1: \documentstyle[12pt,epsf]{article}

2: %\def\baselinestretch{2.0}

3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

4: \setlength{\textheight}{8.4in}

5: \setlength{\topmargin}{-0.3in}

6: \setlength{\textwidth}{6.5in}

7: \setlength{\oddsidemargin}{0.2in}

8: \setlength{\evensidemargin}{\oddsidemargin}

9: %\setlength{\textheight}{22 true cm}

10: %\setlength{\textwidth}{15 true cm}

11: %\hoffset=-0.7 true cm

12: %\voffset=-2.7 true cm

13: \newcommand{\ds }{\displaystyle}

14: \newcommand{\ra}{\rightarrow}

15: \newcommand{\be}{\begin{equation}}

16: \newcommand{\ee}{\end{equation}}

17: \newcommand{\bea}{\begin{eqnarray}}

18: \newcommand{\eea}{\end{eqnarray}}

19: \newcommand{\ci}{\cite}

20: \newcommand{\bi}{\bibitem}

21: \newcommand{\nono}{\nonumber \\}

22: \newcommand{\tr}{{\rm tr}}

23: \newcommand{\p}{{\vec{\bf p}}}

24: \newcommand{\e}{{\rm e}}

25: \newcommand{\R}{{\bf R}}

26: \newcommand{\sF}{{\sin F}}

27: \newcommand{\cF}{{\cos F}}

28: \newcommand{\ssf}{{\sin^2 F}}

29: \newcommand{\ccF}{{\cos^2 F}}

30: \newcommand{\quart}{\frac{1}{4}}

31: \newcommand{\da}{\dagger}

32: \newcommand{\dd}{\partial}

33: \newcommand{\bftau}{\mbox{\boldmath$\tau$}}

34: \newcommand{\bfo}{\mbox{\boldmath$\omega$}}

35: \newcommand{\s}{\sigma}

36: \newcommand{\bfLambda}{\mbox{\boldmath$\Lambda$}}

37: \newcommand{{\bfna}}{\mbox{\boldmath$\vec{\nabla}$}}

38: \newcommand{\epp}{\mbox{\boldmath$\vec\epsilon$}}

39: \newcommand{\half}{\frac{1}{2}}

40: %\newcommand{\ra}{\rangle}

41: \newcommand{\al}{\lambda}

42: \newcommand{\la}{\langle}

43: \newcommand{\jj}{\vec{\bf{j}}}

44: \newcommand{\FF}{\vec{\bf{F}}}

45: \newcommand{\vv}{\vec{\bf{v}}}

46: \newcommand{\JJ}{\vec{\bf{J}}}

47: \newcommand{\rr}{\vec{\bf{r}}}

48: \newcommand{\kk}{\vec{\bf{k}}}

49: \newcommand{\q}{\vec{\bf{q}}}

50: %\newcommand{\az}{\mbox{\!\!}}

51: \newcommand{\ck}{\mbox{\bf{\^k}}}

52: \newcommand{\x}{{\bf x}}

53: \newcommand{\arctanh}{{{\rm arctanh}}}

54: \newcommand{\sech}{{{\rm sech}}}

55: \newcommand{\cl}{\centerline}

56: 

57: \def\dal{\,\lower0.3ex\vbox{\hrule\hbox{\vrule\kern2pt\vbox{\kern4pt\kern4pt}

58: \kern2pt\vrule}\hrule}\,}

59: \def\f{\phi}

60: \def\umu{^\mu m}

61: \def\lmu{_\mu m}

62: \def\s{\sigma}

63: \def\o{\omega}

64: \def\L{{\cal L}}

65: \def\D{_{\rm D}}

66: \def\hrf{\hrulefill}

67: 

68: \begin{document}

69: 

70: 

71: 

72: 

73: \title{\sl The Sine-Gordon Wobble}

74: \vspace{1 true cm}

75: \author{G. K\"albermann$^*$

76: \\Soil and Water dept., Faculty of

77: Agriculture, Rehovot 76100, Israel}

78: \maketitle

79: 

80: \vspace{3 true cm}

81: \begin{abstract}

82: %\baselineskip 1.5 pc

83: 

84: Nonperturbative, oscillatory, winding number one solutions 

85: of the Sine-Gordon equation are presented and studied numerically. 

86: We call these nonperturbative shape modes {\sl wobble} solitons.

87: Perturbed Sine-Gordon kinks are found to decay to {\sl wobble} solitons.

88: 

89: 

90: 

91: \end{abstract}

92: {\bf PACS} 02.30.Jr, 03.40.Kf, 03.50.-z, 03.65.Ge, 03.75.Lm, 05.45.Yv, 

93: 11.10.Lm, 63.20.Pw\\

94: 

95: 

96: 

97: 

98: 

99: $^*${\sl e-mail address: hope@vms.huji.ac.il}

100: 

101: 

102: 

103: \newpage

104: 

105: \section{\sl Introduction}

106: 

107: The Sine-Gordon equation was discovered in the study of constant 

108: negative curvature metric spaces at the end of the 19$^{th}$ 

109: century\ci{eisenhart}. It reappeared in physical problems dealing with

110: one dimensional dislocations

111: \ci{lamb}, long Josephson junctions and in other settings.\ci{rem}

112: 

113: The Sine-Gordon equation possesses solitary wave solutions.

114: These solitary waves are solitons\ci{zabusky}.

115: The equation is completely integrable and has an infinite number of 

116: conserved currents.\ci{lamb, rajaraman}

117: 

118: The solitons of the Sine-Gordon theory carry a topological winding number 

119: {\sl q}. Sine-Gordon solitary waves are topological solitons. 

120: The winding number zero sector {\sl q=0} 

121: supports bound soliton-antisoliton

122: solutions, the breathers, as well as an unbound soliton-antisoliton 

123: pairs\ci{lamb}. 

124: The {\sl q=1} sector solitary wave is the kink soliton.

125: 

126: In the recent past, a controversy has arisen concerning the existence 

127: of oscillatory solutions in the q =1 sector. 

128: Shape modes were predicted by

129: Rice\ci{rice} by means of a collective coordinate method. 

130: Boesch and 

131: Willis\ci{boesch} studied the excitation of this internal quasimode using

132: a more refined collective coordinate approach as well as numerical integration.

133: The predictions of both works are not exact, 

134: or even approximately so, due to the limitations of the collective coordinate 

135: method.

136: In the numerical and collective coordinate treatments, the angular 

137: frequency of the oscillations of the kink

138: soliton width was found to be above the threshold for the production

139: of phonons. Phonons are the solutions of the Klein-Gordon equation, obtained 

140: by linearizing the Sine-Gordon equation\ci{fogel}.

141: Hence, if this oscillation exists, it is embedded in the continuum

142: and must decay, albeit with a small decay constant for angular frequencies

143: near threshold.

144: 

145: Quintero et al.\ci{quintero} have recently contested the existence

146: of this shape mode. They have argued that the numerical solution 

147: of Boesch and Willis\ci{boesch} is incorrect. 

148: Quintero et al.\ci{quintero} suggest that

149: this quasi-mode is nothing more than a numerical effect due to discretization.

150: In discrete nonlinear equations, a mode in the continuum, can sink below 

151: threshold depending on the value of lattice constant\ci{kivshar}.

152: \\

153: \indent

154: In the present work we show analytical nonperturbative solutions to the

155: Sine-Gordon equation that are oscillatory and apparently stable that we call 

156: {\sl the wobble} solitons. The wobble solitons are derived by means of the

157: Inverse Scattering Transform (IST) method following 

158: Lamb\ci{lamb} and Segur\ci{segur}.

159: The IST method produces

160: soliton solutions based on scattering data of 

161: Schr\"odinger-like equations. The data leads to a potential - hence the name

162: inverse scattering transform - from which the soliton is derived. 

163: The expression for the wobble soliton we derive 

164: corrects the one given by Segur\ci{segur}.

165: The solution will be depicted and checked analytically as well as numerically.

166: 

167: The angular frequency of the oscillation of the wobble is found to

168: range between zero and one, where the phonon continuum takes over.

169: There is no gap in the frequency spectrum. A dense set of nonperturbative

170: nonlinear wobbles fills it. We also touch upon the stability issue.

171: 

172: Having shown the existence and probable stability of the wobble, we 

173: connect to the problem of shape oscillations in distorted

174: kinks. We recover the results of Boesch and Willis\ci{boesch} and  

175: point out a probable source 

176: of error in the numerical calculations of Quintero et al.\ci{quintero}

177: The shape modes found in the literature are shown to be

178: an intermediate stage on the way between distorted kinks 

179: and wobbles.

180: 

181: The next section summarizes the IST derivation of the 

182: the wobble\ci{lamb},\ci{segur} and addresses the stability issue.

183: Section 3 deals with the distorted kink problem and the controversy 

184: around the existence of a kink shape mode in the Sine-Gordon equation.

185: Conclusions are presented in section 4.

186: 

187: \section{\sl The wobble by the inverse scattering method}

188: 

189: The powerful technique of the Inverse Scattering Transform\ci{lamb},

190: connects between nonlinear equations, such as the Korteweg-deVries (KdV), 

191: Sine-Gordon, nonlinear Schr\"odinger, modified KdV, etc., and linear

192: eigenvalue equations, such as the Schr\"odinger or Dirac-like two-component

193: equations.

194: The nonlinear equations arise as consistency conditions on the linear

195: equations. The potentials of the linear equations yield solutions

196: of the nonlinear equation. 

197: The method uses the scattering data

198: of the linear problem to predict the nonlinear solution by resorting to

199: an integral equation discovered by Gelfand, Levitan and Marchenko\ci{ablowitz}.

200: 

201: In the case of reflectionless potentials, for which there are only

202: transmitted waves in the linear problem, 

203: the Gelfand-Levitan-Marchenko equations

204: are integrable in closed form. The analytical formulae are given by Lamb

205: \ci{lamb}. Segur\ci{segur} implemented these formulae for the case

206: of what we presently call {\sl the wobble}.

207: The Sine-Gordon soliton composed of a kink and a breather, 

208: the {\sl wobble} $u(x,t)$ with its center at rest, is given by

209: 

210: \be\label{wobble}

211: u(x,t)=4~Im(ln(det(I+iM)),

212: \ee

213: 

214: {\noindent}where $\ds I_{i,j}, i,j=1,3$ is the unit matrix , and $\ds M_{i,j}$

215:  is the matrix containing scattering data of the kink and the 

216: breather.

217: 

218: \bea\label{M}

219: M_{j,k}&=&\frac{-i~m_k}{\zeta_j+\zeta_k}~e^{\theta},\nono

220: \theta&=&-i~(\zeta_j+\zeta_k)\frac{x+t}{2}+i~\frac{x-t}{4~\zeta_k},

221: \eea

222: \noindent

223: where $\ds \zeta_1=i/2,~\zeta_2=\alpha+i\beta, \zeta_3=-\alpha+i\beta$, with

224: $\ds \alpha^2+\beta^2=0.25$ and $\ds \beta~>~0$ are the scattering amplitudes:

225: $\ds \zeta_1$ for the

226: {\sl q=1} kink and $\ds \zeta_{2,3}$ for the {\sl q=0} breather. $\ds m_j$ are 

227: the normalization constants for each matrix element, with $\ds m_1$ real and 

228: $\ds m_2^*=m_3$.

229: The wobble depends on the parameters 

230: $\ds m_2=(m_{2R},m_{2I}),~m_1,~\beta$.

231: A moving soliton can be obtained by boosting 

232: with a Lorentz transformation. We here focus on a {\sl wobble} at rest.

233: \\

234: {\indent}Defining the complex function 

235: \bea\label{F}

236: F=det(I+iM)=V+iW,

237: \eea

238: \noindent

239: 

240: {\noindent} we find

241: \footnote{Eq.(\ref{UW}) corrects the results of Segur\ci{segur}}

242: 

243: \bea\label{UW}

244: V(x,t)&=&1+\frac{|m_2|^2~(\quart-\beta^2)}{\beta^2}e^{4\beta x}-

245: \frac{2~m_1|m_2|(\half-\beta)}{\half+\beta}~e^{x+2\beta x}~cos(2\alpha (t+t_0))

246: ,\nono

247: W(x,t)&=&-\frac{m_1~|m_2|^2~(\half-\beta)^3}{\beta^2(\half+\beta)}

248: e^{x+4\beta x}-~m_1~e^x+2~|m_2|~e^{2\beta x}~cos(2\alpha (t+t_0)),

249: \eea

250: {\noindent}where 

251: 

252: \bea\label{cos}

253: \frac{2}{|m_2|} ~tan(2\alpha t_0)=\frac{m_{2I}\beta+m_{2R}\alpha}

254: {m_{2I}\alpha-m_{2R}\beta}.

255: \eea

256: 

257: 

258: The {\sl wobble} is  obtained by inserting eq.(\ref{UW}) 

259: in eq.(\ref{wobble}).

260: \\

261: {\indent}As shown below, the {\sl wobble}

262: oscillates sweeping over values above $\ds 2\pi$.

263: In numerical codes that limit the inverse tangent to the principal

264: branch, it is imperative to use the 

265: complex natural logarithm expression of eq.(\ref{wobble}), 

266: instead of the translation $\sl Imag(ln(F))=tan^{-1}(\frac{W}{V})$.

267: 

268: The {\sl q=1} kink is recovered from the wobble of 

269: eqs.(\ref{wobble},\ref{UW}) by setting $\ds m_2=(0,0)$.

270: The {\sl q=0} breather requires the substitution $m_1=0$.

271: 

272: The normalization constant $\ds m_1$ determines the location of the center

273: of the wobble, $\ds m_2$ fixes the amplitude of the oscillation and the

274: phase. 

275: The angular frequency of the oscillation is $\omega=2~\alpha$, with

276: upper bound $\omega_{max}=1$, at which the the phonon 

277: spectrum begins\ci{rubinstein},\ci{fogel}.

278: 

279: An alternative simpler form of $F= V+ i~W$ is

280:  

281: \bea\label{UW1}

282: V(x,t)&=&2~e^{2\beta x}\bigg(e^{\mu}cosh(2\beta x+\mu)-|m_2|

283: e^{x+\al-\mu}~cos(2\alpha (t+t_0))\bigg),\nono

284: W(x,t)&=&2~e^{2\beta x}\bigg(-m_1~e^{x+\al}~cosh(2 \beta x+\al)

285: +|m_2|~cos(2\alpha (t+t_0))\bigg),

286: \eea

287: \noindent

288: where $\ds 

289: e^{\al}=\frac{|m_2|}{|\beta|}~\sqrt{\frac{(\half-\beta)^3}{\half+\beta}}$,

290:  $\ds e^{\mu}=\frac{|m_2|}{|\beta|}\sqrt{\quart-\beta^2}$.

291: 

292: It is an arduous but straightforward task to show that the wobble obeys the

293: renormalized Sine-Gordon equation.\footnote{We have verified 

294: eq.(\ref{UW1}) using computerized algebra.}

295: 

296: \be\label{sg}

297: \frac{\dd^2 u}{\dd t^2}-\frac{\dd^2 u}{\dd x^2} + sin(u) = 0,

298: \ee

299: 

300: derivable from the renormalized lagrangian

301: 

302: \bea\label{lag}

303: {\cal L}=\int{~dx~\bigg[\bigg(\frac{\dd u}{\dd t}\bigg)^2

304: -\bigg(\frac{\dd u}{\dd x}\bigg)^2+(cos(u)-1)}\bigg],

305: \eea

306: \noindent

307: with renormalized energy 

308: 

309: \bea\label{energy}

310: E~=~8~+32\beta.

311: \eea

312: 

313: \begin{figure}

314: \epsffile{wob1.eps}

315: \caption{\sl Wobble as a function of distance: $m_1=-1,

316:  m_2=(0.6,0.7),\beta=0.2$} 

317: \label{fig1}

318: \end{figure}

319: The remarkable IST method has yielded a new nonlinear 

320: solution of the Sine-Gordon equation built from a {\sl q=1}

321: kink and a {\sl q=0} breather, a nonperturbative oscillatory shape mode of

322: the kink. 

323: 

324: 

325: \begin{figure}

326: \epsffile{error.eps}

327: \caption{\sl $\epsilon$ as a function of distance for the wobble parameters 

328: of figure 1}

329: \label{fig2}

330: \end{figure}

331: 

332: Figure 1 shows snapshots of the wobble at 

333: times comprising one oscillation period. 

334: As eq.(\ref{UW1}) implies, the curves show  

335: oscillations that are not simple harmonic.

336: The insets in figure 1

337: show the wobble at t=0 and and an identical picture after 19 periods.

338: \\

339: The numerical integration code we used is based

340: on the leapfrog method. The calculation used double precision variables with

341: a fixed time step of dt = 0.01 and a flexible space grid.

342: One measure of the integration accuracy is the

343: conservation of the energy. The error in the energy was demanded 

344: never to exceed 0.2\%. The other measure consisted in an exact match 

345: with the analytical formulae of eq.(\ref{UW1}) for long times.

346: Figure 2 shows one such a comparison for the wobble parameters

347: of figure 1 at t=137 amounting to 13700 iterations of the numerical code. 

348: The ordinate is the absolute value of the percentage 

349: relative deviation of the numerical

350: results $\ds u(x,t)_{num}$ from the analytical

351: formula of eq.(\ref{wobble}) $\ds~u(x,t)_{form}$, 

352: $\ds~\epsilon 

353: =100~abs\bigg(\frac{u(x,t)_{num}-u(x,t)_{form}}{u(x,t)_{form}}\bigg)$.

354: The abscissa spans the region where the results are relevant. Below {\sl x=25}

355: the wobble is negligible and the comparison is irrelevant. The errors very 

356: rarely exceed one percent. 

357: \\

358: {\indent}The stability problem of distorted solitons under large perturbations 

359: has not been settled yet, even for the case of the breather\ci{birnir}.

360: The study of wobble stability can be circumscribed to

361: the analysis of its development for $\ds \alpha,\beta$ 

362: violating the unitarity condition $\ds \alpha^2+\beta^2=\quart$. 

363: These parameters belong to the breather sector of the wobble.

364: 

365: Initially we addressed the stability problem by using a limited set

366: of collective coordinates, promoting the parameters to be time dependent.

367: Unfortunately, the method failed to predict the observed behavior.

368: Even the frequency of the unperturbed breather or wobble cannot be recovered

369: by means of the collective coordinate approach. 

370: 

371: We therefore 

372: proceeded to investigate the question of stability of both the breather and 

373: the wobble by means of numerical simulations.

374: We scanned the parameter ranges $\ds -\infty<\alpha

375: <\infty $, $\ds -0.5<\beta<0.5$, omitting the $|\beta| > 0.5$ 

376: region, for which the distorted

377: breather and wobble decay by emission of soliton-antisoliton pairs.

378: 

379: \begin{figure}

380: \epsffile{brea.eps}

381: \caption{\sl Long time behavior of an initially distorted breather: 

382: $m_1=0,m_2=(0.6,0.7),\beta=0.2,\alpha(t=0)=0.858,~\alpha (t\ra\infty ) =0.314$}

383: \label{fig3}

384: \end{figure}

385: 

386: 

387: Figure 3 shows a typical case for the breather, and figure 4 one for the wobble.

388: After a transient that

389: depends on the magnitude of the distortion, both distorted breathers

390: and wobbles eventually settle down at a nearby, lower energy,  

391: stable breather or wobble. 

392: The excess energy is emitted by means of a trail of phonons both 

393: in the forward and backward directions, as seen in the insets of figures 3 

394: and 4.

395: The trail of phonons resembles an Airy function. 

396: The pictures in figures 3 and 4 repeat themselves

397: for longer and longer times.

398: The relaxation time was found to be of the order of 

399: $\ds \tau=\frac{1}{|\alpha(t=0)-\alpha(t\ra\infty )|}$.

400: For $\ds t<\tau$, the parameters

401: of the wobble and the breather change with time. 

402: At around $t=\tau$, $\alpha$ and $\beta$ obey again the unitarity

403: constraint and the wobble and breather do not decay appreciably any more. 

404: It appears that $\alpha\ra\infty$ depends only on the breather dynamics.

405:  

406: \begin{figure}

407: \epsffile{wob2.eps}

408: \caption{\sl Long time behavior of an initially distorted wobble

409: : $m_1=1, m_2=(0.6,0.7),\beta=0.2,\alpha(t=0)=0.858,

410: ~\alpha (t\ra\infty )= 0.314$}

411: \label{fig4}

412: \end{figure}

413: 

414: 

415: \section{\sl Distorted kinks and the Wobble}

416: 

417: The existence of a shape mode in the {\sl q=1}

418: sector of the Sine-Gordon equation has been surrounded by controversy.

419: The results of the previous section show that there are nonperturbative 

420: oscillatory shape modes in the {\sl q=1} kink sector. 

421: The wobble angular frequency 

422: spectrum fills the gap between the zero mode and the phonon spectrum.

423: The wobble is an exact solution, whereas the phonon spectrum 

424: results from approximate linearized solutions

425: of the Sine-Gordon equation around the kink soliton.

426: The  wobble must play a role in the

427: decay dynamics of distorted kinks. 

428: As we will see below, the shape mode found in the literature,

429:  is an intermediate stage on the way from the distorted kink to a stable wobble.

430: 

431: Rice\ci{rice}, and later Boesch and Willis\ci{boesch} proposed

432: the existence of shape modes for distorted kinks in the Sine-Gordon

433: equation. The angular frequency of the oscillations 

434: of the kink soliton width found by Rice\ci{rice} and Boesch and 

435: Willis\ci{boesch} lies above the phonon threshold of

436: $\ds \omega=1$. The shape mode is therefore unstable to decay into phonons.

437: Recently, Quintero et al.\ci{quintero} have argued that such a shape mode 

438: does not exist. Quintero et al.\ci{quintero} claim that 

439: the behavior observed in the work of Boesch and Willis\ci{boesch} is due

440: to discretization effects in the numerical calculation. 

441: They base their hypothesis on a work by Kivshar et al.\ci{kivshar}. 

442: Kivshar et al.\ci{kivshar} predict the birth of shape modes bifurcating 

443: from the continuum spectrum of phonons that plunge below threshold upon the

444: application of perturbations. 

445: For the unperturbed integrable Sine-Gordon 

446: equation the effect disappears in the continuum, when 

447: lattice spacing tends to zero. Kivshar et al.\ci{kivshar} 

448: state that there are no shape modes for integrable equations.

449: The results of the previous section show that there is

450: a whole set of nonperturbative shape modes in the Sine-Gordon equation.

451: 

452: In order to connect the wobble soliton to perturbed kinks, we

453: followed the evolution of a distorted kink 

454: 

455: \bea\label{kink}

456: u(x,t) = 4~tan^{-1}({e^{\gamma x}}),

457: \eea

458: \noindent

459: with $\ds |\gamma -1|$ the distortion parameter.

460: $\gamma (t)$ is extracted from the data by comparing 

461: to eq.(\ref{kink}). The phenomenological function 

462: \bea\label{fitfun}

463: \gamma(t)=1+a~e^{-t^b c}~cos(2\alpha t+\phi),

464: \eea

465: 

466: {\noindent}captures the broad features of $\ds \gamma(t)$.

467: From eq.(\ref{fitfun}) we obtained the angular frequency of the oscillation,

468: $\ds \alpha$. We found that $\ds \alpha$ varies with time.

469: There is a slow drift of $\alpha$ towards lower values.

470: This is depicted in figure 5. The angular frequency $\ds\alpha$ of 

471: eq.(\ref{fitfun}) drops from $\ds 2\alpha = 1.048 $ at around t=200 

472: to $\ds 2\alpha = 1.0015$ later.

473: Both the amplitude of the oscillation and the frequency diminish gradually.

474: The results of figure 5 

475: agree in general with Boesch and Willis\ci{boesch}.

476: However, contrarily to the predictions

477: of the collective coordinate method of Rice\ci{rice}, the frequency is not 

478: constant.

479: 

480: 

481: \begin{figure}

482: \epsffile{l11new.eps}

483: \caption{\sl Distorted kink width $\gamma$ of eq.(\ref{kink})

484: as a function of time. Fitted values of eq.(\ref{fitfun}), full curve; input

485: data, dashed curve}

486: \label{fig5}

487: \end{figure}

488: 

489: 

490: We consider now the the numerical simulations of Quintero et al.\ci{quintero}.

491: Quintero et al.\ci{quintero} use an insufficient extent for the x axis 

492: $\ds L =| x_{max}|= 100$, that does not prevent the reabsorption of reflected 

493: phonons from the boundary. 

494: These phonons pump back energy into the oscillating

495: soliton and blur the picture.

496: The velocity of propagation of the phonons

497: is $\ds v =\frac{k}{\sqrt{k^2+1}}$, asymptotically tending towards

498: $\ds v =1$. Using this asymptotic value, the reflected

499: phonons collide and feed energy back to the soliton at 

500:  $\ds t\approx 200$. In figure 2 of \ci{quintero}, and at approximately 

501: that time, the decaying single frequency shape mode picture starts to 

502: break down.

503: 

504: Increasing the span of the

505: integration region with time prevents the reabsorption

506: of phonons. The slowly drifting single frequency behavior is seen to persist for

507: longer and longer times. Whether the oscillation frequencies cross the threshold

508: of $\omega=1$, signaling 

509: the transition to a stable shape mode lookes unclear from the previous figures.

510: To accelerate the decay process a very distorted kink is needed. 

511: Figure 6 shows a case

512: with $\gamma(t=0)=0.4$, a distortion of 60\% compared to the

513: unperturbed kink. (The distorted kink energy for this $\gamma$

514: is $ E = 11.6$ still below

515: the threshold for the production of a soliton-antisoliton pair at $E =24$).

516: The angular frequency is now $\ds 2\alpha=0.97$ for

517: short times. 

518: A long wavelength modulation of the amplitude is also noticeable

519: in figure 6.

520: After t=600, the amplitude of the oscillation appears to stabilize.

521: 

522: As $\alpha$ was obtained by means of a phenomenological function, 

523: more convincing evidence of the transition to a wobble-like regime

524: is necessary.

525: 

526: \begin{figure}

527: \epsffile{l04.eps}

528: \caption{\sl Distorted kink width 

529: $\gamma$ of eq.{kink}

530: as a function of time, full curve, fitted values, dashed curve, numerical input}

531: \label{fig6}

532: \end{figure}

533: A distorted kink cannot be put in exact correspondence with the

534: wobble, despite the similarities.

535: The energy of eq.(\ref{energy}) teaches us that $\ds \beta$ is 

536: the relevant parameter for the breather admixture to the kink.

537: Expanding the expression for the wobble of eq.(\ref{wobble}) around 

538: $\ds \beta = 0$ we find $\ds \gamma (t=0)\approx1-8\beta$.

539: The factor of 8, and the unitarity constraint that fixes the angular

540: frequency to be $\ds \alpha=\sqrt{\quart-\beta^2}$ require   

541: an extremely distorted kink in order to reach a fairly visible 

542: frequency below threshold. A stronger distortion than $\ds \gamma(t=0)=0.4$ 

543: as compared to that depicted in figure 6 is necessary.

544: 

545: We therefore considered initial distortions with $\ds \gamma(t=0) < 0.4$.

546: For such a large intial distortion, the decaying kink profile  

547: did not match eq.(\ref{kink}) any more. The extraction

548: of a clean distortion parameter $\ds \gamma(t)$ became impossible. 

549: The time evolution of the highly distorted kinks leads to

550:  a completely different object.

551: Figure 7 shows distorted kinks for $\ds \gamma (t=0)=0.3$ 

552: at around $ \ds t=290$. 

553: The profiles resemble very much the wobbles figure 1.

554: 

555: From the graphs one can read off the value of the angular frequency

556: of the oscillation to be $\ds 2 \alpha\approx 0.92$, well below the phonon

557: threshold.

558: The inset shows phonons receding from the center, 

559: others propagating forwards, are not shown. We performed

560: long time numerical integrations up to t=1000 and did not see any

561: sizeable decay of the amplitude of the wobble-like kink.

562: 

563: The kink appears to be decaying to a wobble. If this is indeed the case, 

564: we should be able to identify the wobble parameters of the decaying 

565: distorted kink.

566: \begin{figure}

567: \epsffile{l03.eps}

568: \caption{\sl Distorted kinks for $\gamma(t=0)=0.3$,

569: as a function of distance, at various times, showing oscillatory behavior}

570: \label{fig7}

571: \end{figure}

572: 

573: To limit the number of free parameters, we first 

574: considered the $\ds x\ra\infty$ region for both the distorted kink

575: and the wobble. The wobble has an asymptotic behavior

576: of $\ds e^{2|\beta |x}$, whereas for the distorted kink

577: it is $\ds e^{\gamma x}$. If the asymptotic

578: behavior does not change with time, we have $\ds \gamma \approx 2\beta$.

579: From this value of $\ds \beta$ and the unitarity constraint, 

580: $\alpha$ is determined. This $\ds \alpha=\frac{\pi}{T}$ with {\sl T} the

581: oscillation period, can be readily compared to the numerical results

582: of figure 7. The agreement is fair, but not satisfactory.

583: Following the reverse path seemed more appropriate.

584: The oscillation period of the numerical data 

585: fixes $\ds \alpha$ and consequently $\ds \beta$ by means of

586: the unitarity constraint.

587: There remained three unknown parameters $\ds m_1,~m_{2R},~m_{2I}$ that

588: were determined using a minimization algorithm.

589: \begin{figure}

590: \epsffile{l03fit.eps}

591: \caption{\sl Distorted kinks at t=291.4, t=294.8 for $\gamma (t=0)=0.3$,

592: as a function of distance. Numerical data of figure 7,broken line; 

593: wobble fit, full line.}

594: \label{fig8}

595: \end{figure}

596: 

597: The results are depicted in figure 8.

598: All the distorted kinks of figure 7 were reproduced with the same parameter set.

599: The agreement with the data is remarkable, especially so in light 

600: of the highly nonlinear wobble function.

601: Long distance discrepancies are due to phonon contributions.

602: 

603: The highly distorted kink has transformed 

604: into a wobble and a wake of phonons.

605: The less distorted cases presumably need a much longer time

606: to reach a wobble. 

607: For small distortions of the kink, 

608: it is hard to discern a clear wobble shape.

609: However, we cannot rule out completely the possibility

610: of a distorted kink decaying to an undistorted kink and phonons.

611: In future work we plan to address this and other related problems.

612: 

613: 

614: \section{\sl Conclusions}

615: 

616: We have shown that there exists a set of wobbling, apparently stable, 

617: nonperturbative solutions to the Sine-Gordon equation in the {\sl q=1} sector. 

618: Highly distorted kinks eventually decay to wobbles and phonons.

619: The results are relevant to the investigation of scattering events

620: of Sine-Gordon kinks from impurities.

621: Other nonlinear equations that support breathers, such as the modified KdV 

622: equation, may bear wobble solutions too.

623: 

624: The existence of the wobble may have technological implications.

625: Wobbles produced in Josephson junctions could carry

626: analogical information with relative stability. 

627: 

628: 

629: \newpage

630: \begin{thebibliography}{99}

631: \bi{eisenhart}L.P. Eisenhart, {\sl A Treatise on the Differential Geometry

632: of Curves and Surfaces}, Dover, New York (1960).

633: \bi{lamb}G. L. Lamb, Jr., {\sl Elements of Soliton theory} Wiley Interscience,

634: New York (1980).

635: \bi{rem} M. Remoissenet, {\sl Waves called solitons : concepts and experiments}

636: , Springer Verlag, Berlin (1996)

637: \bi{zabusky}N. J. Zabusky and M. D. Kruskal, Phys. Rev. 

638: Lett.{\bf 15}, 240(1965)

639: \bi{rajaraman}R. Rajaraman {\sl Solitons and Instantons } North Holland (1987).

640: \bi{rice}M. J. Rice, Phys. Rev. {\bf B 28}, 3587 (1983).

641: \bi{boesch} R. Boesch and C. R. Willis, Phys. Rev. {\bf B42}, 2290 (1989)

642: \bi{fogel} M. B. Fogel, S. E. Trullinger, A. R. Bishop and J. A. Krumhansl,

643: Phys. Rev. {\bf B15}, 1578 (1977).

644: \bi{quintero} N. R. Quintero, A. Sanchez and F. G. Mertens, Phys. Rev. {\bf E62},

645: R60 (2000).

646: \bi{kivshar} Y. S. Kivshar, D. E. Pelinovsky, T. Cretegny and M. Peyrard, Phys.

647: Rev. Lett. {\bf 80}, 5032 (1997).

648: \bi{segur} H. Segur, J. Math. Phys. {\bf 24}, 1439 (1983).

649: \bi{ablowitz}M. J. Ablowitz and H. Segur {\sl Solitons and the Inverse 

650: Scattering Transform}, SIAM, Philadelphia (1981).

651: \bi{rubinstein} J. Rubinstein, J. Math. Phys. {\bf 11}, 258 (1969).

652: \bi{birnir} B. Birnir, H. P. McKean and A. Weinstein, Comm. Pure and Appl. Math.

653: {\bf 47}, 1043 (1994).

654: \end{thebibliography}

655: \end{document}

656: