🔢 mathematics

Algebraic Lattices Arising from Congruence Submodules in Subfields of $p$ -th Cyclotomic Fields

This paper presents new constructions of algebraic lattices derived from congruence submodules within subfields of $p$ -th cyclotomic fields, computing their center density lower bounds and demonstrating that they achieve the best known packing densities in dimensions 2, 3, and 5.

Original authors: Trajano Pires da Nóbrega Neto, Antonio Aparecido de Andrade, Jéfferson Luiz Rocha Bastos, Robson Ricardo de Araujo, José Carmelo Interlando

Published 2026-08-19

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Trajano Pires da Nóbrega Neto, Antonio Aparecido de Andrade, Jéfferson Luiz Rocha Bastos, Robson Ricardo de Araujo, José Carmelo Interlando

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine trying to pack as many identical oranges as possible into a large crate without any of them overlapping. This is the essence of the sphere packing problem, a puzzle that has challenged mathematicians for centuries. While it is easy to see how to stack oranges in a grocery store, finding the absolute most efficient way to do this in higher, abstract dimensions is incredibly difficult. The answer matters far beyond fruit stands; in the world of telecommunications, these packing patterns determine how much data can be sent clearly over a noisy connection. If the "oranges" (which represent signals) are packed too loosely, the message gets lost. If they are packed too tightly, they crash into each other and distort. The goal is to find the perfect arrangement where the signals are as close as possible without touching, maximizing the amount of information that can travel through the air.

To solve this, researchers often look at structures called lattices. Think of a lattice as a grid of points in space, like the corners of a vast, invisible box. If you place a sphere at every single point of this grid, you create a sphere packing. The density of this packing depends entirely on how the grid is shaped. For a long time, mathematicians have known the best ways to pack spheres in just a few specific dimensions, but for most others, the answer remains a mystery. A powerful way to build these grids comes from a branch of mathematics called algebraic number theory, which uses special numbers to create highly structured patterns. These patterns are often derived from cyclotomic fields, which are complex number systems built from the roots of unity, essentially the solutions to equations that describe rotation.

In this new work, a team of researchers has developed a fresh method for constructing these lattices, specifically focusing on a family of subfields within the p-th cyclotomic fields, where p is a prime number. Instead of guessing or searching randomly, the authors created a systematic recipe to build these grids using what are known as congruence submodules. In plain terms, they took the ring of integers from a specific number field and carved out smaller, highly organized subsets based on rules of divisibility by a prime number. They used the geometry of finite vector spaces—essentially grids of numbers that wrap around after reaching a certain limit—to define these subsets. By carefully selecting these subsets, they ensured that the resulting lattices had very specific, desirable properties regarding how tightly they could pack spheres.

The researchers did not just build these lattices; they calculated exactly how dense they could be. They derived a formula to determine the "center density," a measure of how much of the space is actually filled by the spheres. Their calculations showed that for certain dimensions, specifically two, three, and five, the lattices they built achieve the absolute best packing density currently known to science. In dimension two, their construction recreates the famous hexagonal pattern, which is the most efficient way to pack circles on a flat surface. In dimension three, it matches the densest known packing of spheres in space. In dimension five, it reaches the highest density that mathematicians have been able to find so far. This is a significant achievement because it proves that their algebraic method can recover the optimal solutions in these low dimensions without needing to rely on trial and error.

For dimensions higher than five, the paper does not claim to have found the absolute best possible packing, as those answers are still unknown. Instead, the authors provide a reliable lower bound, which is a guaranteed minimum level of efficiency for their new lattices. They showed that for any odd dimension, there is an infinite family of these lattices that can be constructed, offering a vast new toolkit for engineers and mathematicians. The work establishes a solid theoretical foundation, proving that these structures exist and calculating their properties with precision. While the paper stops short of claiming these are the best possible arrangements for every single dimension, it successfully demonstrates that this algebraic approach is a powerful way to generate high-quality lattices, particularly in the dimensions that are most relevant for current communication technologies. The result is a clearer path forward for designing better signal constellations, ensuring that data can be transmitted more efficiently and reliably across the noisy channels of the modern world.

Technical Summary: Algebraic Lattices Arising from Congruence Submodules in Subfields of p-th Cyclotomic Fields

Problem Statement
The classical sphere packing problem seeks to determine the maximum density with which identical non-overlapping $n$ -dimensional spheres can be packed in Euclidean space $\mathbb{R}^n$ . While optimal solutions are known only for dimensions 1, 2, 3, 8, and 24, finding dense lattice packings remains a central challenge in mathematics and telecommunications, particularly for designing signal constellations for Gaussian and Rayleigh channels. Algebraic number theory, specifically through the use of rings of integers in algebraic number fields, provides a structured framework for constructing such lattices via canonical embeddings. This work addresses the need for new explicit constructions of algebraic lattices that achieve high packing densities, particularly in low dimensions, by leveraging congruence submodules within subfields of $p$ -th cyclotomic fields.

Methodology
The authors construct algebraic lattices from specific families of $\mathbb{Z}$ -modules within the ring of algebraic integers $\mathcal{O}_K$ of a subfield $K \subseteq \mathbb{Q}(\zeta_p)$ , where $p$ is a prime and $[K:\mathbb{Q}] = s$ . The methodology proceeds through the following steps:

Field Characterization: The paper establishes that for every divisor $s$ of $p-1$ , there exists a unique cyclic subfield $K$ of the $p$ -th cyclotomic field. The ring of integers $\mathcal{O}_K$ is shown to possess an integral normal basis generated by the trace of a primitive root of unity, $t = \text{Tr}_{L/K}(\zeta_p)$ , and its Galois conjugates.
Trace Form Analysis: The authors derive a closed-form expression for the trace form $\text{Tr}_K(xx)$ restricted to elements $x \in \mathcal{O}_K$ . This expression, $\text{Tr}_K(xx) = p \sum a_i^2 - \frac{p-1}{s} (\sum a_i)^2$ , is critical for determining the minimum norm of lattice vectors and, consequently, the packing radius.
Congruence Submodule Construction: Two parametric families of $\mathbb{Z}$ $Z$ -submodules, denoted $M_p$ $M_{p}$ and $\tilde{M}_p$ $\tilde{M}_{p}$ , are defined within $\mathcal{O}_K$ $O_{K}$ . These modules are constructed by imposing orthogonality conditions on the coefficient vectors of elements in $\mathcal{O}_K$ $O_{K}$ relative to specific vectors in the finite vector space $\mathbb{F}_p^s$ $F_{p}^{s}$ . These vectors are derived from powers of a primitive $s$ $s$ -th root of unity modulo $p$ $p$ .
- $M_p$ is defined by the condition that the coefficient vector is orthogonal to a subspace $W$ spanned by specific powers of the primitive root.
- $\tilde{M}_p$ is a further refinement of $M_p$ imposing an additional parity condition on the sum of coefficients.
Index Computation: Using algebraic properties of submodules and the structure of the finite field $\mathbb{F}_p$ , the authors rigorously compute the algebraic indices $[\mathcal{O}_K : M_p]$ and $[\mathcal{O}_K : \tilde{M}_p]$ .
Density Estimation: By combining the computed indices with the minimum values of the trace form on these submodules, the authors derive lower bounds for the center density $\delta(\sigma(M))$ of the resulting lattices under the canonical embedding.

Key Contributions and Results

Theoretical Framework: The paper provides a complete theoretical foundation for constructing algebraic lattices from congruence submodules in subfields of cyclotomic fields, generalizing previous computational approaches.
Explicit Constructions: The authors present explicit constructions for two families of lattices, $\sigma(M_p)$ and $\sigma(\tilde{M}_p)$ , parameterized by the prime $p$ and the dimension $s$ .
Optimality in Low Dimensions:
- Dimension 2 ( $s=2$ ): For $p=3$ , the constructed lattice $\sigma(M_p)$ attains the optimal center density of $\frac{1}{2\sqrt{3}}$ , corresponding to the hexagonal lattice.
- Dimension 3 ( $s=3$ ): For primes $p \equiv 1 \pmod 3$ (e.g., $p=7$ ), the lattice $\sigma(\tilde{M}_p)$ attains the optimal center density of $\frac{1}{4\sqrt{2}}$ .
- Dimension 5 ( $s=5$ ): For primes $p \equiv 1 \pmod 5$ (e.g., $p=11$ ), the lattice $\sigma(\tilde{M}_p)$ achieves the best known center density of $\frac{1}{8\sqrt{2}}$ .
General Bounds: For general dimensions $s$ , the paper establishes lower bounds for the center density. Specifically, for odd $s$ , the bound is $\delta \geq \frac{1}{2^{(s+2)/2}}$ , and for even $s$ , bounds involving $p$ are derived (e.g., $\delta \geq \frac{1}{2^s\sqrt{p}}$ or $\frac{1}{2^{s/2}\sqrt{p}}$ depending on parity conditions).
Infinite Families: By invoking Dirichlet's Theorem on arithmetic progressions, the authors demonstrate that for every odd dimension $s$ , there exist infinitely many primes $p$ yielding an infinite family of such lattices.

Significance and Scope
The significance of this work lies in its ability to systematically recover the densest known lattice packings in dimensions 2, 3, and 5 using a unified algebraic construction based on congruence submodules. The paper emphasizes that while the construction successfully recovers optimal configurations in these specific low dimensions, the results for general dimensions $s > 5$ serve primarily as theoretical lower bounds and construction methods. The authors explicitly state that they do not claim the optimality of these lattices for arbitrary dimensions $s$ , nor do they analyze the performance of these lattices for $s \notin \{2, 3, 5\}$ . The work lays the groundwork for future research into evaluating these constellations over communication channels and extending the orthogonality-based construction to composite cyclotomic fields.

More like this