The crystal's other native flaw: the self-interstitial
The previous guide showed that every crystal, even at equilibrium, carries a population of vacancies — empty sites — because a few missing atoms buy the crystal a large jump in entropy, and the equilibrium concentration climbs with temperature by a Boltzmann law. A vacancy is a missing atom. Its mirror image is an extra one: the self-interstitial, a host atom of the crystal's own kind crammed into one of the small gaps between the regular lattice sites. Vacancy and self-interstitial are the two intrinsic point defects a pure element can make all by itself, with no foreign atoms involved.
The two are not remotely equal in cost. Prying a hole open and stuffing a whole atom into it squeezes every neighbour outward, so the formation energy of a self-interstitial is huge — roughly 3 to 5 eV in a typical metal, several times a vacancy's ~1 eV. Run both through the same Boltzmann factor n/N = exp(-E/kT) at 1000 K (where kT ≈ 0.086 eV). The vacancy gives exp(-1/0.086) ≈ 9 x 10^-6, about one empty site per hundred thousand. The self-interstitial, with E ≈ 4 eV, gives exp(-4/0.086) ≈ 7 x 10^-21 — fewer than one per 10^20 atoms, which for all practical purposes is none. At thermal equilibrium the vacancy wins by a factor of about 10^15.
An impurity is a point defect made of a stranger
Now let a foreign atom into the crystal. Even a metal refined to 99.9999% purity still hides a stranger among every million host atoms, and each of those strangers is itself a point defect — a single site where the crystal's chemistry departs from the ideal. Most impurities are not accidents to be scrubbed away, though; they are added on purpose, because dissolving the right stranger is exactly how we tune a material. This is the atomic origin of the solid solution: a host (the solvent) with a foreign atom (the solute) dissolved right into its lattice, atom by atom, while the whole thing stays a single crystal structure — sugar stirred into water, but frozen solid.
A dissolved stranger has just two places to live, and which one it picks is the whole story of this guide. It can swap onto a regular lattice site, kicking out one host atom and taking its place — a substitutional impurity — or it can forgo a proper site altogether and squeeze into one of the small gaps between the host atoms — an interstitial impurity. Substitutional is a lodger taking a bedroom; interstitial is a lodger sleeping in the cupboard under the stairs. Size decides almost everything: a stranger close in size to the host takes a seat at the table, while only a genuinely tiny atom can hide in the gaps.
SUBSTITUTIONAL vs INTERSTITIAL (a 2D crystal; host atoms = O)
substitutional solute S interstitial solute i
swaps ONTO a lattice site squeezes INTO a gap
O O O O O O O O
i
O S O O O O O O
i
O O O O O O O O
size ~ host atom size << host atom
set by the Hume-Rothery rules sits in an octahedral /
tetrahedral holeThe interstitial route: tiny atoms in the lattice's holes
Take the interstitial route first, because it is the more demanding. The gaps in a close-packed metal are genuinely small, so only the smallest atoms — hydrogen, carbon, nitrogen, oxygen, boron — can dissolve this way, and when they do the result is an interstitial solid solution. The gaps themselves come in two shapes you already met when we stacked spheres: the roomier octahedral hole, ringed by six host atoms, and the tighter tetrahedral hole, ringed by four. The radius-ratio rule says an octahedral hole comfortably swallows a sphere up to about 0.414 times the host radius, a tetrahedral one only about 0.225 — so the octahedral hole is usually the preferred address.
Steel is this idea's masterpiece. Carbon dissolves interstitially in iron, and how much fits depends violently on which crystal form the iron is in. Face-centred-cubic iron (austenite) dissolves up to about 2.1 wt% carbon; body-centred-cubic iron (ferrite) holds barely 0.02 wt% — a hundredfold difference. Even in the roomier FCC holes carbon is a squeeze: with the iron radius near 1.27 angstrom the octahedral hole is only about 0.53 angstrom across, while a carbon atom is roughly 0.7 to 0.8 angstrom, so every dissolved carbon still shoves its neighbours apart and strains the lattice. That stored strain, released and rearranged by heat treatment, is the lever the entire craft of steel-making pulls on.
The substitutional route and the Hume-Rothery rules
When the stranger is comparable in size to the host, it cannot hide in a hole, so it does the other thing: it evicts a host atom and takes its lattice site, building a substitutional solid solution. Now the question becomes how MUCH will dissolve — a few percent, or all the way to a fifty-fifty alloy? The metallurgist Hume-Rothery distilled the answer into four rules of thumb, and though they are empirical, not laws, they predict solubility remarkably well.
- Size: the atomic radii of solute and solvent should differ by less than about 15%. Beyond that the strain of the misfit becomes too costly and solubility collapses.
- Crystal structure: the two elements should share the SAME crystal structure for complete (all-the-way) solubility to be possible at all.
- Electronegativity: the two should be close in electronegativity. A large difference makes the atoms prefer to form an ordered compound rather than dissolve randomly.
- Valence: similar valence helps; as a rule a metal dissolves more of a higher-valence solute than of a lower-valence one (the relative-valence effect).
Copper and nickel are the textbook success. Their radii differ by only ~2% (1.28 versus 1.25 angstrom), both are FCC, they sit side by side in the periodic table with near-identical electronegativity, and their valences are close — so they dissolve in each other across the ENTIRE composition range, from pure copper to pure nickel, a single continuous solid solution. Contrast copper and zinc (brass): zinc is a touch larger, its structure is HCP not FCC, and its valence is 2 against copper's 1, so only about 35 wt% zinc will dissolve before a second phase must appear. Same host, very different limits — and the four rules tell you why in advance.
Solubility limits, strength, and what comes next
Every solid solution has a ceiling. Push past the solubility limit and the extra solute can no longer be accommodated one atom at a time; the crystal gives up and precipitates a distinct second phase — carbon beyond ferrite's meagre limit forms iron carbide, zinc beyond brass's limit forms a new zinc-rich phase. The limit exists because each dissolved atom carries a strain field, and once enough of them crowd in, it becomes cheaper to gather the excess into a separate structure than to keep straining the host. Where that boundary sits, for every temperature and composition, is exactly what a phase diagram maps.
That same strain is quietly useful. Each solute atom warps the lattice around it, and those little knots of stress get in the way of the dislocations whose gliding lets a metal deform — so dissolving a stranger makes the host harder and stronger. This is solid-solution strengthening, and it is why brass is stiffer than pure copper and why even a pinch of carbon transforms soft iron. Point defects, the smallest imperfections there are, reach all the way up to a bridge girder's yield strength.
Point defects also set how atoms MOVE. A substitutional atom can only shuffle along when a neighbouring vacancy opens up beside it — the vacancy-mediated diffusion you met last guide — which is why substitutional diffusion is sluggish. An interstitial atom needs no such help; it simply hops from one hole to the next, which is why carbon races through iron far faster than iron atoms move through themselves. One last honesty, and a signpost. Everything here assumed a metal, where you can add or remove a single atom freely. In an IONIC crystal you cannot — pull out one ion and you have left a net charge behind — so defects must team up in charge-balanced combinations. That constraint is the subject of the next guides: the paired Schottky and Frenkel defects, the bookkeeping of Kroger-Vink notation, and how the same ideas drift a compound off its ideal formula and even paint crystals with colour.