Reading the Ceramic Systems: Alumina-Silica, Spinel, Zirconia

The Payoff: Three Maps a Ceramist Actually Uses

Four guides ago this rung handed you a small toolkit for reading a map of what is stable at each temperature and mix. You learned to count with the Gibbs phase rule — F = C - P + 1 at fixed pressure — to walk a binary diagram down through its liquidus and solidus, to find the lowest pass where liquid first appears at a eutectic, to weigh how much of each phase is present with the lever rule, and to spot a compound that melts the awkward way at a peritectic. This last guide spends the whole toolkit at once, on the three real systems a ceramist reaches for daily.

The three share a family resemblance worth naming up front. In MgO-Al2O3 an intermediate compound, spinel, sits in the middle of the diagram like a wall and splits one binary into two simpler ones. In Al2O3-SiO2 a single intermediate compound, mullite, carves out the industrially vital field behind every fired clay body. And in the ZrO2 systems the trouble is not a compound at all but a polymorph — zirconia flips between crystal forms as it cools, and we alloy it just to tame that flip. Read all three and you can read almost any ceramic diagram you meet.

Keep one caution in the corner of your eye the whole way. A phase diagram is a map of equilibrium — of what would form given endless time. Real firing is fast, so a fired part almost always strands unmelted grains and a frozen, metastable glass that the map never shows. The diagram is a compass, not a GPS: it tells you which way is downhill, not exactly where your kiln will leave you.

MgO-Al2O3: Spinel Splits the Map in Two

Line up magnesia and alumina and, right in the middle, they lock together into a single stubborn compound: spinel, MgAl2O4, a cubic crystal with Mg2+ in tetrahedral holes and Al3+ in octahedral holes of a close-packed oxygen array. Spinel melts congruently near 2135 degrees C — that is, it melts straight to a liquid of its own composition, cleanly, without decomposing. That single fact is the key to the whole diagram, because a congruently-melting compound behaves like a pure end-member.

So draw a vertical line up through spinel and the messy binary falls into two tidy halves you already know how to read: a simple MgO-spinel eutectic system on the left, with its lowest pass near 2000 degrees C, and a simple spinel-Al2O3 eutectic on the right, dipping to roughly 1925 degrees C. Each half is just the eutectic diagram from guide three, complete with a liquidus you slide down and a lever rule you apply to read how much liquid versus solid you have at any point. Spinel itself also swallows a fair range of excess alumina into a solid solution at high temperature, so its field is a band, not a razor line.

Why does industry care? Because spinel has no destructive low-temperature polymorph and stays solid to more than 2000 degrees C, it makes a superb refractory — the lining that survives inside cement kilns and steel ladles where cheaper bricks would slump. Reading the diagram tells the refractory-maker exactly how much alumina to add to a magnesia brick to grow a strong spinel bond, and how far to stay from each eutectic so that no unwanted liquid appears at service temperature. A phase diagram, put to work, is a recipe for staying solid.

Al2O3-SiO2: Mullite, the Keystone of Fired Clay

If a ceramist could keep only one diagram, it would be this one, because clay is aluminosilicate and firing clay is what humans have done for ten thousand years. Between silica and alumina sits exactly one intermediate compound, mullite, roughly 3Al2O3 . 2SiO2 (about 72 wt% Al2O3), the needle-like crystal that reinforces porcelain like rebar in concrete. The story is dominated by one low number: on the silica-rich side there is a eutectic near 1587 degrees C at only about 6 wt% Al2O3. That low pass is why a clay body can be fired dense in a kiln that need not reach silica's own 1713-degree melting point.

  Al2O3 - SiO2   (the ceramist's keystone diagram)

  T/degC
  2054 |.................................. Al2O3 (corundum) melts 2054
       |                              _.-'
 ~1890 |                        M --'          mullite ~3Al2O3.2SiO2
       |                  _.-'  /  \             (~72 wt% Al2O3)
  1713 * SiO2         _.-'     /    \
       |  \        _.-' LIQUID/      \  mullite + liquid
       |   \     _.-'        /        \
       |    \ _.-'          /          \
  1587 +------E-----------+/            \----- silica-side eutectic
       | cri  |\  mullite + cristobalite \
       | +    | \                         \
       | mul  |  \                          \
       +------+---+--------------------------+----> wt% Al2O3
       0     ~6                             72        100
      SiO2   E                            mullite    Al2O3

  E ~1587 degC: the low pass where a clay body first makes liquid.
  More Al2O3 -> more mullite, less liquid -> higher refractoriness.
  (near the Al2O3 end the books disagree -- see the note below)

Now read the diagram like a refractory buyer. A cheap fireclay brick, low in alumina, sits near that 1587-degree eutectic, so at high temperature the lever rule says a large fraction of it turns to soft liquid and it slumps early. Add alumina and you march up into the mullite field: more mullite crystals, less liquid at any given temperature, and therefore higher refractoriness. Push all the way to a high-alumina or nearly-pure-alumina brick and the corundum liquidus towers to 2054 degrees C. That single trend — more alumina buys more heat resistance — is exactly what the alumina-silica system draws for you, and it is why refractories are sold and priced by their alumina content.

The ZrO2 Systems: Caging a Destructive Flip

Zirconia is the odd one out, because its problem is not mixing but a polymorph. Pure ZrO2 changes crystal form as it cools: cubic (the fluorite structure) down to about 2370 degrees C, then tetragonal down to about 1170 degrees C, then monoclinic all the way to room temperature. The last change, tetragonal-to-monoclinic on cooling, is a sudden diffusionless flip that expands the crystal by roughly 4 to 5 percent — enough to shatter a part from the inside, the way ice cracks a bottle. So a block of pure zirconia simply cannot survive being fired and cooled; on the map of zirconia's polymorphs that harmless-looking phase boundary is a wrecking ball.

The cure is to turn the unary problem into a binary one — alloy the flip away. Dissolve a stabilizer such as Y2O3, MgO, CaO, or CeO2 into the zirconia and you form a solid solution that holds the higher-symmetry cubic or tetragonal form all the way down to room temperature. There is a bonus in the chemistry: because Y3+ carries less charge than the Zr4+ it replaces, every pair of yttrium ions leaves one oxygen site empty, and those vacancies both prop up the fluorite structure and make the material an oxygen-ion conductor. Fully stabilized cubic zirconia at about 8 mol% Y2O3 — yttria-stabilized zirconia — is the electrolyte in oxygen sensors and fuel cells; a leaner mix near 3 mol% Y2O3 holds fine tetragonal grains in a metastable freeze. Where you land is read straight off the ZrO2-Y2O3 binary.

Then comes the beautiful trick: turn the wrecking ball into a shield. In partially-stabilized zirconia those metastable tetragonal grains are poised on a hair trigger. When a crack tries to run through the ceramic, the intense stress at its tip sets the nearby grains off — they flip to monoclinic and swell by that same 4 to 5 percent, and the expansion clamps down on the crack and squeezes it shut. It is an airbag for a crack, and it is why transformation toughening lifts zirconia's fracture toughness to roughly 6 to 12 MPa sqrt(m), several times tougher than ordinary alumina, tough enough for knife blades and dental crowns.

A First Ternary, and Putting the Maps to Work

Real bodies rarely have just two components, so meet the ternary diagram. Draw a triangle: each corner is one pure oxide, each edge is a binary you already know, and every point inside is a three-way mix. Temperature rises out of the page as a surface, drawn with contour lines like a topographic map; the valleys run downhill to a low point where liquid first appears — a ternary eutectic. The porcelain system K2O-Al2O3-SiO2 is the one to remember: its lowest eutectic sits near 985 degrees C, a thousand degrees below pure silica's melting point. That single low valley is the whole reason a pinch of feldspar flux slashes the firing temperature of a clay body.

1. Fix the body composition. Say a fired body is about 60 wt% SiO2 and 40 wt% Al2O3; mark that overall point on the alumina-silica line.
2. Pick the hold temperature and see which two phases coexist. On the silica side of the mullite field, the body is mullite crystals plus a silica-rich liquid — the liquid that will later freeze to glass.
3. Read the two phase compositions off the diagram: mullite at about 72 wt% Al2O3, and the liquid on the liquidus at, say, about 10 wt% Al2O3.
4. Apply the lever rule. Fraction liquid = (72 - 40) / (72 - 10) = 32 / 62, about 0.52 — so roughly 52 percent of the body is liquid at that temperature.
5. Read it as glass. On cooling that 52 percent liquid freezes to a glassy bond phase cementing the mullite together — a quick, honest estimate of a fired body's glass content straight from the map.

Now the payoffs line up. Choosing a refractory means staying clear of low eutectic valleys and deep into a mullite, spinel, or corundum field so no soft liquid forms at service temperature. Predicting glass content is the lever-rule reading you just did. And a flux works precisely because it drops you into a low eutectic: add a little K2O or Na2O and the first liquid appears near 985 rather than 1713 degrees C — the same bridging-oxygen chemistry you met with glass, now read as a low point on a map. That is this rung's real gift. Equilibrium is a compass; you still steer it against the kinetics of real firing, which is exactly where the next rung — the chemistry of defects and diffusion that lets a solid react and densify at all — takes you.