Uniqueness of some Calabi-Yau metrics on \(\mathbf{C}^n\) (Q2216462)

The recent discovery of infinitely many complete Calabi-Yau metrics on \(\mathbb{C}^n\) for \(n\geq 3\) with maximal volume growth, raises the natural uniqueness and classification questions for such metrics. In the simplest and the most generic case of these constructions, the tangent cone at infinity is \(A_1\times \mathbb{C}\), where \(A_1\) stands for the \((n-1)\)-dimensional Stenzel cone. This paper proves that uniqueness holds when the tangent cone is \(A_1\times \mathbb{C}\). In contrast, the conventional uniqueness results in Kähler geometry, exemplified by \textit{R. J. Conlon} and \textit{H.-J. Hein}'s work [Duke Math. J. 162, No. 15, 2855--2902 (2013; Zbl 1283.53045)], typically requires the Calabi-Yau metrics to be asymptotic to a fixed reference metric at infinity with a suitably fast decay rate, and the tangent cone at infinity to have smooth link. In the present case, there is no a priori framing data at infinity (so the huge complex automorphism group of \(\mathbb{C}^n\) becomes a severe issue), and the link of the tangent cone is singular. A rough outline of the proof is: \begin{itemize} \item Produce canonical holomorphic coordinates over large metric balls. This step is achieved through a noncompact variant of Donaldson-Sun theory. \item Produce preferred Kähler potentials on large balls, with close relationship to the distance function. This appeals to the previous work of \textit{G. Székelyhidi} and \textit{G. Liu} [``Gromov-Hausdorff limits of Kähler manifolds with Ricci curvature bounded below'', Preprint, \url{arXiv:1804.08567}]. This is like a gauge fixing step, with the outcome that on the rescaled balls the Kähler potential differs from the model case by a \(C^0\) small amount. \item Show that the \(C^0\)-smallness of the relative Kähler potential has Hölder improvement when passing to half balls, up to adjustments by pluriharmonic functions and complex automorphisms. Using a blow up argument, this requires understanding harmonic functions of at most quadratic growth. One also needs to prevent errors concentrating near the singularity of the link; this is achieved by a barrier construction. These arguments constitute the main bulk of the paper, and draw on inspirations from minimal surface theory. \item In the smooth region, the \(C^0\) smallness of the relative potential can be improved to \(C^\infty\) using the small perturbation theorem of Savin. The desired uniqueness can then be proved by iterating the previous step over an arbitrarily large number of scales. \end{itemize} The paper ends by some interesting speculations on more general tangent cones, and cases beyond \(\mathbb{C}^n\).