Recrystallisation: The Smallest Possible Volume of Hot Solvent

A recrystallisation solvent has to do four things, and the first two are the whole test: **Dissolve the product well when hot and badly when cold.** A hot-to-cold solubility ratio above about five makes a good solvent; below about two it is useless. This is the only property that matters much. **Not react with the product.** Obvious, and worth checking for anything reactive. **Boil below the product's melting point.** If the product melts before it dissolves, it forms an oil and oils do not crystallise cleanly — they trap impurity. "Oiling out" is the classic failure. **Be removable.** Volatile enough to dry off the crystals. **Finding one by test.** Put a few milligrams in a test tube with a few drops of candidate solvent, cold. If it dissolves cold, that solvent is no good. If it does not, heat it: if it dissolves hot and crystallises on cooling, that is the solvent. Five minutes each, and it is the only reliable method. **Mixed solvents** work when no single one does: dissolve in the good solvent hot, then add the poor one dropwise until it just turns cloudy, then a drop of the good one to clear it, and cool. Ethanol and water is the classic pair.

**Dissolve hot, in the minimum.** Add solvent in small portions to the solid in a flask, heating and swirling between additions, and stop the moment everything has dissolved. Every extra millilitre is product that will stay in solution. **Filter hot, only if there is insoluble matter.** Dust, fibres, undissolved impurity. Use a fluted filter and a pre-warmed funnel, and work fast — a cold funnel crystallises the product in the paper, which is the most annoying possible loss. If nothing is floating, skip this step. **Cool slowly.** This is the step that decides crystal quality. Slow cooling grows large, well-formed crystals that trap little; fast cooling gives a fine powder or an amorphous mass full of solvent and impurity. Leave the flask on the bench, covered, and only then move it to ice. **Scratch or seed if nothing happens.** A glass rod on the inner wall, or one seed crystal, starts a solution that is supersaturated but not nucleating. **Filter cold**, wash the crystals with a little ICE-COLD solvent — cold, because warm wash solvent dissolves the crop — and dry. **Keep the mother liquor** until the yield is confirmed. It is the only place the missing material can be.

**Nothing crystallised.** Too much solvent — boil some off and cool again. Or the solution is supersaturated and needs scratching or seeding. Or the solvent was the wrong one and the product is simply soluble cold. **An oil separated instead of crystals.** The product melted before it dissolved, or came out above its melting point. Reheat to dissolve the oil, add a little more solvent, and cool more slowly. A lower-boiling solvent is the real fix. **A fine powder rather than crystals.** Cooled too fast. Redissolve and cool on the bench rather than in ice. **Good crystals but poor recovery.** Too much solvent again. The notebook shows exactly what that costs. **Crystals still coloured or still impure.** Either the impurity is above the level that can stay dissolved, or it co-crystallises. A second recrystallisation helps the first case and not the second. For coloured impurities, a small amount of activated charcoal in the hot solution adsorbs them — a pinch, not a spoonful, because charcoal also adsorbs product. **Everything crystallised, including the impurity.** The two are too alike. Recrystallisation is the wrong method; try chromatography or a different solvent.