Threshold solutions for the nonlinear Schrödinger equation (Q2104845)

In this paper, the authors study the focusing energy-subcritical nonlinear Schrödinger equation (NLS) in \(H^1(\mathbb{R}^N), N \geq 1\) given by \[ \begin{cases} i \partial_t u + \Delta u+|u|^{p-1}\,u=0\,, \\ u(x,0)=u_0(x) \in H^1(\mathbb{R}^N)\,, \end{cases}\tag{1} \] where \[ 1+4/N<p< \begin{cases} 1+4/(N-2)\,, \quad N \geq 3\,,\\ +\infty\,, \quad N=1,2\,. \end{cases} \] One refers to (1) as being in the \textit{intercritical regime} as it is mass-supercritical and energy-subcritical. When \(N \geq 3\), the authors also consider the energy-critical focusing NLS\ \[ \begin{cases} i \partial_t u + \Delta u+|u|^{p_c-1}\,u=0\,, \\ u(x,0)=u_0(x) \in \dot{H}^1(\mathbb{R}^N)\,, \end{cases}\tag{2} \] where \[ p_c:=(N+2)/(N-2)\,. \] For (1), one works in the inhomogeneous Sobolev space \(H^1(\mathbb{R}^N)\) and for (2), one works in the homogeneous Sobolev space \(\dot{H}^1(\mathbb{R}^N)\). For (1), the authors show the existence of special solutions \(Q^{\pm}\) which exponentially approach the standing wave solution \(e^{it}Q\) in the positive time direction, but differ in their behaviour in the negative time direction. Theorems 1.2 and 1.3 give a classification of the solutions at the threshold level. In particular, either blow up occurs in finite (positive or negative) time, or scattering occurs in both time directions, or the solution is equal to one of the special solutions \(Q^{\pm},Q\), up to symmetries of the equation. The result of Theorems 1.2--1.3 was previously known in the case when \(N=3\) and \(p=3\) by the work of \textit{T. Duyckaerts} and \textit{S. Roudenko} [Rev. Mat. Iberoam. 26, No. 1, 1--56 (2010; Zbl 1195.35276)]. Theorems 1.2--1.3 generalise these results to the full intercritical regime. A fundamental challenge is that for \(p<3\), the nonlinearity \(|u|^{p-1}u\) is not a smooth function of \((u,\bar{u})\). Furthermore, the difference \(|u|^{p-1}u-|v|^{p-1}v\) is not polynomial in \((u,\bar{u},v,\bar{v})\). Thus, the estimates from the earlier work [loc. cit.] do not apply. Furthermore, for \(p \leq 2\), the nonlinearity \(|u|^{p-1}u\) is no longer twice real-differentiable. The authors overcome these difficulties by using fractional calculus methods developed by \textit{F. M. Christ} and \textit{M. I. Weinstein} [J. Funct. Anal. 100, No. 1, 87--109 (1991; Zbl 0743.35067)] and \textit{M. Visan} [Duke Math. J. 138, No. 2, 281--374 (2007; Zbl 1131.35081)]. Analogous results are shown in Theorem 1.5--1.6 for the energy-critical problem (2) when \(N \geq 6\). The method developed for the intercritical regime also applies to this setting. The authors hence give an alternative proof of the earlier result of \textit{D. Li} and \textit{X. Zhang} [J. Funct. Anal. 256, No. 6, 1928--1961 (2009; Zbl 1162.35070)], which in turn is a higher-dimensional generalisation of the work of \textit{T. Duyckaerts} and \textit{F. Merle} [Geom. Funct. Anal. 18, No. 6, 1787--1840 (2009; Zbl 1232.35150)].