Bergman kernels over polarized Kähler manifolds, Bergman logarithmic flatness, and a question of Lu-Tian

Basic Information

• Paper ID: 2510.22169
• Title: Bergman kernels over polarized Kähler manifolds, Bergman logarithmic flatness, and a question of Lu-Tian
• Authors: Peter Ebenfelt, Ming Xiao, Hang Xu
• Classification: math.CV (Complex Analysis), math.DG (Differential Geometry)
• Publication Date: October 25, 2025
• Paper Link: https://arxiv.org/abs/2510.22169

Abstract

This paper investigates two classes of Bergman kernels on a complete Kähler manifold $M$ with Kähler metric $g$ induced by a positive line bundle $(L, h) \to M$: the Bergman kernel of the disk bundle of the dual line bundle $(L^*, h^*)$, and the Bergman kernel of the tensor product of the line bundle $(L^k, h^k)$ ($k\geq 1$) with the canonical line bundle of $(M, g)$. The paper first establishes a localization result for the former, then characterizes the absence of logarithmic singularities of this Bergman kernel in terms of the vanishing of coefficients in the Tian-Yau-Zelditch-Catlin type expansion of the latter. This result particularly answers a question posed by Lu and Tian. As an application, it is shown that when $(M, g)$ is a compact locally homogeneous Kähler manifold, the circle bundle of $(L^*, h^*)$ is necessarily Bergman logarithmically flat.

Research Background and Motivation

Core Problems

The core problems addressed in this paper are:

1. Understanding the Fefferman-type expansion of the Bergman kernel of the disk bundle of the dual line bundle
2. Geometric characterization of Bergman logarithmic flatness
3. The Lu-Tian question: characterizing logarithmic singularities using coefficients of the Tian-Yau-Zelditch-Catlin expansion

Significance of the Problems

1. Geometric Significance: Bergman logarithmic flatness is an important geometric property of strongly pseudoconvex real hypersurfaces, closely related to CR geometry
2. Analytic Significance: Asymptotic expansions of Bergman kernels reveal deep information about complex geometric structures
3. Theoretical Value: Connects multiple important concepts in complex analysis, differential geometry, and algebraic geometry

Limitations of Existing Methods

1. Classical Fefferman theory primarily applies to bounded domains, with limited treatment of general manifolds
2. Existing localization results require compact manifolds and cannot handle non-compact cases
3. One direction of the Lu-Tian question (sufficiency) has remained unresolved

Research Motivation

1. Generalize results of Engliš-Zhang on Bergman logarithmic flatness over Hermitian symmetric spaces
2. Develop new techniques for handling non-compact manifolds
3. Establish deep connections between Bergman kernel theory and Kähler geometry

Core Contributions

1. Localization Theorem: Establishes the localization property of the Bergman kernel of the disk bundle of the dual line bundle (Theorem 1.2), extending to non-compact manifolds
2. Fefferman-type Expansion: Establishes complete asymptotic expansion of the disk bundle Bergman kernel (Corollary 1.4)
3. Core Equivalence Theorem: Proves the equivalence between Bergman logarithmic flatness and vanishing of coefficients in the Tian-Yau-Zelditch-Catlin expansion (Theorem 1.11)
4. Complete Resolution of Lu-Tian Question: Provides necessary and sufficient conditions in the compact case (Corollary 1.12)
5. Application Results: Proves Bergman logarithmic flatness of circle bundles over compact locally homogeneous Kähler manifolds (Corollary 1.13)

Methodology Details

Problem Formulation

Study the relationship between two classes of Bergman kernels on polarized manifolds $(M, g; L, h)$:

• Input: Polarized Kähler manifold $(M, g; L, h)$, where $(L, h)$ is a positive line bundle
• Output: Geometric characterization of Bergman logarithmic flatness
• Constraints: $M$ must admit some complete Kähler metric

Core Architecture

1. Construction of Standard Top Form

Define the standard top form of the dual line bundle $(L^*, h^*)$: $T(z, \xi) = (\sqrt{-1})^{(n+1)^2}e^{\phi(z)} \det g(z) dz \wedge d\xi \wedge d\bar{z} \wedge d\bar{\xi}$

where $e^{\phi(z)} = h^*(e_{L^*}(z), e_{L^*}(z))$.

2. Relationship Between Two Classes of Bergman Kernels

Establish the relationship between the Bergman kernel $K_D$ of the disk bundle $D$ and the Bergman kernel $B_k$ of the line bundle $L^k \otimes C_M$: $K_D = \left(\sum_{k=0}^{\infty} \frac{k+1}{2^{n+1}\pi}(1-\rho)^k B_{k+1}\right)T$

where $\rho(v) = 1 - |v|_{h^*}^2$ is the defining function of the circle bundle.

3. Asymptotic Expansion Analysis

Through intricate asymptotic analysis, relate the Fefferman expansion of $K_D$: $K_D = \left(\frac{\Phi}{\rho^{n+2}} + \Psi \log \rho\right)T$

to the Tian-Yau-Zelditch-Catlin expansion of $B_k$: $B_k(z) \sim \left(\frac{k}{\pi}\right)^n \sum_{j=0}^{\infty} \frac{a_j(z)}{k^j}$

Technical Innovations

1. Localization Technique for Non-compact Manifolds

• Utilize $L^2$ estimates to handle non-complete metrics
• Develop new smooth extension methods

2. Explicit Coefficient Computation

Through combinatorial techniques, establish explicit relationships between coefficients $\Phi, \Psi$ and $a_j$: $\alpha_j = \sum_{i=0}^{n+1} \frac{\lambda_{ji} a_i}{(2\pi)^{n+1}}$$\beta_j = \sum_{i=0}^{j} \frac{a_{n+2+i} \tau_{ij}}{(2\pi)^{n+1}}$

3. Universality Results

Prove that coefficients $a_j$ are universal polynomials in curvature and its covariant derivatives, ensuring geometric invariance of results.

Main Theorems

Theorem 1.2 (Localization)

Let $(M, g; L, h)$ be a polarized manifold, and suppose $M$ admits some complete Kähler metric. If $\Omega$ is a relatively compact pseudoconvex domain in $L^*$ with smooth strongly pseudoconvex boundary, $\Omega \subseteq D$, and there exist $p \in S$ and a neighborhood $W$ such that $W \cap \Omega = W \cap D$, then the Bergman kernel difference $K_D - K_{\Omega}$ extends $C^{\infty}$ smoothly to $W \cap \Omega$.

Theorem 1.9 (Core Relationship Theorem)

Under the conditions of Theorem 1.2, there exist unique families of smooth functions $\{\alpha_j\}_{j=0}^{n+1}$ and $\{\beta_j\}_{j=0}^{\infty}$ such that: $\Phi = \sum_{j=0}^{n+1} (\alpha_j \circ \pi)\rho^j + O(\rho^{n+2})$$\Psi = \sum_{j=0}^{N} (\beta_j \circ \pi)\rho^j + O(\rho^{N+1})$

and each $\alpha_j, \beta_j$ can be explicitly computed from $a_0, \ldots, a_{n+2+j}$.

Theorem 1.11 (Main Equivalence Theorem)

Let $U \subset M$ be an open set, and $\Sigma = \pi^{-1}(U) \cap S$. Then the following are equivalent:

1. For all $m \geq 0$, $a_{n+2+m} = 0$ on $U$
2. $\Sigma$ is Bergman logarithmically flat

Application Results

Corollary 1.12 (Compact Case)

For a compact polarized manifold $(M, g; L, h)$, the following are equivalent:

1. Each $a_{n+2+m}$ is constant on $M$
2. $S$ is Bergman logarithmically flat
3. Each $a_{n+2+m}$ is either non-negative or non-positive on $M$

Corollary 1.13 (Locally Homogeneous Case)

If $(M, g; L, h)$ is a compact locally homogeneous polarized manifold, then the circle bundle $S$ of $(L^*, h^*)$ is Bergman logarithmically flat. Furthermore, $S$ is a sphere if and only if $(M, g)$ is locally holomorphically isometric to one of the following complex space forms:

1. $(B^n, \lambda \omega_{-1})$
2. $(CP^n, \lambda \omega_1)$
3. $(C^n, \omega_0)$
4. $(B^l \times CP^{n-l}, \lambda\omega_{-1} \times \lambda\omega_1)$

Technical Details

Proof Strategy

1. Step One: Establish integral relationships between the two classes of Bergman kernels
2. Step Two: Decompose Bergman spaces using the $S^1$ action
3. Step Three: Extract coefficient relationships through asymptotic analysis
4. Step Four: Apply $L^2$ estimates to prove localization properties

Key Lemmas

• Lemma 2.4: Establishes Hilbert space isomorphism between $A^2_{m+1}(D)$ and $A^2(M, L^{m+1} \otimes C_M)$
• Lemma 3.3: Provides combinatorial identities for handling power series expansions
• Lemma 3.5: Analytical lemma concerning asymptotic behavior of functions

Related Work

Historical Background

1. Fefferman Theory: Establishes asymptotic expansions of Bergman kernels for bounded strongly pseudoconvex domains
2. Tian-Yau-Zelditch-Catlin Theory: Develops asymptotic theory of Bergman kernels of line bundles over compact manifolds
3. Engliš-Zhang Work: First constructs non-spherical Bergman logarithmically flat hypersurfaces

Innovations in This Paper

1. Generalizes localization theory to non-compact manifolds
2. Establishes deep connections between two different classes of Bergman kernels
3. Completely resolves the open problem posed by Lu-Tian

Conclusions and Discussion

Main Conclusions

1. Completely characterizes the geometric conditions for Bergman logarithmic flatness of circle bundles over polarized manifolds
2. Establishes new connections between Bergman kernel theory and Kähler geometry
3. Provides systematic methods for constructing non-spherical Bergman logarithmically flat hypersurfaces

Limitations

1. Main results still require the assumption that $M$ admits a complete Kähler metric
2. Complete characterization in the non-compact case remains unclear
3. Coefficient formulas are quite complex when computing specific examples

Future Directions

1. Conjecture 1.14: Study Bergman logarithmic flatness in the non-compact locally homogeneous case
2. Develop Szegő kernel theory on more general manifolds
3. Explore connections with Einstein metric theory

In-Depth Evaluation

Strengths

1. Theoretical Depth: Resolves an important open problem in the field with significant theoretical contributions
2. Technical Innovation: Develops new techniques for handling non-compact manifolds with universal applicability
3. Complete Results: Provides necessary and sufficient conditions with a complete theoretical framework
4. Geometric Insight: Reveals deep connections between complex analysis and differential geometry

Weaknesses

1. Technical Complexity: Proofs involve substantial technical details, affecting readability
2. Limited Scope of Applications: Main results are still restricted to special geometric cases
3. Computational Complexity: Practical verification of Bergman logarithmic flatness involves difficult calculations

Impact

1. Theoretical Impact: Provides new research tools for complex geometry and CR geometry
2. Methodological Value: Localization techniques may apply to other related problems
3. Application Prospects: Offers new approaches for constructing complex manifolds with special properties

Applicable Scenarios

1. Study of complex manifolds with special symmetries
2. Analysis of geometric properties of Bergman kernels
3. Construction of non-spherical hypersurfaces with special CR properties

References

The paper cites 46 important references covering classical and cutting-edge work in Bergman kernel theory, Kähler geometry, CR geometry, and related fields, particularly foundational works by Fefferman 18, Lu-Tian 31, and Engliš-Zhang 17.