One fact, one personality
In the last guide you learned what a ceramic is: an inorganic, non-metallic solid, usually shaped from powder and consolidated by firing. That answered what it is. This guide answers why it acts the way it does — hard, stiff, refractory, chemically calm, and yet heartbreakingly brittle. Remarkably, almost all of that personality flows from a single fact: how the atoms hold hands. In a ceramic, that grip is a mixed ionic-covalent bond.
It helps to line up three ways atoms bond. A metal shares a loose sea of electrons that every atom can wade through; because nothing points in a particular direction, whole planes of atoms can slip past one another and the metal simply bends. A ceramic has no such sea. Its electrons are locked down in two ways at once — partly handed over from a metal-like atom to a non-metal (the ionic share, positive cations and negative anions clamped by electrostatic pull, the picture behind the ionic model), and partly shared along fixed directions between neighbours (the covalent share). Both grips are strong; neither lets atoms quietly change partners.
How much of each? That is set by the electronegativity difference between the two atoms — how unequally they tug on the shared electrons — and it is captured as a percent ionic character. The gap between silicon and carbon is small, so SiC is roughly 90 percent covalent; oxygen pulls far harder than aluminium, so Al2O3 comes out a bit over half ionic (around 60 percent); MgO, with an even wider gap, lands near 70 percent ionic. No ceramic is purely one or the other — the honest word really is mixed, and the exact percentages shift depending on which electronegativity table you use.
Why hard, stiff, and heatproof — the good news
Now the pay-off. If every atom is held by strong bonds pointing in every direction, it takes a great deal of energy to pull any atom out of place — and that one idea explains most of the good news. To melt a ceramic you must shake its ions clean out of their bonds, so the melting point runs high; this is what we mean by refractoriness, the ability to keep its shape red-hot. Alumina (Al2O3) melts near 2054 degrees C, roughly three times the melting point of aluminium metal. The deeper the lattice energy — the energy holding the ionic crystal together — the higher the melting point tends to run.
The same rigid bonds make a ceramic stiff and hard. Stiffness is just how much the bonds resist being stretched: pull on alumina and the bonds barely give, so its Young's modulus is around 380 GPa — stiffer than steel. Hardness is resistance to being scratched or dented, and since the atoms refuse to slide past one another, ceramics sit at the very top of the hardness charts — diamond, boron carbide, silicon carbide. It is exactly why we make cutting tools, grinding wheels, and armour out of them.
Two more traits fall out of the same bond. A ceramic is chemically inert because it is already the burnt-out, fully-reacted product — an oxide is a metal that has already rusted all the way, resting at the bottom of the energy valley with little left to give up, which is why alumina shrugs off acids and stays put in a furnace. And because every electron is pinned into an ionic or covalent bond, none are free to drift and carry current, so the default ceramic is an excellent electrical insulator — the white core of a spark plug is alumina for exactly this reason. (A handful of ceramics are engineered to conduct instead; we will meet those surprises much later.)
How the ions stack — packing and the radius-ratio rule
The ionic picture does more than explain strength — it dictates the very shape of the crystal. Picture the big anions (in an oxide, the O2-) packed like oranges in a crate: this is anion close-packing. The much smaller cations then tuck into the gaps left between them. There are two sizes of gap — a small tetrahedral hole ringed by 4 anions, and a roomier octahedral hole ringed by 6. The count of anions touching a cation is its coordination number, and which hole it chooses comes down to one thing: how big the cation is compared with the anion.
- Find the two sizes. Look up the cation and anion from tables of ionic radius. For MgO, Mg2+ is about 0.072 nm and O2- about 0.140 nm.
- Divide small by large. The radius ratio is r(cation) / r(anion). For MgO that is 0.072 / 0.140 ≈ 0.51.
- Read off the coordination. The critical radius ratios mark the thresholds: 0.225 to 0.414 favours a tetrahedral hole (coordination 4); 0.414 to 0.732 favours an octahedral hole (coordination 6); above 0.732 the cation is big enough for 8 neighbours (cubic). MgO's 0.51 lands squarely in the octahedral band.
- Predict the structure. So each Mg2+ sits in an octahedral hole with 6 oxygen neighbours — exactly the rock-salt structure sketched below. Treat the rule as a guide, not a law: strongly covalent bonds can override it, but for ionic oxides it works remarkably often.
ROCK-SALT (NaCl-type) STRUCTURE - e.g. MgO, 6:6 coordination
Anions (O2-) close-pack on an FCC lattice;
a cation (Mg2+) fills EVERY octahedral hole.
one octahedral hole: the unit cell (a slice):
O O---O---O
| | + | + | O = O2- (large anion)
O---+---O O---O---O + = Mg2+ (small cation)
| | + | + |
O O---O---O
(+ also touched by
one O above and every ion is touched
one O below -> 6) by 6 UNLIKE neighboursThe catch: brittle, with no way to yield
Here is the twist: the bonds that make a ceramic strong also take away its escape hatch. When you overload a metal, tiny line defects called dislocations glide — planes of atoms shuffle past one another, so the metal yields, bends, and soaks up energy instead of snapping. That is ductility. A ceramic cannot do this. Its bonds are directional and charged, so to slide one plane over another you would have to snap many bonds at once and shove like charges together — anion onto anion — which is ruinously expensive. So dislocations barely move at room temperature. With no way to yield, the ceramic stays perfectly rigid right up to the instant it breaks: this is brittleness, and the break itself is brittle fracture.
That inability to flow has a sharp consequence. Any little notch, pore, or scratch — a flaw — acts as a lever that concentrates stress at its tip. A sharp crack can multiply the local stress tens or hundreds of times; this is stress concentration. A metal would simply yield at the tip and blunt it, spreading the load. A ceramic, unable to yield, keeps the tip needle-sharp — so the stress piles up until a bond right at the tip lets go, and then the crack runs. The strength you measure is really the strength of the worst flaw in the piece.
Strength is only as good as the worst flaw
If you could pull on a perfect crystal, you would have to break every bond across a plane at once, and the theoretical strength would be enormous — very roughly one-tenth of the Young's modulus, tens of GPa. Real ceramics break at a few hundred MPa, about a hundred times weaker. The gap is due entirely to pre-existing flaws. Griffith's insight was that a crack runs when the elastic energy released by opening it a little exceeds the energy needed to make the fresh crack surfaces — and the bigger the flaw, the sooner that happens. That is why ceramic strength is flaw-controlled, not a fixed material constant.
Griffith's result boils down to a clean formula: the strength is about K_IC / sqrt(pi x c), where c is the flaw size and K_IC is the fracture toughness — the material's built-in resistance to a running crack. Put in real numbers for alumina: with K_IC = 3 MPa sqrt(m) and a flaw of 30 micron (c = 30 x 10^-6 m), the strength comes out near 3 / sqrt(3.14 x 30 x 10^-6) ≈ 3 / 0.0097 ≈ 310 MPa — right where measured alumina strengths sit. Halve the flaw to 15 micron and the strength jumps by sqrt(2), to about 440 MPa. Polishing a ceramic really does make it stronger, because you are shrinking its worst flaw.
There is one more twist, and it is statistical. Flaws are scattered through the piece at random, so no two specimens carry the same worst flaw — and therefore no two break at the same stress. Ceramic strength is not a single number but a distribution, described by Weibull statistics. The Weibull modulus m measures how tight that scatter is: a typical ceramic has m near 10 (broad scatter), whereas a ductile metal is above 50 (very repeatable). And because the piece fails at its single weakest link — like a chain — a larger part contains more flaws and more chances to hide a bad one, so bigger ceramic parts are, on average, weaker. This counter-intuitive size effect is real, and engineers design around it with proof-testing and by keeping stressed volumes small.
Fighting back — and honest limits
Brittleness is not a life sentence — engineers fight it by raising the fracture toughness. The most elegant trick is transformation toughening in zirconia (ZrO2). Tiny zirconia grains are held, metastably, in a high-temperature crystal form, poised to switch. When a crack approaches, the intense stress at its tip triggers those grains to transform and expand by a few percent, clamping down on the crack and squeezing it shut — an airbag that deploys right where the crack is trying to open. Toughened zirconia can reach K_IC around 10 MPa sqrt(m), several times tougher than plain alumina, which is why it is used for knife blades and hip-joint balls.
Step back and the whole personality lines up behind one fact. The mixed ionic-covalent bond makes ceramics hard, stiff, refractory, inert, and insulating — and the very same rigid, directional bond makes them brittle and flaw-sensitive. But notice where the flaws come from: they are baked in during processing — by the powder, the shaping, and the firing. That is the thread the rest of this rung pulls on. Next we tour the ceramic family — oxides, carbides, nitrides, silicates, and glasses — and then watch how processing, structure, properties, and performance lock together. One bond sets the potential; how you make the part decides how much of it you actually keep.