Universal quadratic forms and Dedekind zeta functions

Basic Information

• Paper ID: 2311.12911
• Title: Universal quadratic forms and Dedekind zeta functions
• Authors: Vítězslav Kala, Mentzelos Melistas
• Classification: math.NT (Number Theory)
• Publication Date: November 21, 2023
• Paper Link: https://arxiv.org/abs/2311.12911
• Funding: Czech Science Foundation (GAČR) grant 21-00420M

Abstract

This paper investigates universal quadratic forms (quadratic forms that represent all natural numbers) on totally real number fields using Dedekind zeta functions. Specifically, under the assumption that the codifferent is generated by totally positive elements, the authors establish explicit upper bounds for the rank of universal quadratic forms over a given number field K. To remove this assumption, the authors also study the problem of determining the minimal number of generators for the totally positive part of ideals in totally real fields.

Research Background and Motivation

Research Problem

Universal quadratic forms are positive definite quadratic forms that can represent all natural numbers. The core question addressed in this paper is: On a totally real number field K, what is the minimum number of variables (i.e., the lower bound on the rank) required for a universal quadratic form?

Importance of the Problem

1. Historical Significance: The study of universal quadratic forms involves work by Diophantus, Fermat, Euler, Gauss, and others. Modern important results include the Conway-Schneeberger 15-theorem and Bhargava-Hanke's 290-theorem.
2. Natural Generalization: Extending from the ring of integers ℤ to the ring of integers O_K of a totally real number field K is a natural problem in number theory.
3. Known Results:
   • Maaß (1941): The sum of three squares is universal over ℚ(√5)
   • Siegel (1945): If arbitrarily many sums of squares are universal over K, then K = ℚ or ℚ(√5)
   • Hsia-Kitaoka-Kneser: Universal quadratic forms exist over every number field
   • Blomer-Kala: For any positive integer N, there exists a real quadratic field such that any universal quadratic form has rank at least N

Limitations of Existing Methods

Although the existence of universal quadratic forms is known, finding the minimum number of variables required for universal quadratic forms over a given number field remains an open and difficult problem. Existing methods lack precise characterization for specific number fields.

Research Motivation

1. Extract more precise information about the rank of universal quadratic forms using Dedekind zeta functions
2. Generalize techniques of Yatsyna and Kala-Yatsyna
3. Explore pathways to remove the technical assumption that "the codifferent is generated by totally positive elements"

Core Contributions

1. Theorem 1.1 (Main Result): Under the assumption that the codifferent O^∨_K is generated by a totally positive element δ, an explicit lower bound for the rank R of universal quadratic forms is provided, depending on the discriminant Δ_K and degree d.
2. Theorem 1.2 (Finiteness Result): For a given degree d and rank R, there exist only finitely many totally real number fields (whose codifferent is generated by totally positive elements) on which there exists a universal quadratic form of rank R, generalizing Yatsyna's Theorem 4.
3. Theorem 2.4 (Lifting Problem): For number fields of degree d ≤ 43 where O^∨_K is a principal ideal satisfying certain conditions, explicit upper bounds on the discriminant are provided.
4. Study of Ideal Positive Parts:
   • Define κ(I) and κ(K) to characterize the minimal number of generators for the totally positive part of ideals
   • Prove the finiteness of κ(K) (Theorem 3.6)
   • Provide explicit bounds related to continued fractions for real quadratic fields (Proposition 3.9)

Detailed Methodology

Task Definition

Input: Totally real number field K, degree d, discriminant Δ_K, rank R of universal quadratic form Q Output: Lower bound inequality that R must satisfy Constraints: K's codifferent O^∨_K = (δ), where δ is totally positive

Core Technical Framework

1. Application of Siegel's Formula

Siegel's formula (Theorem 2.2) establishes the connection between special values of Dedekind zeta functions and arithmetic quantities:

$\zeta_K(-1) = 2^d \sum_{\ell=1}^{r_d} b_\ell(2d) s_\ell^K(2)$

where:

• $r_d = \lfloor d/6 \rfloor$ (when d ≡ 1 mod 6) or $\lfloor d/6 \rfloor + 1$ (otherwise)
• $s_\ell^K(2) = \sum_{\gamma \in O^{\vee,+}_K, \text{Tr}(\gamma)=\ell} \sigma((\gamma)(O^{\vee}_K)^{-1})$
• $b_\ell(2d)$ are rational numbers depending only on d

2. Functional Equation

Utilizing the functional equation of the Dedekind zeta function: $\zeta_K(-1) = (-1)^d |\Delta_K|^{3/2} \left(\frac{1}{4\pi}\right)^d \zeta_K(2)$

3. Short Vector Counting

Using the Regev-Stephens-Davidowitz theorem (Theorem 2.1): For a rank R positive definite ℤ-lattice L, the number of vectors with norm ≤ i satisfies: $N_{\leq i}(L) \leq C(R,i) = 2\binom{R+4i-1}{4i-1} - 1$

4. Key Lemma (Lemma 2.3)

Under the assumption O^∨_K = (δ), for γ ∈ O^∨,+_K with Tr(γ) = ℓ: $\sigma((\gamma)(O^{\vee}_K)^{-1}) \leq g(\ell, \Delta_K)$

where $g(\ell, \Delta_K)$ involves Robin's bounds on the divisor function.

Proof Strategy (Theorem 1.1)

1. Lattice Construction: Given an integral basis ω_1,...,ω_d of O_K, construct a rank Rd quadratic form: $q(x_{11},...,x_{Rd}) = \text{Tr}(\delta Q(x_{11}\omega_1 + \cdots + x_{Rd}\omega_d))$
2. Short Vector Estimation:
   • By universality, q represents all elements in O^∨,+_K with trace ≤ r_d
   • Apply Theorem 2.1 to obtain an upper bound on the number of short vectors
3. Establish Inequality Chain: $C(Rd, r_d) \geq \#\{\gamma \in O^{+,\vee}_K : \text{Tr}(\gamma) \leq r_d\}$
4. Apply Siegel's Formula: Connect the right-hand side to ζ_K(-1) via Lemma 2.3 and Siegel's formula
5. Apply Functional Equation: Finally obtain an explicit inequality involving the discriminant and ζ_K(2)

Technical Innovations

1. Generalization of Kala-Yatsyna Technique: Extend the proof technique from 9, Theorem 5.1 to more general settings
2. Refined Short Vector Analysis: Cleverly combine short vector counting from lattice theory with arithmetic of number fields
3. Systematic Study of Ideal Positive Parts:
   • Introduce the concept of I-indecomposables
   • Define κ(I) and κ(K) to quantify the number of generators
   • Use narrow class groups to prove finiteness
4. Continued Fraction Method: For real quadratic fields, utilize continued fraction expansions to provide explicit bounds on κ(I) (Proposition 3.9)

Experimental Setup

Dataset

This is a pure theoretical mathematics paper with no traditional experiments or datasets.

Computational Verification

The paper provides concrete examples:

• Example 3.7: ℚ(√2) has narrow class number 1, therefore κ(ℚ(√2)) = 1
• Example 3.10: Detailed calculation proves κ(ℚ(√3)) = 2

Mathematical Tools

• LMFDB Database: Used to query fundamental invariants of number fields (class numbers, fundamental units, etc.)
• Continued Fraction Expansion: Used for computations in real quadratic fields

Experimental Results

Main Theoretical Results

Explicit Inequality from Theorem 1.1

For a universal quadratic form of rank R: $2\binom{Rd + 4r_d - 1}{4r_d - 1} - 1 > \frac{G(\Delta_K)}{B(d)2^d} |\Delta_K|^{3/2} \left(\frac{1}{4\pi}\right)^d$

where:

• $G(\Delta_K) = \min\{1/g(\ell, \Delta_K) : \ell \leq r_d\}$
• $B(d)$ involves the coefficients $b_\ell(2d)$ from Siegel's formula

Significance: The right-hand side grows with |Δ_K|, so larger discriminants require larger rank R.

Finiteness from Theorem 1.2

For fixed d and R, there are only finitely many number fields satisfying the conditions.

Proof Strategy:

• From Theorem 1.1, the discriminant has an upper bound
• Apply the Hermite-Minkowski theorem: given degree and discriminant bound, only finitely many number fields exist

Discriminant Bound from Theorem 2.4

For d ≤ 43 and number fields satisfying certain conditions, if a universal ℤ-type exists: $|\Delta_K| < |b_{r_d}(2d)(4\pi^2)^{dd}|^{2/3}$

Results on Ideal Positive Part Generators

Propositions 3.5 and Theorem 3.6

• For any ideal I, κ(I) is finite
• κ(K) = max_I κ(I) is finite (using finiteness of narrow class group)

Proposition 3.9 (Real Quadratic Fields)

For K = ℚ(√D), let ξ_D = u_0, u_1,..., u_s be the continued fraction expansion, then: