One Component: The Silica Polymorphs

One component, and what the map can show

Guide 1 handed us a counting machine, the Gibbs phase rule, F = C - P + 2. Now we point it at the simplest possible system: one component, C = 1, a single pure substance whose whole map is drawn against just two knobs, temperature and pressure. The rule collapses to F = 1 - P + 2 = 3 - P, and that little equation tells you at a glance what every feature of the diagram is. Where one phase is stable — quartz alone, say — F = 2, so you can wander over an area, changing T and P freely and still keep that one crystal. That region is a field on the map, not a line.

Push two phases into coexistence — quartz touching its melt, or one crystal touching the next — and F drops to 1. You have spent a degree of freedom, so the two phases can only share a line on the T-P plane: pick the temperature and the pressure is fixed, or vice versa. Force three phases together and F = 0, an invariant point pinned to one exact temperature and pressure — the triple point. Areas, lines, points: on a unary diagram the phase rule is not abstract bookkeeping, it is literally the dimensionality of what you are looking at.

For firing, we usually work at one fixed pressure — the atmosphere in the kiln — so we slice the map along the 1 atm line and read straight up the temperature axis. Fixing pressure spends one more degree of freedom, giving the condensed rule F = C - P + 1 = 2 - P. Now a single phase owns a temperature range (F = 1), but the boundary where two crystals swap, or where a crystal melts, is a single sharp temperature (F = 0). That is the deep reason a melting point or an inversion temperature is one crisp number rather than a smear: at fixed pressure, two coexisting phases have used up every degree of freedom there is.

Silica's family portrait

Silica is the perfect first specimen because one composition, SiO2, crystallises into several different solids. That is polymorphism — same recipe, different architecture — and its cast here is the silica polymorphs: quartz, tridymite, and cristobalite. Every one is built from the same brick you met on the silicate rung, the SiO4 tetrahedron, each silicon caged by four oxygens. And every one is a fully corner-shared framework, so every oxygen bridges two tetrahedra and the ratio is exactly Si:O = 1:2. What differs is only how those identical tetrahedra hook together — the twist of the links, the size of the rings, and above all how tightly or loosely the framework is packed.

SiO2 at 1 atm  ->  one composition, several crystals

  polymorph      stable range (equilibrium)   density     framework
  -----------    --------------------------   ---------   ---------------
  quartz         up to   867 C                2.65 g/cc   tight, twisted
  tridymite       867 - 1470 C                2.26 g/cc   open, layered
  cristobalite   1470 - ~1723 C               2.32 g/cc   open, cubic-ish
  liquid         above ~1723 C                  melt      network fluid
  (fused silica = the melt quenched to glass)  2.20 g/cc   frozen liquid

  each crystal ALSO has a fast low<->high (alpha<->beta) twin:
     quartz       alpha<->beta   573 C     ~0.8 %  volume   "quartz inversion"
     cristobalite alpha<->beta  ~220 C     a few % volume   (worst for brick)
     tridymite    alpha<->beta  117,163 C   small

  hotter crystals are MORE OPEN (lower density): atoms want room when hot

The whole silica story on one card: three crystals in their temperature bands, their densities, and the fast alpha-beta twin each one carries. Notice the high-temperature forms are the roomier, lower-density ones.

Read the density column and a pattern jumps out: the hotter a form is stable, the more open and less dense it is — quartz 2.65, tridymite 2.26, cristobalite 2.32 grams per cubic centimetre. That is exactly what you would guess from thermal expansion writ large: heat makes atoms jostle harder and demand more elbow room, so the framework equilibrium relaxes into airier arrangements. It also means every jump from one form to the next carries a volume change, and volume changes inside a rigid, brittle solid are precisely what we will have to manage when we fire real silica.

Two kinds of change: rebuild versus tilt

Silica's transformations come in two utterly different flavours, and telling them apart is the master idea of this guide. A reconstructive change — quartz to tridymite to cristobalite — must actually break Si-O bonds and stitch the framework back together in a new pattern. That is demolishing a house and rebuilding it from the bricks: it costs a huge activation energy, crawls along, and in practice needs very high temperature, long soaks, and often a trace flux (a mineraliser) to happen at all. A displacive change is the opposite. No bond is broken; the tetrahedra merely tilt and rotate a little on their shared corners, like the same house leaning slightly in the wind. It is instant, effortless, and perfectly reversible.

The famous displacive one is the low-to-high inversion of quartz at 573 degrees C — every crystallographer's landmark, and every potter's, too. Cross 573 on the way up and each quartz grain snaps from its low (alpha) form to its slightly more open high (beta) form, expanding about 0.8 percent by volume all at once; cross it on the way down and every grain contracts just as abruptly. Because no bonds break, this happens on every single heat and cool of the kiln, sharp and unavoidable at that one temperature. The change is benign if the whole piece crosses together and slowly; it turns vicious when one region is expanding while its neighbour has not yet reached 573.

Why real silica ignores its own diagram

String the equilibrium fields together and the 1 atm map of silica reads simply: quartz up to 867 degrees C, then tridymite to 1470, then cristobalite to about 1723, then liquid. But that is the equilibrium story, and equilibrium is exactly what the reconstructive steps are too lazy to reach. Heat ordinary quartz sand quickly and it will sail straight past 867 and 1470 without ever becoming tridymite or cristobalite, then melt directly from quartz — the phase diagram promised conversions that kinetics never delivered. This is the honest, load-bearing caveat of every phase diagram: it shows what is stable, never how fast the system gets there.

The same laziness rules the melt. Cool liquid silica and, given the chance, it should crystallise back to cristobalite — but the reconstructive rebuild is so sluggish that an ordinary cooling rate freezes the tangled liquid network in place instead, giving fused silica glass: a liquid caught mid-freeze, not a substance that ever crystallised. Two honest points follow. A glass is not a slow-flowing liquid at room temperature — window panes do not sag over centuries; it is a rigid solid whose atoms are simply frozen in a liquid-like pattern, still built of the same SiO4 tetrahedra with sharp short-range order. And held hot for long enough, silica glass will finally give in and grow cristobalite crystals — the unwanted crystallisation called devitrification. Metastable silica is the rule in a real kiln, not the exception.

Off the bottom of an ordinary firing chart lie two more polymorphs that need the other knob — pressure. Squeeze silica to a few gigapascals and it forms coesite (density near 2.9); squeeze past roughly 10 GPa and you get stishovite (density about 4.3), where the change is dramatic: silicon abandons its four-oxygen cage for a six-oxygen one, exactly the coordination jump the radius-ratio rule taught you, now driven by pressure instead of ionic size. These are the crystals of meteorite-impact craters and the deep mantle. They are a vivid reminder that a diagram drawn at 1 atm is only one slice of a bigger map, and that both temperature and pressure decide which architecture wins.

Putting the polymorphs to work

All this pays off the moment you fire silica for real. A silica brick is the classic refractory for the crowns of glass tanks and the roofs of coke ovens, and above about 700 degrees C it is superb — dimensionally steady, load-bearing to nearly its melting point. Its weakness is entirely down low, where the fast displacive inversions live. Cross 573 (quartz) or, worse, the fat inversion near 220 (cristobalite) too quickly, and different parts of the brick expand or contract out of step; the abrupt volume mismatch sets up thermal stress that a brittle, tension-hating solid answers with a crack. The same trap catches the free quartz grains in an ordinary clay body: every kiln cycle marches them through 573, and a careless cool cracks the ware — the potter's dreaded dunting.

Below about 500 degrees C, ramp freely — no silica inversion is in play yet, so speed here costs nothing.
Approaching 573, slow right down (a common rule of thumb is 50 to 100 degrees C per hour) through the quartz inversion, so every grain makes its ~0.8 percent jump gently and together, not as a shockwave.
Above roughly 600, the inversion is done; open the throttle again and climb to the peak soak.
On cooling, slow through 573 a second time — the contraction is just as abrupt, and a fast cool here is the classic cause of dunting.
If the body was high-fired and now holds cristobalite, creep even more gently through its ~220 inversion, whose several-percent jump is the bigger hazard of the two.

Step back and see what one component has already taught us. A single map, read through the phase rule, distinguishes fields, lines, and an invariant point; a single composition can wear several crystalline coats; and the speed of change — lazy reconstructive versus snappy displacive — decides whether a fired body ends up at equilibrium or full of metastable survivors we must fire around. Every idea here returns, enriched, once we add a second component in the next guide: the map gains a whole new axis of composition, two-phase fields open up, and a low-melting eutectic explains why a pinch of flux can slash a firing temperature by hundreds of degrees.