[Myrddin]

A molecular sieve is what most places use now. Cyclohexane is to nasty to use. Very flammable. I used to work at a plant that made ethanol, and we had a few contractors that would sneak the stuff straight out of the sieves. Gave them one heck of a hangover.

My lab skills are a bit dated. A molecular sieve wasn't common in labs in the mid-70's. The closest approximation was Sephadex gels used to separate salt solutions from larger proteins. The Sephadex gels cause a larger included volume to be experienced by smaller salt particles. Proteins zip on past the outside of the gels due to larger physical size. I usually finished the desalting process with reverse osmosis via semi-permeable membranes.

[redgolum]

This is on a large scale with zeolyte (sp?) beads with porous structures. Superheated vapor is passed through the bed of beads, and the water gets "stuck" in the pores. The water is then forced out by recircing some of the finished product back through the beds. It works ok, but makes for a control headache.