Curves on Compact Arithmetic Quotients of Hyperbolic 2-ball

Zhehao Li · 2025

Read Curves on Compact Arithmetic Quotients of Hyperbolic 2-ball by Zhehao Li (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-2502-11582.txt.

Introduction

Zhehao Li studies low genus curves on compact Shimura surfaces in a two thousand twenty-five mathematics paper. Shimura surfaces arise as quotients of complex hyperbolic balls. They encode moduli problems for abelian varieties with special endomorphism types.

Li compares two families, Picard and Hilbert modular surfaces. Fix a bound on the genus, he shows, and take the defining discriminant large enough. Then the compact Picard surfaces carry no low genus curves at all. On Hilbert surfaces such curves may persist, but only special ones parametrizing extra endomorphisms. The proof leans on how special subvarieties repel each other and on volume estimates in hyperbolic geometry. The result mirrors uniform boundedness phenomena for abelian varieties over function fields.

Catalog record

CREATORZhehao Li
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE78 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-11582

Browse the full library · About the Tower of Babel

Full text

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

Curves on Compact Arithmetic Quotients of Hyperbolic 2-ball Zhehao Li Abstract 1. Introduction Complex hyperbolic n -ball \mathbb{B}^{n} is a higher dimensional analogue of the unit 1-ball \mathbb{B}^{1} in hyperbolic geometry. It is the domain \mathbb{B}^{n}=\{z\in\mathbb{C}^{n}:|z|<1\} , or equivalently \mathbb{P}V_{-} the projectivization of the negative cone V_{-}=\{u\in V:h(u,u)<0\} for V\cong\mathbb{C}^{n+1} equipped with a hermitian form h with signature (n,1) . In this paper we study the following compact arithmetic quotient of \mathbb{B}^{2} . We start from K=\mathbb{Q}(\sqrt{D}) a quadratic field with discriminant D>0 and a CM extension F=K(\sqrt{\alpha}) for some totally negative \alpha\in K . We have 2 pairs of embeddings of F\to\mathbb{C} : \sigma_{1},\bar{\sigma}_{1},\sigma_{2},\bar{\sigma}_{2} . Consider (V,h) where V\cong F^{\oplus 3} is an F -vector space equipped with a hermitian form h with signature (2,1) under the first pair of embeddings and definite under the other pair of embeddings. Then the special unitary group \Gamma:=SU(h^{\sigma_{1}},\mathcal{O}_{F}) that preserves h and has entries in \mathcal{O}_{F} gives an arithmetic lattice, and its action on (F\otimes\mathbb{C},h^{\sigma_{1}}) induces a compact arithmetic quotient of \mathbb{B}^{2} . Denote D^{\prime} to be the norm of the relative discriminant of the CM extension F/K . We proved the following. Theorem 1.1 . For any genus g , there exists D_{0}(g)>0 and D^{\prime}_{0}>0 such that for D>D_{0}(g) and D^{\prime}>D^{\prime}_{0} , the corresponding \Gamma\backslash\mathbb{B}^{2} defined as above has no complex curves of genus g . This also has a moduli interpretation in terms of abelian sixfolds with \mathcal{O}_{F} -endomorphism. See Section 5.2 for detail. We say that a skew-Hermitian form T over \mathbb{C} has signature (m,n) if the Hermitian form i^{-1}T has signature (m,n) . Theorem 1.2 . Fix a skew-Hermitian form T over F such that T is of signature (2,1) under a pair of complex embeddings of F and is definite under the other pair. For any complex curve C of genus g , consider abelian schemes f:\mathcal{A}\to C of dimension 6 with \mathcal{O}_{F} -endomorphism of type ( \mathcal{O}_{F}^{3} , T ). Then there exists D_{0}(g)>0 and D^{\prime}_{0}>0 such that for D>D_{0}(g) and D^{\prime}>D^{\prime}_{0} , such type of abelian schemes must be isotrivial. The proof uses the method of volume estimates, which has been used in, for example, [ HT02 , HT12 , BT13 , BT18 , BT19 ] to prove interesting results in algebraic geometry. In contrast to existing literature, we use it to understand the geometry of quotient singularities. Compared to similar objects such as Hilbert modular surfaces, notable new phenomena for arithmetic quotient of 2-balls include the existence of 1-dimensional quotient singularities given by some special curves, and non-cyclic quotient singularities at the intersection of these curves. This requires a volume estimate to give a bound on the intersection number between these stacky special curves and any other complex curves. Compared to the volume estimates in [ Mem23 ] , our work focuses on the geometry near the quotient singularities instead of the boundaries. 2. Background 2.1. Complex Hyperbolic 2-ball In this section we define complex hyperbolic n -balls and summarize useful facts about the n=2 case. See [ Gol99 ] or [ Par03 ] for detail. 2.1.1. The projective model The unit ball in \mathbb{C}^{n} and the Bergman metric, which is a natural metric with constant negative holomorphic sectional curvature, give a model for complex hyperbolic n -space \mathbb{B}^{n} . When n=1 , this is the usual Poincaré unit disc model of the hyperbolic plane. We now give a brief review of \mathbb{B}^{n} , especially the case where n=2 . A Hermitian form on \mathbb{C}^{n} is a \mathbb{C} -valued \mathbb{R} -bilinear form h:\mathbb{C}^{n}\times\mathbb{C}^{n}\to\mathbb{C} such that for any fixed w\in\mathbb{C}^{n} , v\mapsto h(v,w) is \mathbb{C} -linear, and for any v,w\in\mathbb{C}^{n} , h(v,w)=\overline{h(w,v)} . Alternatively, if we fix a basis of \mathbb{C}^{n} , so v,w can be considered as column vectors, then a Hermitian form h is given by (v,w)\mapsto v^{H}J_{h}w for some Hermitian matrix J_{h} . Here the superscript H means the Hermitian conjugate, i.e. the transpose of the element-wise complex conjugation of the matrix, and thus a matrix J is Hermitian if J=J^{H} . Let V be \mathbb{C}^{n+1} equipped with a Hermitian form \langle\cdot,\cdot\rangle of signature (n,1) . One example is the standard Hermitian form h_{std}(u,v)=u_{1}\bar{v}_{1}+\cdots+u_{n}\bar{v}_{n}-u_{n+1}\bar{v}_{n+1} . Since \langle u,u\rangle is real for any u\in V , we can define subsets: negative \displaystyle V_{-} \displaystyle=\{u\in V:\langle u,u\rangle<0\} null \displaystyle V_{0} \displaystyle=\{u\in V\backslash\{0\}:\langle u,u\rangle=0\} positive \displaystyle V_{+} \displaystyle=\{u\in V:\langle u,u\rangle>0\} and for u\in V we denote u<0 (resp. u>0 ) if u\in V_{-} (resp. V_{+} ). Since \langle\lambda u,\lambda u\rangle=|\lambda|^{2}\langle u,u\rangle for any scalar \lambda\in\mathbb{C}^{*} and any u\in V , the projectivization map \mathbb{P}:\mathbb{C}^{n+1}\to\mathbb{P}_{\mathbb{C}}^{n} preserves the sign of \langle u,u\rangle and thus \mathbb{P}V_{-} , \mathbb{P}V_{0} , and \mathbb{P}V_{+} are well-defined. The projective model of complex hyperbolic n -space \mathbb{B}^{n} is defined to be the negative lines in V , i.e. \mathbb{B}^{n}:=\mathbb{P}V_{-} . Thus, its boundary \partial\mathbb{B}^{n}=\mathbb{P}V_{0} . We can define a natural metric on \mathbb{B}^{n} called the Bergman metric. First we denote the “tance” function for any u,v\in V_{-}\cup V_{+} : \mathrm{ta}(u,v):=\frac{\langle u,v\rangle\langle v,u\rangle}{\langle u,u\rangle\langle v,v\rangle}\in\mathbb{R} It has the following useful properties: (i) \mathrm{ta}(u,v)=\mathrm{ta}(v,u) for any u,v\in V_{-}\cup V_{+} ; (ii) \mathrm{ta}(\lambda u,v)=\mathrm{ta}(u,v) for any \lambda\in\mathbb{C}^{*} and u,v\in V_{-}\cup V_{+} , so \mathrm{ta}(\cdot,\cdot) is well-defined on \mathbb{P}V_{-}\cup\mathbb{P}V_{+} ; (iii) \mathrm{ta}(u,v)\geq 1 if u,v\in V_{-} ; \mathrm{ta}(u,v)\geq 0 if u,v\in V_{+} ; \mathrm{ta}(u,v)\leq 0 if u\in V_{+} and v\in V_{-} . For the proof, only \mathrm{ta}(u,v)\geq 1 for u,v\in V_{-} is not trivial. It can be explicitly proved in the case of the standard Hermitian form and then extended to the general case by Cayley transforms which will be defined below. Now the distance function d(\cdot,\cdot) on \mathbb{B}^{n} given by the Bergman metric is defined to be \cosh^{2}\left(\frac{d(u,v)}{2}\right):=\mathrm{ta}(u,v):=\frac{\langle u,v\rangle\langle v,u\rangle}{\langle u,u\rangle\langle v,v\rangle} for u,v\in\mathbb{B}^{n} , or alternatively, ds^{2}=\frac{-4}{\langle u,u\rangle^{2}}\det\begin{pmatrix}\langle u,u\rangle&\langle u,du\rangle\\ \langle du,u\rangle&\langle du,du\rangle\end{pmatrix}. The -4 normalizes its holomorphic sectional curvature to be -1 . The Bergman metric is invariant to the choice of the Hermitian form in the sense that for different choices of Hermitian forms, the complex hyperbolic n -spaces are isometric. More specifically, we have the standard Hermitian form h_{std}(u,v)=u_{1}\bar{v}_{1}+\cdots+u_{n}\bar{v}_{n}-u_{n+1}\bar{v}_{n+1} corresponding to the diagonal matrix J_{std} with diagonal entries 1,\cdots,1,-1 . Then consider J any Hermitian matrix of signature (n,1) , so J can be unitarily diagonalized with real eigenvalues, i.e. we can find a matrix T\in\mathrm{GL}_{n+1}(\mathbb{C}) such that J=T^{H}J_{std}T . Then we have an isometry called Cayley transform (\mathbb{C}^{n+1},J)\to(\mathbb{C}^{n+1},J_{std}),\ \ u\mapsto Tu . When we consider h_{std}(u,v)=u_{1}\bar{v}_{1}+\cdots+u_{n}\bar{v}_{n}-u_{n+1}\bar{v}_{n+1} , any point [u_{1},\cdots,u_{n+1}]\in\mathbb{P}V_{-} corresponds to a point (u_{1}/u_{n+1},\cdots,u_{n}/u_{n+1}) inside the open unit ball in \mathbb{C}^{n} , so we also call \mathbb{B}^{n} the complex hyperbolic n -ball. 2.1.2. Isometries Consider V\cong\mathbb{C}^{n+1} equipped with a Hermitian form h(\cdot,\cdot)=\langle\cdot,\cdot\rangle of signature (n,1) . If we fix a basis of V , the Hermitian form is represented by a Hermitian matrix J , and linear transformations M on V preserving the Hermitian form satisfy: \langle Mu,Mv\rangle=v^{H}M^{H}JMu=\langle u,v\rangle=v^{H}Ju. for any u,v\in V , so it is equivalent to requiring M^{H}JM=J . Thus they define the unitary group: U(J):=\{M\in M_{n+1}(\mathbb{C}):\langle Mu,Mv\rangle=\langle u,v\rangle\ \ \forall u,v\in V\}\cong U(n,1). The last isomorphism is induced by Cayley transform T : U(J)\to U(J_{std})=:U(n,1),\ \ M\mapsto TMT^{-1} . Since \det(M^{H}JM)=\det(J) , we also have |\det(M)|=1 . Since scalar multiplication \lambda I_{n+1} for any \lambda\in\mathbb{C}^{*} commutes

Read the complete text