Sphere-Packing in Atomic Assembly

Updated 7 January 2026

Sphere-packing is a geometric framework that models atoms as hard spheres, elucidating configurations, phase transitions, and packing densities.
It classifies densest packings in both monodisperse and binary systems, revealing how periodic and layered arrangements influence material properties.
The approach integrates statistical mechanics, entropic analysis, and jamming metrics to predict mechanical stability, self-assembly, and structural diversity.

The sphere-packing description of atomic assembly provides a rigorous geometric and statistical mechanical framework for understanding atomic configurations, phase transitions, and structural diversity in condensed matter. In this view, atoms are modeled as hard spheres with diameter constraints, and crystal lattices, glassy aggregates, and more complex alloys emerge as geometric arrangements that maximize packing density or optimize entropic and energetic criteria. The approach unifies mathematical, physical, and materials–science perspectives, with explicit solutions known in key settings, and underpins much of contemporary theory in atomic self-assembly, statistical mechanics, and crystallography (Mazel et al., 2023, Logan et al., 2021, Hopkins et al., 2011, Torquato, 2018, Hopkins et al., 2011).

1. Hard-Sphere Models and Their Statistical Mechanics

The hard-sphere model idealizes atoms or colloidal particles as non-overlapping spheres. The fundamental object is a configuration $\omega\subset\mathcal{L}$ (typically $\mathbb{Z}^3$ or $\mathbb{R}^3$ for lattice or continuum models, respectively) with a minimum Euclidean distance $D>0$ between centers: $\forall\, x,y\in\omega,\ x\neq y\ \implies\ \|x-y\|_2\geq D$ This exclusion condition defines the set of $D$ -admissible configurations. The associated Hamiltonian for configuration $\omega$ is

$H(\omega)=-|\omega|\ln u$

where $u$ is the fugacity (exponential of chemical potential). The partition function $Z_{\Lambda}(u,D)$ sums over all $D$ -admissible configurations in a domain $\Lambda$ with weights $u^{|\omega|}$ , leading to the pressure (per-site free energy) in the thermodynamic limit (Mazel et al., 2023).

At large $u$ , the dominant contributions are from densest packings, corresponding to atomic crystals in real materials. In this high-density regime, all admissible arrangements maximize the number of occupied lattice sites while respecting the hard-core constraint.

2. Classification and Densities of Densest Packings

The maximal density $\delta(D)$ for $D$ -admissible configurations on $\mathbb{Z}^3$ has been classified for multiple values $D^2\in\{2,3,4,5,6,8,9,10,11,12\}$ and for $D^2=2\ell^2$ . Each value produces either a unique class of densest periodic lattices or a countable family due to sliding or layering degeneracies. Table 1 summarizes representative cases (Mazel et al., 2023):

$D^2$	Structure	Index in $\mathbb{Z}^3$	$\delta(D)$
2	FCC ( $A_3$ )	2	1/2
3	BCC	4	1/4
4	$2\mathbb{Z}^3$ (sliding family)	8	1/8
5	Layered HCP family	9	1/9
6	Layered families	12	1/12
8	$2 \cdot FCC$	16	1/16
9	Deformed FCC	20	1/20
10	Deformed FCC	26	1/26
11	Sliding non-layered	32	1/32
12	$2 \cdot BCC$	32	1/32

For $D^2=2\ell^2$ , the densest packings are embeddings of $D\cdot A_3$ (scaled FCC) sublattices with index $2\ell^3$ , giving $\delta(D)=1/(2\ell^3)$ for $D=\ell\sqrt2$ . Kepler’s conjecture, asserting that FCC is the densest sphere packing in $\mathbb{R}^3$ , underpins this result for all $\ell$ (Mazel et al., 2023).

Ground states are classified into: (i) finitely many periodic packings (e.g., FCC, BCC, deformed FCC), (ii) countably many layered periodic ground states (e.g., at $D^2=6$ ), and (iii) countably many sliding or non-layered periodic ground states (e.g., at $D^2=4,11$ ).

3. Entropic and Topological Description of Sphere Packings

Beyond geometric density, statistical mechanics models incorporate entropic and topological contributions. For $N$ sticky hard spheres of mean diameter $a$ in $V$ , two key thermodynamic variables are the packing (volume) fraction $\phi=(N \cdot v_p)/V$ and mean coordination number $z=2Z/N$ , with $Z$ the total number of contacts. The total packing entropy per particle, $S(\phi, z)$ , decomposes into: $S_{\text{pack}}(z) = \langle S_{\text{geo}} \rangle_Z + S_{\text{topo}}(z)$ $S_{\text{geo}}$ counts the geometric configurations for a given contact network, while $S_{\text{topo}}$ is the entropy from the number of distinct contact graphs.

In the thermodynamic limit, isostatic packings (coordination $z=6$ in 3D) dominate, yielding $s^*_{\text{geo}} \approx 0.35$ , $s^*_{\text{topo}} \approx 3.02$ , so $S_{\text{pack}}(\phi,6) \approx 3.37N$ . Including short-range attractions mimicking chemical bonding introduces a per-contact free energy $\epsilon$ , leading to: $f(\phi, z) = \mu(\phi, z) = (\epsilon/2) z - [s^*_{\text{geo}} + s^*_{\text{topo}}]$ Minimizing $\mu$ at fixed $\phi$ determines equilibrium $z(\epsilon, \phi)$ . For strong attraction $\epsilon \gg 1$ , $z \to 6$ and FCC/HCP-like crystals prevail; weaker attraction stabilizes glassy, floppy clusters with $z < 6$ , corresponding to under-coordinated, kinetically arrested assemblies (Logan et al., 2021).

4. Binary Sphere Packings and Alloy Structural Diversity

Binary atomic assemblies are modeled by mixtures of hard spheres with radii $R_S < R_L$ . The parameter space is spanned by size ratio $\alpha = R_S/R_L$ and concentration $x = N_S/(N_S+N_L)$ . The Torquato–Jiao (TJ) linear-programming algorithm systematically enumerates densest periodic packings for each $(\alpha,x)$ (Hopkins et al., 2011, Hopkins et al., 2011).

Key results include:

For $\alpha = 1$ , monodisperse (FCC/HCP) Barlow packings yield the Kepler density $\phi_{\mathrm{Barlow}} = \pi/\sqrt{18}$ .
For $\alpha < \sqrt{2}-1$ , “interstitial” alloys such as NaCl (rock-salt, XY $_1$ ), CsCl, and more complex packings (XY $_n$ ) emerge as the densest arrangements at specific stoichiometries (e.g., x = ½, ¾, etc.).
Novel structures with ratios 7:3, 5:2, and higher (“magic” interstitials) appear, revealing previously unrecognized classes of mechanically stable, strictly jammed binary lattices (Hopkins et al., 2011).

Many binary metallic, ionic, and hydride crystals map directly to these sphere-packings, both at ambient and high-pressure conditions, underpinning a global entropic principle for phase stability and composition–structure relationships.

5. Jamming, Order Metrics, and Physical Properties

Sphere-packing models are classified not only by density but also by jamming categories—local, collective, and strict jamming—each with implications for mechanical stability and response to perturbation (Torquato, 2018). The geometric-structure “order map” representation (packing density $\phi$ vs. scalar order metric $\psi$ ) organizes all configurations, with the maximally random jammed (MRJ) state at minimal disorder and an intermediate density ( $\phi_{\rm MRJ} \approx 0.64$ in 3D, $Z \approx 6$ ), and the densest (FCC) and lowest-density strictly jammed states as boundary points.

Jamming categories correlate with bulk and shear moduli, while hyperuniformity—vanishing structure factor at small $k$ —is associated with suppressed long-wavelength density fluctuations and anomalous transport or wave properties. Extensions to polydisperse spheres and nonspherical shapes further enrich the phase space, with the sphere-packing framework providing predictions for exotic structural and mechanical properties (Torquato, 2018).

6. Sphere Packing and Atomic Assembly: Synthesis and Implications

In the atomic-assembly interpretation, the sphere diameter $D$ is set by atomic or ionic size, and the underlying lattice by the crystallographic environment. Fugacity $u$ plays the role of the chemical potential. Each periodic ground state $\omega_0$ yields a crystalline “parent phase” at high fugacity, with rare defects or fluctuations (Peierls contours) described via Pirogov–Sinai theory (Mazel et al., 2023). The mean density in each phase saturates at $\delta(D)$ , and the free-energy admits an analytic expansion: $p(u, D) = \delta(D) \ln u + \sum_{\Gamma} \varphi(\Gamma) u^{-\Delta(\Gamma)}$ Kepler’s theorem, ensuring FCC-optimality in the continuum, provides the geometric endpoint for lattice-based classifications. The sphere-packing paradigm, therefore, delivers a unified, quantitative, and physically grounded description of atomic assemblies—determining which structures are possible, which are optimal, and how entropic and energetic considerations interplay across condensed matter systems.