Bravais-Lattice Test: Criteria and Applications

Updated 11 July 2026

Bravais-Lattice Test is a collection of diagnostic criteria that evaluate a lattice’s local stability, optimality at fixed density, symmetry classification, and dynamic compatibility.
These tests employ methodologies such as radial interaction energies, theta functions, Hessian analyses, and metric reduction (in G⁶ or S⁶) to yield concrete measures like neighbor distances and pressure.
Applications range from confirming triangular lattice optimality in 2D and cubic lattice stability via Hessian tests to error-stable lattice classification in crystallography and dynamic transport diagnostics.

“Bravais-Lattice Test” is best understood as an Editor’s term for a family of criteria that decide whether a Bravais lattice satisfies a prescribed property: local stability under finite perturbations, global optimality at fixed density, membership in a Bravais symmetry class, or compatibility with a transport or deformation mechanism. In the literature, such tests are formulated through radial interaction energies, theta functions, Hessians on fixed-density lattice manifolds, metric-tensor reduction in $G^6$ or Selling space $S^6$ , and rank-one or dynamical signatures (Bétermin, 2015, Bétermin, 2015, Bétermin, 2016, Andrews et al., 2012, Andrews et al., 2023, Mukhopadhyay et al., 2019, Ball, 2023).

1. Formal settings and principal meanings

A Bravais lattice in $d$ dimensions is written as

$L=\bigoplus_{i=1}^d \mathbb Z u_i$

or, equivalently, $L=A\mathbb Z^d$ for an invertible matrix $A$ . In fixed-density problems the relevant scalar is the covolume $|\det A|$ , with density $1/|\det A|$ . In crystallographic determination, the same object is represented by a Gram matrix $G$ , by a $G^6$ metric vector

$S^6$ 0

or by six Selling scalars in a Cartesian space $S^6$ 1 (Bétermin, 2015, Bétermin et al., 2017, Andrews et al., 2012, Andrews et al., 2023).

The phrase covers several non-equivalent tasks. In some works it means a sufficient condition for local minimality of a given lattice under a family of radial potentials; in others it means a criterion for triangular optimality among two-dimensional Bravais lattices; in crystallography it means a reduction-theoretic decision rule for Bravais type; and in dynamics or mechanics it means using transport or rank-one compatibility to infer lattice geometry or deformation modes (Bétermin, 2015, Bétermin, 2015, Oishi-Tomiyasu, 2023, Mukhopadhyay et al., 2019, Ball, 2023).

Context	Tested object	Representative criterion
Energy stability	$S^6$ 2 under finite perturbations	small pressure, fast decay, local convexity
Fixed-density optimality	$S^6$ 3 or $S^6$ 4	complete monotonicity or nonnegative theta kernel
Cubic local optimality	Hessian of $S^6$ 5 at SC, FCC, BCC	positive definiteness or saddle structure
Bravais determination	reduced metric in $S^6$ 6 or $S^6$ 7	distance to boundary polytopes or Delone-type projectors
Dynamics and mechanics	current direction or interface compatibility	ballistic attractors or rank-one reflection criteria

2. Compact local minimality as an energy-stability test

For radial interactions, one formulation starts from a $S^6$ 8-admissible potential $S^6$ 9 whose lattice sums for $d$ 0, $d$ 1, and $d$ 2 are absolutely convergent. The energy per particle is

$d$ 3

and the relevant perturbations are not full lattice deformations but $d$ 4-compact perturbations of a finite set $d$ 5. If $d$ 6, the energy variation is

$d$ 7

A lattice is an $d$ 8-compact local minimum if every perturbation of at most $d$ 9 particles, for sufficiently small $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 0, satisfies $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 1 (Bétermin, 2015).

The decisive quantities are the first-neighbor distance $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 2, the second-neighbor distance $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 3, and the pressure

$L=\bigoplus_{i=1}^d \mathbb Z u_i$ 4

The sufficient condition is expressed through an $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 5-family $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 6 satisfying three hypotheses: a small-pressure bound $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 7, fast decay

$L=\bigoplus_{i=1}^d \mathbb Z u_i$ 8

and uniform local convexity $L=\bigoplus_{i=1}^d \mathbb Z u_i$ 9 near $L=A\mathbb Z^d$ 0. Under these assumptions, for every $L=A\mathbb Z^d$ 1 there exists $L=A\mathbb Z^d$ 2 such that $L=A\mathbb Z^d$ 3 is an $L=A\mathbb Z^d$ 4-compact local minimum for every $L=A\mathbb Z^d$ 5, and the maximal perturbation radius can be taken as $L=A\mathbb Z^d$ 6. The proof is an explicit first- and second-variation estimate,

$L=A\mathbb Z^d$ 7

so nearest-neighbor convexity dominates long-range corrections for small $L=A\mathbb Z^d$ 8 (Bétermin, 2015).

This test is strictly stronger than local minimality under uniform dilations. The same framework defines the isothermal compressibility through

$L=A\mathbb Z^d$ 9

and shows that zero pressure together with positive compressibility implies local minimality among dilations. The paper explicitly stresses that local minimality among dilations does not imply $A$ 0-compact local minimality, and gives a one-dimensional counterexample (Bétermin, 2015).

3. Fixed-density optimality tests in two dimensions

A second major meaning of the term concerns deciding which Bravais lattice minimizes an interaction energy at fixed area $A$ 1. For a radial potential $A$ 2, the energy per point is

$A$ 3

and the key auxiliary object is the lattice theta function

$A$ 4

Montgomery’s theorem states that, among Bravais lattices of fixed area, the triangular lattice $A$ 5 uniquely minimizes $A$ 6 for every $A$ 7. This leads to a practical criterion: if $A$ 8 is admissible and has inverse Laplace transform $A$ 9, then

$|\det A|$ 0

Hence the kernel

$|\det A|$ 1

The same logic extends to spatially extended particles. For a radially symmetric probability measure $|\det A|$ 7 and interaction $|\det A|$ 8, the energy is

$|\det A|$ 9

Two regimes are isolated. If the mass distribution is radial and sufficiently concentrated, there exists $1/|\det A|$ 0 such that for $1/|\det A|$ 1 the triangular lattice is the unique minimizer at unit density. If, more strongly, both $1/|\det A|$ 2 and the radial density $1/|\det A|$ 3 of $1/|\det A|$ 4 admit completely monotone Laplace representations, then the triangular lattice is the unique minimizer for all unit-density Bravais lattices, independent of concentration scale (Bétermin et al., 2017).

For concrete pair potentials, the conclusions become density dependent. In the two-dimensional Lennard–Jones case,

$1/|\det A|$ 5

one sufficient high-density condition is

$1/|\det A|$ 6

which guarantees that $1/|\det A|$ 7 is the unique minimizer among lattices of area $1/|\det A|$ 8. By contrast, the triangular lattice is not optimal at sufficiently small density, and a necessary condition is obtained from

$1/|\det A|$ 9

For the Thomas–Fermi interaction,

$G$ 0

the triangular lattice is the unique minimizer for every fixed area because the kernel representation is a positive integral of theta terms. The same paper also gives an explicit convex decreasing positive potential,

$G$ 1

for which the triangular lattice is not a minimizer for some areas, showing that convexity and monotonicity alone are insufficient (Bétermin et al., 2014, Bétermin, 2015).

4. Hessian and local-optimality tests for cubic lattices

In three dimensions, a Bravais-lattice test is often a Hessian test at a high-symmetry lattice of fixed volume. Using Ennola’s five-parameter description $G$ 2, the energy

$G$ 3

is differentiated with respect to shape parameters at fixed covolume $G$ 4. For every $G$ 5, the simple cubic lattice $G$ 6, the face-centered cubic lattice $G$ 7, and the body-centered cubic lattice $G$ 8 are critical points. At SC the Hessian is block diagonal in $G$ 9 and $G^6$ 0; at FCC it reduces to three scalar lattice sums involving

$G^6$ 1

where $G^6$ 2 and $G^6$ 3. This reduction is the local-optimality test in its most explicit form: the sign of a small set of scalar combinations determines minimum, maximum, or saddle behavior (Bétermin, 2016).

For theta energies $G^6$ 4, the classification is sharp. SC is a saddle point for every $G^6$ 5. FCC is a saddle point for sufficiently small $G^6$ 6 and a local minimizer for sufficiently large $G^6$ 7; by theta duality the corresponding statements for BCC exchange the small- $G^6$ 8 and large- $G^6$ 9 regimes. For Lennard–Jones-type energies

$S^6$ 00

the test becomes density dependent: FCC and BCC are local minimizers at high density, local maximizers at low density, and saddles in an intermediate interval. For the classical $S^6$ 01– $S^6$ 02 case, the paper gives

$S^6$ 03

for FCC and BCC, and

$S^6$ 04

for the interval in which SC is a local minimizer; outside that interval SC is a saddle point. The same analysis extends to the unconstrained case by first optimizing the scale $S^6$ 05 in

$S^6$ 06

and shows that the optimal FCC and BCC dilations are locally minimizing among all Bravais lattices, whereas the optimal SC dilation is not (Bétermin, 2016).

5. Reduction-theoretic Bravais-type determination

In crystallography, the Bravais-lattice test is a classification problem for an experimentally determined metric. One classical approach uses Niggli reduction in the six-dimensional metric space $S^6$ 07, where

$S^6$ 08

The Niggli cone is the region of reduced metrics satisfying a set of linear inequalities. Its boundary is stratified into 216 boundary polytopes: 15 five-dimensional boundary polytopes, 53 four-dimensional polytopes, 79 three-dimensional polytopes, 55 two-dimensional polytopes, and 14 one-dimensional polytopes. All primitive lattice types can be represented as combinations of the 15 five-dimensional boundary polytopes, while all non-primitive lattice types can be represented as combinations of those 15 together with 7 special-position subspaces. In this setting, a Bravais-lattice test amounts to determining which equalities and intersections are satisfied, or how close a point in $S^6$ 09 lies to the corresponding subspaces (Andrews et al., 2012).

A second reduction-theoretic formulation is explicitly error stable. In two dimensions, the observed metric tensor $S^6$ 10 is first Gauss reduced so that

$S^6$ 11

The true lattice type is then inferred by projecting $S^6$ 12 onto symmetry subspaces such as

$S^6$ 13

and analogous centered-rectangular subspaces, under an error hypothesis $S^6$ 14. The output is a finite list of basis changes $S^6$ 15 and projected metrics $S^6$ 16, and the decision is made from the minimal projection error within each Bravais class. This yields a mathematically proved error-stable Bravais lattice determination algorithm for 2D lattices (Oishi-Tomiyasu, 2023).

SELLA replaces $S^6$ 17 by Selling space $S^6$ 18. After Selling reduction, the fundamental region is the all-negative orthant. Delone types become polytopes defined by zero coordinates and equal-coordinate multiplets, and orthogonal projectors $S^6$ 19 onto these polytopes are precomputed. The corresponding “perps” $S^6$ 20 provide a clear metric of fit: for a reduced Selling vector $S^6$ 21, the distance to a type is $S^6$ 22. When all symmetry-related representatives are generated by reflections and boundary transforms, SELLA obtains 239 non-triclinic projectors and uses them as a complete, closed solution for Bravais lattice determination (Andrews et al., 2023).

6. Dynamical and mechanical diagnostics

A different meaning of the term appears in nonequilibrium dynamics. For underdamped particles moving in a two-dimensional periodic potential built from Gaussian barriers on a Bravais lattice and driven by an unbiased ac force

$S^6$ 23

the direction of transport can be controlled by lattice geometry as well as by the strength and orientation of the oscillating drive. In rectangular and square lattices, reflection symmetries constrain axial or lateral currents; in oblique lattices, where no reflection lines exist, standard symmetry arguments fail. The paper shows that the decisive objects are ballistic attractors in phase space. Because the geometry selects a discrete set of possible attractor velocities, systematic scans of transport direction versus drive parameters can serve as a Bravais-lattice test that distinguishes square, rectangular, and oblique geometries and can even reconstruct lattice-vector orientations and anisotropies (Mukhopadhyay et al., 2019).

In crystal mechanics, slip and twinning provide a rank-one compatibility test for cubic Bravais lattices. With $S^6$ 24 a lattice basis and $S^6$ 25 the Ericksen energy-well set, planar interfaces correspond to rank-one connections

$S^6$ 26

A slip is a lattice-invariant shear; a twin is a rank-one connection involving a nontrivial reflection of the lattice across some plane. The paper defines Type 1, Type 2, and compound twins in these terms and rigorously calculates the slips and twins minimizing shear magnitude for simple cubic, bcc, and fcc lattices. It also proves that rank-one connections for the dual lattice can be obtained explicitly from those for the original lattice, so that the rank-one connections for fcc can be obtained explicitly from those for bcc (Ball, 2023).

7. Scope, limitations, and broader extensions

The literature does not support a single universal Bravais-lattice test. Rather, it supports several domain-specific criteria with different strengths and blind spots. In atomistic/continuum analysis, the relevant test is structural: interpolation kernels on $S^6$ 27 are checked for regularity, locality, affine reproduction, nodality, norm equivalence, and polynomial reproduction. Those properties determine whether discrete $S^6$ 28 norms and continuous $S^6$ 29 norms of interpolants are comparable and whether continuum error estimates transfer back to lattice functions (Ortner et al., 2012).

In lattice Boltzmann analysis on triangular meshes, the same phrase points to a consistency test for whether the scheme truly exploits Bravais-lattice structure. On a triangular Bravais lattice, the global opposite-direction property

$S^6$ 30

supports standard streaming and high-order Taylor analysis for D2T7. On a constant-degree but non-Bravais triangular mesh, D2T4 still admits a discrete-particle formulation, but numerical experiments show that formal high-order equivalent equations may collapse to second-order convergence in realistic bounded problems, which the paper presents as a limit of validity of Taylor-expansion analysis and as a source of questions about possible super-convergence (Dubois et al., 2014).

Phase diagrams can also function as Bravais-lattice tests. In the Jaynes–Cummings–Hubbard model, the Mott lobes and the critical hopping are not scalable only for the FCC lattice, whereas in the large excitation number regime the critical hopping is scalable for all the lattices and it does not depend on the detuning. This makes the scaling structure of the Mott–superfluid boundary itself a lattice-sensitive diagnostic (Gomes et al., 2012).

By contrast, some observables are almost insensitive to Bravais geometry. For the dilute Bose–Hubbard gas on a three-dimensional Bravais lattice with finite-range positive hopping and on-site repulsion,

$S^6$ 31

and the leading-order energy is universal: the lattice geometry affects the microscopic dispersion relation, but it enters the leading order asymptotics only through the lattice scattering length. This suggests a limit case in which a Bravais-lattice test cannot distinguish lattice types beyond a single effective parameter (Mokrzański et al., 18 Feb 2026).

Taken together, these formulations show that “Bravais-Lattice Test” denotes not a single theorem but a technically diverse class of diagnostics. Their common structure is the replacement of raw lattice data by a reduced or projected object—an energy variation, a theta-function kernel, a Hessian, a reduced metric, a transport attractor, or a rank-one connection—on which Bravais-specific conclusions become decidable.