Nonexistence of small, odd breathers for a class of nonlinear wave equations
Extracted source text ยท arXiv 1607.06421
\documentclass[10pt]{amsart}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{graphicx}
\usepackage{amsthm,graphicx,color,yfonts}
\usepackage{pdfsync}
\usepackage{epstopdf}%
\usepackage{pifont}
\usepackage{bbm}
%\usepackage{showkeys}
%\usepackage{hyperref}
%\usepackage{amsmath,amssymb,amsthm,amsfonts,graphicx,color,yfonts}
\usepackage[colorlinks=true]{hyperref}
\hypersetup{linkcolor=red,citecolor=blue,filecolor=dullmagenta,urlcolor=darkblue} % coloured links
%\hypersetup{linkcolor=black,citecolor=black,filecolor=black,urlcolor=black} % black links, for printed output
\addtolength{\hoffset}{-2.5cm}
\addtolength{\textwidth}{4cm}
\addtolength{\voffset}{-1cm}
\addtolength{\textheight}{1.5cm}
\setcounter{MaxMatrixCols}{30}
%TCIDATA{OutputFilter=latex2.dll}
%TCIDATA{Version=5.50.0.2960}
%TCIDATA{CSTFile=40 LaTeX article.cst}
%TCIDATA{Created=Friday, October 08, 2010 08:02:03}
%TCIDATA{LastRevised=Sunday, March 06, 2011 19:46:59}
%TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
%TCIDATA{<META NAME="SaveForMode" CONTENT="1">}
%TCIDATA{BibliographyScheme=Manual}
%TCIDATA{<META NAME="DocumentShell" CONTENT="Standard LaTeX\Blank - Standard LaTeX Article">}
%TCIDATA{Language=American English}
%TCIDATA{ComputeDefs=
%$\phi\left( v_{u_{1}},h_{u_{1}}\right) -\phi\left( v_{u_{2}},h_{u_{2}%
%}\right) $
%}
%BeginMSIPreambleData
%\providecommand{\U}[1]{\protect\rule{.1in}{.1in}}
%EndMSIPreambleData
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%\newcommand{\YM}{\color{magenta}}
\newcommand{\YM}{\color{black}}
\newcommand{\D}{\displaystyle}
\newcommand{\ttt}{\tilde }
\newcommand{\wt}{\widetilde}
\newcommand{\LL}{{\tt L} }
\newcommand{\LLL}{{\mathcal L} }
\newcommand{\nn}{ {\nabla} }
\newcommand{\inn}{ \quad\hbox{in } }
\newcommand{\onn}{ \quad\hbox{on } }
\newcommand{\ass}{ \quad\hbox{as } }
\newcommand{\bz}{ {\bar z } }
\newcommand{\pp}{ {\partial} }
\newcommand{\vv}{ \mathbf{v} }
\newcommand{\ww}{{\tt w} }
\newcommand{\www}{{\bf w} }
\newcommand{\pff}{{\bf Proof.} }
\newcommand{\A}{\mathcal{A}}
\newcommand{\vp}{\varphi}
\newcommand{\B}{\beta }
\newcommand{\rR}{{\mathcal R}}
\newcommand{\SSS}{{\mathcal S}}
\newcommand{\RR}{{{\mathbb R}}}
\newcommand{\RRN}{{{\mathbb R}^3}}
\newcommand{\NN}{ \mathbb N}
\newcommand{\NNN}{ \mathbf N}
\newcommand{\C}{{\mathbb C}}
\newcommand{\R} {\mathbb R}
\newcommand{\cuad}{{\sqcap\kern-.68em\sqcup}}
\newcommand{\abs}[1]{\mid #1 \mid}
\newcommand{\DD}{{\mathcal D}}
\newcommand{\dist}{{\rm dist}\, }
\newcommand{\foral}{\quad\mbox{for all}\quad}
\newcommand{\ve}{\varepsilon}
\newcommand{\ts}{{\tilde s}}
\newcommand{\be}{\begin{equation}}
\newcommand{\ee}{\end{equation}}
\newcommand{\Lam}{\Lambda}
\newcommand{\la}{\lambda}
\newcommand{\ptl}{{\partial}}
\newcommand{\Om}{{\Omega}}
\newcommand{\uu }{{\bf u}}
\newcommand{\equ}[1]{(\ref{#1})}
\newcommand{\dx}{\; dx}
\newcommand{\dd}[2]{\frac{d#1}{d#2}}
\newcommand{\cal}{\mathcal}
\newcommand{\st}{{\tt t}}
\newcommand{\sy}{{\tt y}}
\newcommand{\pD}{{\mathring D_\tau}}
\newcommand{\I}{{I_{\delta_0/\ve}}}
\definecolor{darkgreen}{rgb}{0.2,0.7,0.1}
\newcommand{\green}{\color{blue}}
\newcommand{\blue}{\color{blue}}
\newcommand{\sech}{\mathop{\mbox{\normalfont sech}}\nolimits}
\newcommand{\ver}{\varsigma}
\newcommand{\const}{\mathrm{const}}
\newcommand{\sgn}{\mathrm{sgn}}
\renewcommand{\Re}{\mathrm{Re}}
\renewcommand{\Im}{\mathrm{Im}}
\newcommand{\re}{\operatorname{Re}}
\newcommand{\ima}{\operatorname{Im}}
\newcommand{\interior}{\mathrm{int}}
\newcommand{\meas}{\operatorname{meas}}
\newcommand{\N}{\mathbb{N}}
\newcommand{\Z}{\mathbb{Z}}
\newcommand{\T}{\mathbb{T}}
\newcommand{\Q}{\mathbb{Q}}
\newcommand{\Com}{\mathbb{C}}
\newcommand{\pd}{\partial}
\newcommand{\al}{\alpha}
\newcommand{\bt}{\beta}
\newcommand{\ga}{\gamma}
\newcommand{\de}{\delta}
\newcommand{\te}{\theta}
\newcommand{\arctanh}{\operatorname{arctanh}}
%\newcommand{\tanh}{\operatorname{tanh}}
\newcommand{\spawn}{\operatorname{span}}
\newcommand{\diam}{\operatorname{diam}}
\newcommand{\supp}{\operatorname{supp}}
\newcommand{\pv}{\operatorname{pv}}
\def\bm{\left( \begin{array}{cc}}
\def\endm{\end{array}\right)}
% \newcommand{\Dom}{\operatorname{Dom}}
% \newcommand{\Rec}{\operatorname{Rec}}
% \newcommand{\Int}{\operatorname{Int}}
% \newcommand{\Tr}{\operatorname{tr}}
%\left[ \left( \newcommand{\meas}{\operatorname{meas}}
% \newcommand{\vol}{\operatorname{vol}}
\providecommand{\abs}[1]{\left|#1 \right|}
\providecommand{\norm}[1]{\left\| #1 \right\|}
\newcommand{\sublim}{\operatornamewithlimits{\longrightarrow}}
% \newcommand{\pot}[1]{\emph{Proof of Theorem #1. }}
% \newcommand{\notac}{{\sf Notation}:\ }
% \newcommand{\rem}{{\sf Remark}:\ }
\newcommand{\ex}{{\bf Example}:\ }
% \newcommand{\obser}{{\sf Observation}:\ }
\newcommand{\fin}{\hfill$\blacksquare$\vspace{1ex}}
\newcommand{\ba}{\begin{equation*}}
\newcommand{\ea}{\begin{equation*}}
\newcommand{\bea}{\begin{eqnarray}}
\newcommand{\eea}{\end{eqnarray}}
\newcommand{\bee}{\begin{eqnarray*}}
\newcommand{\eee}{\end{eqnarray*}}
\newcommand{\ben}{\begin{enumerate}}
\newcommand{\een}{\end{enumerate}}
\newcommand{\nonu}{\nonumber}
\setlength\arraycolsep{3pt}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%\newtheorem{theorem}{Theorem}[section]
%\newtheorem{acknowledgement}[theorem]{Acknowledgement}
%\newtheorem{algorithm}[theorem]{Algorithm}
%\newtheorem{axiom}[theorem]{Axiom}
%\newtheorem{case}[theorem]{Case}
%\newtheorem{claim}[theorem]{Claim}
%\newtheorem{conclusion}[theorem]{Conclusion}
%\newtheorem{condition}[theorem]{Condition}
%\newtheorem{conjecture}[theorem]{Conjecture}
%\newtheorem{corollary}[theorem]{Corollary}
%\newtheorem{criterion}[theorem]{Criterion}
%\newtheorem{definition}[theorem]{Definition}
%\newtheorem{example}[theorem]{Example}
%\newtheorem{exercise}[theorem]{Exercise}
%\newtheorem{lemma}[theorem]{Lemma}
%\newtheorem{notation}[theorem]{Notation}
%\newtheorem{problem}[theorem]{Problem}
%\newtheorem{proposition}[theorem]{Proposition}
%\newtheorem{remark}[theorem]{Remark}
%\newtheorem{solution}[theorem]{Solution}
%\newtheorem{summary}[theorem]{Summary}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\long\def\red#1{{\color{red}#1}}
\def\mm{m}
\numberwithin{equation}{section}
\newtheorem{theorem}{Theorem}[section]
\newtheorem*{theorem*}{Theorem}
\newtheorem{proposition}{Proposition}[section]
\newtheorem{corollary}{Corollary}[section]
\newtheorem{lemma}{Lemma}[section]
\newtheorem{definition}{Definition}[section]
\newtheorem{example}{Example}[section]
\newtheorem{problem}{Problem}[section]
\newtheorem{claim}{Claim}[section]
\theoremstyle{remark}
\newtheorem{remark}{Remark}[section]
%\newcommand{\remark}{{\bf Remark.}}
%%%%
\title[Nonexistence of small, odd breathers]{Nonexistence of small, odd breathers for a class of nonlinear wave equations}
\author{Micha{\l} Kowalczyk}
\address{Departamento de Ingenier\'{\i}a Matem\'atica and Centro
de Modelamiento Matem\'atico (UMI 2807 CNRS), Universidad de Chile, Casilla
170 Correo 3, Santiago, Chile.}
\email {kowalczy@dim.uchile.cl}
\thanks{M. Kowalczyk was partially supported by Chilean research grants FONDECYT 1130126, Fondo Basal CMM-Chile.}
\author{Yvan Martel}
\address{CMLS, \'Ecole polytechnique, CNRS, Universit\'e Paris Saclay, 91128 Palaiseau Cedex, France}
\email{yvan.martel@polytechnique.edu}
\author{Claudio Mu\~noz}
\address{CNRS and Departamento de Ingenier\'{\i}a Matem\'atica and Centro
de Modelamiento Matem\'atico (UMI 2807 CNRS), Universidad de Chile, Casilla
170 Correo 3, Santiago, Chile.}
\email{claudio.munoz@math.u-psud.fr, cmunoz@dim.uchile.cl}
\thanks{C. Mu\~noz was partly funded by Chilean research grants FONDECYT 1150202, Fondo Basal CMM-Chile, and Millennium
Nucleus Center for Analysis of PDE NC130017}
\thanks{Part of this work was done while the authors were hosted at the IHES during the program \emph{Nonlinear Waves 2016}. Their stay was partly funded by the grant ERC 291214 BLOWDISOL}
\renewcommand{\subjclassname}{\textup{2000} Mathematics Subject Classification}
\subjclass{35J61}
%\newenvironment{proof}[1][Proof]{\noindent\textbf{#1.} }{\ \rule{0.5em}{0.5em}}
\begin{document}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{abstract}
In this note, we show that for a large class of nonlinear wave equations with {\it odd} nonlinearities, any globally defined odd solution which is small in the energy space decays to $0$ in the local energy norm. In particular, this result shows nonexistence of small, odd breathers for some classical nonlinear Klein Gordon equations such as the sine Gordon equation and $\phi^4$ and $\phi^6$ models. It also partially answers a question of Soffer and Weinstein in \cite[p. 19]{MR1681113} about nonexistence of breathers for the cubic NLKG in dimension one.
\end{abstract}
\maketitle
\section{Introduction}
{\YM\subsection{Main result}
In this note, we consider the following class of nonlinear wave equations,
\begin{equation}\label{wave ac}
\partial_t^2 u - \partial_x^2 u = \mm u +f(u), \quad (t,x)\in \RR\times\RR,
\end{equation}
where $\mm\in \R$ and the nonlinearity $f\colon \R\to \R$ is a $C^1$, \emph{odd} function such that for some $p>1$,
\[
|f'(u)|\lesssim |u|^{p-1}, \qquad \forall |u|<1.
\]
Denote $F(u)=\int_0^u f$.
An important property of (\ref{wave ac}) is the conservation of \emph{energy}
\be\label{Energy0}
E(u,\partial_t u):=\int \frac{1}{2}\partial_t u^2+\frac{1}{2}\partial_x u^2 - \frac{\mm}2 u^2- F(u),
\ee
along the flow. In particular, $H^1\times L^2$ perturbations of the zero solution are referred as \emph{perturbations in the energy space}. For many standard examples, the Cauchy problem for the model \eqref{wave ac} is globally well-posed
for initial data
$(u(0),\partial_t u(0)) = (u_1^{in},u_2^{in}) \in H^1\times L^2$ small enough.
Here, we do not make any {\it a priori} assumption on the sign of $\mm$ nor the nature of $F(u)$ that would guarantee that $E(u)$ is coercive and controls the $H^1\times L^2$ norm of some solutions. Instead, for general nonlinearity $f$ as above, we consider global solutions whose energy norm is uniformly small in time.
Another important property of \eqref{wave ac} is that for odd initial data the associated solution is also odd for all time. For the rest of this paper, we work in such framework, and we consider only odd perturbations in the energy space. Note that \eqref{wave ac} is invariant under space translation and under the Lorentz transformation but since we consider only odd solutions, these invariances are irrelevant here.
Set
\be\label{eqvarphi0}
u_1 =u ,\quad u_2= \partial_t u ,
\ee
so that in terms of $(u_1, u_2)$, equation (\ref{wave ac}) becomes
\begin{equation}\label{eqvarphi}\left\{\begin{aligned}
& \partial_t u_1 = u_2 \\ & \partial_t u_2 = \partial_{x}^2 u_1 +\mm u_1 + f(u_1). \end{aligned}\right.\end{equation}
Our main result is the following property for small, odd solutions of (\ref{eqvarphi}).
\begin{theorem}\label{TH1}
There exists $\ve>0$ such that any odd global $H^1\times L^2$ solution $(u_1, u_2)$ of (\ref{eqvarphi}) such that \[
\sup_{t\geq 0}\|(u_1(t),u_2(t))\|_{H^1\times L^2}< \ve,
\]
satisfies
\be\label{Conclusion_0}
\lim_{t \to +\infty} \|(u_1(t),u_2(t))\|_{H^1(I)\times L^2(I)} =0,
\ee
for any bounded interval $I\subset \RR$.
In particular, no odd, small breather solutions exist for (\ref{wave ac}).
\end{theorem}
By \emph{breather}, we mean a solution in $H^1\times L^2$ which is periodic in time, up to the symmetries of the equation. Several integrable equations have \emph{stable} breather solutions, including the sine-Gordon model \cite{Lamb}
\be\label{SG}
\phi_{tt}-\phi_{xx}=-\sin \phi.
\ee
See also \cite{AM,AM1,AM2} and references therein for more details on the literature.
%\texttt{May be another older reference, like a standard book ?}
\medskip
Our paper was motivated by (and gives a partial answer to) a conjecture of Soffer and Weinstein \cite{MR1681113} which says that no small amplitude breathers should exist for
\[
\phi_{tt}-\phi_{xx}=-\phi+\phi^3.
\]
Our assumptions allows other classical examples such as the $\phi^4$ model \cite{MR2282481,MR2318156,MR1402248}
\[
\phi_{tt}-\phi_{xx}-\phi+\phi^3=0,
\]
the $\phi^6$ model \cite{Lohe},
\[
\phi_{tt}-\phi_{xx}=-\phi+4\phi^3-3\phi^5,
\]
and the sine-Gordon equation \eqref{SG}.
\medskip
Let us discuss the sine-Gordon model \eqref{SG} in more details.
The energy of the sine Gordon equation is
\[
E(\phi, \phi_t)=\int \frac{1}{2} \phi_t^2+\frac{1}{2} \phi_x^2 +(1-\cos \phi).
\]
Recall that an explicit family of breathers is known \cite{Lamb}
\[
B_{\bt}(t,x)= 4\arctan \Big( \frac{\bt}{\al} \frac{\cos(\al t)}{ \cosh(\bt x)}\Big), \qquad \al^2 +\bt^2 =1.
\]
The function $B_{\bt}$ has arbitrarily small energy norm provided $|\bt|$ is small.
Such breathers oscillate around the constant, finite energy state $u=0$, which is linearly stable.
Note that
for all $t$, $B_{\bt}(t)$ is an even function, which proves in some sense the necessity of the oddness assumption to obtain a general result of asymptotic stability of the zero solution such as Theorem~\ref{TH1}.
\medskip
At a more technical level, the oddness assumption allows to avoid an {\it even} resonance of the linear operator, supporting the belief that the existence of breathers is partly related to the presence of resonances.
However, the case of the $\phi^4$ model suggests that the situation is more subtle in general. For this model, in the presence of a resonance, a formal argument due to Kruskal and Segur \cite{kruskal_segur} excludes the existence of breathers of small amplitude oscillating around $\pm 1$, which are linearly stable, constant states. As we mentioned above Theorem~\ref{TH1} implies nonexistence of odd breathers of the $\phi^4$ model oscillating about $0$, however our result does not apply for the case considered in \cite{kruskal_segur} since the nonlinearity written for $u=\phi\pm 1$ is not odd.
\medskip
For the proof of Theorem \ref{TH1}, we follow a simplified version of the method developed in our recent work \cite{KMM} on asymptotic stability of the kink under odd perturbations for the $\phi^4$ model.
The main idea in \cite{KMM} and in this paper is the introduction of a generalized \emph{virial identity}, suitably constructed for each considered problem.
Such approach is inspired by previous works by the second author and Merle \cite{MR1753061,zbMATH01631995,MM3} for the generalized Korteweg-de Vries equations and by Merle and Rapha\"el \cite{MR2150386} for nonlinear Schr\"odinger equations. Unlike in those works, here the analysis is performed for solutions in the neighborhood of $0$ and the spectral analysis reduces to classical operators.
\subsection{Previous results}
Rigorous proofs of nonexistence of breathers solutions in scalar field equations date back to the work by Coron \cite{Coron}. He showed that under the assumption of vanishing energy and $L^\infty$ norm of the solution at infinity in space, and an $L^1_{tx}$ integrability condition, breathers cannot have arbitrarily small periods. Later, Vuillermot \cite{Vuillermot} showed that under some growth and convexity assumptions on the nonlinearity, there are no breather solutions for suitable scalar field equations. Some formal arguments for existence and nonexistence of breather solutions can be found in Kichenassamy \cite{Kich}, and Birnir-McKean-Weinstein \cite{Birnir}, respectively. Finally, Denzler \cite{denzler} provided a proof of nonexistence of breather solutions for nonlinearities that are small, complex-analytic perturbations of the sine-Gordon case. From his result, it is concluded that the existence of breathers is a very rare property, probably related to the integrability of the equation.
\medskip
As in the present paper, the nonexistence of breather solutions for \eqref{wave ac} can also be seen as a consequence of the asymptotic stability of the vacuum solution in some topology. From this point of view, we refer to the original works of Buslaev and Perelman \cite{bus_per1,bus_per2}, and Buslaev-Sulem \cite{buslaev_sulem} on the NLS case, and Soffer-Weinstein \cite{MR1681113} on nonlinear Klein-Gordon models in 3D. See also the works by Delort \cite{Delort,Delort_fourier}, Lindblad-Soffer \cite{LS1,LS2,LS3}, Bambusi and Cuccagna \cite{Bam_Cucc} and Sterbenz \cite{Ste} on decay of small solutions for nonlinear Klein-Gordon equations in one dimension. This method has been pushed forward by considering different nonlinearities and lower dimensions. For recent results in this area, see e.g. \cite{MR1664792,Tsai_Yau, Rod_Soffer_Schlag, MR2373326,cuccagna_3, cuccagna_4, MR2835867,Kr_Sch, Bet_Gra_Smets1,Lin_Tao}.
}
\section{Proof of the theorem}\label{VIRIAL}
\medskip
\noindent
{\underline{Step 1: Virial identity.}}
Recall the set of coupled equations \eqref{eqvarphi}. For a smooth and bounded function $\psi$ to be chosen later, let
\begin{equation}\label{defI}
\mathcal I (u) := \int \psi (\partial_x u_1) u_2 + \frac 12 \int \psi' u_1 u_2 = \int \left(\psi \partial_x u_1 + \frac 12 \psi' u_1\right) u_2.
\end{equation}
Let $(u_1,u_2)$ be a solution of (\ref{eqvarphi}). {\YM Since for any $v\in H^1$, $\int (\left(\psi v_x+\frac{1}{2} \psi' v\right) v=0$, one computes}
\begin{equation}\label{Ione}
\begin{aligned}
\frac{d}{dt} \mathcal I (u) & = \int \left(\psi u_{1,x}+\frac{1}{2} \psi' u_1\right) u_{1,xx}
+ \int \left(\psi u_{1,x}+\frac{1}{2} \psi' u_1\right) f(u) \\
& =-\mathcal B(u_1)-\int \psi'\left[ F(u_1)-\frac{1}{2} u_1 f(u_1)\right],
\end{aligned}
\end{equation}
where we have denoted
\[
\mathcal B(u_1)=-\int \left(\psi u_{1,x}+\frac{1}{2} \psi' u_1\right) u_{1,xx}=\int \psi' (\partial_x u_1)^2 -\frac{1}{4}\int\psi'''u_1^2.
\]
Therefore,
\begin{equation}
\label{moondawn}
-\frac d{dt} {\mathcal I} =\mathcal B(u_1) + \int \psi'\left[ F(u_1)-\frac{1}{2} u_1 f(u_1)\right].
\end{equation}
\medskip
\noindent
{\underline{Step 2: Coercivity of the bilinear form $\mathcal B$.}}
Now we choose a specific function $\psi$ and we consider the question of the coercivity of the bilinear form $\mathcal B$.
Let $\lambda>0$ be fixed. We set
\begin{equation}\label{defpsi}
\psi(x) : = \lambda\tanh\left(\frac{x}{\lambda}\right),
\end{equation}
in the definition of $\mathcal I$.
Note that $\zeta>0$ everywhere. Let $w$ be the following auxiliary function
\be\label{def w}
w:=\zeta u_1;
\quad \zeta(x) :=\sqrt{\psi'(x)}=\sech \left(\frac{x}{\lambda}\right).
\ee
First, note that by integration by parts,
\begin{align*}
\int w_x^2
&= \int (\zeta \partial_x u_1 +\zeta' u_1 )^2
= \int \psi' (\partial_x u_1)^2 + 2 \int \zeta \zeta' u_1 (\partial_x v_1) + \int (\zeta')^2 u_1^2 \\
& = \int \psi' (\partial_x u_1)^2 - \int \zeta \zeta'' u_1^2 \\
& = \int \psi' (\partial_x u_1)^2 - \int\frac{\zeta''}{\zeta} w^2.
\end{align*}
Thus,
\begin{equation}\label{idgrad}
\int \psi' (\partial_x u_1)^2 = \int w_x^2 + \int\frac{\zeta''}{\zeta} w^2.
\end{equation}
Second,
$$
\int\psi'''u_1^2 = \int \frac{(\zeta^2)''}{\zeta^2} w^2
= 2 \int \left(\frac {\zeta''}{\zeta}+ \frac{(\zeta')^2}{\zeta^2} \right) w^2.
$$
Therefore,
\begin{equation}\label{change}\begin{aligned}
\mathcal B(u_1) &= \int \psi' (\partial_x u_1)^2 -\frac{1}{4}\int\psi'''u_1^2\\
& = \int w_x^2 + \frac 12 \int \left( \frac {\zeta''}{\zeta} -\frac{(\zeta')^2}{\zeta^2} \right) w^2.
\end{aligned}\end{equation}
Set
\[
{\mathcal B}^\sharp(w):= \int \left(w_x^2 - V w^2\right),\quad \hbox{where} \quad
V:= - \frac 12 \left( \frac {\zeta''}{\zeta} -\frac{(\zeta')^2}{\zeta^2} \right),
\]
so that
\be\label{B sharp}
{\mathcal B}^\sharp(w)=\mathcal B(u_1).
\ee
Note that by \eqref{defpsi} and direct computations,
\be\label{V}
V(x)=
\frac{1}{2\lambda^2}\sech^2\left(\frac{x}{\lambda}\right).
\ee
{\YM
Recall that the index of the operator
$
-\frac{d^2}{dx^2}-V
$
associated to $\mathcal B^\sharp$ is $1$, which means that this operator has only one negative, discrete eigenvalue with an even corresponding eigenfunction. This follows from the well known fact (see e.g. Titchmarsh \cite[\S 4.19]{Tit} and \cite[p. 55]{Goldman}) that for any $\lambda>0$, the index $\kappa$ of the operator
\[
-\frac{d^2}{dx^2}-\frac{V_0}{\lambda^2}\sech^2\left(\frac{x}{\lambda}\right), \quad V_0\in [0,+\infty),
\]
is the largest integer such that
\[
\kappa<\frac{1}{2}\sqrt{4 V_0+1}+\frac{1}{2}.
\]
Here, our claim concerning $\mathcal B^\sharp$ follows from
$
1< \frac{\sqrt{3}+1}{2}<2.
$
In fact, we claim the following more precise result.
\begin{lemma}\label{le:posVir}
For any $\lambda>0$ and for any odd function $w\in H^1$,
\begin{equation}\label{posVir1}
{\mathcal B}^\sharp(w)
\geq \frac{3}{4} \int w_x^2.
\end{equation}
\end{lemma}
\begin{proof}
We write
\begin{equation}
\mathcal B^\sharp(w)=\frac 34 \int w_x^2+\frac 14 \int \left(w_x^2 - \frac{2}{\lambda^2}\sech^2\left(\frac{x}{\lambda}\right) w^2\right).
\label{decomp pos Vir}
\end{equation}
Since $ \frac{1}{2}\sqrt{4.2+1}+\frac{1}{2}=2$,
the second term in the right-hand side above is nonnegative for any odd function $w$, and the result follows.
\end{proof}}
\medskip
\noindent
{\underline{Step 3: Control of the error terms and conclusion of the Virial argument.}}
{\YM As in statement of Theorem \ref{TH1}, we consider an odd solution $(u_1(t),u_2(t))$ of \eqref{eqvarphi} global in $H^1\times L^2$ for $t\geq 0$ and satisfying, for all $t\geq 0$,
\begin{equation}\label{smallphi}
\|(u_1(t),u_2(t))\|_{H^1\times L^2} < \varepsilon.
\end{equation}
We define}
\begin{equation}\label{nloc}
\|u_1\|_{H^1_{\omega}}^2 :=
\int \left( |\partial_x u_1|^2 + u_1^2 \right) \sech\left(x\right),\quad
\|u_2\|_{L^2_{\omega}}^2 :=
\int u_2^2 \sech\left(x\right),
\end{equation}
and
\begin{equation}\label{nloc2}
\|(u_1,u_2)\|_{H^1_{\omega}\times L^2_{\omega}}^2 :=\|u_1\|_{H^1_{\omega}}^2 + \|u_2\|_{L^2_{\omega}}^2.
\end{equation}
The key ingredient of the proof of asymptotic stability in the energy space is the following result.
\begin{proposition}\label{pr:11}
For $\varepsilon>0$ small enough,
\begin{equation}\label{11un}
\int_{0}^{+\infty} \|(u_1(t),u_2(t))\|_{H^1_{\omega}\times L^2_{\omega}}^2 dt
\lesssim \varepsilon^2 .
\end{equation}
\end{proposition}
\begin{proof}[Proof of Proposition \ref{pr:11}]
We consider the virial-type quantity $\mathcal I(t)$ defined in \eqref{defI} with $\lambda=100$.
\medskip
The proof of \eqref{11un} is based on the following two estimates, which hold for some constant $C_1, C_2>0$:
\begin{align}
-\frac d{dt} \mathcal I
& \geq C_1 \|u_1\|_{H^1_{\omega}}^2 ,
\label{10un} \\
\frac d{dt} \int \sech\left(x\right) u_1 u_2
&\geq \|u_2\|_{L^2_{\omega}}^2
-C_2 \|u_1\|_{H^1_{\omega}}^2 .
\label{10trois}
\end{align}
\medskip
{\YM
First we prove \eqref{11un} assuming \eqref{10un} and \eqref{10trois}.
Integrating \eqref{10un} on $[0,t_0]$, using the bound \eqref{smallphi}, and passing to the limit as $t_0\to +\infty$, we find
\begin{equation}\label{11unbis}
\int_{0}^{+\infty} \|u_1(t)\|_{H^1_{\omega}}^2 dt
\lesssim \varepsilon^2 .
\end{equation}
Then, using
\eqref{10trois} similarly, we obtain \eqref{11un}.}
\medskip
Thus, to finish the proof, we only have to prove \eqref{10un} and \eqref{10trois}. We begin with the proof of \eqref{10un}. We have from \eqref{moondawn} and \eqref{B sharp},
{\YM
\[
-\frac d{dt} {\mathcal I} = \mathcal B (u_1) - \int \psi'\left[ F(u_1)-\frac{1}{2} u_1 f(u_1)\right] .
\]
Recall from Section \ref{VIRIAL} the notation $w = u_1 \zeta=u_1 \sech\left(\frac x{100}\right)$ (see \eqref{def w}).
From Lemma \ref{le:posVir}, we have
\[
\mathcal B (u_1)= \mathcal B^\sharp(w) \geq \frac 34 \int w_x^2 \quad \text{and equivalently}\quad
\int w_x^2 \geq \frac {1}{5.10^3} \int \sech^2\left(\frac{x}{100}\right) w^2.
\]
Thus,
\begin{equation}\label{h1}
\mathcal B (u_1)\gtrsim\|\partial_x w\|_{L^2}^2\gtrsim
\int \sech^2\left(\frac {x}{100}\right) w^2=
\int \sech^4\left(\frac {x}{100}\right) u_1^2\gtrsim
\int \sech\left({x}\right) u_1^2 .
\end{equation}
Next, we have
\begin{align*}
\|\partial_x w\|_{L^2}^2 &\gtrsim \int \sech^2\left(\frac {x}{100}\right) |\partial_x w|^2
= \int \sech^2\left(\frac {x}{100}\right) |\zeta \partial_x u_1 + \zeta' u_1 |^2\\
& \gtrsim \int \sech^4\left(\frac {x}{100}\right) |\partial_x u_1|^2
+ 2 \int \sech^2\left(\frac {x}{100}\right) \zeta \zeta' (\partial_x u_1) u_1 + \int \sech^2\left(\frac {x}{100}\right) (\zeta')^2 u_1^2\\
& \gtrsim \int \sech\left(x\right)|\partial_x u_1|^2
+ \int u_1^2\left( - \left(\sech^2\left(\frac {x}{100}\right) \zeta \zeta'\right)' + \sech^2\left(x\right) (\zeta')^2\right).
\end{align*}
Thus, using \eqref{h1},
\[
\int \sech (x )|\partial_x u_1|^2
\lesssim \|\partial_x w\|_{L^2}^2+\int \sech^2 (x ) u_1^2
\lesssim \|\partial_x w\|_{L^2}^2 \lesssim \mathcal B (u_1).
\]
This implies
\begin{equation}\label{rwv}
\mathcal B (u_1)\gtrsim\|\partial_x w\|_{L^2}^2 \gtrsim \|u_1\|_{H^1_{\omega}}^2.
\end{equation}}
{\YM Now, we claim that for any $q>0$,
\begin{equation}\label{SFcinq}
\int \psi' |u_1|^{2+q} \lesssim \|u_1\|_{L^\infty}^q \|\partial_x w\|_{L^2}^2 \lesssim \varepsilon^q \|\partial_x w\|_{L^2}^2.
\end{equation}
Indeed, by parity, the definition of $\psi$ \eqref{defpsi} and $w$ \eqref{def w}, we have (with $\lambda=100$)
\begin{align*}
\int \psi' |u_1|^{2+q}
\lesssim \int_0^{+\infty} e^{-\frac {2 x}{{\lambda }}} |u_1|^{2+q}
\lesssim \int_0^{+\infty} e^{q \frac x{\lambda}} |w|^{2+q}.
\end{align*}
Integrating by parts and using $w(0)=0$ (the function $w$ is odd)
\begin{align*}
\int_0^{+\infty} e^{q \frac x{\lambda}} |w|^{2+q}
&= - \frac{\lambda}q \int_0^{+\infty} e^{q \frac x{\lambda}} \partial_x(|w|^{2+q})
= - \frac {2+q}q \lambda \int_0^{+\infty} e^{q\frac x{\lambda}} (\partial_x w) w |w|^q\\
&\leq C \|u_1\|_{L^\infty}^{\frac q2} \int_0^{+\infty} e^{q \frac x{2 \lambda}} |\partial_x w| |w|^{1+\frac q2}
\leq C^2 \|u_1\|_{L^\infty}^q \int_0^{+\infty} |\partial_x w|^2 +\frac 14 \int_0^{\infty} e^{q\frac x{\lambda}} |w|^{2+q}.
\end{align*}
Thus,
\[
\int_0^{+\infty} e^{q \frac x{\lambda}} |w|^{2+q}
\lesssim \|u_1\|_{L^\infty}^q \|\partial_x w\|_{L^2}^2,
\]
and \eqref{SFcinq} is proved. Since, for some $p>1$,
\[
\left| F(u_1)-\frac{1}{2} u_1 f(u_1) \right |\lesssim |u_1|^{p+1},
\]
we estimate by \eqref{SFcinq},
\begin{equation}\label{SFsix}
\left|\int \psi'\left[ F(u_1)-\frac{1}{2} u_1 f(u_1)\right]\right| \lesssim \varepsilon^{p-1} \|\partial_x w\|_{L^2}^2.
\end{equation}}
Combining \eqref{rwv} and \eqref{SFsix} proves \eqref{10un} for $\ve$ small enough.
%In conclusion, \eqref{lll} is a consequence of \eqref{rwv}, \eqref{SFun}, \eqref{SFdeux}, \eqref{SFtrois}, \eqref{SFquatre}, \eqref{SFcinq} and \eqref{SFsix}.
%\medskip
Finally, we show \eqref{10trois}. From \eqref{eqvarphi}, we compute
\[
\begin{aligned}
\frac{d}{dt}\int \sech(x) u_1 u_2&=\int\sech(x) u_2^2-\int \sech(x) u_{1,x}^2+\int \left[a\sech(x)+\frac{1}{2}\sech''(x)\right] u_1^2 + \int \sech(x) u_1 f(u_1).
\end{aligned}
\]
From this, \eqref{10trois} follows readily by definition of the norm in $H^1_\omega\times L^2_\omega$ and (\ref{SFsix}).
This ends the proof of Proposition \ref{pr:11}.
\end{proof}
\medskip
\noindent
{\underline{Step 4: Conclusion of the proof of Theorem \ref{TH1}.}}
Let
\begin{equation}\label{defH}
\mathcal H(t):= \int \sech\left(x\right)\left[u_{1,x}^2 +u_1^2 + u_2^2\right](t) .
\end{equation}
Then, using \eqref{eqvarphi}, we have {\YM
\begin{equation}\label{Kone}\begin{aligned}
\frac{d}{dt}{\mathcal H} &= 2 \int \sech\left(x\right) \left(u_{1,xt} u_{1,x} + u_{1,x} u_1 + u_{2,t} u_2\right)\\
& = 2 \int \sech\left(x\right) \left[ u_{2,x} u_{1,x} + u_2 u_1 + \left(u_{1,xx} +\mm u_1+f(u_1)\right)u_2 \right]\\
& = 2 \int \sech\left(x\right) \left[\left(1+\mm\right)u_1+f(u_1)\right]u_2 - 2 \int \sech'(x) u_2 u_{1,x} ,
\end{aligned}\end{equation}
and so
\begin{equation}\label{Ktwo}
\left| \frac{d}{dt}{\mathcal H} \right|
\lesssim \int \left(u_{1,x}^2+u_1^2+{u_2^2} \right) \sech\left(x\right) \lesssim \|u(t)\|_{H^1_{\omega}\times L^2_{\omega}}^2.
\end{equation}}
From \eqref{11un} there exists a sequence $t_n\to +\infty$ such that $\mathcal H(t_n)\to 0$.
%From \eqref{Kone}, \eqref{Ktwo}, we have
%\[
%|\dot {\mathcal H}| \lesssim \|v(t)\|_{H^1_{\omega}\times L^2_{\omega}}^2 .
%\]
Let $t\in \RR$. Integrating on $[t,t_n]$ and passing to the limit as $n\to +\infty$ we obtain
\[
\mathcal H(t) \lesssim \int_{t}^{+\infty} \|u(t)\|_{H^1_{\omega}\times L^2_{\omega}}^2 dt.
\]
From \eqref{11un} it follows that $\lim_{t\to +\infty} \mathcal H(t)=0.$
Thus, $\lim_{t\to + \infty} \|u\|_{H^1_{\omega}\times L^2_{\omega}}=0.$
This implies (\ref{Conclusion_0}) and finishes the proof of the Theorem.
%\bibliography{biblio_24_05_2015}
%\bibliographystyle{amsplain}
\providecommand{\bysame}{\leavevmode\hbox to3em{\hrulefill}\thinspace}
\providecommand{\MR}{\relax\ifhmode\unskip\space\fi MR }
% \MRhref is called by the amsart/book/proc definition of \MR.
\providecommand{\MRhref}[2]{%
\href{http://www.ams.org/mathscinet-getitem?mr=#1}{#2}
}
\providecommand{\href}[2]{#2}
\begin{thebibliography}{1}
\bibitem{AM} M.A. Alejo, and C. Mu\~noz, \emph{Nonlinear stability of mKdV breathers}, Comm. Math. Phys. (2013), Vol. 324, Issue 1, pp. 233--262.
\bibitem{AM1} M.A. Alejo, and C. Mu\~noz, \emph{Dynamics of complex-valued modified KdV solitons with applications to the stability of breathers}, Anal. and PDE. \textbf{8} (2015), no. 3, 629--674.
\bibitem{AM2} M.A. Alejo, C. Mu\~noz, and J. M. Palacios, \emph{On the Variational Structure of Breather Solutions}, preprint arXiv:1309.0625.
{\color{black}
\bibitem{Bam_Cucc}
D.~Bambusi, and S.~Cuccagna, \emph{On dispersion of small energy solutions to the nonlinear Klein Gordon equation with a potential}, Amer. J. Math. \textbf{133} (2011), no. 5, 1421--1468.
}
{\color{black}
\bibitem{Bet_Gra_Smets1}
F.~B\'ethuel, P.~Gravejat, and D.~Smets, \emph{Asymptotic stability in the energy space for dark solitons of the Gross-Pitaevskii equation}, Ann. Sci. \'Ec. Norm. Sup\'er. (4), in press, 2014.
}
\bibitem{Birnir} B. Birnir, H. McKean, and A. Weinstein, \emph{The rigidity of Sine-Gordon breathers}, Comm. Pure and Applied Math., Vol. XLVII, 1043--1051 (1994).
\bibitem{bus_per1}
{V}ladimir {B}uslaev and {G}alina {P}erelman, \emph{Scattering for the
nonlinear {S}chr\"odinger equations: states close to a soliton}, St.
Petersburgh Math. J. \textbf{4} (1993), no.~6, 1111--1142.
\bibitem{bus_per2}
\bysame, \emph{{O}n the stability of solitary waves for nonlinear
{S}chr{\"o}dinger equations}, {\color{black} Nonlinear evolution equations, 75--98, Amer. Math. Soc. Transl. Ser. 2, 164, Amer. Math. Soc., Providence, RI, 1995. }%Trans. Amer. Math. Soc \textbf{164} (1995), no.~2, 75--98.
\bibitem{buslaev_sulem}
{V}ladimir {B}uslaev and {C}atherine {S}ulem, \emph{On asymptotic stability of
solitary waves for nonlinear {S}chr\"odinger equations}, Ann. Inst. H.
Poincar\'e Anal. Non Lin\'eaire \textbf{20} (2003), no.~3, 419--475.
\bibitem{Coron} J.-M. Coron, \emph{P\'eriode minimale pour une corde vibrante de longueur infinie}, C.R. Acad. Sc. Paris S\'erie vol. 294 (18 Janvier 1982), p. 127.
\bibitem{MR2373326}
Scipio Cuccagna, \emph{On asymptotic stability in 3{D} of kinks for the
{$\phi^4$} model}, Trans. Amer. Math. Soc. \textbf{360} (2008), no.~5,
2581--2614 \MR{2373326 (2009e:35181)}.
\bibitem{cuccagna_3}
{S}cipio {C}uccagna, \emph{{T}he {H}amiltonian structure of the nonlinear
{S}chr{\"o}dinger equation and the asymptotic stability of its ground
states}, Comm. Math. Phys. \textbf{305} (2011), no.~2, 279--331.
\bibitem{cuccagna_4}
\bysame, \emph{{O}n asymptotic stability of moving ground states of the
nonlinear {S}chr{\"o}dinger equation}, Trans. Amer. Math. Soc \textbf{366}
(2014), no.~2827--2888.
{\color{black}
\bibitem{Delort}
Jean-Marc Delort, \emph{Existence globale et comportement asymptotique pour l'{}\'equation de Klein-Gordon quasi lin\'eaire \`a donn\'ees petites en dimension 1}, Ann. Sci. \'Ecole Norm. Sup. \textbf{34(4)} (2001) pp. 1--61.
}
{ \YM
\bibitem{Delort_fourier}
Jean-Marc Delort, \emph{Semiclassical microlocal normal forms and global solutions of modified one-dimensional KG equations}, Annales de l'Institut Fourier \textbf{66} (2016) 1451--1528.
.}
\bibitem{denzler}
{J}. {D}enzler, \emph{{N}onpersistence of breather families for the
perturbed sine {G}ordon equation}, Comm. Math. Phys. \textbf{158} (1993),
397--430.
\bibitem{Goldman} Gol{'}dman, I. I. and Krivchenkov, V. D. and Ge{\u\i}likman, B. T. and Marquit, E. and Lepa, E., \emph{Problems in quantum mechanics}, Authorised revised ed. Edited by B. T. Geilikman; translated from the Russian by E. Marquit and E. Lepa, Pergamon Press, 1961.
\bibitem{MR678151}
Daniel~B. Henry, J.~Fernando Perez, and Walter~F. Wreszinski, \emph{Stability
theory for solitary-wave solutions of scalar field equations}, Comm. Math.
Phys. \textbf{85} (1982), no.~3, 351--361 \MR{678151 (83m:35131)}.
\bibitem{Kich} Kichenassamy, \emph{Breather Solutions of the Nonlinear Wave Equation}, Comm. Pure and Applied Math., Vol. XLIV, 789-818 (1991).
\bibitem{MR2835867}
E.~Kopylova and A.~I. Komech, \emph{On asymptotic stability of kink for
relativistic {G}inzburg-{L}andau equations}, Arch. Ration. Mech. Anal.
\textbf{202} (2011), no.~1, 213--245. \MR{2835867 (2012h:35333)}.
\bibitem{KMM} M. Kowalczyk, Y. Martel, and C. Mu\~noz, \emph{Kink dynamics in the $\phi^4$ model: asymptotic stability for odd perturbations in the energy space}, preprint arXiv:1506.07420 (2015).
{\color{black}
\bibitem{Kr_Sch} J.~Krieger, and W.~Schlag, \emph{Stable manifolds for all monic supercritical focusing nonlinear Schr\"odinger equations in one dimension}, J. Amer. Math. Soc. \textbf{19} (2006), no. 4, 815--920.
}
\bibitem{kruskal_segur}
M.D {K}ruskal and Harvey Segur, \emph{{N}onexistence of small-amplitude
breather solutions in $\phi^4$ theory}, Phys. Rev. Lett. \textbf{58} (1987),
no.~8, 747--750.
\bibitem{Lamb} G. L. Lamb, \emph{Elements of Soliton Theory}, Pure Appl. Math., Wiley, New York, 1980.
{\color{black}
\bibitem{LS1}
H.~Lindblad and A.~Soffer, \emph{A Remark on long range scattering for the nonlinear Klein-Gordon equation}, J. Hyperbolic Differ. Equ. \textbf{2} (2005), no. 1, 77--89.
\bibitem{LS2}
\bysame, \emph{A remark on asymptotic completeness for the critical nonlinear Klein-Gordon equation}, Lett. Math. Phys. \textbf{73} (2005), no. 3, 249--258.
\bibitem{LS3}
\bysame, \emph{Scattering for the Klein-Gordon equation with quadratic and variable coefficient cubic nonlinearities}, (2013) preprint arXiv:1307.5882.
}
\bibitem{Lin_Tao} H. Lindbland, and T. Tao, \emph{Asymptotic decay for a one-dimensional nonlinear wave equation}, Anal. and PDE Vol. 5 (2012), No. 2, 411--422
DOI: 10.2140/apde.2012.5.411.
\bibitem{Lohe} M. A. Lohe, \emph{Soliton structures in $P(\phi)_2$}, Phys. Rev. D. Vol. 20, no. 12, 15 dec. 1979, 3120--3130.
\bibitem{MR2068924}
Nicholas Manton and Paul Sutcliffe, \emph{Topological solitons}, Cambridge
Monographs on Mathematical Physics, Cambridge University Press, Cambridge,
2004. \MR{2068924 (2006d:58020)}.
\bibitem{MR1753061}
Yvan Martel and Frank Merle, \emph{A {L}iouville theorem for the critical
generalized {K}orteweg-de {V}ries equation}, J. Math. Pures Appl. (9)
\textbf{79} (2000), no.~4, 339--425. \MR{1753061 (2001i:37102)}.
\bibitem{zbMATH01631995}
\bysame, \emph{Asymptotic stability of solitons for subcritical generalized {K}d{V}
equations}, Arch. Ration. Mech. Anal. \textbf{157} (2001), no.~3, 219--254.
{
\bibitem{MM3}
\bysame, \emph{Asymptotic stability of solitons for subcritical gKdV equations revisited}.
Nonlinearity, 18 (2005), no.~1, 55-80. }
\bibitem{MR2150386}
Frank Merle and Pierre Raphael, \emph{The blow-up dynamic and upper bound on
the blow-up rate for critical nonlinear {S}chr\"odinger equation}, Ann. of
Math. (2) \textbf{161} (2005), no.~1, 157--222. \MR{2150386 (2006k:35277)}.
\bibitem{MR1402248}
Michael~E. Peskin and Daniel~V. Schroeder, \emph{An introduction to quantum
field theory}, Addison-Wesley Publishing Company, Advanced Book Program,
Reading, MA, 1995, Edited and with a foreword by David Pines. \MR{1402248
(97j:81001)}.
{\color{black}
\bibitem{Rod_Soffer_Schlag}
I. Rodnianski, W. Schlag, and A. Soffer, \emph{Asymptotic stability of N-soliton states of NLS}, preprint arXiv:math/0309114.
}
\bibitem{MR708660}
Harvey Segur, \emph{Wobbling kinks in {$\varphi ^{4}$} and sine-{G}ordon
theory}, J. Math. Phys. \textbf{24} (1983), no.~6, 1439--1443. \MR{708660
(84i:35130)}.
\bibitem{MR1664792}
A.~Soffer and M.~I. Weinstein, \emph{Time dependent resonance theory}, Geom.
Funct. Anal. \textbf{8} (1998), no.~6, 1086--1128. \MR{1664792 (99k:81329)}
\bibitem{MR1681113}
\bysame, \emph{Resonances, radiation damping and
instability in {H}amiltonian nonlinear wave equations}, Invent. Math.
\textbf{136} (1999), no.~1, 9--74. \MR{1681113 (2000k:37119)}.
{\color{black}
\bibitem{Ste}
J.~Sterbenz, \emph{Dispersive Decay for the 1D Klein-Gordon Equation with Variable Coefficient Nonlinearities}, preprint (arXiv:1307.4808).
}
\bibitem{Tit}
E.~C. Titchmarsh, \emph{{E}igenfunction {E}xpansions {A}ssociated {W}ith
{S}econd-order {D}ifferential {E}quations}, Oxford University Press, 1946.
\bibitem{MR2282481}
Tanmay Vachaspati, \emph{Kinks and domain walls}, Cambridge University Press,
New York, 2006, An introduction to classical and quantum solitons.
\MR{2282481 (2008k:82003)}.
{\color{black}
\bibitem{Tsai_Yau} T.-P. Tsai, and H.T. Yau, \emph{Asymptotic dynamics of nonlinear Schr\"odinger equations: resonance-dominated and dispersion-dominated solutions}, Comm. Pure Appl. Math. \textbf{55} (2002), no. 2, 153--216.
}
\bibitem{Vuillermot} P.-A. Vuillermot, \emph{Nonexistence of spatially localized free vibrations for a class of nonlinear wave equations}, Comment. Math. Helvetici \textbf{64} (1987) 573--586.
\bibitem{MR2318156}
Edward Witten, \emph{From superconductors and four-manifolds to weak
interactions}, Bull. Amer. Math. Soc. (N.S.) \textbf{44} (2007), no.~3,
361--391 (electronic). \MR{2318156 (2008e:58021)}.
\end{thebibliography}
\end{document}