Grains, Grain Boundaries, and Porosity

From powder particles to a mosaic of grains

Guide 1 argued that in ceramics the microstructure is the product. Now we open it up. Fire a powder compact — the snowman firming up, particles welding at their necks while pores shrink, all without melting — and the solid you pull out is not one crystal but a dense mosaic of them. Each little crystal tile is a grain: a region where the atoms line up in one continuous orientation, a single crystal a few microns across. A finger-sized alumina part packs in millions of them. Meet the first actor, the grain, the basic building block you will read, measure, and design for the rest of this rung.

Picture a dry-stone wall, or a frozen foam of soap bubbles: the grains are the stones or the bubble bodies, packed to fill space with almost no gaps. But do not confuse a grain with the powder particle you started from. During firing, atoms hop across from one grain to another, small grains dissolve away and big ones fatten, and grain boundaries sweep through the solid. So the grains in the finished part have grown, changed shape, and re-sorted; the microstructure you see is what firing made, not what you poured in.

The grain boundary: the seam where two crystals meet

Two neighbouring grains almost never share the same orientation — one is tilted this way, its neighbour that way. Where they meet, the atoms cannot keep both lattices happy at once, so a thin layer of squashed, mismatched atoms forms between them: the grain boundary, the second actor. It is only a fraction of a nanometre to a few nanometres wide, but it is a genuine feature of the solid — a region of disorder and higher energy, like the grout line between two tiles or the seam where two sheets of wallpaper don't quite line up.

That looseness has consequences. Because the atoms are loosely packed and higher in energy, a boundary is a fast lane for diffusion — atoms scoot along it far quicker than through the crystal interior, which is exactly how firing moves matter to shrink pores. It is also the favourite parking spot for impurities and, above all, for the thin glassy film that liquid-phase sintering leaves behind — the quiet governor of high-temperature strength and creep that guide 3 is devoted to. And where three grains meet, their boundaries share a line called a triple junction; the angle the boundary makes as it dips into a pore or a second phase there, the dihedral angle, decides whether that pocket wets and spreads or stays balled up.

So a boundary is double-edged. It can help: it scatters heat-carrying phonons, so a fine-grained ceramic conducts heat a little less, and it can pin a would-be crack briefly. It can also hurt: at high temperature the glassy film there softens and lets grains slide past each other, and boundaries are where corrosion and electrical breakdown often begin. Whether boundaries are your friend or your enemy depends on what you are asking the part to do — the recurring theme of this rung.

Porosity: the leftover that almost never fully leaves

The third actor is the empty space. A powder compact starts around 55 to 60 percent of its theoretical density — nearly half of it is air between particles. Firing drives that air out: the body shrinks about 15 to 20 percent in each linear dimension as the density climbs past 98 percent of theoretical. But the last couple of percent of pores are stubborn, and almost every ceramic keeps some. That residual empty space is the porosity. It is quoted as a fraction: a part at 3.90 g/cm3 against alumina's theoretical 3.99 g/cm3 is 3.90/3.99 = 97.7 percent dense, so 2.3 percent porous — and you measure it simply, by weighing the part dry, wet, and suspended in water (Archimedes).

Not all pores are alike, and the difference matters. Early in firing the empty space is one connected sponge of channels that all reach the outer surface — open porosity. Water can soak in, gas can flow through, and a fresh terracotta pot drinks like this. As densification runs on, those channels pinch off; somewhere around 90 to 95 percent of theoretical density the sponge breaks up into isolated sealed bubbles — closed porosity, like the trapped bubbles in a chocolate bar. A body with only closed pores is watertight even though it is not fully dense, which is why a glazed, high-fired porcelain rings and holds liquid while low-fired earthenware seeps.

  PORE TYPE   WHERE IT SITS              CAN IT STILL LEAVE?
  --------------------------------------------------------------
  open        connected channel that     yes - vents to the outside;
              reaches the outer surface  the part soaks up water

  closed      sealed bubble sitting on   slowly - trapped gas must
              a boundary / triple point  diffuse out along the wall

  trapped     swallowed INSIDE a grain   almost never - only by slow
              (a boundary ran past it)   lattice diffusion: near-permanent

Why do the last pores dig in? Here is the twist that catches beginners: hotter and longer is not always better. A shrinking pore normally clings to a grain boundary and rides along as the boundary migrates — the pore-boundary interaction — and there it drains away easily down the fast boundary path. But if the boundary moves too fast, as when a few grains grow abnormally and run away (guide 4's subject), it breaks free and leaves the pore stranded deep inside a grain. Cut off from the boundary, that trapped pore can only shrink by sluggish lattice diffusion, so it is essentially frozen in forever. Overfiring can therefore coarsen the grains faster than it removes the pores and lock in the very porosity you were trying to burn out.

How grains, boundaries, and pores read out as properties

Now the payoff: this three-actor picture is not decoration, it is what the property meter reads. Start with porosity, because empty space carries no load and no stiffness. A workable rule for the elastic modulus is E = E0 times (1 - 1.9P + 0.9P^2): at 10 percent porosity (P = 0.1) that is 1 - 0.19 + 0.009 = 0.82, so the ceramic is already about 18 percent floppier than the pore-free ideal. Strength falls even faster, roughly as strength = strength0 times exp(-bP) with b around 4 to 7; at 10 percent porosity that can cost you 40 percent of the strength, because each pore is not just missing material but a stress raiser. This is the front line of the microstructure-property relationship.

Grain size sets strength too, through the same flaw logic from the mechanics rung. A ceramic is only as strong as its worst flaw, and the largest flaw usually scales with the largest grain — a big grain is a big natural crack. Because the Griffith criterion makes strength go as 1/sqrt(flaw size), halving the grain size raises strength by roughly the square root of two, and in practice a fine 1 to 2 micron alumina reaches 500 to 600 MPa where a coarse 50 micron one manages barely 200 MPa. Be honest about the caveat: this is statistical, not deterministic — the strength you get is a Weibull distribution, and a bigger part, holding more flaws, is weaker on average. Finer, more uniform grains give a higher and more reliable strength.

Boundaries and pores also gate light and current. Ordinary alumina is white and opaque because every pore and every boundary is a tiny lens that flings light sideways — frosted glass, not window glass. Scrub the porosity below about 0.1 percent and keep the grains fine and clean and the same alumina turns into translucent alumina, the milky-clear envelope of a sodium street lamp. On the electrical side the boundary can be engineered to block current (a varistor clamps a voltage spike right at its boundaries) or to carry it (an oxygen sensor conducts ions through them). Same three actors, dialled to a different property.

Seeing it: the craft of ceramography

You cannot design what you cannot see, and grains and boundaries are microns wide, buried inside an opaque solid. Revealing them is a craft of its own, ceramography — sister to the metallographer's, but harder, because ceramics are so hard and brittle that careless grinding tears grains out and leaves pits you might mistake for pores. The goal is a flat, mirror-polished section that shows the true microstructure, then an image sharp enough to measure.

1. Section and mount. Cut a slice with a diamond saw, then set it in an epoxy puck so its edges and pores are supported and won't crumble.
2. Grind flat. Work down through coarse-to-fine diamond abrasive to take out the saw damage and get a truly flat face.
3. Polish to a mirror. Finish with fine diamond paste, down to a micron or less, until the surface is featureless and reflective.
4. Etch to reveal boundaries. A mirror shows nothing yet. Attack the higher-energy boundaries — briefly heat the sample (thermal etch) or use a chemical etch — so they groove and stand out, uncovering the grain mosaic.
5. Image it. Look under an optical microscope or, for the fine detail, a scanning electron microscope (SEM), whose depth of field and resolution make grains, pores, and any second phase pop.

From that image you read the whole cast: the grains and their sizes, the boundaries and what decorates them, and the pores and where they sit. Guide 4 turns the reading into a number with the linear-intercept method for grain size, and guide 3 zooms in on the second phases and the glassy boundary film. The mindset all five guides build toward is microstructural design — choosing grain size, density, and phase distribution on purpose to hit the strength, transparency, or conductivity you need. See the microstructure clearly and you can steer it; leave it to chance and it steers you.