CTAB has two ends, a lipophilic (cetyl chain) and a hydrophilic-hygroscopic (trimethylammonium + Br-). When it is crystallized from a protic solvent, a hydrophilic part in the crystal is oriented to surface of the crystal. Then comes the interaction of hydrophilic part with air moisture and the crystal is dissolved. In the crystallization carried out from an apolar solvent the hydrophobic cetyl chains are at the crystal surface and moist air dissolution does not occur.
On the other hand, CTAB is very poorly soluble in nonpolar solvents.
This discrepancy can be solved, so that CTAB is dissolved in a minimum amount of a polar solvent and the solution is then poured into a large excess of non-polar solvent with vigorous stirring (or better under ultrasound action). It is obvious that both solvents must be miscible. For example, the couple EtOH-Et2O, iPrOH-hexane, EtOAc-MeOH, AcOH-toluene, ...
Nota bene: This procedure is common for successful crystallization of most organic onium salts (ammonium, phosphonium, sulfonium, quaternized N-heterocycles etc.).
1. It is very difficult to filter CTAB precipitate due to structure of this (from EtOH:H2O)
2. much better to collect precipitate by centrifugation (we used 1 L bottles)
3. then we tried to wash wet CTAB precipitate with 10 V Et2O it melted with residual H2O and than mixed with Et2O (residual color present in final mixture)
4. we've mixed it again with 5V H2O and have putted to freezer for precipitation :)