How to temper chocolate

Melt fully, cool while stirring in tempered seed, then hold at working temperature. A thermometer is not optional.

intermediate

If you read nothing else

Use the seeding method, use couverture, and use a thermometer. Two of the three most common failures are materials problems rather than technique problems.

The method, in four steps

Melt two thirds of the chocolate to 45–50°C for dark, 40–45°C for milk and white — all crystal structure must be destroyed. Remove from the heat and stir in the remaining third, finely chopped and already in temper. Keep stirring until it cools to 31–32°C for dark, 30–31°C for milk, 29–30°C for white. Test by spreading a little on a cold knife: in temper it sets hard and glossy within about three minutes.

Why the temperatures differ by type

Milk fat interferes with cocoa butter crystallisation, so milk and white chocolate temper cooler than dark. A single quoted temperature is wrong for two of the three types, which is why generic instructions so often fail.

Use couverture

Eating bars carry less cocoa butter, so they set softer, coat thicker and release less cleanly. If your tempering 'works' but the result is dull and fingerprints easily, the material is the problem. Callets are the most practical form: uniform pieces melt evenly and are already in temper, making them ideal seed.

Room temperature matters

The set test is reliable below about 20°C ambient and unreliable above about 24°C. In a warm kitchen a correctly tempered batch will seem to fail. Cool the room, or test against a genuinely cold surface.

Scale is a real constraint

Below about 300g a batch loses heat faster than the method assumes, and hand tempering becomes considerably harder regardless of skill. That is physics rather than technique — work larger and set the surplus aside as bark.

Covered in this guide

Sources

  • ChocolateHQ editorial synthesisChocolateHQ(citation identity confirmed; passage not re-read)
  • The Science of ChocolateRoyal Society of Chemistry(citation identity confirmed; passage not re-read)