Distribution of toric periods of modular forms on definite quaternion algebras

TL;DR

This paper investigates the distribution of toric periods of algebraic modular forms on definite quaternion algebras, demonstrating non-vanishing for many cases and infinitely many sign changes, with numerical evidence and conjectures.

Contribution

It proves non-vanishing of toric periods for a positive proportion of imaginary quadratic fields and establishes infinitely many sign changes, advancing understanding of modular form coefficients.

Findings

Non-vanishing of toric periods for positive proportion of imaginary quadratic fields

Infinitely many sign changes in toric period sequences

Numerical experiments leading to new conjectures

Abstract

Let $D$ be a definite quaternion algebra over $\mathbb{Q}$ and $\mathcal{O}$ an Eichler order in $D$ of square-free level. We study distribution of the toric periods of algebraic modular forms of level $\mathcal{O}$. We focus on two problems: non-vanishing and sign changes. Firstly, under certain conditions on $\mathcal{O}$, we prove the non-vanishing of the toric periods for positive proportion of imaginary quadratic fields. This improves the known lower bounds toward Goldfeld's conjecture in some cases and provides evidence for similar non-vanishing conjectures for central values of twisted automorphic $L$-functions. Secondly, we show that the sequence of toric periods has infinitely many sign changes. This proves the sign changes of the Fourier coefficients $\{a(n)\}_n$ of weight 3/2 modular forms, where $n$ ranges over fundamental discriminants. In the final section, we present numerical experiments in some cases and formulate several conjectures based on them.