Author: Xin Fu
Abstract
Let $(X, L)$ be a polarized Calabi-Yau variety (or canonical polarized variety) with crepant singularities. Suppose $\omega_{KE} \in c_1(L)$ (or $\omega_{KE} \in c_1(K_X)$) is the unique Ricci-flat current (or Kähler-Einstein current with negative scalar curvature) with local bounded potential. We show that the local tangent cone at any point $p \in X$ of the metric $\omega_{KE}$ is unique.
Introduction
In geometric analysis and algebraic geometry, the study of "optimal" geometric shapes often leads to the investigation of Kähler-Einstein (KE) metrics. These metrics represent spaces with constant Ricci curvature and serve as a gold standard for understanding the geometry of complex manifolds. However, many significant spaces—such as Calabi-Yau varieties used in string theory—possess singularities: points where the standard smooth structure of the manifold breaks down.
To understand the behavior of a metric near a singularity, mathematicians employ a technique called "rescaling." By zooming in on a singular point $p$ by a factor $1/r_i$ as $r_i \to 0$, one obtains a tangent cone, which serves as a local model of the singularity. A fundamental question in the field is whether this limit is unique. While uniqueness was previously established for cases where the singular space is a limit of smooth manifolds (notably by Donaldson and Sun), this paper addresses the problem for spaces that are intrinsically singular, specifically those with crepant singularities.
The Problem of Uniqueness
In the study of differential geometry, the limit of a sequence of rescaled metrics is not guaranteed to be unique; theoretically, different sequences of scales could yield different tangent cone shapes. Crepant singularities are a class of "mild" singularities that allow for a volume-preserving resolution, making them a natural focus for this inquiry. The challenge lies in proving that the infinitesimal shape of the variety at a singular point is a well-defined algebraic object that does not depend on the specific sequence of scales used to reach the limit.
Methodology
Direct Rescaling and Analytic Tools
Unlike previous methodologies that relied on approximating singular spaces with smooth ones, Fu works directly on the singular variety $(X, \omega_{KE})$. This approach requires advanced analytic tools adapted for singular settings, most notably $L^2$-estimates for the $\bar{\partial}$-operator. These estimates allow for the construction of "peak sections"—holomorphic functions concentrated near the singularity—which are used to probe the local geometry.
The H-Condition and Codimension 4 Regularity
The author adapts the "H-condition" from the Donaldson-Sun framework. This technical condition ensures that the local geometry is stable enough to produce a well-defined algebraic structure. A crucial observation supporting this is the Codimension 4 Regularity: for these KE metrics, the singular set has a Hausdorff codimension of at least 4. This high codimension ensures that the "bad" parts of the space are small enough that analytic tools can function similarly to how they would on a smooth manifold.
K-Stability and Hilbert Schemes
The proof culminates in an algebraic argument. By demonstrating that the tangent cone is K-polystable—a notion of stability derived from the Minimal Model Program—the author utilizes the theory of Hilbert schemes to prove that any two potential tangent cones must be isomorphic as polarized affine varieties. This confirms the uniqueness of the limit.
Key Results
The paper establishes two primary theorems regarding the uniqueness of tangent cones:
- Theorem 1.1 (Calabi-Yau Case): For a polarized Calabi-Yau variety with crepant singularities, the local tangent cone at any point $p$ is unique.
- Theorem 1.2 (Canonical Case): The same uniqueness holds for canonical polarized varieties (those with negative scalar curvature).
Furthermore, the tangent cone is proven to be a normal affine variety with Kawamata log terminal (klt) singularities. These are standard, well-behaved objects in algebraic geometry, reinforcing the idea that the local structure of these singular KE metrics is highly rigid.
Discussion and Future Work
Fu's work bridges the gap between hard analysis and algebraic geometry, showing that the metric completion of the smooth part of these varieties is homeomorphic to the original algebraic variety. This ensures that the analytical results align with algebraic intuition.
However, limitations remain. The current proof relies heavily on the properties of crepant singularities. Extending these results to more "violent" singularities, where the canonical bundle changes significantly upon resolution, remains an open challenge. Additionally, the Fano case (positive Ricci curvature) presents further difficulties due to potential automorphisms that allow the space to "drift," complicating the uniqueness of limits. Finally, while this paper bypasses the need for smooth approximations, the exact relationship between the direct singular approach and smooth approximation methods remains a fertile ground for future research.
Summary updated