Computation

Overview of the crystal structure calculation process for organic solids

In 1988 Maddox famously described the ongoing inability to predict the structure of molecular solids from a knowledge of their chemical structure as a ‘continuing scandal’.1 Since that time there has been significant progress in hard crystal structure prediction, much of it by the use of ever more sophisticated and intensive computational methods and hence the term crystal structure calculation is perhaps more appropriate. Crystal structure calculation is a task that is severely complicated by the occurrence of polymorphism, solvates/co-crystals and crystal structure with Z′ > 1. It requires a good knowledge of the most stable or most likely molecular structure which in itself can be difficult for conformationally flexible molecules, particularly since crystal packing effects can stabilise conformers that are relatively unstable in solution. The relative stability of a series of candidate crystal structures is also not only potentially dependent on lattice energy but also at real temperatures (i.e. above 0 K) on lattice entropy considerations such as molecular vibration. Moreover, crystal structure is highly dependent on nucleation and growth kinetic factors and morphological considerations (very anisotropic shapes such as thin plates are less likely to be stable). Despite these difficulties the prediction of crystal structure is a much sought-after goal. Ultimately all the materials properties of molecular solids stem from their crystal structure and hence the ability to know ab initio what polymorphic forms are to be expected, and what their solubility, hardness, magnetic and optical properties are likely to be before even synthesising a molecule would be highly desirable. There is now a tremendous amount of data available on crystal structures, particularly from the Cambridge Structural Database, which forms an invaluable tool for organisation and statistical analysis of experimental structure data for comparison with prediction and calculation. Despite the fact that, in general, crystal structures are not predictable, a number of attempts, some of them increasingly successful, have been made to address the problem from a computational standpoint. Crystal structure calculation is particularly useful in tandem with additional data from other (usually experimental) sources.

The calculation of hydrate crystal structure is particularly complicated by the need to determine the positions of both water and organic components simultaneously. The CPOSS grouping, particularly the work of Price at University College London, has reported some progress recently: "Which, if any, hydrates will crystallise? Predicting hydrate formation of two dihydroxybenzoic acids", D. E. Braun, P. G. Karamertzanis and S. L. Price, Chem. Commun., 2011, 47, 5443. There is also a report of the structure calculation of a dihydrate structure implicated in ultrasound-induced gel formation: "Structure Calculation of an Elastic Hydrogel from Sonication of Rigid Small Molecule Components", K. M. Anderson, G. M. Day, M. J. Paterson, P. Byrne, N. Clarke and J. W. Steed, Angew. Chem., Int. Ed., 2008, 47, 1058–1062.

1. Maddox, J. Crystals from 1st Principles. Nature 335, 201-201, (1988).