Rules governing the crystal packing of mono- and dialcohols

A new program, Mercury, has been used to survey 144 monoalcohol (C(m)H(n)OH) and 101 dialcohol [C(m)H(n)(OH)(2)] crystal structures. Results show that their hydrogen-bonding patterns are strongly correlated with steric effects. Primary monoalcohols have a strong preference to form infinite ...OH...OH... chains. Secondary monoalcohols form chains and rings of hydrogen bonds with about equal facility. Tertiary monoalcohols very often form isolated OH...O hydrogen bonds or structures containing no OH...O hydrogen bonds at all. In the latter case, however, the structures almost invariably contain CH...O and/or OH...pi interactions. Substitution on the beta-carbon(s) of monoalcohols has a profound effect on packing patterns, with increased substitution disfavouring chains and rings. Dialcohols show a much stronger preference for chains of hydrogen bonds, compared with monoalcohols. This is particularly so when at least one of the hydroxyl groups is primary, in which case chains are overwhelmingly preferred. Once again, substitution on the beta-C atoms is influential, heavy substitution tending to lead to packing arrangements that involve isolated or intramolecular OH...O hydrogen bonds. Dialcohols almost never crystallize without at least some OH...O hydrogen-bond formation. In both monoalcohols and dialcohols, chains show a stronger preference to be helical (usually threefold helices) as steric hindrance increases. Hydrogen-bonded rings usually contain four OH...O hydrogen bonds. It is possible that empirical observations such as these may aid crystal-structure prediction.