Sphere packing is the mathematical study of arranging non-overlapping spheres in a space so that they occupy as much volume as possible. The classical formulation considers congruent spheres in three-dimensional Euclidean space, but the topic naturally generalizes to circles in the plane, spheres of different sizes, higher-dimensional Euclidean spaces, and curved spaces such as the sphere or hyperbolic space. Problems in this area ask for the densest possible arrangements, typical densities for random or jammed packings, and combinatorial counts such as how many equal spheres can simultaneously touch a given central sphere.

Basic concepts and measures

Central concepts include:

  • Packing: a collection of non-overlapping closed balls (spheres with their interior) placed in a space.
  • Density: the fraction of space covered by the spheres. In infinite spaces this is defined as an average or asymptotic density taken over larger and larger regions; see general discussions of density.
  • Monodisperse vs. polydisperse: whether all spheres have the same radius (monodisperse) or vary in size (polydisperse).
  • Lattice and periodic packings: packings that are invariant under a lattice of translations or that repeat a finite motif periodically; many optimal candidates are of this type.

In two dimensions the densest packing of equal circles is the hexagonal (triangular) lattice. In three dimensions the best-known dense arrangements are the face-centered cubic and hexagonal close packing structures, often described together as close-packed lattices; both attain the same maximum density for equal spheres. Exact algebraic expressions for these densities appear in standard references on geometry and on the theory of spheres.

Densest packings and low-dimensional results

The three-dimensional densest-packing problem was conjectured in the 17th century and became known as the Kepler conjecture. Its resolution relied on a combination of geometric reasoning and extensive computation. In higher dimensions the situation is more varied: in dimensions 8 and 24 special lattices (the E8 lattice and the Leech lattice) have been proved to achieve optimality among all packings in their dimensions by analytic methods developed in recent decades.

Methods and bounds

Techniques used to study sphere packing include constructive approaches (building explicit lattice or periodic arrangements), analytic bounds (linear programming and harmonic analysis methods), and computational verification. Lower bounds arise from explicit constructions, while upper bounds are obtained by global inequalities or optimization problems that restrict how densely spheres can be placed. For very high dimensions, asymptotic estimates and probabilistic constructions give insight, but exact optimality is known in only a few dimensions.

Closely connected problems include the kissing number problem, which asks how many equal spheres can touch a central one; this is solved in several dimensions but remains open in general. Other variants consider packing non-equal spheres, coverings (how spheres can cover space with minimal overlap), and packing on curved surfaces. Empirical studies distinguish crystalline dense packings from disordered dense states: the concept of random close packing or jammed packings captures physically observed dense arrangements of many particles that are not perfectly periodic.

Applications

Sphere packing has practical significance beyond pure mathematics. In communications theory, high-dimensional packings model signal constellations and error-correcting codes, where maximizing minimal distances corresponds to efficient, robust transmission. In physics and materials science, packing models help explain crystal structures, granular media behavior, and how atoms or colloidal particles organize. Computational methods for packing aid in simulations of porous or particulate materials and in optimization problems across engineering disciplines.

Open questions and further reading

Many directions remain active: determining optimal packings in other specific dimensions, understanding polydisperse packings and phase transitions between ordered and disordered states, and extending results to non-Euclidean geometries. Surveys, textbooks and expository articles provide introductions and detailed accounts; readers can consult standard geometry references for background, accessible expositions on spheres and packing problems, technical treatments that discuss bounds and proofs, and resources on measures of density and computational methods. A variety of online and printed sources are available for both introductory material and current research summaries (Euclidean-space overviews and specialized surveys).

For further study, pursue texts on discrete geometry and harmonic analysis applied to packing problems, research articles describing the resolution of classical conjectures, and computational literature on algorithms used to explore dense and random packings.