Why one ion cannot leave alone
The last two guides built the whole zoo of point defects in a metal: a vacancy where an atom is missing, a self-interstitial squeezed into a gap, and impurity atoms sitting substitutionally or interstitially to make a solid solution. In a metal all of that is cheap, because every atom is electrically the same — pull one out to the surface and the crystal shrugs. An ionic crystal like common salt plays by a stricter rule. Its lattice is a checkerboard of alternating charges held together by the ionic bond: Na+ on the red squares, Cl- on the black. You cannot just lift one Na+ out. Do it and the crystal is left with a net negative charge, and the electrostatic penalty for charging up a whole macroscopic solid is astronomical.
So the ionic crystal faces a puzzle. Guide 1 proved that defects MUST exist above absolute zero — they raise the entropy, so free energy always favours a few of them. Yet here the obvious way to make one is forbidden by charge. Nature's escape is elegant: make defects in charge-neutral bundles. There are exactly two ways to do it. Either remove one positive and one negative ion together, so the books still balance — the Schottky defect — or don't remove anything at all, just push one ion off its proper seat into a nearby gap, leaving a hole behind but changing no charge — the Frenkel defect. Every intrinsic defect in a pure ionic crystal is one of these two moves.
One handy piece of bookkeeping you will lean on throughout. A missing cation is a missing positive charge, so a cation vacancy behaves as if it carries an EFFECTIVE NEGATIVE charge relative to the perfect lattice; a missing anion behaves as an effective POSITIVE; an interstitial cation drops an extra positive into a spot that should be neutral, so it is effectively positive. Nothing is really charged — these are only charges relative to the flawless background — but tracking them is exactly how you check that a defect reaction stays neutral. Guide 4's Kroger-Vink notation turns this bookkeeping into a formal algebra.
The Schottky defect: a vacancy pair
The Schottky defect is the vacancy of the last two guides, made legal by working in pairs. In rock salt — Na+ and Cl- in the rock-salt structure, each ion sitting in an octahedral hole with a coordination number of 6 — a Schottky event pulls one Na+ and one Cl- out of the interior and re-plants them on the crystal surface. What is left behind is a cation vacancy (effective minus) and an anion vacancy (effective plus), sitting somewhere in the bulk. The two effective charges cancel, so the crystal stays neutral, and because you removed one of each kind, the ratio of Na to Cl is untouched — the compound is still perfectly NaCl. That last point matters: a Schottky defect preserves stoichiometry.
ROCK-SALT SLICE + cation - anion o vacancy * interstitial
PERFECT SCHOTTKY PAIR FRENKEL (cation)
+ - + - + - + - + - + - + - + - + -
- + - + - + - o - + - + - + - o - +
+ - + - + - + - + - + - + - + * - +
- + - + - + - + - o - + - + - + - +
+ - + - + - + - + - + - + - + - + -
one + and one - one + leaves its seat and
move to the surface squeezes into a gap nearby:
-> V_cation (eff -) -> vacancy (eff -) PLUS
V_anion (eff +) interstitial (eff +)
charge 0 stoichiometry KEPT same # of ions; net charge 0How many Schottky pairs are there? The same Boltzmann argument from guide 1 carries over, with one twist. Minimising the free energy of a crystal that can host these pairs gives the equilibrium concentration n/N = exp(-E_s / 2kT), where E_s is the energy to form ONE pair, k is Boltzmann's constant, and T the absolute temperature. The exponent carries E_s divided by 2, not the bare E_s, and the reason is honest and physical: one Schottky event creates TWO vacancies — one on the cation sublattice, one on the anion sublattice — so the entropy you buy per unit of energy spent is doubled. That factor of 2 is the fingerprint of a paired defect.
Put real numbers through it. For NaCl the Schottky pair costs roughly E_s = 2.3 eV, and kT is 0.026 eV at 300 K but 0.086 eV at 1000 K (near the melting point). At room temperature the exponent is -2.3/(2 times 0.026) = -44, so n/N = exp(-44), about 1 in 10^19 — utterly negligible. At 1000 K the exponent is only -2.3/(2 times 0.086) = -13, so n/N = exp(-13), about 2 in a million. Heating from room temperature to red heat multiplies the thermal vacancy population by something like 10^13. That ferocious, exponential rise with temperature is the whole reason ionic crystals diffuse and conduct so much better when hot.
The Frenkel defect: an ion steps aside
The Frenkel defect takes the opposite route: it moves nothing to the surface at all. One ion simply leaves its regular lattice seat and lodges in a nearby interstitial site — one of the octahedral or tetrahedral gaps you met when we packed spheres. Picture a car pulling out of its numbered bay and double-parking in the aisle: the bay is now empty (a vacancy) and the aisle is blocked (a self-interstitial), but no car has left the garage. The displaced ion and the vacancy it left behind carry equal and opposite effective charges, so the pair is automatically neutral, and since not a single ion was added or removed, the stoichiometry is again untouched.
Whether a Frenkel defect is affordable comes down to one thing: is there room in the gaps? In a tightly packed crystal the interstitial holes are small, so cramming a full-sized ion into one costs a lot of strain energy, and Frenkel defects are rare. That is why classic close-packed rock-salt crystals like NaCl and MgO prefer Schottky. But two situations open the door. First, when the cation is small and squishy — silver in silver bromide, AgBr, is the textbook case — the little Ag+ slips into an interstitial with modest cost, giving a CATION Frenkel defect. Second, when the anion sublattice itself has generous open holes: the fluorite structure of CaF2, UO2 and ZrO2 has a large empty cage at the centre of the cell, so it is the ANION that goes interstitial (an anion, or anti-, Frenkel defect).
Which one wins, and how many
A pure ionic crystal always has BOTH kinds present — thermodynamics never lets a defect population fall to exactly zero. What differs is which one dominates, and that is decided by which has the lower formation energy in that particular structure. The winner outnumbers the loser by a colossal margin, because the population depends on energy through an exponential: even a 0.5 eV edge in formation energy, at 1000 K, changes the count by exp(0.5/0.086) which is about a factor of 300. So we speak of 'a Schottky crystal' (NaCl, KCl, MgO) or 'a Frenkel crystal' (AgBr, CaF2) even though the other defect is always lurking at a far lower level.
- Look at the two ion sizes and how open the structure is — how big are the interstitial gaps compared with the ions?
- Similar-size ions in a close-packed frame (NaCl, MgO)? The gaps are tiny, so interstitials are costly and SCHOTTKY (vacancy pairs) wins.
- One ion small and easily polarised (Ag+ in AgBr)? It fits an interstitial cheaply -> cation FRENKEL wins.
- Big open holes in the anion sublattice (fluorite CaF2, UO2)? The anion goes interstitial -> anion (anti-)FRENKEL wins.
- Both are always present; the one with the lower formation energy simply outnumbers the other, more and more heavily the colder it gets.
One crucial caveat before you over-trust these thermal numbers. Everything so far is INTRINSIC — the pristine, thermally generated equilibrium of a chemically perfect crystal, the intrinsic defect equilibrium. Real crystals are never perfectly pure, and impurities of the wrong charge force their own vacancies to keep the crystal neutral. Dissolve a little CaCl2 into NaCl (a substitutional solid solution from guide 2): each Ca2+ replaces one Na+ but carries a double positive charge, so the lattice must create one extra cation vacancy per Ca2+ to stay balanced. These EXTRINSIC vacancies do not care about temperature, so at low T they swamp the tiny intrinsic population. Only when you heat the crystal enough does the exp(-E_s/2kT) term overtake them — the crossover from extrinsic to intrinsic behaviour.
Why these defects run the show
These charge-neutral defects are not just curiosities — they are the moving parts that make ionic solids DO things. An ion cannot travel through a perfect crystal; it needs somewhere to go. A Schottky vacancy gives it exactly that: an ion hops into the neighbouring empty seat, the vacancy hops the other way, and repeated over billions of jumps this is vacancy-mediated diffusion. A Frenkel interstitial offers a second highway, the ion threading from gap to gap. Because the defect population itself climbs as exp(-E/2kT), and the hopping over each barrier is Arrhenius too, ionic diffusion and ionic conductivity both rise steeply with temperature. This is why AgBr, riddled with mobile silver interstitials, is a fast ion conductor and the light-sensitive heart of old photographic film.
Point defects can even paint a crystal. Take a colourless alkali-halide crystal and heat it in its own metal vapour, or blast it with radiation, and it takes on a rich colour. The mechanism is beautifully simple: an anion vacancy is effectively positive, so it can trap a stray electron to neutralise itself. That trapped electron sits in a tiny box of quantised energy levels, and the gaps between those levels happen to fall in the visible range, so it absorbs certain colours of light. This is a color centre, the F-centre (from the German Farbe, colour). It is a startling demonstration that a bare vacancy — a piece of nothing — can be the thing you SEE. Guide 5 follows this thread all the way to nonstoichiometry.