Tempering and Age Hardening

Two Ways to Tame a Trapped State

Everything in this rung has pointed the same way: the useful structure is the one you trap by cooling fast, not the one the equilibrium diagram promises. In the last guide you met the extreme case — martensite, a diffusionless, carbon-stuffed crystal so hard it can reach about 65 on the Rockwell C scale and so brittle a dropped bearing can crack. A trapped state that hard is almost useless as-quenched. This guide is about the second step: a gentle reheat that lets a controlled little bit of the frozen reaction finally happen, tuning the metal into something you would actually build with. Think of it as shocking steel glass-hard, then relaxing it just enough to be tough.

There are two great members of this family, and they run in opposite directions. Tempering works on quench-hardened steel: you always give hardness away to buy back toughness. Age hardening (also called precipitation hardening) works on alloys like aluminum: you build hardness up from a soft start. They look like different tricks, but underneath they are the same idea — trap a non-equilibrium solid solution, then reheat just enough to let it unmix a controlled amount.

Tempering: Trading Hardness for Toughness

Take that as-quenched martensite. Reheat it below the eutectoid line — anywhere from about 150 to 650 degrees C — hold for an hour or so, and air-cool. Martensite's body-centered-tetragonal cell is a cage with carbon wedged in, straining it out of shape; that trapped carbon is exactly what makes it so hard and so brittle. Warm it and the carbon finally gets the mobility to leave, precipitating as a fine dust of cementite (Fe3C) particles scattered through a soft ferrite matrix. This new structure is tempered martensite, and the process is tempering.

Now the payoff, and the price. The more you heat and the longer you hold, the more carbon escapes and the coarser those carbide particles grow — so the steel gets steadily softer but tougher. A low temper near 200 degrees C keeps most of the hardness (around 60 HRC), perfect for a knife edge or a bearing that must resist wear. A mid temper near 400 degrees C lands around 45 HRC, springy and resilient. A high temper near 600 degrees C settles around 30 to 35 HRC, giving up hardness for the toughness and ductility a structural part needs. You dial hardness against toughness simply by choosing the temperature.

1. Austenitize: heat the steel above its eutectoid line (roughly 760 to 850 degrees C for plain-carbon steel) so it becomes uniform austenite, the parent phase you met in the iron-carbon guide.
2. Quench: cool fast enough to dodge the pearlite and bainite noses on the TTT/CCT diagram, freezing the austenite into hard, brittle martensite. This is the diffusionless step from the last guide.
3. Temper: reheat to your chosen temperature (150 to 650 degrees C), hold, and air-cool. Higher temperature equals softer and tougher. Pick the number, and you have picked the final property.

Be honest about two things. First, tempered martensite is not a consolation prize — that fine, uniform sprinkle of carbides in ferrite gives the best strength-plus-toughness combination in all of steel, better than coarse pearlite of the same hardness, because the obstacles are smaller and more evenly spread. Second, tempering has forbidden zones: certain windows (a brittleness near 260 to 320 degrees C, and slow-cooling many alloy steels through about 375 to 575 degrees C) actually make the steel brittle again, so metallurgists deliberately temper around them rather than through them.

Age Hardening: Strength From Nothing You Can See

Now switch metals. Pure aluminum is a soft thing — a yield strength of maybe 10 to 30 MPa, easily bent by hand. Yet aircraft skins and rivets are aluminum alloys that yield above 400 or even 500 MPa. That leap is age hardening, and it is why the aluminum alloys 2024 and 7075 exist. The magic is a scatter of precipitates too fine to see in an ordinary microscope, each one a tiny snag for gliding dislocations — the same precipitation strengthening principle you can now recognize, delivered on purpose.

The classic is aluminum with about 4% copper (old duralumin). Heat it near 500 to 540 degrees C and all the copper dissolves into a single aluminum solid solution — that is the solution heat treatment. Then quench it in water: too fast for the copper to precipitate, so it stays trapped, wildly supersaturated at room temperature. Finally age it — hold at room temperature for days, or at about 150 to 190 degrees C for hours. The copper atoms, restless with a little mobility, gather into Guinier-Preston zones: coin-shaped, copper-rich discs only a few atoms thick, coherent with the lattice so they strain it locally. That strain field, and the fine precipitates that follow it, are what pin the dislocations and make the metal hard.

1. Solution treat: heat to a single-phase field (near 500 to 540 degrees C for Al-Cu) so every solute atom dissolves into one uniform solid solution.
2. Quench: cool fast to trap that solution supersaturated at room temperature, giving the solute no chance to precipitate on the way down.
3. Age: hold at a low temperature (room temperature or roughly 150 to 190 degrees C) so the trapped solute precipitates as fine zones and particles — and stop at peak hardness.

The Peak and the Overage

Here is the twist that catches beginners: with age hardening, more time is not more strength. Hardness climbs to a peak, then falls. Early on, the precipitates are fine, coherent, and packed close together, so a dislocation cannot slip far without meeting one — that is peak-aged, the strongest state. Keep going and the particles coarsen and their spacing widens (the big ones eat the small ones). Now a dislocation can simply bow out and loop around the widely spaced obstacles instead of fighting through them, and the alloy softens. That downhill slope is called overaging.

  Age-hardening curve:  hardness vs. aging time (log scale)

  hardness
    ^                _ _peak_ _
    |             _/           \_ _
    |           _/                 \_ _ _   OVERAGED
    |         _/                          \_ _ _
    |       _/  under-aged                      \_ _ _
    | _ _ _/   (GP zones forming)
    |  as-quenched (soft, all solute trapped)
    +--------------------------------------------------> time (log)
       SSSS  ->  coherent GP zones  ->  peak  ->  coarse CuAl2

  Hotter aging: the peak arrives SOONER but sits LOWER.
  (SSSS = supersaturated solid solution)

Compare this with tempering and the mirror image is clear. Tempering only ever slides downhill in hardness — more heat or time means softer, always. Age hardening climbs to a peak first, then slides down. Both are the same controlled decomposition of a trapped, supersaturated solid, just turning the hardness knob in opposite directions. And in both, temperature trades against time: age hotter and you reach peak sooner but a lower peak; the metallurgist picks the point on the curve that the part actually needs.

One Idea, Honest Limits

Step back and both treatments prove this rung's thesis. Tempered martensite and Guinier-Preston zones appear on no equilibrium phase diagram — they are metastable, engineered states, reached only by trapping something out of equilibrium and then letting it relax a measured amount with time and temperature. This is the metallurgist's real lever of control, and it is why the same two elements can be a soft wire or a hardened tool depending entirely on their thermal history.

Now the honest caveats. Heat treatment moves strength, not stiffness: pure aluminum has a Young's modulus near 70 GPa, and so does fully peak-aged 7075 — the strength jumps twentyfold while the modulus barely twitches, because stiffness is set by atomic bonding, not by a few percent of precipitate. Steel sits near 200 GPa at every temper. And you never get something for nothing: the strength-ductility tradeoff means a peak-aged or lightly-tempered part is strong but less forgiving, so you pick where on that seesaw the job belongs.

Two last honest limits. Age hardening is not universal: it only works when solid solubility shrinks sharply as the alloy cools, so that quenching genuinely supersaturates the solution — an alloy with a flat solubility line has nothing to precipitate and cannot age-harden. And hardness is not hardenability: hardness is a single number you press into one spot, while hardenability (the Jominy idea from the TTT/CCT guide) is how deep into a thick bar martensite forms at all. A steel can be very hard at the surface yet have poor hardenability in the core — different questions, both worth asking. Carry these two habits — trap it, then relax it just enough — into every alloy family ahead.