Arithmetic properties and zeros of the Bergman kernel on a class of quotient domains

Luke D. Edholm and Vikram T. Mathew · 2025

Read Arithmetic properties and zeros of the Bergman kernel on a class of quotient domains by Luke D. Edholm and Vikram T. Mathew (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-20489.txt.

Introduction

Luke Edholm and Vikram Mathew wrote a 2025 mathematics paper on the Bergman kernel, a key function in complex analysis. It is studied on regions of space, and one open problem asks when the kernel has zeros inside its region. Zeros can change how the region's geometry behaves.

The authors work on a class of quotient domains with a curved, ring-like shape. For rational shape values above one, they write an exact formula for the kernel. The formula turns on number traits of the shape value. They then build a chain of values nearing one where each kernel has zeros, though the limit case has none. That chain settles the open question for this class.

Catalog record

CREATORLuke D. Edholm and Vikram T. Mathew
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE71 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-20489

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Arithmetic properties and zeros of the Bergman kernel on a class of quotient domains Thanks: The first author was supported in part by Austrian Science Fund (FWF) grants: DOI 10.55776/I4557 and DOI 10.55776/P36884. Thanks: 2020 Mathematics Subject Classification: 32A25 (Primary); 32A60, 30C15 (Secondary) Luke D. Edholm and Vikram T. Mathew Abstract . An effective formula for the Bergman kernel on ℍ γ = { | z 1 | γ < | z 2 | < 1 } {\mathbb{H}}_{\gamma}=\{|z_{1}|^{\gamma}<|z_{2}|<1\} is obtained for rational γ = m n > 1 \gamma=\frac{m}{n}>1 . The formula depends on arithmetic properties of γ \gamma , which uncovers new symmetries and clarifies previous results. The formulas are then used to study the Lu Qi-Keng problem. We produce sequences of rationals γ j ↘ 1 \gamma_{j}\searrow 1 , where each ℍ γ j {\mathbb{H}}_{\gamma_{j}} has a Bergman kernel with zeros (while ℍ 1 {\mathbb{H}}_{1} is known to have a zero-free kernel), resolving an open question on this domain class. 1. Introduction Given a domain (an open connected set) \Omega\subset{\mathbb{C}}^{n} , the Bergman space A^{2}(\Omega) is the subspace of L^{2}(\Omega) consisting of holomorphic functions. The Bergman space is itself a Hilbert space, so there is a canonical orthogonal projector P_{\Omega}:L^{2}(\Omega)\to A^{2}(\Omega) , called the Bergman projection. This projection is carried out by integrating against the Bergman kernel, the reproducing kernel function K_{\Omega}:\Omega\times\Omega\to{\mathbb{C}} of A^{2}(\Omega) : P_{\Omega}f(z)=\int_{\Omega}K_{\Omega}(z,w)f(w)\,dV(w),\qquad f\in L^{2}(\Omega). (1.1) The Bergman kernel is uniquely determined by the following three properties: for each fixed w\in\Omega , the function K_{\Omega}(\cdot,w)\in A^{2}(\Omega) ; it is conjugate symmetric, i.e. K_{\Omega}(z,w)=\overline{K_{\Omega}(w,z)} ; and it reproduces holomorphic functions, i.e. for f\in A^{2}(\Omega) , P_{\Omega}f=f . In addition, K_{\Omega} satisfies many useful extremal and invariance properties; see e.g. [ 4 , 27 , 25 , 29 ] . In Section 2.2 we recall those properties relevant to the present article. For \gamma>0 , define a family of bounded, non-smooth, weakly pseudoconvex domains called generalized Hartogs triangles by setting {\mathbb{H}}_{\gamma}=\{(z_{1},z_{2})\in{\mathbb{C}}^{2}:|z_{1}|^{\gamma}<|z_{2}|<1\}. (1.2) When \gamma is rational, {\mathbb{H}}_{\gamma} can be realized as a quotient of {\mathbb{D}}\times{\mathbb{D}}^{*} induced by a monomial mapping (see Section 1.2 ). Many aspects of Bergman theory on these domains have been studied in recent years; see [ 10 , 16 , 17 , 18 , 19 , 8 , 26 , 9 , 3 , 23 , 24 , 11 ] and the references therein. The purpose of this paper is two fold: (1) to give a full and effective formula for Bergman kernel of {\mathbb{H}}_{\gamma} for every rational \gamma>1 , and (2) to use the formulas to exhibit new symmetries and locate zeros of the kernel in {\mathbb{H}}_{\gamma}\times{\mathbb{H}}_{\gamma} . The question on the existence/non-existence of zeros of Bergman kernel (the Lu Qi-Keng problem) goes back several decades and was originally motivated by the global well-definedness of Bergman representative coordinates; see e.g., [ 31 ] , and the survey [ 5 ] . A domain \Omega is said to be a Lu Qi-Keng domain when its Bergman kernel is non-vanishing. Appropriately interpreted, the zero set of the Bergman kernel is an analytic variety on \Omega\times\Omega , and transformation law ( 2.6 ) shows this variety is a biholomorphic invariant of \Omega . Simple domains such as the ball and polydisc have non-vanishing Bergman kernel, and it is an interesting open problem to ask when the taking of quotients of these and other Lu-Qi Keng domains induces zeros in the quotient Bergman kernels. This has been investigated in special cases (certain ball quotients [ 6 ] , symmetrized polydiscs [ 20 , 32 ] , the tetrablock [ 37 ] and quotients of higher dimensional Cartan type IV domains [ 21 ] ), but a unified approach to this problem is only in its early stages. Since the zero set of the Bergman kernel is a biholomorphic invariant, a preliminary step to the Lu Qi-Keng problem on the domain family \{{\mathbb{H}}_{\gamma}:\gamma>0\} would be to identify its biholomorphic equivalence classes. In Theorem 2.19 and Corollary 2.29 below, we prove that for distinct \gamma,\eta\geq 1 , the domains {\mathbb{H}}_{\gamma} and {\mathbb{H}}_{\eta} are never biholomorphic. On the other hand, given \gamma\in(0,1) , there is a unique \widetilde{\gamma}\geq 1 such that {\mathbb{H}}_{\gamma} is biholomorphic to {\mathbb{H}}_{\widetilde{\gamma}} , and therefore the problem can be reduced to understanding the situation for \gamma\geq 1 . Up to the time of writing of this paper, the Lu Qi-Keng problem on {\mathbb{H}}_{\gamma} has stood as follows: (a) The Bergman kernel on {\mathbb{H}}_{1} (the classical Hartogs triangle) is non-vanishing; this can be seen directly from formula ( 2.7 ) below. (b) The Bergman kernel on {\mathbb{H}}_{\gamma} has zeros for \gamma\in[2,\infty) . This was shown first for integers \gamma\geq 2 in [ 16 ] , then for all other \gamma>2 in [ 18 ] . (c) It has remained open as to whether the Bergman kernel of {\mathbb{H}}_{\gamma} has zeros for \gamma\in(1,2) . In [ 18 ] , when Edholm and McNeal found zeros of {\mathbb{H}}_{\gamma} for \gamma\geq 2 , it was also observed that the same approach fails to address \gamma\in(1,2) ; see Section 1.3 below for more details. The question has remained open since: (Q1) For which \gamma\in(1,2) does the Bergman kernel have zeros? A weaker version of (Q1) has also been raised; see [ 15 ] . Since the topologically equivalent domains {\mathbb{H}}_{\gamma_{j}} tend to {\mathbb{H}}_{1} as \gamma_{j}\searrow 1 , Ramadanov’s theorem (more precisely, an extended version of the more classical result; see [ 28 ] ) shows that the Bergman kernels K_{\gamma_{j}}(\cdot,\cdot) of {\mathbb{H}}_{\gamma_{j}} tend to the Bergman kernel K_{1}(\cdot,\cdot) of {\mathbb{H}}_{1} locally normally on {\mathbb{H}}_{1}\times{\mathbb{H}}_{1} . As K_{1} is known to be non-vanishing, it is interesting to ask (Q2) Is there some \epsilon>0 so that K_{\gamma}(\cdot,\cdot) is non-vanishing on {\mathbb{H}}_{\gamma}\times{\mathbb{H}}_{\gamma} for \gamma\in[1,1+\epsilon) ? This paper addresses both questions. In Section 4 we exhibit infinitely many rational \gamma\in(1,2) for which K_{\gamma}(\cdot,\cdot) has zeros inside {\mathbb{H}}_{\gamma}\times{\mathbb{H}}_{\gamma} , marking the first progress made on (Q1). These \gamma values accumulate at 1, which answers (Q2) by proving no such \epsilon can exist. 1.1. An effective kernel formula While the existence of the Bergman kernel is guaranteed by abstract Hilbert space theory, concrete information about its properties as a function K_{\Omega}:\Omega\times\Omega\to{\mathbb{C}} (e.g. its asymptotics near the boundary, whether it has zeros inside \Omega\times\Omega , and if so, properties of the zero set) requires hard analysis. Our first main result is a new and effective formula for the Bergman kernel of {\mathbb{H}}_{m/n} , valid for any rational \frac{m}{n}>1 . Previous work (see [ 18 , 9 , 1 ] ) has expressed these kernels (which have been known since [ 18 ] to be rational functions) in one form or another, but until now a description of the numerator polynomial from which it is possible to study the zero set of the kernel when \gamma\in(1,2) has remained elusive. For comparison, these past results are recalled in Section 1.3 . Theorem 1.3 . Let m,n\in{\mathbb{Z}}^{+} with m>n and \gcd(m,n)=1 . Given points z=(z_{1},z_{2}) and w=(w_{1},w_{2}) in {\mathbb{H}}_{m/n} , let s=z_{1}\overline{w}_{1} , t=z_{2}\overline{w}_{2} . The Bergman kernel of {\mathbb{H}}_{m/n} is K_{m/n}(z,w)=\frac{P_{m,n}(s,t)}{m\pi^{2}(1-t)^{2}(t^{n}-s^{m})^{2}}, (1.4a) where the numerator is a polynomial of degree 2m-1 , given explicitly by P_{m,n}(s,t)=p_{0}(s,t)+p_{1}(s,t)+p_{2}(s,t)+p_{3}(s,t)+p_{4}(s,t), (1.4b) where \displaystyle p_{0}(s,t) \displaystyle=m^{2}s^{m-1}t^{n}, (1.4c) \displaystyle p_{1}(s,t) \displaystyle=\sum_{j=0}^{m-2}(j+1)(\kappa(j)+1)s^{j}t^{2n-L(j)}, (1.4d) \displaystyle p_{2}(s,t) \displaystyle=\sum_{j=0}^{m-2}(j+1)(m-\kappa(j)-1)s^{j}t^{2n+1-L(j)},

Read the complete text