: \end{equation}
188: where the integer $T$ is given by
189: \begin{equation}
190: T=\frac {1} {4 \pi}\int_{-\infty}^{\infty} {\int_{0}^{\infty} {{\bf n^*}\cdot\left[\frac {\partial {\bf n^*}} {\partial r}\times \frac {\partial {\bf n^*}} {\partial z} \right]} d{r}}d{z}.\label{eq:TNumber}
191: \end{equation}
192:
193: To solve the variational problem numerically by the
194: finite difference method, the energy functional was
195: replaced by a function of the projections of the vector ${\bf n^*}$
196: at the numerical-grid nodes. Since the field configuration
197: for negative $z$ values is determined by the symmetry
198: of the problem, the discretization domain was the
199: unit square $0 \leq r \leq 1,0 \leq z \leq 1$.
200:
201: The expected domain of soliton localization should
202: be about the typical scale
203: \begin{equation}
204: l_0=\sqrt{\frac{\alpha}{\beta}}.\label{eq:L0}
205: \end{equation}
206:
207: The parameters $\alpha$ and $\beta$ were chosen to ensure that $1/l_0\sim\sqrt{D}$, where $D$ is the number of grid nodes in one
208: direction. The key point of this method is to ensure that
209: the soliton is entirely contained in the simulation
210: domain, but is sufficiently large to be adequately
211: described by the discrete grid. In addition, to improve
212: the performance, the numerical grid was nonuniform
213: with the density of nodes increasing toward the coordinate
214: origin. This means that the inhomogeneity was
215: assumed to be localized near the center, while the field
216: quickly reaches the asymptotic uniform state at infinity.
217: The vector ${\bf n_{0}}=(0,0,1)$ at the boundaries $z=1$ or $r=1$
218: of the domain corresponds to the ground state.
219:
220: The energy with an additive penalty function was
221: minimized by the conjugate gradient method. A quadratic
222: penalty function imposes a constraint on the
223: norm of the vector ${\bf n^*}$ which is unity at any point. Integral~(\ref{eq:Nintegral}) is discretized to a linear form, and the imposed
224: constraint is taken into account by applying a linear
225: operator~\cite{bib:PD} in the calculation of the direction toward
226: the minimal objective function for each iteration.
227:
228: A feature of our numerical analysis was the choice
229: of the initial state, i.e., the trial field configuration from
230: which the minimization begins. The vector corresponded
231: to the ground state in the entire domain of simulation
232: excluding a small quadrant near the coordinate
233: origin. The vectors inside the quadrant lied in the
234: orthogonal plane, and their directions varied randomly
235: from point to point. Thus, the initial configuration is to
236: a certain extent chaotic and does not even obey the criteria
237: for the discretization of any smooth function.
238:
239: The result was directly tested in addition to the
240: observed approach of the energy to its asymptotic value
241: after many iterations. The frequency $\omega$ was calculated
242: using Eq.~(\ref{eq:omega2}). The second derivatives and the discrepancy
243: of Landau-Lifshitz equation~(\ref{eq:LLeq}) were numerically
244: evaluated at each internal point of the simulation
245: domain.
246:
247: For the solitons presented below, the numerical calculations
248: based on Eq.~(\ref{eq:TNumber}) yielded $T = 1$. In this case,
249: according to Eq.~(\ref{eq:HopfIndex}), the Hopf index $H = Q$. Topological
250: solitons of other types will be described elsewhere.
251: The results of our present study are discussed below.
252:
253: Figure \ref{F:fig1} shows a typical configuration of the field $\bf{n}$
254: for $H = 3$. Here, the coordinate axes are shifted for clarity
255: and illustrate only the spatial orientation of the soliton.
256: It is seen that the linking number of the preimages
257: of two points of the $S^2$ sphere is equal to three. The toroidal
258: surface corresponds to $\theta=\pi/2$. Note that ${\bf n}={\bf n}_0$ on
259: the symmetry axis $r = 0$ and ${\bf n}\rightarrow{\bf n}_0$ for $r^2+z^2\rightarrow\infty$.
260:
261: The contours of the angles parameterizing the vector
262: ${\bf n^*}$ are plotted in Fig. \ref{F:fig2}. The curve $\theta=\pi/2$ is a section
263: of an axisymmetric surface of kind 1 (toroidal surface)
264: in which a large fraction of the soliton energy is
265: contained. The common point of the lines $\phi=const$ is
266: the south pole of the sphere $S^2$, at which $\theta=\pi$.
267:
268: Analysis showed that it is most difficult to simulate
269: solitons with small H values. A series of efficacious
270: numerical results were obtained for $H = 3$ and $4$ using a
271: 600x400 grid. However, in the case where $H = 2$, a
272: plausible result was obtained only for a 1600x800
273: grid. The cause is an interesting feature of the structure
274: of the studied class of objects. This feature becomes
275: apparent in numerical calculations of the normalized
276: energy density
277: \begin{equation}
278: \rho_E=\frac{1}{E}\int_{-\infty}^{\infty} w_E d{z}.\label{eq:EnergyDensity}
279: \end{equation}
280: The dependences of $\rho_E$ on $r$ for various values of $H$ are
281: plotted in Fig. \ref{F:fig3}. The maximum $\rho_E$ values are reached in the vicinity of $\theta=\pi/2$.
282: With a decrease in $H$, the localization
283: domain corresponding to the first maximum of
284: $\rho_E(r)$ becomes narrower and the maximum becomes
285: steeper. The soliton structure becomes narrower, so that
286: a finer grid with more nodes should be used.
287:
288: Three-dimensional solitons with $H = 0$ exist for $\omega\leq0.915\omega_0$ ~\cite{bib:IvKos1}, where $\omega_0$ is the ferromagnetic-resonance frequency:
289: \begin{equation}
290: \omega_0=\frac {2 \mu_0 M_0 \beta} {\hbar}\label{eq:omega0}
291: \end{equation}
292: An increase in the precession frequency results in the
293: compression of the soliton (Fig.\ref{F:fig4}) implying the possible
294: existence of certain threshold ratio $\omega/\omega_0$ similar to
295: the nontopological soliton~\cite{bib:IvKos1}.
296:
297: Figure \ref{F:fig5} shows the dimensionless energy $\varepsilon={E}/(\alpha l_0)$
298: as a function of $\omega/\omega_0$. It is seen that the soliton energy decreases
299: with an increase in the precession frequency.
300: A similar dependence for a magnon drop, calculated by
301: the aforementioned procedure, is shown in this figure
302: for a comparison.
303:
304: The discovered solitary structures with typical sizes
305: of a few to a few tens of $l_0$ are much smaller than cylindrical
306: magnetic domains and, therefore, can potentially
307: be used for information recording and storage if mechanisms
308: for their generation and control will be developed
309: in the same way as it is done now for magnetic
310: vortices~\cite{bib:OBS}. This would provide the possibility of
311: recording information in three-dimensional samples.
312:
313: \begin{thebibliography}{99}
314: $ $
315: \bibitem{bib:BW}
316: R. Bott and L.\,W. Too, Differential Forms in Algebraic Topology, 1982, New York.\\
317:
318: \bibitem{bib:LD}
319: L.\,D. Faddeev, IAS Princeton, IAS-Report No.75-QS70, (1975).\\
320:
321: \bibitem{bib:LN}
322: \bibinfo{author}{L.~D. Faddeev},
323: \bibinfo{author}{A.~J. Niemi},
324: \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{387}},
325: \bibinfo{pages}{58} (\bibinfo{year}{1997}).\\
326: \url{http://arxiv.org/abs/hep-th/9610193}.\\
327:
328: \bibitem{bib:LN1}
329: \bibinfo{author}{L.~D. Faddeev},
330: \bibinfo{author}{A.~J. Niemi},
331: \bibinfo{journal}{hep-th/9705176},
332: \bibinfo{pages}{} (\bibinfo{year}{1997}).\\
333: \url{http://arxiv.org/abs/hep-th/9705176}.\\
334:
335: \bibitem{bib:BS}
336: \bibinfo{author}{R.~A. Battye},
337: \bibinfo{author}{P.~M. Sutcliffe},
338: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{81}},
339: \bibinfo{pages}{4798} (\bibinfo{year}{1998}).\\
340: \url{http://prola.aps.org/abstract/PRL/v81/i22/p4798_1},\\
341: \url{http://arxiv.org/abs/hep-th/9808129}.\\
342:
343: \bibitem{bib:Glad}
344: \bibinfo{author}{J. Gladikowski},
345: \bibinfo{author}{M. Hellmund},
346: \bibinfo{journal}{Phys. Rev. D} \textbf{\bibinfo{volume}{56}},
347: \bibinfo{pages}{5194} (\bibinfo{year}{1997}).\\
348: \url{http://prola.aps.org/abstract/PRD/v56/i8/p5194_1},\\
349: \url{http://arxiv.org/abs/hep-th/9609035}.\\
350:
351: \bibitem{bib:VM}
352: \bibinfo{author}{G.~E. Volovik},
353: \bibinfo{author}{V.~P. Mineev},
354: \bibinfo{journal}{Sov. Phys. JETP} \textbf{\bibinfo{volume}{46}},
355: \bibinfo{pages}{401} (\bibinfo{year}{1977}).\\
356:
357: \bibitem{bib:DI}
358: \bibinfo{author}{I.~E. Dzyloshinskii},
359: \bibinfo{author}{B.~A. Ivanov},
360: \bibinfo{journal}{JETP Lett.} \textbf{\bibinfo{volume}{29}},
361: \bibinfo{pages}{540} (\bibinfo{year}{1979}).\\
362: \url{http://jetpletters.ac.ru/ps/1455/article_22156.shtml}\\
363:
364:
365: \bibitem{bib:Derrick}
366: \bibinfo{author}{G.~M. Derrick},
367: \bibinfo{journal}{J. Math. Phys} \textbf{\bibinfo{volume}{5}},
368: \bibinfo{pages}{1252} (\bibinfo{year}{1964}).\\
369: \url{http://link.aip.org/link/?JMAPAQ/5/1252/1}.\\
370:
371: \bibitem{bib:KBK}
372: \bibinfo{author}{A.~M. Kosevich},
373: \bibinfo{author}{B.~A. Ivanov},
374: \bibinfo{author}{A.~S. Kovalev},
375: \bibinfo{journal}{Phys. Rep.} \textbf{\bibinfo{volume}{194}},
376: \bibinfo{pages}{117} (\bibinfo{year}{1990}).\\
377: \url{http://dx.doi.org/10.1016/0370-1573(90)90130-T}.\\
378:
379: \bibitem{bib:PT}
380: \bibinfo{author}{N. Papanicolaou},
381: \bibinfo{author}{T.~N. Tomaras},
382: \bibinfo{journal}{Nucl. Phys. B} \textbf{\bibinfo{volume}{360}},
383: \bibinfo{pages}{425} (\bibinfo{year}{1991}).\\
384: \url{http://dx.doi.org/10.1016/0550-3213(91)90410-Y}.\\
385:
386:
387: \bibitem{bib:OBS}
388: \bibinfo{author}{T. Okuno},
389: \bibinfo{author}{K. Mibu},
390: \bibinfo{author}{T. Shinjo},
391: \bibinfo{journal}{J. Appl. Phys} \textbf{\bibinfo{volume}{95}},
392: \bibinfo{pages}{3612} (\bibinfo{year}{2004}).\\
393: \url{http://link.aip.org/link/?JAPIAU/95/3612/1}.\\
394:
395: \bibitem{bib:Volovik1}
396: \bibinfo{author}{G.~E. Volovik},
397: \bibinfo{journal}{cond-mat/0701180},
398: \bibinfo{pages}{} (\bibinfo{year}{2007}).\\
399: \url{http://arxiv.org/abs/cond-mat/0701180}.\\
400:
401: \bibitem{bib:Volovik2}
402: \bibinfo{author}{Yu.~M. Bunkov},
403: \bibinfo{author}{G.~E. Volovik},
404: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{98}},
405: \bibinfo{pages}{265302} (\bibinfo{year}{2007}).\\
406: \url{http://link.aps.org/abstract/PRL/v98/e265302},\\
407: \url{http://arxiv.org/abs/cond-mat/0703183}.\\
408:
409: \bibitem{bib:Volovik3}
410: \bibinfo{author}{Yu.~M. Bunkov},
411: \bibinfo{author}{G.~E. Volovik},
412: \bibinfo{journal}{J. Low Temp. Phys} \textbf{\bibinfo{volume}{150}},
413: \bibinfo{pages}{135} (\bibinfo{year}{2008}).\\
414: \url{http://www.springerlink.com/content/a55n5722617l68k8},\\
415: \url{http://arxiv.org/abs/0708.0663}.\\
416:
417: \bibitem{bib:Volovik4}
418: \bibinfo{author}{Yu.~M. Bunkov},
419: \bibinfo{author}{G.~E. Volovik},
420: \bibinfo{journal}{Physica C} \textbf{\bibinfo{volume}{468}},
421: \bibinfo{pages}{609} (\bibinfo{year}{2008}).\\
422: \url{http://dx.doi.org/10.1016/j.physc.2007.11.026},\\
423: \url{http://arxiv.org/abs/0710.3448}.\\
424:
425: \bibitem{bib:Borisov}
426: \bibinfo{author}{A.~B. Borisov},
427: \bibinfo{journal}{ JETP Lett.} \textbf{\bibinfo{volume}{76}},
428: \bibinfo{pages}{84} (\bibinfo{year}{2002}),\\
429: \url{http://www.springerlink.com/content/g48t670u4906110h}.\\
430:
431: \bibitem{bib:Cooper}
432: \bibinfo{author}{N.~R. Cooper},
433: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{82}},
434: \bibinfo{pages}{1554} (\bibinfo{year}{1999}),\\
435: \url{http://link.aps.org/abstract/PRL/v82/p1554},\\
436: \url{http://arxiv.org/abs/cond-mat/9901037}.\\
437:
438: \bibitem{bib:IvKos1}
439: \bibinfo{author}{B.~A. Ivanov},
440: \bibinfo{author}{A.~M. Kosevich},
441: \bibinfo{journal}{JETP Lett.} \textbf{\bibinfo{volume}{24}},
442: \bibinfo{pages}{495} (\bibinfo{year}{1976}).\\
443: \url{http://jetpletters.ac.ru/ps/1816/article_27757.shtml}.\\
444:
445: \bibitem{bib:Sut2001}
446: \bibinfo{author}{T. Ioannidou},
447: \bibinfo{author}{P.~M. Sutcliffe},
448: \bibinfo{journal}{Physica D} \textbf{\bibinfo{volume}{150}},
449: \bibinfo{pages}{118} (\bibinfo{year}{2001}).\\
450: \url{http://dx.doi.org/10.1016/S0167-2789(00)00221-9},\\
451: \url{http://arxiv.org/abs/cond-mat/0101129}.\\
452:
453: \bibitem{bib:Sut2007}
454: \bibinfo{author}{P.~M. Sutcliffe},
455: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{76}},
456: \bibinfo{pages}{184439} (\bibinfo{year}{2007}).\\
457: \url{http://link.aps.org/abstract/PRB/v76/e184439},\\
458: \url{http://arxiv.org/abs/0707.1383}.\\
459:
460: \bibitem{bib:KUR}
461: \bibinfo{author}{A. Kundu},
462: \bibinfo{author}{Y.~P. Rybakov},
463: \bibinfo{journal}{ J. Phys. A} \textbf{\bibinfo{volume}{15}},
464: \bibinfo{pages}{269} (\bibinfo{year}{1982}).\\
465: \url{http://www.iop.org/EJ/abstract/0305-4470/15/1/035}.\\
466:
467: \bibitem{bib:Tjon}
468: \bibinfo{author}{J. Tjon},
469: \bibinfo{author}{J. Wright},
470: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{15}},
471: \bibinfo{pages}{3470} (\bibinfo{year}{1977}).\\
472: \url{http://link.aps.org/abstract/PRB/v15/p3470}.\\
473:
474: \bibitem{bib:PD}
475: B.\,N. Pshenichnyi and Yu.\,M. Danilin, Numerical Methods in Extremal Problems (Nauka, Moscow, 1975) [in Russian].
476:
477: \end{thebibliography}
478:
479: \end{document}
480: