To understand how particles fill space we treat them as identical hard spheres and arrange them as compactly as possible — close packing. In a single layer (layer A) each sphere touches six neighbours, giving the maximum two-dimensional packing (hexagonal close packing of one layer).
Stacking layers leads to two important three-dimensional patterns:
- Hexagonal close packing (hcp) — the third layer lies directly above the first, giving the repeating sequence ABAB... (e.g. Mg, Zn).
- Cubic close packing (ccp), identical to face-centred cubic (fcc) — the third layer occupies new positions, giving the sequence ABCABC... (e.g. Cu, Ag, Au).
Both hcp and ccp give the same maximum coordination number of 12 (the number of nearest neighbours touching a given sphere) and the same packing efficiency of 74%. A bcc structure has coordination number 8 and a simple cubic structure has coordination number 6.
The empty spaces left between packed spheres are voids (interstitial sites). There are two kinds: a tetrahedral void is surrounded by 4 spheres (smaller, radius ratio $r/R=0.225$) and an octahedral void is surrounded by 6 spheres (larger, $r/R=0.414$). For $N$ close-packed spheres there are $2N$ tetrahedral voids and $N$ octahedral voids.
Packing efficiency is the percentage of total space actually occupied by spheres: $$\text{Packing efficiency}=\frac{z\times\tfrac{4}{3}\pi r^3}{a^3}\times100$$ Using the edge–radius relations:
- Simple cubic: $a=2r$ gives 52.4%.
- Body-centred cubic: $\sqrt{3}\,a=4r$ gives 68%.
- Face-centred (ccp) / hcp: $\sqrt{2}\,a=4r$ gives 74% — the most efficient.
Finally, the density of a unit cell connects the microscopic structure to a measurable bulk property:
$$d=\frac{z\,M}{a^3\,N_A}$$ where $z$ = number of atoms (formula units) per unit cell, $M$ = molar mass, $a$ = edge length and $N_A=6.022\times10^{23}$ is Avogadro's number. When $a$ is in cm and $M$ in g/mol, $d$ comes out in $\text{g cm}^{-3}$. This single relation lets us find any one of $d,z,M,a$ when the other three are known, and is the workhorse of all numerical problems in this chapter.