Iron–Carbon Phase Diagram

What the diagram tells you

Pure iron is allotropic — it changes its crystal structure with temperature. Below 912 °C it is body-centred cubic (BCC), called ferrite or α-iron. Between 912 °C and 1394 °C it is face-centred cubic (FCC), called austenite or γ-iron. The FCC lattice has larger interstitial gaps, so it dissolves far more carbon than the BCC lattice can. The entire iron-carbon diagram is a consequence of that single fact: heating opens up room for carbon, cooling squeezes it back out, and the carbon has to go somewhere — into a separate carbide phase.

The chart we use in practice is the metastable Fe–Fe₃C diagram, plotting temperature on the vertical axis against weight-percent carbon on the horizontal, from 0% (pure iron) to 6.67% (the composition of cementite, the iron carbide Fe₃C). Three invariant reactions anchor it:

Peritectic  (1495 °C, 0.17% C):  δ-ferrite + Liquid → austenite
Eutectic    (1147 °C, 4.30% C):  Liquid → austenite + cementite   (ledeburite)
Eutectoid   ( 727 °C, 0.76% C):  austenite → ferrite + cementite   (pearlite)

For everyday steel work, the eutectoid reaction is the one that matters. It is the lowest-temperature, all-solid transformation, and it is the reaction every heat treatment is built around.

The four players

Four constituents do almost all the work on the steel side of the diagram. Three are true phases; pearlite is a two-phase microconstituent.

Constituent	Crystal structure	Carbon range	Hardness (approx.)	Character
Ferrite (α)	Body-centred cubic	0 – 0.022% C	~80 HB	Soft, ductile, magnetic
Austenite (γ)	Face-centred cubic	0 – 2.14% C	~150 HB (at temp)	Soft at temperature, non-magnetic, only stable hot (or stabilized)
Cementite (Fe₃C)	Orthorhombic carbide	6.67% C (fixed)	~800–1000 HV	Hard, brittle, reinforcing phase
Pearlite	Lamellar α + Fe₃C	0.76% C (eutectoid)	~200–300 HB	Tough layered composite of the above two
Martensite*	Body-centred tetragonal	inherits parent C	up to ~700 HV (65 HRC)	Non-equilibrium; from quenching, not on the diagram

*Martensite is shown for context. It does not appear on the equilibrium iron-carbon diagram because it forms by a diffusionless shear when austenite is quenched too fast for carbon to diffuse out — a TTT/CCT-curve phenomenon, not an equilibrium one.

The eutectoid reaction: austenite → pearlite

Take a steel of exactly the eutectoid composition, 0.76% carbon, and hold it at 800 °C. It is single-phase austenite: every carbon atom sits comfortably in the interstices of the FCC lattice. Now cool it slowly past 727 °C. The iron wants to become BCC ferrite, but ferrite can hold only 0.022% carbon — about a thirtieth of what is dissolved. The excess carbon has nowhere to stay, so it diffuses sideways and precipitates as cementite. The transformation runs as a coupled front:

γ (0.76% C, FCC)  →  α (0.022% C, BCC)  +  Fe₃C (6.67% C)
        austenite             ferrite              cementite

Layer thickness ratio (lever rule across the eutectoid line):
  cementite fraction = (0.76 − 0.022) / (6.67 − 0.022) ≈ 0.111  (≈ 11% Fe₃C)
  ferrite   fraction = (6.67 − 0.76 ) / (6.67 − 0.022) ≈ 0.889  (≈ 89% α)

Carbon only has to travel a fraction of a micron sideways, so the phases grow as alternating plates — the lamellar microstructure called pearlite, named for the mother-of-pearl sheen the etched layers give under a microscope. The interlamellar spacing is set by how fast you cool through 727 °C: cool slowly (a furnace cool / full anneal) and you get coarse pearlite with spacings around 0.4–1 µm, softer and easier to machine; cool faster (air cool / normalize) and you get fine pearlite, spacings nearer 0.1 µm, stronger and harder because the dense cementite plates block dislocations more effectively.

Hypoeutectoid and hypereutectoid steels

Most steels are not at the eutectoid composition. Their cooling story has an extra first act.

• Hypoeutectoid (C < 0.76%): a typical structural steel like AISI 1018 (0.18% C). On cooling, when the alloy crosses the upper boundary (the A₃ line) the first thing to appear is proeutectoid ferrite, nucleating at austenite grain boundaries. As temperature drops, more ferrite forms and the remaining austenite enriches in carbon, sliding along the A₃ line until it reaches 0.76% C at 727 °C — at which point it all converts to pearlite. The room-temperature structure is soft proeutectoid ferrite grains outlined by islands of pearlite.
• Hypereutectoid (C > 0.76%): a high-carbon tool or spring steel like AISI 1095 (0.95% C). Crossing the A_cm line, the first phase out is proeutectoid cementite, which forms as a brittle network along the austenite grain boundaries. The leftover austenite loses carbon, slides down the A_cm line to 0.76% C at 727 °C, and the rest becomes pearlite. That grain-boundary cementite network is exactly why hypereutectoid steels are spheroidized (heat-soaked to ball up the carbides) before machining — otherwise the brittle shell cracks.

Worked example: lever rule on a 0.40% carbon steel

How much of the room-temperature structure of a 0.40% C steel (AISI 1040, a common axle and gear steel) is pearlite, and how much is free ferrite? Apply the lever rule on the tie-line just below the eutectoid temperature, between ferrite (0.022% C) and pearlite (treated as 0.76% C):

Overall carbon:  C₀ = 0.40%
Ferrite limit:   Cα = 0.022%
Eutectoid line:  Cp = 0.76%   (composition of the just-formed pearlite)

Fraction pearlite        = (C₀ − Cα) / (Cp − Cα)
                         = (0.40 − 0.022) / (0.76 − 0.022)
                         = 0.378 / 0.738
                         ≈ 0.51   →  51% pearlite

Fraction proeutectoid α  = (Cp − C₀) / (Cp − Cα)
                         = (0.76 − 0.40) / (0.738)
                         ≈ 0.49   →  49% ferrite

So a 1040 steel cooled slowly is about half soft ferrite, half pearlite — a balance that gives it ~620 MPa tensile strength and decent toughness, which is why it is a workhorse for shafts and gears. Push the carbon to 0.76% and the structure becomes 100% pearlite, raising strength toward 900 MPa but cutting ductility.

Reading heat treatments off the diagram

The horizontal lines you austenitize above and the rate you cool through them define the classic steel heat treatments. All of them start by heating above the upper transformation line into the austenite field, then differ in cooling:

• Full anneal: austenitize, then furnace-cool very slowly. Equilibrium wins — coarse pearlite plus soft ferrite. Maximum machinability and ductility, lowest strength.
• Normalize: austenitize, then still-air cool. Faster than annealing, so finer pearlite and finer grains — higher strength and a more uniform structure, the standard delivery condition for many forgings.
• Harden (quench): austenitize, then quench in oil or water. Cooling is so fast that carbon cannot diffuse to form pearlite at all; the austenite shears into hard, brittle martensite instead. This is an escape from the equilibrium diagram, mapped instead by the steel's CCT curve.
• Temper: reheat the quenched martensite to 150–650 °C to precipitate fine carbides, trading some hardness for toughness — the difference between a glass-brittle quenched file and a usable spring.

Steel versus cast iron

The line at 2.14% carbon is the technical boundary between steel and cast iron, and it is more than a label. Below 2.14% the alloy can be heated entirely into the single-phase austenite field, where it is soft and uniform — that is what makes it forgeable, rollable and weldable. Above 2.14%, the alloy passes through the eutectic at 1147 °C, solidifying as the austenite-plus-cementite mixture ledeburite, and can never be made fully austenitic below the solidus. Cast irons therefore cannot be wrought; they are shaped by pouring. Whether their carbon ends up as brittle cementite (white iron) or soft graphite flakes/nodules (grey or ductile iron) depends on cooling rate and silicon content, which shifts the system toward the stable Fe–graphite diagram rather than the metastable Fe–Fe₃C one.

	Low-carbon steel	Eutectoid steel	High-carbon / tool steel	Cast iron
Carbon	0.05 – 0.25%	~0.76%	0.6 – 1.4%	2.14 – 4.5%
Slow-cool structure	Ferrite + minor pearlite	100% pearlite	Pearlite + grain-boundary cementite	Pearlite/graphite + ledeburite
Example grade	AISI 1018	AISI 1080	AISI 1095 / W1	Grey iron GG25
Typical use	Car body, structural sections	Rail, wire rope	Springs, knives, files	Engine blocks, manhole covers
Forgeable?	Yes	Yes	Yes (with care)	No — cast only
Hardenable to martensite?	Poorly (too little C)	Yes	Yes (best)	Surface only / specialized

Failure modes and trade-offs

• Grain-boundary cementite embrittlement. In hypereutectoid steel cooled slowly, proeutectoid cementite forms a continuous brittle shell around former austenite grains, providing an easy crack path. Spheroidizing or controlled normalizing breaks the network.
• Quench cracking. Cool faster than the steel's critical rate and you form martensite, but martensite is ~4% larger in volume than austenite; an abrupt, uneven quench locks in tensile residual stress that can split the part. Higher-carbon steels are most prone — which is why 1095 is oil-quenched, not water-quenched, where geometry allows.
• Decarburization. Heating in air strips carbon from the surface, leaving a soft ferritic skin that under-hardens — fatal for a spring or a bearing race. Protective atmospheres or carbon-potential furnaces prevent it.
• Banding / segregation. Real ingots are not uniform; carbon and alloy segregation during casting produces alternating ferrite-rich and pearlite-rich bands that show up as directional properties. The diagram describes equilibrium, not the chemistry gradients left by solidification.
• The strength–ductility trade. Every step up in carbon adds cementite, raising strength and hardness but cutting elongation and impact toughness. The diagram does not pick the right point for you — the application does. A crash structure wants low carbon and ductility; a file wants high carbon and hardness.

Where the diagram stops being enough

The Fe–Fe₃C diagram is an equilibrium map: it assumes infinitely slow cooling so diffusion always keeps up. Real processing is faster, so engineers pair it with two non-equilibrium tools. TTT (time-temperature-transformation) diagrams show what forms when you hold austenite at a fixed temperature — bainite appears below the pearlite nose, martensite below the M_s line. CCT (continuous-cooling-transformation) diagrams overlay actual cooling curves to predict the final mix from a real quench. Alloying elements shift everything: adding manganese, chromium or nickel moves the eutectoid composition and temperature, drops it (e.g. the eutectoid carbon falls well below 0.76% with chromium additions), and slows transformation enough that even thick sections harden. The iron-carbon diagram remains the indispensable starting point — it just is not the whole story once the clock and the alloy box come into play.