# Heat treating **Heat treating** is the controlled heating and cooling of a metal to change its [[Microstructure|microstructure]] without changing its shape or its composition. Nothing is added and nothing removed: the same piece of [[Steel|steel]], with the same [[Carbon|carbon]] content, can be made soft enough to machine with a hand tool or hard enough to cut other steels, and the only difference is the temperature history it was given. That is why the process exists — it lets a part be formed while soft and used while hard.[^manuf-ch6] In the microsim below the reader has one control, the cooling rate. The background is the isothermal transformation map of a eutectoid steel over log time and temperature, with the pearlite nose near 550 °C at about one second and the martensite start line `M_s` at 220 °C; the reader's cooling curve is drawn across it and either crosses the nose or misses it. The readout gives the phase mixture and a [[Hardness|Rockwell]] hardness, from roughly 15 HRC for a fully annealed structure to about 65 HRC when the curve escapes the nose, and the presets are the four shop procedures the Portal Book describes: anneal, normalize, quench and temper.[^manuf-ch6][^ttt-cn] On the [[Materials_science]] flagship this page serves Part IV, Fundamentals › Processing, in the section *Processing*, where the phase diagram of Part V stops being a map of equilibrium and becomes a set of instructions — every useful heat treatment being a deliberate failure to reach equilibrium. ## Physical processes Three things can happen to a solid when it is heated and cooled, and heat treating uses all of them. The first is a change of [[Phase_transition|phase]]. Iron is body-centred cubic below 912 °C and face-centred cubic above it, and the face-centred form, [[Austenite|austenite]], dissolves far more carbon — up to 2.14 % by weight at 1147 °C, against 0.022 % in ferrite at 727 °C. Heating a steel into the austenite field takes carbon into [[Solid_solution|solution]]; cooling it forces the carbon back out. What it comes out as is the whole subject. The second is [[Diffusion|diffusion]]. Separating one phase into two requires atoms to move, and how far they move in a given time follows [[Fick's_laws_of_diffusion|Fick's laws]] with a diffusivity that rises steeply with temperature in the [[Arrhenius_equation|Arrhenius]] manner. Cooling fast enough means giving the carbon too little time to go anywhere: a race between falling temperature and falling diffusion distance. The third is recovery of the defects deformation left behind. A cold-rolled or forged metal is full of [[Dislocation|dislocations]]; warming it lets them annihilate and rearrange, then lets new strain-free [[Crystallite|grains]] nucleate and consume the deformed ones — [[Recrystallization_(metallurgy)|recrystallization]] — and then lets those coarsen by [[Grain_growth|grain growth]]. This route needs no phase change, which is how a [[Copper|copper]] or [[Aluminium|aluminium]] sheet is softened between passes. The sim is about the first two. Its vertical axis is temperature, its horizontal axis is time on a logarithmic scale because diffusion distances change by orders of magnitude across a few hundred degrees, and its curves are the loci on which a given fraction of austenite has transformed. ## Effects of composition Which transformations are available depends on how much carbon the steel carries, and the dividing line is the eutectoid composition: 0.76 % carbon by weight, at which austenite transforms completely at a single temperature, 727 °C, with no two-phase interval on either side.[^atomsfirst-10-6] The sim uses a eutectoid steel because its map is the simplest — one transformation, one temperature — and because every other plain carbon steel is a variation on it. Alloying elements move the lines. Almost everything added to steel — manganese, chromium, molybdenum, nickel — lowers the eutectoid carbon content and shifts the transformation curves to longer times, which is what makes an alloy steel hardenable in a thick section where a plain carbon steel would transform before the interior had cooled. ### Eutectoid alloys At 0.76 % carbon the whole of the austenite transforms at once into a lamellar mixture of ferrite and [[Cementite|cementite]], Fe₃C, alternating on a spacing of a fraction of a micrometre. The mixture is called [[Pearlite|pearlite]], and the arithmetic of it is fixed: ferrite holds almost no carbon and cementite holds 6.67 %, so a 0.76 % alloy is about 89 % ferrite and 11 % cementite by weight. Because ferrite is soft and cementite is hard and brittle, the properties of pearlite are set by how finely the two are interleaved, which in turn is set by the temperature at which it formed — the higher the temperature, the more time for diffusion, and the coarser the plates. ### Hypoeutectoid alloys Below 0.76 % carbon the austenite begins to reject ferrite before it reaches the eutectoid temperature. Cooling slowly, ferrite nucleates on the [[Grain_boundary|grain boundaries]] and grows, enriching the remaining austenite in carbon until what is left reaches 0.76 % and transforms to pearlite at 727 °C. The result is a network of soft proeutectoid ferrite with islands of pearlite between, and the proportions follow the [[Lever_rule|lever rule]]: a 0.4 % steel ends as roughly half pearlite. Most structural steel is hypoeutectoid, which is why ordinary mild steel is soft, weldable and tough. ### Hypereutectoid alloys Above 0.76 % carbon it is cementite that comes out first, and because cementite nucleates on austenite grain boundaries it forms a continuous brittle shell around each grain. That network is the reason a high-carbon steel cooled too slowly is weak in tension despite its hardness, and the reason such steels are normalized rather than fully annealed, or spheroidized so that the cementite gathers into isolated round particles instead of a film. Tool steels live here, and so do the carbon contents used for files, dies and bearing races. ## Effects of time and temperature The equilibrium [[Phase_diagram|phase diagram]] says what a steel becomes if given forever, and nothing about the second that counts. The map that does is the [[Isothermal_transformation_diagram|isothermal transformation diagram]], which E. S. Davenport and Edgar Bain built in 1930 by quenching small eutectoid-steel specimens into baths at fixed temperatures and measuring how much austenite had transformed after each interval.[^bain1930] Each isotherm gives a sigmoid of the [[Avrami_equation|Avrami]] form `X = 1 − exp(−k·t^n)`, transformation proceeding by [[Nucleation|nucleation]] and growth; joining the 1 % and 99 % times produces the C-shaped pair that is the sim's background.[^avrami1939][^johnson-mehl1939] The C shape comes of two opposing tendencies: just below 727 °C the driving force is tiny and nucleation slow, so the reaction takes hours; far below, the driving force is large but diffusion sluggish, so it is slow again. Between them lies a nose of maximum speed, near 550 °C at about one second.[^ttt-cn] Above the nose the product is pearlite, coarse near 700 °C and fine near 600 °C; below it diffusion is too slow for lamellae and bainite forms. Below `M_s` = 220 °C the remaining austenite stops waiting for diffusion and shears into martensite, a supersaturated body-centred tetragonal structure holding all the carbon in solution; the change is athermal, so the fraction transformed depends on how far below `M_s` the steel goes, not how long it waits.[^koistinen1959] That is where the reader's control bites. A cooling curve falls across the map, and everything depends on whether it clears the nose. Furnace cooling crosses the upper diagram and gives coarse pearlite at about 15 HRC; still air gives fine pearlite; an oil quench may clip the nose. A curve reaching 220 °C untouched gives fully [[Martensite|martensitic]] steel at about 65 HRC. The critical cooling rate follows from the nose: 727 °C to 550 °C in under a second is roughly 180 K/s, which is why plain carbon steel must be quenched violently and thin, and why an alloy steel with the nose pushed to ten seconds hardens in oil. At the limit, a metal whose nose is missed at a few kelvin per second freezes as glass — [[Amorphous_metal|amorphous metal]]. The sim's curves are continuous while the map is isothermal, so the real boundaries sit lower and to the right; the readout treats that shift as a fixed offset, which is ILLUSTRATIVE rather than measured. ## Types of heat treatment The named treatments of the shop floor are just points on the map above, with the rate chosen to land in a particular product. The Portal Book's chapter organises them as hardening, tempering, annealing and normalizing, adding that the shop's real constraints are the quench tank, the ventilation and the time the furnace takes to come up.[^manuf-ch6] Every one of them begins by holding the part long enough at temperature for the whole section to reach it and for the carbon to dissolve, which is why soak times scale with thickness rather than with mass. ### Annealing [[Annealing_(materials_science)|Annealing]] means heating above the transformation temperature and cooling as slowly as possible, usually by shutting the furnace down with the work inside. The slow curve stays far to the right of the nose, so the product is coarse pearlite: the softest and most ductile condition, with the lowest strength and the largest grain size. It is done to make a part machinable, to erase the effects of previous work, and to relieve the whole of its stored energy at once.[^manuf-ch6] ### Normalizing Normalizing uses the same soak but cools in still air, a rate perhaps a hundred times faster. The curve crosses the diagram lower, so the pearlite is finer and the grains smaller, and the result is harder and stronger than annealed material while remaining tough. It is the standard condition for a structural steel and the usual remedy for the coarse, uneven structure left by [[Forging|forging]], [[Casting|casting]] or [[Welding|welding]].[^manuf-ch6] ### Stress relieving Stress relieving heats the part to a few hundred degrees below the transformation temperature and holds it, so that no phase change occurs at all. What it removes is elastic strain locked in by machining, welding or an earlier quench, which relaxes as dislocations recover. The dimensions and the hardness are left essentially unchanged, which is the point: a part that is already at final hardness can be stabilised before finish grinding. ### Aging Some alloys harden not by transformation but by precipitation. Solution-treating dissolves an alloying element, quenching traps it in a supersaturated [[Solid_solution|solid solution]], and holding at a moderate temperature lets it come out as fine particles that obstruct dislocations — [[Precipitation_hardening|precipitation hardening]]. Alfred Wilm found the effect in an [[Aluminium_alloy|aluminium]]–copper–magnesium alloy that hardened over days at room temperature, and it is now the basis of aerospace aluminium and of the [[Superalloy|superalloys]].[^wilm1911] The strength passes through a maximum and falls again as the particles coarsen, so an ageing schedule is a time as much as a temperature. ### Quenching [[Quenching|Quenching]] is cooling fast enough to miss the nose. The medium sets the rate: brine is fiercest, then water, then oil, then polymer solutions, then forced air, and the choice is a compromise, because the same rapid contraction that traps the carbon also generates thermal stresses that distort parts and crack them. Section thickness matters as much as medium, since the interior of a thick bar cannot cool faster than heat can leave it — the reason hardenability, measured by an end-quench test, is reported as a depth rather than a hardness.[^jominy1938] Continuing below room temperature into [[Cryogenics|cryogenic]] treatment converts austenite that was retained above `M_f`. ### Tempering As-quenched martensite is hard and almost useless: it is brittle and carries enormous internal stress. [[Tempering_(metallurgy)|Tempering]] reheats it below the transformation temperature so the trapped carbon precipitates as fine carbides, trading hardness for toughness along a curve that the tempering temperature selects.[^manuf-ch6] Almost every hardened steel part in service is quenched and tempered, and the pair is one operation in practice: the quench is what makes the hardness available and the temper is what makes it usable. The [[Ductility|ductility]] recovered is visible as area under the [[Stress–strain_curve|stress–strain curve]]. ## Specification of heat treatment A drawing cannot specify a microstructure, so it specifies a measurable outcome: a hardness, a depth at which it is reached, and sometimes the process itself. The Portal Book's companion unit covers the measurement — the Rockwell test, in which an indenter is pressed into the surface under a minor load and then a major one and the extra depth is read directly as a hardness number, the C scale using a diamond cone for hardened steel.[^manuf-hardness] Hardness is quick, cheap and nearly non-destructive, and correlates well enough with tensile strength to serve as the acceptance criterion for a heat treatment. ### Case hardening Case hardening produces a hard surface on a tough core, which is what a gear tooth or a camshaft lobe needs: wear resistance where the contact is, and the ability to absorb shock underneath. [[Carburizing|Carburizing]] does it by adding carbon — holding the part in a carbon-rich atmosphere in the austenite range so that carbon diffuses in, then quenching so that only the enriched skin becomes martensite. Since diffusion distance grows as the square root of time, case depth scales as `√(D·t)`: four times the furnace time buys twice the case. Nitriding achieves the same end below the transformation temperature and so distorts less, while induction and flame hardening heat only the surface and leave the composition alone. ### Through hardening Through hardening treats the whole section to the same structure, and whether it can be done is a question of hardenability rather than of hardness. A plain carbon steel bar more than a few millimetres thick cannot cool at its centre fast enough to miss the nose, so its core transforms to pearlite however violent the quench; an alloy steel with the nose pushed to longer times will harden right through. Specifying through hardening on a heavy section is therefore a specification of the alloy as much as of the furnace. ## Types of furnaces A furnace is judged on three things: whether it holds the set temperature uniformly across the work, how fast it can be loaded, and whether it can keep the atmosphere off the steel. The last is least obvious and most expensive to get wrong, because a hot steel surface in air loses carbon to it. Decarburization leaves a soft skin exactly where hardness was wanted, and the remedies are a protective or carbon-controlled atmosphere, a vacuum, a salt bath, or a machining allowance ground away afterwards. ### Batch furnaces A batch furnace holds a fixed charge for one cycle and is the general-purpose tool of a small shop. The variants are named for how the work gets in — box, bell, pit, car-bottom and elevator — and the choice is set by the size and the shape of the load rather than by anything thermal, since a long shaft is best treated hanging in a pit while a heavy die is best rolled in on a car. ### Salt bath furnaces A bath of molten salt heats by conduction rather than radiation, so a part reaches temperature several times faster and more evenly than in a gas-fired box, and the liquid excludes air completely, which prevents both scaling and decarburization. They are the traditional choice for tool steels and for selective heating, where only the end of a part is dipped, and they demand care: the salts are corrosive and hot, and water carried in on a wet part flashes to steam.[^manuf-ch6] ### Fluidised bed furnaces A fluidised bed suspends fine ceramic particles in an upward gas flow so the bed behaves as a liquid, giving heat-transfer rates approaching a salt bath's without the chemistry. The gas doubles as atmosphere control, so one vessel runs neutral or carburizing, and the work comes out clean. ## See also - [[Quenching]] — the variant sim, the cooling curve alone - [[Tempering_(metallurgy)]] - [[Isothermal_transformation_diagram]] — the variant sim, the bare map - [[Annealing_(materials_science)]] - [[Martensite]] - [[Amorphous_metal]] — the limit in which the nose is missed entirely - [[Eutectic_system]] - [[Pearlite]] ## References [^manuf-ch6]: Virasak, LamNgeun. *Manufacturing Processes 4-5* (2019), Ch. 6 Heat Treating, pp. 147–150 (hardening, tempering, annealing and normalizing as shop procedures, with quench media, ventilation and safety) (page to pin). https://open.umn.edu/opentextbooks/textbooks/manufacturing-processes-4-5 [^manuf-hardness]: Virasak (2019), Ch. 6 Unit 2 Hardness Testing, pp. 151–154 (the Rockwell test: minor and major loads, depth read directly as a hardness number; calibration) (page to pin). [^atomsfirst-10-6]: Flowers, P.; Neth, E.; Robinson, W. et al. *Chemistry: Atoms First*, 2nd ed. (2019), OpenStax, Ch. 10 Liquids and Solids, §10.6 Lattice Structures in Crystalline Solids (the body- and face-centred cubic cells behind ferrite and austenite, and the interstitial sites carbon occupies), pp. 475–544 (page to pin). https://open.umn.edu/opentextbooks/textbooks/chemistry-atoms-first [^bain1930]: Davenport, E. S.; Bain, E. C. (1930). "Transformation of Austenite at Constant Subcritical Temperatures." *Transactions of the American Institute of Mining and Metallurgical Engineers* 90: 117–144 (volume and pages to pin; no DOI asserted). [^avrami1939]: Avrami, M. (1939). "Kinetics of Phase Change. I. General Theory." *Journal of Chemical Physics* 7 (12): 1103–1112 (DOI to pin). [^johnson-mehl1939]: Johnson, W. A.; Mehl, R. F. (1939). "Reaction Kinetics in Processes of Nucleation and Growth." *Transactions of the American Institute of Mining and Metallurgical Engineers* 135: 416–458 (volume and pages to pin). [^koistinen1959]: Koistinen, D. P.; Marburger, R. E. (1959). "A general equation prescribing the extent of the austenite–martensite transformation in pure iron–carbon alloys and plain carbon steels." *Acta Metallurgica* 7 (1): 59–60 (DOI to pin). [^wilm1911]: Wilm, Alfred (1911). "Physikalisch-metallurgische Untersuchungen über magnesiumhaltige Aluminiumlegierungen." *Metallurgie* 8: 225–227 (volume and pages to pin; the discovery of age hardening, made in 1906). [^jominy1938]: Jominy, W. E.; Boegehold, A. L. (1938). "A Hardenability Test for Carburizing Steel." *Transactions of the American Society for Metals* 26 (volume, pages and year to pin). [^ttt-cn]: *Citation needed.* The transformation boundaries the sim draws for a eutectoid steel — the pearlite nose near 550 °C at about one second, `M_s` = 220 °C, and the annealed-to-quenched hardness span of roughly 15 to 65 HRC — are the values supplied by the M19 sim row. No Portal Book carries a transformation diagram; the ASM *Atlas of Time–Temperature Diagrams for Irons and Steels* (or an equivalent published isothermal diagram for AISI 1080) should be pinned here, and the 180 K/s critical cooling rate computed on this page from the nose position pinned with it. The tempering hardness curve is described qualitatively for the same reason. [^earle-ch3]: Earle, Steven. *Physical Geology* (2015), BCcampus Open Education, Ch. 3 Intrusive Igneous Rocks, pp. 67–92 (cooling rate against crystal size — the same competition between nucleation and growth, in rock) (page to pin). https://open.umn.edu/opentextbooks/textbooks/physical-geology ## Further reading - Virasak, LamNgeun. *Manufacturing Processes 4-5* (2019) — Ch. 6 and its hardness-testing unit, the shop-floor account this page's process names follow. - Flowers, P.; Neth, E.; Robinson, W. et al. *Chemistry: Atoms First*, 2nd ed. (2019), OpenStax — Ch. 10 §10.6, for the cubic cells behind ferrite, austenite and martensite. - Earle, Steven. *Physical Geology* (2015), BCcampus — Ch. 3, for the same cooling-rate argument written about magma.[^earle-ch3] <!-- MATTERSIM:BEGIN g24 — Matter & Energy Cluster microsim (framework build, specs/sims/Heat_treating.json); do not hand-edit inside --> **Microsim — three.js (Wikitube framework), pending deploy:** *Heat treating* will play here once `https://wikitube-3d-microsims.netlify.app/matter/Heat_treating.html` is live. <!-- pending: <div class="wt-sim" data-src="https://wikitube-3d-microsims.netlify.app/matter/Heat_treating.html" data-title="Heat treating"></div> --> <!-- MATTERSIM:END --> ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Heat_treating) : [Wikitube](https://en.wikitube.io/wiki/Heat_treating) · pinned revision [1368183355](https://en.wikipedia.org/w/index.php?oldid=1368183355) · 2026-09-11 ## Previous hub tags Hubs: `Life_Physics`. Portals: [[PORTAL_Materials_science]]. --- *Matter & Energy Cluster child articles, wave 1 · 2026-09-11 · drafted · Materials_science row M19 · sim pending (matter/Heat_treating).*