Dual lattice — where it appears
Named by 15 essays across one field — each of them below, with the objects they name alongside it.
The dual lattice, as a construction
The reciprocal lattice is usually introduced as a formula and then used as a fact. Building it instead — one point per family of lattice rows, at the inverse of the spacing — makes every property it has obvious rather than memorable.
Every plane lattice is its own dual
The dual of a lattice has the inverse Gram matrix, and in two dimensions the inverse is the adjugate over the determinant — which is what one particular change of basis does to a Gram. So a plane lattice's dual is the lattice itself, turned through a right angle and scaled, for every lattice with no exception. In three dimensions it is a condition, and the face-centred and body-centred cubic lattices are duals of each other rather than of themselves.
One perfect form in space
Which lattice packs spheres most densely is a question about a maximum over a continuum, and Voronoi turned it into a rank calculation and a sign check. A lattice is a local maximum exactly when its shortest vectors pin its shape down completely and its inverse can be written over them with positive coefficients. Searching every reduced integer form of minimum two finds one such lattice in the plane and one in space.
The sum that turns a lattice into its dual
Put a Gaussian on every point of a lattice and add them up. The answer equals the same sum over the dual lattice with the width inverted and the covolume divided out — exactly, to the last bit a double carries, on lattices with no symmetry in them. The identity is free and the reason to have it is that the two sides do not cost the same: at one end of the range the direct sum needs forty thousand terms and the dual sum needs a hundred and twenty-five.
The sum whose answer depends on the shape
Give the points of a cubic lattice alternating signs and add up one over the distance. Added over expanding cubes the total creeps towards 1.747565; added over expanding spheres it does not converge at all, landing on both sides of that number and never settling. The terms are identical and only the order differs. Splitting the sum in two with the theta transformation gives it a value — ten decimal places from a few thousand terms.
No lattice in space wins at every width
Spread a Gaussian over every point of a lattice and add them up. In the plane one lattice gives the smallest sum at every width — the hexagonal one — and that is a theorem. In space no lattice does, and the proof is two lines: at large widths the winner must be the densest lattice, at small widths it must be the densest lattice's dual, and in space those are face- and body-centred cubic. They change places at exactly the width where each is the other's dual.
In eight dimensions one lattice wins at every width
In space no lattice has the smallest Gaussian sum at every width, because the densest lattice wins when the Gaussian is narrow and its dual wins when it is wide, and fcc is not bcc. The argument fails exactly when the densest lattice is its own dual. In eight dimensions it is: E₈, with 240 shortest vectors, beats D₈, its dual, the cubic lattice and a sum of two four-dimensional lattices at every width tried, by a factor of at least 1.31, and its sum is mirror-symmetric in the logarithm of the width.
The two lattices where a four-dimensional descent stops short
In four dimensions the argument that rules out a single best lattice in space has nothing to act on, so whether one lattice wins at every width is left to be measured. A search over every four-dimensional lattice finds D₄ from almost anywhere. The exceptions are the finding: two lattices, A₄ and its dual, where a descent can stop short, each at one end of the widths and neither in the middle, for the same reason Voronoi found four dimensions has two perfect forms.
The body-centred hypercube is the checkerboard lattice
In three dimensions the centring letters come out as the subspaces a holohedry keeps modulo a prime, and only two and three make new lattices. In four dimensions the same census, run on the six lattices whose symmetry groups are maximal, finds a prime that the three-dimensional argument never could: five centres A₄ into its dual. It also finds that the most familiar centring does not survive the extra dimension. The body-centred hypercubic lattice is not a centred hypercube at all: its body centre is exactly as far away as an edge, and it is D₄, with three times the symmetry.
The self-dual width is not flat
A search over four-dimensional lattices found that A₄ and its dual stop being traps for the Gaussian sum almost exactly at the self-dual width, and read that as a flat of the sum sitting on the width itself. Computed exactly rather than by finite differences, the curvature there is not zero. It is six parts in a million of the other curvature, positive, and the same for both lattices, because duality makes their Hessians one matrix at that width. Both are minima, in a window four parts in a hundred thousand wide, with basins a few millionths across.
![Every four-dimensional lattice with a five-fold rotation, on one interval. The lattices with a five-fold rotation in four dimensions are one ring, ℤ[ζ₅], with the form Tr(a·x·x̄), and the number t = log(a/σ(a)) / 2 log φ, folded into the interval from 0 to 1, says which lattice a is. The dots are the members with a = u + v√5 for u up to 9, 8 distinct lattices. The two ends are A₄* and A₄, with 240 symmetries each; every point between has 20. The arcs join each member to the one a centring by fifths makes of it, its reflection t ↦ 1 − t.](/f/every-four-dimensional-lattice-with-a-five-fold-rotation-on-one-interval-cc0b2586.svg)
A five-fold lattice has one free number
In four dimensions a lattice can have a five-fold rotation, and A₄ and its dual both do. Every other lattice with one turns out to be the same ring, the integers of the fifth roots of unity, with a single number deciding its shape. Folded onto an interval, the number runs from A₄* at one end to A₄ at the other, and every lattice strictly between has twenty symmetries. Centring by fifths does not make a new lattice. It reflects the interval end to end, exchanging A₄ with A₄* and pairing every other lattice with a partner, and no lattice sits at the middle to be its own.
![Every six-dimensional lattice with a seven-fold rotation, on one torus. The lattices with a rotation of order seven in six dimensions are the ring ℤ[ζ₇] with three positive weights, one on each complex line it spans. Up to scale and isometry the weights live on a torus, drawn as a sixty-degree rhombus with opposite sides glued, shaded by how well each lattice packs spheres. The Galois automorphism turns the torus a third of a turn about three points, on the long diagonal: A₆* with 10080 symmetries, A₆ with 10080, and the ideal (1 − ζ)² with 672, the densest lattice in the family. The small dots are integral members, each with 28 symmetries.](/f/every-six-dimensional-lattice-with-a-seven-fold-rotation-on-one-torus-551a6153.svg)
A seven-fold lattice lives on a folded torus
In six dimensions a lattice can have a seven-fold rotation, and every one that does is the integers of the seventh roots of unity with three weights on it. Up to scale the weights live on a hexagonal torus that a Galois symmetry folds by a third-turn, and the fold has three fixed points: A₆*, A₆, and a third lattice that packs better than either. Centring by sevenths is no longer a reflection, as fifths was in four dimensions. It moves every lattice a third of the way along the torus and carries the three special ones round in a cycle.
The densest eleven-fold lattice is not its own dual
Every ten-dimensional lattice with an eleven-fold rotation is one ring with five weights, and its shapes form a four-dimensional torus folded by a Galois rotation of order five. The seven-fold essay predicted five special lattices, the powers of 1 − ζ, cycled by centring, and all of that holds. What does not carry over is its best result. In six dimensions the densest seven-fold lattice was the one duality fixes. In ten, duality fixes (1 − ζ)³, and the densest is (1 − ζ)², whose dual is (1 − ζ)⁴.
Thirteen-fold, and nothing is its own dual
Every twelve-dimensional lattice with a thirteen-fold rotation is one ring with six weights, and its shapes form a five-dimensional torus folded six ways. The fold has six fixed points, the powers of 1 − ζ, and duality fixes none of them, as in four dimensions and unlike six and ten: it pairs them off. The eleven-fold essay asked whether a longer cycle would let a higher power take over as the densest. It does not. (1 − ζ)² is the densest member at seven, at eleven and at thirteen, whatever duality does.
At seventeen the densest power of 1 − ζ moves
Three primes running, the densest lattice with a p-fold rotation sat at the second power of 1 − ζ. At seventeen, in sixteen dimensions, and at nineteen, in eighteen, it sits at the third. The move is not a surprise once the minima are read: every power up to the fourth reaches Craig's bound exactly, which turns each density into a formula and the competition between powers into an inequality. The second power wins exactly while (3/2) raised to half of p − 1 stays below p, and the formula's next prediction, the fourth power at twenty-three, comes true by three per cent.
Named alongside it
The objects these essays reach for when they reach for this one.
Bravais latticeKissing numberCentringTheta seriesGram matrixHigher-dimensional latticeHolohedryLatticeReciprocal latticeLattice sumPerfect formPoisson summation