[Paper Review] A complete isometry classification of 3-dimensional lattices

Vitaliy Kurlin|arXiv (Cornell University)|Jan 25, 2022

Silicone and Siloxane Chemistry4 citations

TL;DR

This paper presents a complete, continuous isometry classification of 3-dimensional lattices using a novel set of homogeneous invariants derived from square roots of scalar products of four vectors summing to zero with non-acute pairwise angles. The root invariants provide a bijective, continuous, and computable parameterization of the Lattice Isometry Space (LIS), resolving long-standing discontinuities in prior classification methods and enabling robust geometric comparison of crystal lattices under perturbations.

ABSTRACT

A periodic lattice in Euclidean 3-space is the infinite set of all integer linear combinations of basis vectors. Any lattice can be generated by infinitely many different bases. This ambiguity was only partially resolved, but standard reductions remained discontinuous under perturbations modelling crystal vibrations. This paper completes a continuous classification of 3-dimensional lattices up to Euclidean isometry (or congruence) and similarity (with uniform scaling).The new homogeneous invariants are uniquely ordered square roots of scalar products of four superbase vectors whose sum is zero and all pairwise angles are non-acute. These root invariants continuously change under perturbations of basis vectors. The geometric methods extend the work of Delone, Conway and Sloane.

Motivation & Objective

• To resolve the discontinuity problem in existing lattice classification methods under perturbations modeling crystal vibrations.
• To provide a complete and continuous invariant mapping from the Lattice Isometry Space (LIS) in R³ to a simpler parameter space.
• To ensure the invariant is computable from any lattice basis and allows explicit reconstruction of the lattice.
• To extend the continuous classification framework from 2D lattices (previously solved) to 3D lattices, which presents significantly greater geometric complexity.
• To establish a foundation for continuous metrics on lattices and future work on orientation-aware equivalences and similarity classifications.

Proposed method

• Define an obtuse superbase of four vectors in R³ summing to zero, with all pairwise scalar products non-positive (non-acute angles).
• Construct root invariants as the square roots of the absolute values of these scalar products, forming a set of up to six parameters.
• Order the root invariants to ensure uniqueness and continuity under basis perturbations in the Minkowski metric.
• Use column sums of conorms (squared distances) in a coform matrix to derive invariants that are invariant under index permutations up to equivalence.
• Prove that the root invariants form a complete and continuous invariant by showing they distinguish non-isometric lattices even under small perturbations.
• Leverage geometric techniques inspired by Delone, Conway, and Sloane, extending them to ensure continuity and completeness in 3D.

Experimental results

Research questions

• RQ1 Can a complete and continuous isometry classification of 3D lattices be achieved, avoiding the discontinuities of prior methods?
• RQ2 Do the proposed root invariants uniquely determine lattice isometry classes and vary continuously under basis perturbations?
• RQ3 Can the invariants be computed explicitly from any lattice basis and used to reconstruct the original lattice?
• RQ4 How do the root invariants distinguish lattices that have identical distance distributions to nearest neighbors?
• RQ5 Can the invariants be extended to define continuous metrics on the Lattice Isometry Space for applications in crystallography and materials science?

Key findings

• The root invariants—defined as ordered square roots of scalar products of four vectors summing to zero with non-acute angles—form a complete and continuous invariant for 3D lattices up to isometry.
• The invariants are independent of the choice of basis and remain unchanged under Euclidean isometries, satisfying invariance and completeness (no false positives or negatives).
• The method resolves the discontinuity issue in earlier lattice classification schemes, such as those based on conorms or Voronoi types, by ensuring continuous variation under perturbations.
• Even when two lattices have identical distance distributions to their seven nearest neighbors (DC⁷ = 0), the root invariants can distinguish them if their conorm column sums differ under index permutations.
• The root invariants are computable from any basis and allow explicit reconstruction of the lattice, satisfying the inverse design condition.
• The framework provides a foundation for future continuous metrics on LIS(R³), enabling the definition of chiralities and real-valued measures of symmetry deviation in crystal lattices.