Molecular Packing in Liquids, Solutions, and Crystals: Acetic Acid as a Test Case

Abstract

Several papers in recent times [1-4] have discussed the general difficulties encountered in the completely ab initio prediction of full crystal structures by molecular mechanics calculations. There is almost general consensus that this is possible only for rigid molecules, and only for crystals with one single molecular entity in the asymmetric unit. Even for this severely restricted subset of organic molecular crystalline systems, success is measured by different degrees of ambition and is sometimes more serendipitous than robust and reproducible. Current methods cannot in fact go beyond the generation of a number of plausible crystal structures, always very similar in packing energy and density, among which usually, but not always, the observed crystal structure is found. The essential difficulty is just that close-packing is necessary but not sufficient, while the discriminating power of current force fields is still low. Prediction of the structure when cell parameters and space group are known — i.e., solution of the phase problem in X-ray crystallography — is more within reach of computer simulations, but, alas, the task can be much more comfortably and quickly carried out by direct methods.