Steel and heat treatment
Make metal hard enough to cut other metal.
Everything for this step
- Field guide
- Smithing → 6 sections Properties of materials → 2 sections
- Builds
- Blacksmith's forge
- Reference sheets
- None.
- Track
- Metals, level 2 of 4 · next: Blast furnace and cheap iron
Why this matters
Wrought iron is tough but soft; cast iron is hard but brittle. Steel — iron with a controlled fraction of a percent of carbon, then heat treated — is both. Without it there are no cutting tools, no springs, no files, and therefore no machine tools.
The shape of it
- Understand the variable. Carbon content is nearly everything: around 0.1% gives soft wrought iron; 0.6–1.2% gives tool steel; above about 2% gives brittle cast iron. Your entire job is controlling that number without being able to measure it directly.
- Cementation: pack wrought iron bars in charcoal inside a sealed clay box and hold at red heat for days. Carbon diffuses inward. The result is blister steel — carburised on the outside, softer within.
- Forge-weld and fold the blister steel repeatedly to distribute the carbon evenly. This is what makes shear steel, and the folding is about homogeneity, not mysticism.
- Crucible steel is the better route once you have refractories: melt blister steel in a sealed clay crucible so the carbon distributes completely in the liquid. It gives genuinely uniform steel and is the basis of all fine tool making before Bessemer.
- Harden by quenching: heat to a bright even red — past the point where it stops being magnetic, which is a free and accurate indicator — then quench in water, brine or oil. Water is fast and risks cracking; oil is gentler.
- Then temper, always. Fully hardened steel is glass-brittle and will shatter. Reheat gently and read the oxide colours: pale straw (~230 °C) for a razor or file, brown for a chisel, blue (~300 °C) for a spring. Tempering trades a little hardness for the toughness that stops it snapping.
- Case-harden where you only need a hard skin: pack a low-carbon part in charcoal and bone, heat, and quench. Cheap, and correct for gears and wear surfaces.
- Test destructively and write it down. Break samples, look at the grain, and record the process settings against the result. This is a proper experimental programme, and it is how you actually converge.
Doing it, step by step
The summary above says what has to happen. This says how, with the numbers that decide whether it works.
1 · Judge carbon without instruments — the spark test
Carbon content is nearly everything: about 0.1% is soft wrought iron, 0.6–1.2% is tool steel, above about 2% is brittle cast iron. You cannot measure it directly, but you can read it from sparks.
Hold the metal against a grindstone — ideally a fast one — in dim light, beside reference samples you have kept:
- Wrought iron: long, straight, straw-coloured streaks, few or no bursts.
- Medium carbon: streaks with forked bursts.
- High carbon tool steel: dense, bushy, bright bursting stars, many per streak.
Keep labelled reference bars for every material you make.
2 · Make blister steel by cementation
Pack bars of wrought iron in crushed charcoal inside a sealed clay box, the bars not touching each other or the walls, the lid luted with clay. Hold at bright red to orange heat — around 1,000 °C — for days: a week or more for thick bars. Carbon diffuses in from the surface.
Cool slowly and break a test bar. The surface is covered in blisters — hence blister steel — and the fracture shows a fine-grained, high-carbon rim round a coarser, softer core. Longer soaking pushes the carbon deeper.
3 · Fold it for uniformity, or melt it in a crucible
Shear steel: cut blister steel into lengths, stack, forge-weld them together, draw out, fold and weld again — see blacksmithing. Each fold halves the scale of the variation. Folding is about homogeneity, not mysticism.
Crucible steel: melt blister steel in a sealed clay crucible at around 1,500 °C so the carbon distributes completely in the liquid. It needs good firebrick, a strong draught and a great deal of fuel, and it gives genuinely uniform steel — the basis of all fine toolmaking before the Bessemer converter.
4 · Harden: heat just past non-magnetic, then quench
Heat evenly to a bright red, just past the point where a magnet no longer sticks — around 770–800 °C for tool steel. The magnet test is free and accurate: touch a magnet to the steel as it heats, and heat a little beyond the moment it lets go.
Quench immediately, plunging straight down and moving it up and down, never side to side:
- Water or brine — fastest, hardest, and most likely to crack thin or complex parts.
- Oil warmed to about 50 °C — gentler, less cracking, slightly less hard.
A file skates off fully hardened steel without biting.
5 · Temper — always, and immediately
Fully hardened steel is glass-brittle and will shatter, possibly into someone. Temper within an hour of quenching. Polish a strip of the part bright, then reheat gently and watch the oxide colours move across the bright surface:
- Pale straw, ~220 °C — razors, scrapers, files.
- Straw, ~230–240 °C — drills, punches, knives.
- Brown, ~250–260 °C — chisels, axes.
- Purple, ~270–290 °C — hammer faces.
- Blue, ~300 °C — springs, saws.
Quench the moment the colour arrives at the working edge. Tempering trades a little hardness for the toughness that stops it snapping.
6 · Case-harden parts that only need a hard skin
For gears, pins, bearing surfaces and wear faces: pack a low-carbon iron part in charcoal and charred bone or leather inside a sealed box, hold at bright red — around 900 °C — for several hours, then quench straight from the box. The skin takes up carbon to a depth of roughly half to one millimetre and hardens; the core stays tough.
It is cheap, uses wrought iron you already have, and is the right treatment for anything that must resist wear but also take shock.
7 · Read the fracture grain
Break a test piece after every change and look at the grain on the broken face, in good light:
- Fine, grey, silky — correct hardening temperature.
- Coarse, glittering, crystalline — overheated. The grain has grown and the steel is weak; reheat to just past non-magnetic and quench again, or it stays brittle.
- Fibrous and torn, bending before breaking — soft or unhardened.
Keep the broken pieces labelled with the process that made them.
8 · Run it as an experiment, and write every setting down
The process conditions that produce good steel from your iron and your charcoal are not in any book — you have to find them. Record every batch: source iron, carbon method, soak time and temperature, fold count, hardening temperature and quench medium, temper colour, and the test result.
Change one variable at a time, and keep the conditions that work. The test that you have arrived: a file that cuts mild steel, and a spring that returns to shape after repeated flexing.
Build it
Dimensioned drawings, bills of materials and construction procedures for the things this step asks for.