The Fork in the Road: What Austenite Becomes
In the last three guides you watched the same lever appear again and again: cooling rate. Guide 3 drew it as a curve on the continuous-cooling diagram — a line racing down out of the austenite field, and where it crosses decides everything. Now stand at that fork and name what waits down each road. Cool a piece of hot austenite slowly and its carbon has all the time it needs to sort itself out by diffusion, separating into soft ferrite and hard cementite laid down as coarse pearlite — the equilibrium product you met last rung.
Speed the cooling up and the same reaction still runs, but with less time the layers come out finer and harder — fine pearlite. Faster still, dropping the austenite to a lower temperature before it can transform, and diffusion gets so sluggish that the ferrite and cementite can no longer grow as neat parallel plates; they form a feathery, needle-like mixture called bainite, harder and tougher again. Every one of these is still a diffusional product — carbon atoms are migrating to build separate phases, just over shorter and shorter distances. The whole menu is a race between the clock and the atoms.
Diffusionless: A Transformation by Shear, Not by Sorting
What does diffusionless actually mean? Every transformation so far has worked by sorting: atoms hop, one at a time, over the free-energy barriers of diffusion, migrating until carbon-rich and carbon-poor regions have separated into distinct phases. That sorting takes time, and time is exactly what a fast quench refuses to give. The martensitic transformation does something completely different: instead of atoms wandering to new addresses, the whole lattice shears over in one coordinated, almost military move, every atom shifting less than a single bond-length in step with its neighbours.
Picture a deck of cards pushed sideways: no card leaves the deck, none swaps places with another, yet the whole stack tilts into a new shape in an instant. That is the martensitic shear. Because no atom has to diffuse anywhere, the transformation waits for nothing — a martensite plate grows at nearly the speed of sound, appearing faster than you can measure. And because nobody moved to a new neighbour, the new phase has exactly the same chemical composition as the austenite it came from. This is the deep difference: pearlite and bainite are the same iron and carbon re-sorted into two phases; martensite is the same iron and carbon, un-sorted, forced into one new crystal.
That same-composition fact is the whole reason martensite is a problem waiting to become useful. Hot face-centered cubic austenite is a roomy crystal that happily dissolves a lot of carbon in its interstitial gaps. The room-temperature equilibrium form, body-centered cubic ferrite, is a cramped crystal that holds almost none. Slow cooling lets the excess carbon diffuse out to build cementite. But the martensitic shear gives it no exit: the lattice flips toward BCC while every carbon atom is still frozen in place, wildly over-stuffed. What that trapped carbon does to the crystal is the next section.
The Body-Centered Tetragonal Trap
If BCC ferrite cannot hold the carbon and the shear does not let it leave, something has to give — and it is the crystal's shape. The trapped carbon atoms sit in interstitial sites lined up along one axis of the would-be cube, and they prop that axis open, stretching the cube taller than it is wide. The result is not quite BCC: it is body-centered tetragonal (BCT), a cube pulled into a slightly oblong box. Martensite is really just ferrite's crystal, over-stuffed with carbon it never had room for and held open under the strain — a supersaturated interstitial solid solution, frozen far past its solubility limit.
FCC AUSTENITE --(shear, no diffusion, ~speed of sound)--> BCT MARTENSITE
BCC ferrite (equilibrium, ~no C) BCT martensite (C trapped inside)
c = a c > a (stretched)
+--------+ +----------+
| | a | | c <- longer axis
| Fe | | Fe o | propped open
| | | C | a by trapped C
+--------+ +----------+
a a
c/a ratio: 1.00 (no C) --> ~1.02 at 0.4%C --> past 1.05 near 1%C
MORE carbon => MORE tetragonal => MORE distortion => HARDER, brittlerThe numbers make the trap concrete. With no carbon the c/a ratio is exactly 1.00 (plain BCC); it climbs roughly as 1 + 0.045 times the weight-percent carbon, so a 0.4 percent carbon martensite sits near 1.02 and a near-1 percent steel pushes past 1.05. Every bit of that stretch is stored strain, and stored strain is stored resistance to the shear of a gliding dislocation. More carbon means a taller cell, more distortion, and a harder phase — which is why, unlike almost everything else in metallurgy, martensite's hardness is set almost entirely by one number: its carbon content.
Why It's Hard — and Why It Depends on Temperature, Not Time
Martensite's legendary hardness is not one trick but three, stacked. First and biggest, the trapped carbon distorts the lattice so severely that dislocations can barely move — an extreme form of solid-solution strengthening, every carbon atom a boulder in the road. Second, martensite forms as a myriad of tiny plates or laths, and all those internal boundaries block slip just as fine grains do. Third, the shear itself leaves behind a colossal dislocation density, a tangle so dense it jams further motion — the same work-hardening pile-up you met before, but built in at birth. Stack the three and a fully martensitic 0.6 percent carbon steel reaches around 65 on the Rockwell C scale, roughly file-hard, several times harder than the pearlite the same steel would form if cooled slowly.
Now the strangest thing about martensite, and the reason it needs its own kind of line on the diagram. Because no diffusion is involved, the reaction does not need time — it needs cold. Martensite starts forming the instant the austenite drops below a fixed temperature called Ms (martensite-start), and more of it forms only as you cool further, finishing near Mf (martensite-finish). Hold the steel dead still at some temperature between Ms and Mf and, for ordinary carbon steels, the transformation simply stops — waiting longer does not make more. That is why Ms and Mf are drawn as flat horizontal lines across the bottom of the TTT and CCT diagrams, utterly unlike the time-curving noses above them. The amount of martensite is read off the thermometer, not the clock.
Hard, but Too Brittle to Use: The Case for Tempering
There is a catch, and it is a serious one. As-quenched martensite is not just hard — it is glass-brittle, and it comes out of the quench full of locked-in stresses from the sudden shear and the fast, uneven cooling. Drop a freshly quenched blade and it can shatter; sometimes it cracks in the quench tank all by itself. Hardness this extreme has bought its strength by spending nearly all of the steel's ductility — a stark case of the strength-ductility tradeoff you have met before. Raw martensite is a phase you almost never put to work as-is.
- Austenitize: heat the steel into the austenite field and hold it, so the carbon fully dissolves into that roomy FCC crystal.
- Quench: cool fast enough to dive past the diagram's nose before any pearlite or bainite can form, dropping below Ms so the austenite shears into hard martensite.
- Temper: reheat to a moderate temperature (roughly 150 to 650 degrees C) and hold, letting a little carbon finally diffuse out of the strained BCT lattice into fine carbides.
- Choose the trade: hotter or longer tempering sheds more hardness but buys back more toughness — you dial the balance to the job, from a razor edge to a shock-tough gear.
That controlled softening is tempering: shocking the steel glass-hard, then relaxing it just enough to be tough — the everyday quench-then-temper behind every knife, spring, and gear. Two honest reminders as you go. First, tempering changes strength and hardness enormously but barely touches stiffness: Young's modulus stays near 200 GPa through the whole cycle, because it is set by iron's atomic bonding, not by microstructure. Second, this diffusionless-then-reheat trick is one branch of a bigger family — aluminium and other alloys reach their strength by a cousin route, age hardening, where a phase is dissolved, quenched to trap it in supersaturation, then aged to precipitate it fine. Guide 5 takes up both tempering and age hardening in full.