The Plastic Family: Powder Plus a Bit of Glue
The first three guides gave us two ways to force a powder into shape. Guide 2 rammed dry powder into a rigid die by pressing, fast but fighting wall friction; guide 3 let a plaster mould suck the water out of a runny slip. This guide takes up a third family, the one that sits between bone-dry powder and a pourable liquid: mix the powder with just enough liquid and organic binder to make a soft, stiff dough — a paste — that flows when you push it and freezes into shape when you stop. This is plastic forming, and its oldest member is the lump of wet clay on a potter's wheel.
What turns loose powder into a workable paste is plasticity — the dough must yield and flow under the tool, then stand up under its own weight the instant the tool lets go. In the rheologist's terms it wants a yield stress (a threshold force below which it behaves like a solid) and it should be shear-thinning (thinner while it is being worked, stiffer at rest). Wet clay gets this for free: its platy clay particles slide on a film of water like wet playing cards, gliding past one another yet clinging when still. Non-clay powders — alumina, zirconia, silicon nitride — have no such gift, so we buy plasticity in a bottle, blending in an organic binder to hold the particles together and a plasticizer to keep the mix supple.
Extrusion: A Stiff Paste Through a Die
Extrusion is the Play-Doh Fun Factory of ceramics. A stiff, dough-like paste is forced by a screw (an auger) or a piston through a shaped opening — the die — and comes out as a continuous length of whatever cross-section the die carves: a solid rod, a hollow tube, a square bar, a fluted profile. Because the die sets the profile and the machine just keeps pushing, extrusion is wonderfully suited to anything long and uniform. Cut the endless extrudate to length and you have made pipes, rods, electrical insulators, and — the reason your car is not poisoning the street — the honeycomb catalyst support.
That honeycomb is extrusion's showpiece. A single cordierite die extrudes a monolith threaded with hundreds of parallel channels — commonly 400 to 900 cells per square inch, with walls only about 0.1 mm thick — so that exhaust gas meets an enormous catalyst-coated surface inside a small block. No other forming method makes such a thing in one cheap, continuous pass. At the other end of the value scale the same idea shapes the humblest ceramics of all: building brick, roof tile, clay pipe, and blocks of refractory for furnace linings are almost all extruded from stiff clay paste, then cut to length with wires like blocks of cheese.
Extrusion works only if the paste has the split personality we met in section 1: fluid enough to flow through a narrow die under pressure, yet with enough yield stress to hold its shape the instant it emerges, before it has dried at all. Get that balance wrong and the classic flaws appear — the auger can fold the paste into spiral laminations (weak internal seams that peel apart on firing), and drag at the die wall can leave the surface moving slower than the core, tearing S-shaped cracks. And because an extruded body still holds a lot of water or binder, drying is a slow, nervous business: dry it too fast and it warps or cracks as the wet core shrinks against a stiff, already-dry skin. As always, a lamination you extrude in is a flaw you will find again after firing.
Ceramic Injection Molding: Small, Complex, and the Debinding Tax
For a part that is small but geometrically fiendish — a fibre-optic ferrule, a turbocharger rotor, an orthodontic bracket, a tangle of undercuts a die could never press — ceramists borrow the plastics industry's most productive machine. Ceramic injection molding (CIM) mixes fine powder with a large charge of thermoplastic binder, so much that the whole blend melts and flows like hot plastic. This feedstock is heated to a syrup, injected under high pressure into a cold steel mould where it freezes to shape, and popped out seconds later as a finished-looking green part — complex, precise, and made by the thousands from one mould. The catch is hidden in that word large.
- Compound the feedstock: blend the ceramic powder with roughly 40 to 50 vol% thermoplastic binder (waxes and polymers) into uniform pellets.
- Inject: melt the feedstock and shoot it into a cool mould cavity, exactly as a plastic part is moulded; it solidifies in seconds.
- Debind: slowly remove that huge binder fraction by heat, solvent, or catalyst — the delicate, days-long step where the part is most fragile.
- Sinter: fire the fragile brown body to full density, where it shrinks dramatically to its final size.
Everything hard about CIM lives in step 3. All that binder — up to half the volume — has to come out of a solid part without wrecking it, and it cannot leave all at once. Heat it too fast and the trapped organics boil into gas that has nowhere to go, blistering the part in a defect aptly called bloating; so debinding is deliberately crawled through over many hours or days, the most expensive and failure-prone stage of the whole route. Then comes the second tax: because the feedstock was only half powder, the space the binder vacated must all be closed by densification, so a CIM part shrinks a hefty 15 to 20 percent on every dimension during firing — predictable, but unforgiving of any density variation. This is why CIM rules the world of small precision parts and never the large ones: the bigger the part, the longer the debind and the more that a percent of uneven shrinkage throws it out of tolerance.
Gelcasting: Freeze the Shape in the Mould
Gelcasting is the clever child of slip casting and CIM, keeping the best of each. Like a slip, it starts from a fluid, high-solids suspension you can pour into any mould, even a complex non-porous one; unlike a slip, it does not wait for a plaster wall to slowly wick the water away. The trick is that the liquid carries a dissolved organic monomer. Add an initiator, and the monomer polymerizes in place — the whole body sets from a pourable slurry into a firm, rubbery gel in minutes, the way liquid gelatine sets into jelly. Gelcasting freezes the shape everywhere at once, so it fills intricate moulds that slip casting is too slow for and CIM is too binder-heavy for.
The payoff is in the numbers. Where CIM needs its binder to fill half the volume, gelcasting's gel network needs only a few weight percent of organics — often just 3 to 4 percent — because the monomer is not there to make the whole thing flow, only to lock the already-dense particle packing in place. So the green density is high and uniform, the firing shrinkage is modest, and — the property that made gelcasting famous when Oak Ridge developed it around 1990 — the green strength is high enough to pick up, handle, and even machine the unfired part on a lathe. You can turn a green blank to near-net shape with ordinary tools, then fire it. Little binder to remove, little to shrink, and a strong green body: it comes close to the ceramist's ideal, when it works.
Old Hands and New Machines: From the Wheel to the Printer
Before any of this, there were hands. The whole traditional industry rests on one happy accident of nature — that wet clay is plastic — and on the machines that exploit it. Throwing on a potter's wheel is plastic forming in its purest form, the potter's fingers serving as a living die. For flatware at scale the die is mechanized: in jiggering, a lump of clay is slapped onto a spinning plaster mould that shapes one face while a profiled tool lowered from above shapes the other, stamping out plate after identical plate. Jiggering and its cousin jolleying are why your dinner plates and mugs are cheap; they are extrusion's and pressing's low-tech siblings, and they work only because clay forgives.
At the far new end sit the printers. Ceramic additive manufacturing builds a part layer by layer straight from a CAD file, with no die and no mould at all — so it makes shapes none of the older methods can, from lattices to internal channels. The routes rhyme with the ones we have met: stereolithography cures a powder-loaded photopolymer resin with light, layer by layer; binder jetting glues a powder bed one printed layer at a time; and robocasting, or direct ink writing, simply extrudes a shear-thinning paste along a programmed path — the section-2 extruder, freed from its fixed die. The freedom is real, but so is the price: printing is slow, and every layer interface is a potential weak plane, a built-in flaw waiting to delaminate.
FORMING METHODS AT A GLANCE (each makes a GREEN BODY -> then sintering)
method liquid / organic green result signature flaw
------ ---------------- ------------ --------------
die pressing ~0-2% (dry) ~55-60% TD density gradients
cold isostatic ~0-2% (dry) ~60% TD, even slow; rubber bag
slip casting ~25-35% water ~60-65% TD slow; wall gradient
tape casting solvent + binder thin flat sheet warping / camber
extrusion ~15-25% paste ~55-60% TD laminations, S-crack
injection mold 40-50 vol% binder fills the mould bloating; long debind
gelcasting ~3-4% monomer ~55-60%, strong O2-inhibited skin
3D printing route-dependent any shape layer delamination
best for: pressing = simple/flat CIP = uniform blanks slip = hollow/complex
tape = substrates/MLCC extrusion = tube/honeycomb/brick
CIM = small precision gelcasting = machinable green
3D printing = lattices, channels, one-offsStep back and the sprawl of methods collapses into one picture. Whether you press dry powder, cast a slip, extrude a paste, inject a feedstock, gel a slurry, throw on a wheel, or print layer by layer, you are doing the same job by different means: arranging particles into a shape and holding them there as a green body. The choice among them is a trade of shape freedom against the organic you must later remove and the shrinkage you must later manage. And every one of them is haunted by the same ghost — the uniformity of that green body. A density gradient from the die, a lamination from the auger, a bubble from the binder, a weak layer from the printer: each is a nonuniformity locked in now that firing will faithfully preserve. That green body — how dense it is, how even it is, and why that decides the fate of the fired part — is exactly where guide 5 takes us.