Quench and temper, explained: martensite, retained austenite and the temper

Hardening a steel is three events in quick succession — soak it to austenite, quench it to martensite, temper it back to something tough. The middle step is where the physics lives, and it is almost entirely decided by chemistry and how cold you take it.

The quench: martensite start

Martensite does not form gradually as the part cools — it waits until the temperature crosses a threshold, the martensite start (Ms), then forms more the further below Ms you go. Andrews' equation sets that threshold from composition:

Ms(°C) = 539 − 423·C − 30.4·Mn − 17.7·Ni − 12.1·Cr − 7.5·Mo

Carbon's coefficient — 423 — dwarfs the rest: every 0.1% carbon drops Ms about 42°C. A 1018 mild steel sits near 440°C; AISI 4140 near 330°C; a 0.95% carbon blade steel down near 125°C. The martensite start temperature calculator runs the equation and the retained-austenite step below from a composition.

The leftover: retained austenite

The transformation never fully finishes at room temperature. The fraction that has turned to martensite by the time the part reaches the quench temperature follows Koistinen-Marburger:

martensite fraction = 1 − exp(−0.011·(Ms − Tq)) · retained austenite = the balance

For 4140 quenched to 20°C that leaves about 3.3% retained austenite — small, handled by a normal temper. But push the carbon to 0.95% and Ms drops to 125°C: now the same room-temperature quench leaves about 31% retained austenite, soft and unstable, liable to transform later and move the part. That is the high-carbon tool-steel problem in one number.

What cryo buys

Retained austenite is a function of how far below Ms you stop. Cooling 4140 from 20°C to −80°C drives its retained austenite from 3.3% down to 1.1%; on a high-carbon steel the swing is far larger. That is the entire case for a sub-zero or cryogenic soak — it is not magic, it is just continuing down the Koistinen-Marburger curve.

The temper

As-quenched martensite is hard and brittle. Tempering trades a little hardness for toughness and relieves the quench stress; time and temperature interchange through the Larson-Miller parameter, so a hotter, shorter temper can match a cooler, longer one. The case-hardening side — driving carbon in before the quench — is the carburizing case depth calculator, the austenitizing hold is the soak time calculator, and the whole sequence is laid out in the heat treatment planning suite.

Common mistakes

• Sizing the quench on hardness alone. Two steels at the same hardness can carry very different retained austenite; the one with the lower Ms is the one that moves in service.
• Skipping cryo on high-carbon tool steel. 25–35% retained austenite is common in untreated high-carbon grades — it transforms slowly at room temperature and shifts dimensions on precision parts.
• Treating Andrews as exact. It is a composition-only screen calibrated for low-alloy steel to ~0.6% C. Cooling rate, grain size and austenitizing temperature all move the real Ms; a dilatometer is the measured answer.
• Tempering before the part is fully cool. Tempering while austenite is still untransformed defeats the purpose — let the quench finish (or run the sub-zero step) first.

Frequently asked questions

What is the martensite start temperature?

The temperature, on cooling, where austenite first begins shearing into martensite. It is set almost entirely by composition — carbon most of all. The Andrews equation gives Ms(°C) = 539 − 423·C − 30.4·Mn − 17.7·Ni − 12.1·Cr − 7.5·Mo (weight percent), so AISI 4140 comes out near 330°C and a 1018 mild steel near 440°C.

Why does high-carbon steel keep more retained austenite?

Because carbon drives Ms down — about 42°C for every 0.1% C — and the lower Ms sits, the less of the transformation finishes by the time the part reaches room temperature. A 0.4% carbon 4140 retains about 3% austenite quenched to 20°C; a 0.95% carbon steel with Ms near 125°C retains roughly 30%.

What does a cryogenic or sub-zero treatment actually do?

It lowers the quench temperature, which drives more austenite to martensite. By Koistinen-Marburger the transformed fraction is 1 − exp(−0.011·(Ms − Tq)), so cooling 4140 from 20°C to −80°C cuts its retained austenite from about 3.3% to 1.1%. For high-carbon and high-alloy tool steels with much more retained austenite, that step matters far more.

Why temper after quenching?

As-quenched martensite is hard but brittle and full of residual stress. Tempering trades a little hardness for toughness, and a low-temperature temper also conditions retained austenite. Temper time and temperature interchange through the Larson-Miller parameter, so a shorter hold at higher temperature reaches the same condition as a longer hold cooler.