The rigidity theorems of self shrinkers via Gauss maps (Q323663): Difference between revisions

A self-shrinker is a submanifold \(X:M^n\to \mathbb{R}^{n+m}\) whose mean curvature vector \(H\) satisfies \[ H= -\frac 12 X^{N}, \] where \(X^N\) denotes the normal component of the position vector. The self-shrinkers give rise to the self-similar solutions of the mean curvature flow, and play an important role in the singularity analysis of the mean curvature flow. Thus it is an important topic to classify the self-shrinkers under some natural assumptions. It is well known that the self-shrinkers have close relationship with minimal submanifolds, which are other special solutions to the mean curvature flow and satisfy \(H=0\) identically. The method in the well-developed theory for minimal submanifolds then can be used to study the self-shrinkers. The main purpose of this paper is to study the rigidity of self-shrinkers via the restriction of the image of the Gauss map, and try to obtain some results that are similar for minimal submanifolds. The contributions of this paper are the following. {\parindent=6mm \begin{itemize}\item[(1)] Let \(M^n\) be a complete properly immersed self-shrinker hypersurface in \(\mathbb{R}^{n+1}\). The authors show that if the image of the Gauss map is contained in \(S^n\setminus \bar{S}_+^{n-1}\), then \(M\) is a hyperplane. Here, \(\bar{S}_+^{n-1}\) denotes the closure of the \((n-1)\)-dimensional hemisphere. Moreover, if the image of the Gauss map of a self-shrinker hypersurface \(M\) is contained in a closed hemisphere, then \(M\) is either a hyperplane or a \(N\times \mathbb{R}\), where \(N\) is an \((n-1)\)-dimensional self-shrinker in \(\mathbb{R}^n\). \item[(2)] For the higher-codimensional case, the target manifold of the Gauss map is the Grassmannian manifold. The authors prove that if a self-shrinker \(M^n\) is given by the graph of a smooth vector-valued function \(u=(u^1,\dots, u^m): \mathbb{R}^n\to \mathbb{R}^m\), and if the slope of \(u\) satisfies \[ \Delta_u:=\det\left(\delta_{ij}+\sum_{\alpha}u_i^{\alpha}u_j^{\alpha}\right)^{1/2}< 3, \] then the function \(u\) must be linear and \(M\) is a linear subspace. \end{itemize}} The proof uses the method developed in [Calc. Var. Partial Differ. Equ. 47, No. 3--4, 711--737 (2013; Zbl 1275.58012)] by \textit{J. Jost}, the second and third author for studying the higher-codimension Bernstein problem.