Sharp stability of Alexandrov's theorem for $C^1$ domains in the small-excess regime

Abstract

We prove a sharp quantitative stability result for Alexandrov's theorem in arbitrary dimension for bounded $C^1$ open sets in a small-excess regime. More precisely, if $E\subset {\mathbb{R}}^n$ is a bounded $C^1$ open set with the same volume as the unit ball $B$, small excess, and scalar distributional mean curvature ${\mathcal{H}}_{\partial E}\in L^2(\partial E)$, then, up to a translation, $$\mathrm{Exc}(E)+|E\Delta B{|}^2+|\mu -(n-1){|}^2\le C(n)\Vert {\mathcal{H}}_{\partial E}-\mu {\Vert }_{L^2(\partial E)}^2\forall \mu \in \mathbb{R}.$$ In other words, both the excess and the symmetric difference from the ball are controlled by the optimal $L^2$-oscillation of the mean curvature. This yields a sharp stability estimate in a genuinely non-parametric regime. The proof combines a $BV$ version of Fuglede's spectral-gap argument, a star-shaped rearrangement for sets of finite perimeter, quantitative estimates for the part of the boundary contained in the tentacles, and a polyhedral approximation argument for the non-graphical region. We note that the $C^1$ regularity assumption enters only as a qualitative technical ingredient of the proof, but all constants in the final estimate depend only on the dimension.

59 Pages, 1 figure