← Back to arXiv
arXivNumber TheoryarXiv:2608.17056

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

The sphere packing problem asks how efficiently identical spheres can be arranged in space so they take up as much room as possible. One powerful approach to this problem uses mathematical objects called lattices, which are regular, grid-like arrangements of points in space. When the centers of packed spheres form a lattice, the problem becomes more tractable because the regular structure allows precise mathematical analysis. Researchers can build useful lattices from number theory, specifically by taking sets of algebraic integers from special number systems called cyclotomic fields, which are built by extending the ordinary rational numbers using roots of unity.

The paper focuses on a particular family of these number systems, specifically subfields of cyclotomic fields associated with prime numbers. Within these fields, the authors construct specific collections of algebraic integers called congruence submodules, essentially sets of numbers satisfying certain divisibility-like conditions. By mapping these collections into ordinary geometric space through a standard mathematical procedure called the canonical embedding, they obtain concrete lattices whose geometric properties, particularly how densely the associated spheres can pack, can be carefully computed.

The main results are new lattice constructions that achieve provably good sphere packing densities. The authors calculate lower bounds on a key measurement called center density, which captures how efficiently the lattice packs spheres. Notably, their constructions recover the best currently known packing arrangements in two, three, and five dimensions. While optimal packing is only fully proven in dimensions two, three, and eight (and recently twenty-four), achieving best-known results in dimension five is meaningful progress, and the algebraic framework provides a unified, systematic way to generate competitive lattice packings across multiple dimensions.

Read original →