1: %\documentstyle[preprint,pre,aps,epsfig]{revtex}
2: \documentstyle[pre,multicol,aps,epsfig]{revtex}
3:
4: \begin{document}
5:
6: \draft
7:
8: \tighten
9:
10: \title{On the existence of internal modes of sine-Gordon kinks}
11:
12: \author{Niurka R.\ Quintero$^{*}$ and Angel S\'anchez$^{\dag}$}
13:
14: \address{Grupo Interdisciplinar de Sistemas
15: Complicados (GISC), Departamento de Matem\'aticas,
16: Universidad Carlos III de Madrid,\\
17: Avenida de la Universidad 30, E-28911 Legan\'{e}s, Madrid, Spain}
18:
19: \author{Franz G.\ Mertens$^{\ddag}$}
20:
21: \address{Physikalisches Institut, Universit\"at Bayreuth,
22: D-95440 Bayreuth, Germany}
23:
24: \date{\today}
25:
26: \maketitle
27:
28: \begin{abstract}
29: We study whether or not sine-Gordon kinks exhibit internal modes or
30: ``quasimodes.'' By considering the response of the kinks to ac forces
31: and initial distortions, we show that neither intrinsic
32: internal modes nor ``quasimodes'' exist in contrast to previous reports.
33: However, we do identify a
34: different kind of internal mode
35: bifurcating from the bottom edge of the phonon band
36: which arises from the discretization of the system in the numerical
37: simulations, thus confirming
38: recent predictions.
39: \end{abstract}
40:
41: \pacs{PACS numbers: 05.45.Yv, 02.30.Jr, 03.50.-z, 63.20.Pw}
42:
43: %\narrowtext
44: \begin{multicols}{2}
45:
46: \section{Introduction}
47: \label{intro}
48:
49: Solitons, which were originally a concept arising in the study of integrable
50: systems \cite{scott}, have rapidly become a key paradigm in perturbed,
51: non-integrable
52: systems describing very many physical applications \cite{yurirev}. A very
53: succesful picture of soliton dynamics under different perturbations has
54: resulted from the use of collective coordinate (CC) techniques \cite{siam}: In many
55: instances, solitons behave basically like point-like particles, and therefore
56: their time evolution can be predicted by equations of motion which
57: are ordinary differential equations.
58: Nevertheless, along the last decade it has been progressively realized that
59: internal degrees of freedom of the solitons play a crucial role in a number of
60: problems: They govern resonant soliton (or solitary wave) collisions
61: \cite{David}, give rise to nontrivial soliton-impurity interactions
62: \cite{Yura}, and can be excited both by ac forces \cite{prl}
63: and by thermal noise \cite{grant}. Therefore,
64: it is very important to assess whether or not physically relevant solitons
65: possess internal modes.
66:
67: In this Rapid Communication we study solitons of
68: the sine-Gordon (sG) equation, which arises in a diverse range of areas of
69: physics, covering from crystal dislocation theory to Josephson junctions
70: \cite{scott,yurirev,siam}. In spite of the fact that linear stability
71: analysis leads to only one eigenvalue in the discrete spectrum,
72: corresponding to a zero frequency Goldstone mode \cite{yurirev},
73: Boesch and Willis \cite{willis} claimed that they found
74: an internal ``quasimode'' above but close to the lower phonon band edge.
75: They described this ``quasimode'' as a long-lived oscillation of the
76: width of the sG kink.
77: The possibility of such a mode had been suggested earlier by Rice
78: \cite{Rice}, by means of a variational approach that reproduced with
79: great accuracy the internal mode frequency of the $\phi^4$ model
80: \cite{scott}
81: and predicted the existence of a similar mode for the sG problem.
82: To our knowledge, Ref.\ \cite{willis}
83: is the only paper reporting the observation of
84: this mode, although hints of its existence are scattered all over
85: the soliton literature \cite{mas}. A few years later, it was
86: theoretically found \cite{yuri}, within the Rice approach, that
87: the internal mode of the sG and the $\phi^4$ kinks could be
88: excited by a constant external force or by an initial distortion;
89: however, no numerical evidence
90: was presented to confirm these results.
91: Therefore, we decided to revisit the
92: issue of the existence of the internal mode of the sG equation in order
93: to give a definitive answer to it. Our approach, based in part on our
94: earlier findings \cite{prl}, overcomes the difficulty
95: that the internal ``quasimode'' would lie within the phonon band by allowing
96: us to look at driving frequencies well in the gap.
97: As we will show below, this permits us to establish clearly that
98: this putative internal mode does not exist. Furthermore, we do observe
99: the excitation of a discreteness induced internal mode \cite{Braun,kivshar},
100: whose frequency is in excellent agreement with recent predictions by
101: Kevrekidis and Jones \cite{jones}. We substantiate
102: these claims by briefly recalling our main analytical results in Sec.\
103: \ref{colco}, collecting our numerical results in Sec.\ \ref{numso},
104: and discussing our conclusions in Sec.\ \ref{conclu}.
105:
106: \section{Resonances due to ac forces}
107: \label{colco}
108:
109: We begin by considering the perturbed sG equation
110: \begin{equation}
111: \begin{array}{c}
112: \phi_{tt} - \phi_{xx} = - \sin\phi +
113: \epsilon \sin(\delta t + \delta_{0}).
114: \label{ecua1}
115: \end{array}
116: \end{equation}
117: where $\epsilon$, $\delta$, and $\delta_0$ are the amplitude, frequency,
118: and phase of the external periodic force.
119: In \cite{prl} we analyzed analytically and numerically the above problem
120: in the more general context of nonlinear (possibly damped)
121: Klein-Gordon systems,
122: which includes Eq.\ (\ref{ecua1}) as well as the $\phi^4$ model.
123: Particularizing our results, obtained by means of the generalized
124: travelling wave {\em Ansatz} \cite{franz} combined with the Rice {\em Ansatz}
125: \cite{Rice}, we find that
126: the momentum $P(t)$ and the width of the kink $l(t)$
127: (of a general nonlinear Klein-Gordon equation; specific constants for
128: the sG case will be given below) obey the equations
129: \begin{eqnarray}
130: \frac{dP }{dt} &=& - q\, \epsilon \sin(\delta t + \delta_{0}), \quad
131: P(t) \equiv \frac{M_{0} l_{0} \dot{X}}{l(t)},
132: \label{ecua23}\\
133: \alpha \left(\dot{l}^{2} - 2 l \ddot{l} \right) &=&
134: \frac{l^{2}}{l_{0}^{2}} \left (1 + \frac{P^{2}}{M_{0}^{2}} \right) - 1,
135: \label{ecua24}
136: \end{eqnarray}
137: where X(t) is the kink position. For
138: sG, $q=2\pi$, $M_{0}=8$, $l_{0}=1$ and $\alpha=\pi^2/12$.
139: The equation for the momentum is linear and can be solved exactly, yielding
140: \begin{equation}
141: P(t) = \left[P(0)-
142: \frac{q\,\epsilon
143: \cos(\delta_{0})}{\delta} \right]+
144: \frac{q\,\epsilon}{\delta} \,
145: \cos(\delta t + \delta_{0})
146: \label{ecua4}
147: \end{equation}
148: We thus see that the term $P(t)^2$ in Eq.\ (\ref{ecua24}) involves two
149: frequencies, $\delta$ and $ 2 \delta$.
150: As shown in \cite{prl}, for the $\phi^4$ kink this leads to
151: dramatic resonance effects when the ac driving frequency $\delta$ is
152: half the frequency $\Omega_i$ of the internal mode.
153: Without resonant excitation of kink modes,
154: the kink oscillates around its initial position as a pure point-like
155: particle, as shown analytically and numerically in \cite{EPJBus}.
156: However, as shown in \cite{prl} for the above parameters, in case
157: an internal mode is excited as in the $\phi^4$ system, the system energy
158: grows and eventually leads to chaotic motion of the kink.
159: In addition, the resonance at $\Omega_i/2$ turns
160: out to be much stronger than that at $\Omega_i$, which
161: makes it a convenient way to probe those modes:
162: In case the sG
163: kink would possess such a degree of freedom, we should observe similar
164: phenomena.
165: In \cite{willis},
166: the frequency of the numerically observed
167: ``quasimode'' was $\omega_s=1.004\pm 0.001$ in our
168: units, whereas Rice's theoretical
169: analysis leads to \cite{Rice} $\Omega_R=1.103$.
170: In both cases, the corresponding half frequencies lie well within the
171: gap; we can thus force the system at the predicted resonances without
172: exciting much radiation which could mask any internal mode effects.
173:
174: \section{Numerical results}
175: \label{numso}
176:
177: In this Section, we present the results of our numerical search for
178: internal modes or ``quasimodes'' of the sG equation.
179: We have computed the numerical solution of the perturbed
180: sG equation (\ref{ecua1})
181: by using the Strauss-V\'azquez scheme \cite{stvaz},
182: the total length of the system being $L=100$, the steps being
183: $\Delta x=0.05$, $\Delta t=0.005$ (or other values when indicated).
184: Free boundary conditions were also used, and
185: the final time in the simulations was $25 \, 000$, enough to detect any
186: possible resonance \cite{prl}.
187:
188: \subsection{ac force}
189: \label{acfor}
190:
191: First, we have perturbed our system with an ac force of amplitude
192: $\epsilon = 0.01$ and phase $\delta_{0} = \pi/2$.
193: We have used a
194: sG kink at rest as initial condition. We
195: have computed the energy (as defined in \cite{stvaz}) during the evolution,
196: and then the mean value of the energy in the time interval
197: $10\,000 \le t \le 25\,000$, studying a range of
198: values of $\delta$ around half of the possible internal mode frequencies.
199: {}Figure \ref{f8} collects our results for the mean energy as a function
200: of the frequency. We clearly see that the mean energy rapidly
201: increases for
202: some values of $\delta$ which do not coincide with any of the proposed
203: internal modes or ``quasimodes.''
204: The first (leftmost) maximum lies below half the frequency of the
205: lower phonon band edge; therefore, this one can not be associated with
206: a phonon mode and we will discuss its origin in detail below.
207: As for the other peaks, Table \ref{tab2.1} shows that
208: we have been able to identify them with half of the values of the first
209: frequencies of the phonon band, $\omega_n$, which are given by
210: \begin{equation}
211: \omega_{n}=\sqrt{1+\left[{(2 \pi n )}/{L}\right]^{2}},$
212: $n=1,2,3,\ldots, N=L/\Delta x.
213: \label{fon}
214: \end{equation}
215: A comparison of
216: the third and fourth columns of Table \ref{tab2.1},
217: where we list
218: the half values of the frequencies corresponding to the first
219: $8$ radiational modes and the numerical peak
220: frequencies ${\tilde{\omega}_{n}}$ in
221: Fig.\ \ref{f8}, counting from the second maximum from the left,
222: allows to verify that these values are nearly identical in all cases
223: (see also the relative differences
224: $|\omega_{n}/2-{\tilde{\omega}_{n}}|$ in Table \ref{tab2.1}).
225: We stress that identification of radiation modes as the origin of
226: these resonances makes perfect physical sense: Indeed, an extended
227: CC approach including radiation modes leads to the conclusion that
228: the radiation modes should also be excited parametrically by ac
229: drivings.
230: In view of all this, the main conclusion we can
231: draw from Fig.\ \ref{f8} is the absence of
232: any resonance due to
233: the internal quasi-mode in the phonon band
234: reported by Boesch and Willis
235: \cite{willis}, which, in view of the frequency they reported,
236: would lead to an extra peak located between the two rightmost
237: ones in Fig.\ \ref{f8}. In this regard, it is important to note that
238: the accuracy of our simulations in detecting every radiation mode
239: rules out the possibility that we have missed this quasi-mode
240: (at either $\omega_s$ or $\Omega_R$).
241:
242: Interestingly, we do observe an internal mode: Let us recall that the
243: first maximum of the energy occurs at
244: $\delta=\omega_{d}=0.4999$, i.e., $2 \omega_{d}=0.9998$ is below
245: the lower phonon frequency $\omega_{ph}=1$ as we mentioned above.
246: The reason for this peak is a
247: novel internal mode
248: of the sG kink, that appears as a bifurcation from the lowest phonon of
249: the continuum equation, of frequency $\omega_{ph}=1$,
250: due to the effect of discreteness \cite{Braun,kivshar,jones}.
251: As has been recently shown \cite{jones}, the frequency
252: of this new internal mode is
253: $\omega_{i} \approx \sqrt{1-(4/2025) (\Delta x)^4}$ (the $(\Delta x)^4$
254: dependence was
255: already observed in \cite{Braun}).
256: In our simulations, $\Delta x=0.05$, hence
257: $\omega_{i}=0.999998$, approximately equal to $2 \omega_{d}$.
258: Therefore, the first peak in Fig.\ \ref{f8} corresponds to the parametric
259: resonance at half the value of the discreteness induced internal mode,
260: very much like the resonance with the $\phi^4$ kink internal mode reported
261: earlier \cite{prl}.
262:
263: \subsection{Deformed initial kink}
264: \label{confo}
265:
266: In view of our negative result about the internal quasi-mode of the sG kink,
267: we carried out more numerical simulations specifically designed to find it.
268: Following \cite{willis}, we chose
269: initial conditions given by
270: \begin{equation}
271: \begin{array}{l}
272: {\displaystyle{
273: \phi(x,0)=4\,\mbox{arctan}\left(
274: \exp\left[\frac{x-X(0)}{l(0)}\right]\right)}},\\
275: \\
276: {\displaystyle{
277: \phi_t(x,0)=2\,\frac{\displaystyle{\left[
278: -\frac{u(0)}{l(0)}-\frac{x-X(0)}{l^2(0)}
279: \dot{l}(0)\right]}}{\displaystyle\mbox{cosh}\left[\frac{x-X(0)}{l(0)}\right]}
280: }},
281: \end{array}
282: \label{val-ini-sG}
283: \end{equation}
284: which correspond to a deformed kink if $l(0)\neq 1$ or $\dot{l}(0)\neq 0$.
285: By starting their simulations with
286: such initial conditions, Boesch and Willis claimed that they were able
287: to excite the internal quasi-mode, and hence we hope that, if it is
288: indeed present, we should find it.
289:
290: At this point, we have to recall the work of
291: Majern\'\i kov\'a, Gaididei and Braun
292: \cite{yuri}, who, as mentioned in the introduction,
293: theoretically considered the problem of a deformed kink
294: given by the expressions above driven by a external constant force.
295: When applied to this problem, our
296: CC approach yields the same
297: two ordinary differential equations as in \cite{yuri}
298: for $X(t)$ and $l(t)$,
299: [Eqs.\ (\ref{ecua23})-(\ref{ecua24}) without the sine term, i.e., with
300: only a constant force $\epsilon$].
301: We have solved these equations exactly for arbitrary initial conditions
302: \cite{usi}
303: in terms
304: of the Whittaker functions, concluding
305: that the kink width oscillates only if $l(0) \ne l_{s}$
306: (the natural width of the undistorted kink, $1/\sqrt{1-u(0)^2}$)
307: or ${\dot{l}}(0) \ne 0$ in (\ref{val-ini-sG}), whereas
308: external force and dissipation (if present) only damp out
309: those oscillations. Hence, kink width oscillations
310: should be easiest to detect without force and damping and, consequently,
311: here we consider only free propagation of an initially distorted kink
312: (a general analysis of the constant force problem will be presented
313: elsewhere \cite{usi}). As a check of the validity of this procedure to
314: probe internal modes, we
315: have numerically verified that the $\phi^4$ kink indeed
316: behaves in this fashion due to the excitation of the internal mode,
317: exhibiting oscillations in the kink width very close to those predicted
318: by the CC approach. We then turned to
319: the sG case, and we found that the
320: oscillations are not correctly described
321: by the CC equations. In Fig.\
322: \ref{f3.8} we can see the evolution of the kink width computed from
323: numerical simulation of the sG system, with $l(0)=1$, $u(0)=0$, and
324: ${\dot{l}}(0)=0.3$. It is clear that $l(t)$ is not a simple oscillatory
325: function as in the $\phi^4$ case, and
326: Fourier analysis yields
327: the following frequencies:
328: $\omega_{d}=0.9983$, ${\tilde{\omega}}_{1}=1.0034$,
329: ${\tilde{\omega}}_{2}=1.0083$, ${\tilde{\omega}}_{3}=1.0184$,
330: ${\tilde{\omega}}_{4}=1.0335$, ${\tilde{\omega}}_{5}=1.0511$,
331: ${\tilde{\omega}}_{6}=1.0712$, ${\tilde{\omega}}_{7}=1.0963$.
332: As before, all these frequencies are very close to the
333: discreteness induced internal mode and the first few radiational
334: modes, with no evidence supporting the existence of the internal
335: quasi-mode above the phonon band edge.
336:
337: \section{Conclusions}
338: \label{conclu}
339:
340: The first conclusion of this work is that the existence of the
341: proposed internal quasi-mode of the sG kink above the phonon
342: band edge \cite{willis} is very unlikely. We have found
343: neither resonances nor long-lived kink width oscillations
344: arising from such a mode when the kink is subject
345: to ac driving or to initial deformations, respectively. We are confident
346: that we have carefully explored all the range of relevant frequencies,
347: as we have accurately detected all possible resonances within that
348: range. In addition, the comparison with the phenomenology observed
349: for the $\phi^4$ kink, that possesses a true internal mode, shows
350: that if the sG kink would have such an internal mode, we would have
351: detected its influence through numerical simulations as for the
352: $\phi^4$ equation \cite{prl,usi}. In fact, now that we have
353: seen that radiation modes are parametrically excited by their
354: corresponding half-frequency drivings, we believe we can suggest an
355: explanation for the results of Boesch and Willis.
356: Notice that they used Eq.\ (\ref{fon}) to calculate
357: the lower phonon frequency, obtaining that
358: $\omega_{1}=\sqrt{1+(2 \pi/1000)^2} = 1.00002$, whereas
359: the internal ``quasimode'' frequency that they found
360: in the numerical simulations was $\omega_{s}=1.004$. As those two
361: values were largely different, they concluded that the quasi-mode was
362: well above the lowest phonon frequency and different from it.
363: However, as we have discussed in the previous sections,
364: they used initial conditions in which the initial kink
365: velocity was not zero; in this case, it can be easily shown that
366: the corresponding frequencies of the phonons are given by
367: $$
368: {\displaystyle{
369: \bar{\omega}_{k}=\frac{\omega_{k}-k\,u(0)}{\sqrt{1-u^2(0)}}, \,\ \,\
370: \omega_{k}=\sqrt{1+k^2}.
371: }}
372: $$
373: If we now insert the parameters
374: used in \cite{willis} to perform a numerical experiment similar to the
375: one discussed here, we find
376: $\omega_{1}=1.004$, i.e., the resonance observed by Boesch and Willis
377: took place in fact with the lowest frequency phonon {\em in the presence
378: of a moving kink}
379: and not with any internal quasi-mode.
380:
381: The second conclusion of the present paper relates to the recently
382: proposed internal modes induced by perturbations, in our case by
383: discreteness \cite{Braun,kivshar,jones}. By the two different types of
384: numerical simulations used in our research, we have detected the
385: internal mode bifurcating from the edge of the phonon band
386: (of the continuum problem). The frequency of that mode agrees qualitatively
387: very well with the one predicted in \cite{jones}. Therefore, our research
388: fully confirms the results of those papers and, in addition, shows that
389: these perturbation induced internal modes behave very much like intrinsic
390: ones.
391: Our result, along with the finding of similar
392: internal modes in other nonintegrable sG systems \cite{nueva},
393: strongly supports the generality of the phenomenon predicted
394: in \cite{kivshar}.
395: Finally, the fact that we
396: are able to detect this novel mode reinforces our previous conclusion
397: of the non-existence of the internal quasi-mode, as indeed the sG kink
398: responds to the ac driving or to deformations as it should if it had
399: an internal mode, but for the discreteness induced one only.
400: We thus believe that the present work definitely rules out the
401: possibility of internal modes or ``quasimodes'' of sG kinks close to
402: and above the lower phonon band edge.
403:
404: \vspace*{-1cm}
405:
406: \section{Acknowledgments}
407:
408: We thank Yuri Kivshar and Yuri Gaididei for discussions.
409: Work at GISC (Legan\'es) has been supported by
410: DGESIC (Spain) grant PB96-0119. Travel between Bayreuth and Madrid has been
411: supported by ``Acciones Integradas Hispano-Alemanas'', a joint program of
412: DAAD (Az.\ 314-AI) and DGESIC.
413:
414: \begin{references}
415:
416: \bibitem[*]{1} kinter@math.uc3m.es
417:
418: \bibitem[\dag]{2} anxo@math.uc3m.es
419:
420: \bibitem[\ddag]{3} franz.mertens@theo.phy.uni-bayreuth.de
421:
422: \bibitem{scott} A.\ C.\ Scott, {\em Nonlinear Science} (Oxford University,
423: Oxford, 1999).
424:
425: \bibitem{yurirev} Yu.\ S.\ Kivshar and B.\ A.\ Malomed, Rev.\ Mod.\ Phys.\
426: {\bf 61}, 763 (1989).
427:
428: \bibitem{siam} A.\ S\'anchez and A.\ R.\ Bishop, SIAM Rev.\ {\bf 40}, 579
429: (1998).
430:
431: \bibitem{David} M.\ Peyrard and D.\ K.\ Campbell, Physica D {\bf 9}, 33 (1983);
432: D.\ K.\ Campbell, J.\ F.\ Schonfeld, and C.\ A.\ Wingate, {\em ibid.} {\bf 9},
433: 1 (1983); D.\ K.\ Campbell, M.\ Peyrard, and P.\ Sodano, {\em ibid.} {\bf 19},
434: 165 (1986).
435:
436: \bibitem{Yura} Yu.\ S.\ Kivshar, F.\ Zhang, and L.\ V\'azquez, Phys.\
437: Rev.\ Lett.\ {\bf 67}, 1177 (1991); Phys.\ Rev.\ A {\bf 46}, 5214 (1992).
438:
439: \bibitem{prl} N.\ R.\ Quintero, A.\ S\'anchez, and F.\ G.\ Mertens,
440: Phys.\ Rev.\ Lett.\ {\bf 84}, 871 (2000).
441:
442: \bibitem{grant} S.\ Habib, G.\ Lythe, and F.\ G.\ Mertens, unpublished.
443:
444: \bibitem{willis} R.\ Boesch and C.\ R.\ Willis, Phys.\ Rev.\ B {\bf 42},
445: 2290 (1990).
446:
447: \bibitem{Rice} M.\ J.\ Rice, Phys.\ Rev.\ B {\bf 28},
448: 3587 (1983).
449:
450: \bibitem{mas} As a few examples, see
451: P.\ Tchofo Dinda. R.\ Boesch, E.\ Coquet and C.\ R.\ Willis, Phys.\
452: Rev.\ B {\bf 46}, 3311 (1992);
453: Zhang Fei, Vladimir V.\ Konotop, Michel Peyrard and Luis V\'azquez,
454: Phys.\ Rev.\ E {\bf 48}, 548 (1993);
455: Fei Zhang, Phys.\ Rev.\ E {\bf 54}, 4325 (1996);
456: Physica D {\bf 110}, 51 (1997).
457:
458: \bibitem{yuri} E.\ Majern\'\i kov\'a, Yu.\ B.\ Gaididei, and O.\ M.\ Braun,
459: Phys.\ Rev.\ E {\bf 52},
460: 1241 (1995).
461:
462: \bibitem{Braun} O.\ M.\ Braun, Yu.\ S.\ Kivshar, and M.\ Peyrard, Phys.\
463: Rev.\ E {\bf 56}, 6050 (1997).
464:
465: \bibitem{kivshar} Yu.\ S.\ Kivshar, D.\ E.\ Pelinovsky, T.\ Cretegny, and
466: M.\ Peyrard, Phys.\ Rev.\ Lett.\ {\bf 80}, 5032 (1998).
467:
468: \bibitem{jones} P.\ G.\ Kevrekidis and C.\ K.\ R.\ T.\ Jones,
469: Phys.\ Rev.\ E {\bf 61}, 3114 (2000).
470:
471: \bibitem{franz} F.\ G.\ Mertens, H.\ J.\ Schnitzer, and A.\ R.\ Bishop,
472: Phys.\ Rev.\ B {\bf 56}, 2510 (1997).
473:
474: \bibitem{EPJBus} N.\ R.\ Quintero and A.\ S\'anchez,
475: Eur.\ Phys.\ J.\ B {\bf 6}, 133 (1998).
476:
477: \bibitem{stvaz} W.\ A.\ Strauss and L.\ V\'azquez,
478: J.\ Comput.\ Phys.\ {\bf 28},
479: 271 (1978).
480:
481: \bibitem{usi} N.\ R.\ Quintero, A.\ S\'anchez, and F.\ G.\ Mertens,
482: unpublished. We note that in \cite{yuri} only solutions for a special,
483: restricted set of initial conditions were found.
484:
485: \bibitem{nueva} S.\ F.\ Mingaleev, Yu.\ B.\ Gaididei, E.\ Majern\'\i kov\'a,
486: and S.\ Shpyrko, Phys.\ Rev.\ E {\bf 61}, 4454 (2000).
487: \end{references}
488:
489: \begin{figure}
490: \epsfig{figure=fig2.5.7a.eps, width=2.3in, angle=-90}
491: \caption{Mean energy of the
492: sG system, in dimensionless units, when $\delta$ is close to $\Omega_{R}/2$.
493: Points have been calculated from numerical simulations for
494: $10\,000 \le t \le 25\,000$, with
495: $\epsilon=0.01$, $\delta_{0}=\pi/2$, $u(0)=0$, $X(0)=0$.
496: The parameters of the discretization are
497: $\Delta x=0.05$, $\Delta t=0.005$ and $L=100$.}
498: \label{f8}
499: \end{figure}
500: \begin{figure}
501: \epsfig{file=f3.8b.eps,width=2.3in,angle=-90}
502: \caption{Evolution of the kink width, in dimensionless units,
503: obtained from numerical simulations starting from a free, deformed kink, with
504: $u(0)=0$, $l(0)=1$ and ${\dot{l}}(0)=0.3$. The parameters
505: of the discretization are $L=100$, $\Delta x=0.1$ and $\Delta t=0.01$.
506: }
507: \label{f3.8}
508: \end{figure}
509: \begin{table}[h]
510: \begin{center}
511: \begin{tabular}{|c|c|c|c|c|}
512: Mode($n$) & $\omega_{n}$ & $\omega_{n}/2$ &
513: $\widetilde{\omega}_n$ &
514: $\left|\omega_{n}/2-\widetilde{\omega}_n \right|$
515: \\
516: \hline
517: 1 & 1.0019 & 0.5009 & 0.5009 & 8.6 10${}^{-5}$ \\
518: \hline
519: 2 & 1.0079 & 0.5039 & 0.5040 & 7 10${}^{-5}$ \\
520: \hline
521: 3 & 1.0176 & 0.5088 & 0.5090 & 1.9 10${}^{-4}$ \\
522: \hline
523: 4 & 1.0311 & 0.5155 & 0.5161 & 5.5 10${}^{-4}$ \\
524: \hline
525: 5 & 1.0482 & 0.5241 & 0.5250 & 9.1 10${}^{-4}$ \\
526: \hline
527: 6 & 1.0687 & 0.5343 & 0.5356 & 1.3 10${}^{-3}$ \\
528: \hline
529: 7 & 1.0924 & 0.5462 & 0.5479 & 1.7 10${}^{-3}$ \\
530: \hline
531: 8 & 1.1192 & 0.5596 & 0.5618 & 2.2 10${}^{-3}$
532: \end{tabular}
533: \end{center}
534: \caption{Comparison of the half of the first $8$ frequencies
535: $\omega_{n}/2$ of
536: radiational modes with the computed frequencies
537: $\widetilde{\omega}_n$
538: from numerical simulations of (\ref{ecua1}) at which peaks in
539: $E_{m}$ arise (see Fig.\ \ref{f8}).}
540: \label{tab2.1}
541: \end{table}
542:
543: \end{multicols}
544:
545: \end{document}
546: