Uniqueness of conical singularities for mean curvature flows (Q6065785)

The mean curvature flow in codimension \(k\) is a geometric heat equation which evolves a family of immersions \(\mathbf{x}: M^n \times I \rightarrow \mathbb{R}^{n+k}\) such that \[ \partial_t \mathbf{x} = \mathbf{H}, \] where \(\mathbf{H}\) is the mean curvature vector of the immersed submanifold \(M^n_t := \mathbf{x}\left(M^n,t\right)\). Due to the nonlinearities of this PDE, one frequently encounters finite-time singularities. Understanding the formation of these singularities is of vital importance when using the mean curvature flow to prove geometric/topological results. In order to study the singularity formation, one takes ``blow-up limits''/``tangent flows'' around a singular point by parabolically rescaling. However, the limit one gets could depend on the subsequence of rescaling parameters one passes to. To prove that the blow-up limit does not depend on the subsequence is thus an important problem. This paper proves such a uniqueness result for so-called asymptotically self-shrinkers of mean curvature flows in high codimensions. A self-shrinker \(\Sigma\) is defined by the equation \[ \mathbf{H} = \frac{x^\perp}{2}. \] By \textit{G. Huisken}'s monotonicity formula [J. Differ. Geom. 20, 237--266 (1984; Zbl 0556.53001)], self-shrinkers frequently arise as singularity models of the mean curvature flow. Asymptotically conical self-shrinkers are self-shrinkers which are (smoothly) asymptotic to some cone \(\mathcal{C}\). The main tool in showing uniqueness of such tangent flows is a ``Łojasiewicz-Simon inequality''. The techniques used in this paper are based on those in the work of \textit{O. Chodosh} and \textit{F. Schulze} [Duke Math. J. 170, No. 16, 3601--3657 (2021; Zbl 1489.53123)], who proved asymptotically conical tangent flows are unique in codimension \(1\). Extending the techniques of Chodosh-Schulze to higher codimension uses some standard techniques, so the reader may wish to consult the Chodosh-Schulze paper for more details about some of the arguments. In some parts, however, new ideas are needed to compensate for the unavailability of certain tools in high codimension, such as maximum principles. The analysis is based on defining appropriate weighted spaces which consist of entire vector fields with appropriate decay rates. In particular, three different Hölder spaces are used: \begin{itemize} \item[1.] Homogeneous spaces which reflect certain polynomial decay properties. \item[2.] Anisotropic Hölder spaces which require stronger decay in the radial direction. \item[3.] Conical Hölder spaces which reflect the conical structure at infinity. \end{itemize} Sobolev spaces with respect to a Gaussian weighted volume measure are also defined and it is shown that the stability operator \[ L := \Delta^\perp_\Sigma - \frac{1}{2} \nabla^\perp_{x^T} + \frac{1}{2} + \sum_{k,\ell} \left<\cdot, A_{k\ell}\right> A_{k\ell} \] is Fredholm between these weighted Sobolev spaces, a property which is crucially needed when proving the Łojasiewicz-Simon inequality. The Łojasiewicz-Simon inequality is first proved globally, under the assumption that \(M\) can be written as an entire graph over \(\Sigma\) of some vector field with sufficiently small norm in the conical Hölder space mentioned above. However, when studying blow-up limits, such a graphicality assumption will only be satisfied on a (possibly large) compact set. The estimate thus needs to be localized and the resulting error terms need to be dealt with. To do this, the authors, inspired by the work of Chodosh-Schulze, define certain scales (shrinker, rough conical, conical) in the high codimension setting. Some technical lemmas about the relationships between these scales allow for the localization of the Łojasiewicz-Simon inequality to occur, and from there the uniqueness of the tangent flows follows along the lines of Chodosh-Schulze's argument. The one exception is a certain curvature bound, which the authors of this paper prove using a different approach. As an aside, Łojasiewicz-Simon inequalities have been used in other situations to prove uniqueness and dynamical stability results which may be of independent interest to the reader. A non-exhaustive list is: [\textit{A. Deruelle} and \textit{T. Ozuch}, Calc. Var. Partial Differ. Equ. 62, No. 3, Paper No. 84, 60 p. (2023; Zbl 1521.53073); \textit{R. Haslhofer} and \textit{R. Müller}, Math. Ann. 360, No. 1--2, 547--553 (2014; Zbl 1300.53064); \textit{K. Kröncke}, Calc. Var. Partial Differ. Equ. 53, No. 1--2, 265--287 (2015; Zbl 1317.53086); \textit{L. Simon}, Ann. Math. (2) 118, 525--571 (1983; Zbl 0549.35071)].