I am trying to accurately model conformational equilibria of acyclic molecules with multiple hydroxy groups, in methanol, but it is proving difficult. I'm doing conformer generation with MD, that is probably OK. (I suspect) the main problem is with obtaining relative energies of the conformers.

If I use implicit solvent, then any conformer with intramolecular hydrogen bonding is overstabilised (to the extent where the most stable conformer is one where the molecule has curled up into a ball to maximise intramolecular HB), but if I were to use explicit solvent I'm rather worried that the energies will be dictated by the geometries and hydrogen bonding of the methanol molecules rather than the solute.

Do any of you have experience with solvents like methanol or water?

I tried looking in the literature, but as far as I can tell people just use implicit solvent... but they work with cyclic stuff like sugars which don't really have the same conformational flexibility.