How do you prove crystals are pure after re-crystallisation?
If perhaps your compound is unknown, and thus are unable to obtain a reference value, you could obtain melting point of the sample, and then re-crystallize a few more times, obtaining a new melting point each time, until it is unchanged by re-crystallizing. This will of course decrease your yield, but if there is little fluctuation in your series of melting points, you can be sure you have a relatively pure sample.
