Chisels That Chip: Hardening 0.8% Carbon

Shank cracking (seven parts). Three factors combine. First, the steel was overheated: a bright yellow-orange is roughly 1,050–1,200 °C, far above the 780–800 °C this steel needs, and it coarsens the austenite grain and encourages decarburisation. Second, cold water is a far too severe quench for 0.78% C - water has a severity of cooling around 1.0 against roughly 0.3 for oil, and a high-carbon steel quenched in it transforms to untempered martensite with very high internal stress. Third, the quench is only 40 mm deep and vertical, so there is a sharp boundary between transformed martensite and untransformed hot steel; the two shrink differently and the triaxial stress across the shank cracks it. Cracks appearing within a day are consistent with delayed quench cracking, and because tempering is done with a torch after the part has cooled, there is a window where the part sits untempered.

Edge chipping (four parts). At 62–64 HRC the edge is harder than a cold chisel needs and is brittle; the correct range for this application is about 56–60 HRC. Torch tempering to a colour is not a controlled time-temperature treatment, so the tip may be effectively untempered in places and over-tempered in others. Overheating in the forge also burns and decarburises the edge, leaving a weak skin that spalls away under load. A brittle, coarse-grained, over-hard edge on mild steel will chip rather than deform.

Permanent bending. Quenching the whole chisel hardens the shank as well as the tip; 55 HRC at mid-shank is far too hard for a slender 8 mm hex shank under hammer shock. The shank should be tough, in the region of 40–45 HRC. A hard, coarse-grained shank has little capacity to absorb impact, so it either takes a permanent set or snaps - and with the grain coarsened by overheating, both are likely.

1. Normalise the forged blanks. Heat to 840–860 °C, soak 10–15 minutes for an 8 mm section, and air cool. Forging at a yellow heat has left a coarse, uneven grain; normalising refines it before hardening and is the step that most directly addresses both the chipping and the cracking.

2. Austenitise. 780–800 °C in a muffle furnace with a thermocouple, not an open forge. At 0.78% C the steel is eutectoid, so 780–800 °C puts it fully into austenite without grain growth. Soak about 10 minutes at temperature; longer adds nothing but decarburisation, so protect the surface or use a controlled atmosphere.

3. Quench. Quench only the cutting end, 30–40 mm, vertically, in warm oil at 50–80 °C (or a polymer quench), agitating up and down. Do not use cold water on 0.78% C; if oil is absolutely unavailable, use warm water at 40–50 °C with vigorous agitation, and expect more rejects. Keep the transition between hardened tip and soft shank as gradual as possible.

4. Temper immediately. Temper before the part cools below about 100 °C - ideally straight from the quench - to avoid delayed cracking. Oven temper at 200–220 °C for one hour, twice, cooling between. That gives a pale-straw to straw oxide and an edge hardness of about 58–60 HRC.

5. Differential temper on the shank. With the tip wrapped in a wet rag or stood in water, bring the shank to about 300–320 °C (blue) so the shank finishes at 40–45 HRC. The shank must be tough and the tip must be hard; one uniform temper cannot do both.

6. Finish. Grind the edge after tempering, with coolant, at 60–70° included angle, to avoid overheating the hardened tip.

Temper colours as a rough guide: pale straw about 220 °C, dark straw about 240 °C, brown about 255 °C, purple about 280 °C, blue about 300 °C.

Verification available in the toolroom:

• Process record first. A furnace with a thermocouple and a chart or log is the control; if there is no furnace, buy or borrow one before making another chisel. The treatment is only as good as the record of temperature and time.
• File test. A new file should skate across the cutting edge without biting (indicating roughly 55 HRC or harder) and should cut the shank readily (indicating 45 HRC or softer). This is a crude but genuinely useful two-point check and takes seconds per part.
• Hardness files or a portable hardness tester. A set of comparison files at 40, 45, 50, 55 and 60 HRC, or a rebound tester, lets you confirm the edge and shank targets rather than guessing.
• Fracture test on one sample per batch. Nick the edge, break it, and read the grain: fine, silky grey is correct; bright, coarse, crystalline means overheating. This is the cheapest check on austenitising temperature you have.
• Magnet and decalescence. Steel loses magnetism at about 770 °C, so a magnet is a rough indicator that the part is above the critical range; watching for the shadow that travels across the surface as it is heated is a better in-forge guide than colour alone.
• Spark and file check for decarburisation. A soft skin on a hard part shows as a file that bites at the surface but skates underneath.

Why temper colour is unreliable. The oxide colours are a time-temperature effect on a clean, polished surface, viewed in good light, and they depend on how long the surface has been at temperature, not just how hot it is. A thin section reaches straw colour at a lower temperature than a thick section, the colours continue to darken while the part sits hot, and a torch gives a steep gradient, so the colour you see is an average of a range. On a forged and ground chisel the surface is neither clean nor uniform, and the torch flame itself oxidises the surface. Colour can confirm that tempering happened; it cannot certify the temperature or the resulting hardness.