Divisibility of the coefficients of modular polynomials

Basic Information

• Paper ID: 2509.06423
• Title: Divisibility of the coefficients of modular polynomials
• Author: Florian Breuer
• Classification: math.NT (Number Theory)
• Publication Date: September 8, 2025 (arXiv preprint, revised October 16, 2025)
• Paper Link: https://arxiv.org/abs/2509.06423

Abstract

Let $N>1$, and let $\Phi_N(X,Y)\in\mathbb{Z}[X,Y]$ be the modular polynomial, which vanishes at pairs $(j_1,j_2)$ of $j$-invariants of elliptic curves if and only if these elliptic curves are connected by a cyclic isogeny of degree $N$. This paper investigates the divisibility properties of coefficients of the polynomial $\Phi_N(X+J, Y+J)$ for specific algebraic numbers $J$ (particularly $J=0$ and other singular moduli). The study reveals that these coefficients exhibit high divisibility at small primes where $J$ is supersingular.

Research Background and Motivation

1. Problem Background: Modular polynomials $\Phi_N(X,Y)$ are important objects in number theory and algebraic geometry, encoding information about isogeny relations between elliptic curves. Although the coefficients of these polynomials are typically very large, they also exhibit high divisibility by small primes.
2. Practical Value:
   • Important applications in cryptography and computational number theory
   • Useful for studying reduction types of elliptic curves
   • Helpful for optimizing algorithms based on the Chinese Remainder Theorem
3. Limitations of Existing Research: Wang proposed conjectures for $p \leq 5$ in 23, but complete proofs and more general results were lacking.
4. Research Motivation:
   • Provide lower bounds for divisibility of modular polynomial coefficients
   • Extend to the general case of singular moduli
   • Provide theoretical support for computational algorithms

Core Contributions

1. Main Theorem 1.1: For the modular polynomial $\Phi_N(X,Y) = \sum_{0\leq i,j\leq\psi(N)} a_{i,j}X^iY^j$, when $i+j < \psi(N)$, precise lower bounds for the $p$-adic valuation of coefficients $a_{i,j}$ are provided.
2. Main Theorem 1.2: The results are extended to the case of rational singular moduli $J \in \mathbb{Z}$, proving that coefficients of $\Phi_N(X+J, Y+J)$ exhibit stronger divisibility properties.
3. Computational Optimization: Through predicted divisibility properties, the space required to store modular polynomial coefficients can be significantly reduced (e.g., 43% savings for $N=5$).
4. Theoretical Framework: A general framework connecting deformation theory of elliptic curves with divisibility of modular polynomial coefficients is established.

Methodology Details

Problem Definition

Study lower bounds for the $p$-adic valuation $v_p(a_{i,j})$ of coefficients of $\Phi_N(X+J, Y+J)$ at prime $p$, where $J$ is a specific algebraic number (particularly singular moduli).

Core Technical Methods

1. Interpolation Lemma (Lemma 3.3)

A key interpolation lemma is constructed that, through properties of Vandermonde determinants, relates the valuation of polynomial coefficients to the valuation of function values.

2. Elliptic Curve Deformation Theory

• Supersingular Case: Utilizes Lubin-Tate theory to parametrize deformations of $p$-divisible groups through one-parameter families $\Gamma(t)$
• General Case: Uses Serre-Tate parameters $q \in 1 + \pi A$ to parametrize deformations of elliptic curves

3. Main Technical Tool (Proposition 3.4)

Constructs families of elliptic curves $E_k/K$ satisfying specific conditions:

• Each $E_k$ has good reduction
• $v(j(E_k)-J) = v(j(E_k)-j(E_l)) = n$ for all $k \neq l$
• Appropriate valuation conditions for elliptic curves $\tilde{E}_k$ connected to $E_k$ by cyclic isogenies of degree $N$

Technical Innovations

1. Unified Framework: Wang's conjectures and the case of singular moduli are unified within a single theoretical framework.
2. Precise Bounds: Not only existence results are provided, but precise numerical bounds are given, and these bounds are proven to be optimal in most cases.
3. Computational Feasibility: The method is computationally feasible and can be practically applied to modular polynomial computation and storage optimization.

Experimental Setup

Data Sources

• Global minimal models of elliptic curves obtained from the online database LMFDB
• Verification of theoretical results using Sutherland's modular polynomial database

Verification Methods

• Computational verification for $N \leq 400$ and many large primes $N$
• Special verification of 13 rational singular moduli cases

Computational Examples

The paper provides complete coefficient factorization of $\Phi_5(X,Y)$ for $N=5$, demonstrating the divisibility properties predicted by the theory.

Experimental Results

Main Results Verification

1. Verification of Theorem 1.1:
   • For $p=2,3,5$: Verified inequalities such as $v_2(a_{i,j}) \geq \frac{1}{5}(\psi(N)-i-j)$ respectively
   • For $p \geq 11, p \equiv 2 \pmod{3}$: Verified more general bounds
2. Storage Optimization Effects:
   • $N=5$: Reduced from 523 digits to 298 digits, saving 43%
   • $N=101$: Reduced from 6,383,216 digits to 5,606,370 digits, saving 12%
3. Special Case Analysis:
   • Table 1 lists exceptional valuation cases for 13 rational singular moduli
   • Table 2 shows complete factorization of coefficients of $\Phi_5(X,Y)$

Accuracy of Theoretical Predictions

Computations show that the values of $n_p$ given in Theorem 1.2 are optimal in most cases, with room for improvement only in the cases $D=-12$ and $D=-27$.

Related Work

1. Gross-Zagier Theory: This paper is a variant of the theory of differences of singular moduli, inheriting the research tradition pioneered by Gross-Zagier.
2. Wang's Work: Directly responds to and proves Wang's conjectures for $p \leq 5$ in 23.
3. Computational Aspects: Related to research on modular polynomial computation algorithms by Bröker-Sutherland and others, providing theoretical support for CRT-based algorithms.
4. Elliptic Curve Reduction Theory: Closely related to research on reduction types of elliptic curves, particularly applications in 24.

Conclusions and Discussion

Main Conclusions

1. Proves that modular polynomial coefficients exhibit expected high divisibility
2. Establishes complete theory for the case of singular moduli
3. Provides theoretical foundation for computational optimization

Limitations

1. Cases with $p=2,3$ require more complex technical treatment
2. Certain special discriminants ($D=-12,-27$) still have room for improvement
3. Storage optimization effects diminish for large $N$

Future Directions

1. Further optimize bounds for $D=-12,-27$ cases
2. Extend to non-rational singular moduli
3. Develop more efficient partial factorization storage methods

In-Depth Evaluation

Strengths

1. Theoretical Depth: Cleverly combines deformation theory of elliptic curves, modular form theory, and $p$-adic analysis
2. Practical Value: Provides practical optimization schemes for modular polynomial computation and storage
3. Completeness: Not only proves existence but provides precise numerical bounds
4. Sufficient Verification: Theoretical results are verified through extensive computational examples

Weaknesses

1. Technical Complexity: Cases with $p=2,3$ require additional technical tools, increasing proof complexity
2. Optimization Effects: Storage optimization effects are limited for large $N$
3. Special Cases: Certain discriminant cases have not yet achieved optimal bounds

Impact

1. Theoretical Contribution: Provides new perspectives and tools for modular polynomial theory
2. Computational Applications: Significant importance for elliptic curve cryptography and isogeny computation
3. Methodological Innovation: Application of deformation theory to modular polynomial research is illuminating

Applicable Scenarios

1. Isogeny computation in elliptic curve cryptography
2. Efficient storage and transmission of modular polynomials
3. Optimization of CRT-based modular polynomial computation algorithms

References

The paper cites 24 related references covering important works in multiple fields including modular form theory, elliptic curve theory, quaternion algebras, and computational number theory, reflecting the interdisciplinary nature and theoretical depth of the research.