Crystal structure prediction search method

Abstract Methods, evolutionary and systematic search approaches, and applications of crystal structure prediction of closest-packed and framework materials are reviewed. Strategies include developing better cost functions, used to assess the quality of the candidate structures that are generated, and ways to reduce the set of candidate structures to be assessed. The crystallographic coordinates for new materials, available only as a powder sample, are often intractable from diffraction data alone. In recent years, steady progress has been made in the ability to solve previously unknown crystal structures of such compounds, the generation of known structures (inferring more confidence in such approaches) and the prediction of hypothetical yet-to-be-synthesised structures. 

There have been significant recent advances in both the model potentials that can be used in molecular crystal structure prediction and new methods of searching for hypothetical structures. We may expect that these will come together over the next decade, so that the prediction of a molecular crystal structure from first principles becomes possible. This should lead to an understanding of polymorphism. 

The prediction of crystal structures involves the identifieation of minima on the potential energy landscape, regardless of the method used to calculate this landscape. Owing to the many degrees of freedom involved, this process is not trivial, in particular for co-crystals and salts. This section provides an overview of search methods used in crystal structure prediction and the most important pitfalls are discussed. 

One can distinguish two phases in a crystal structure prediction, both quite difficult. The first phase is to make a list of structures that are not entirely unreasonable. This list was in older work rather short and obtained by hand, by comparing the molecule with related ones and guessing what kind of interactions could be possible and what kind of structures looked probable. A review of that type of work can be found in the book by Pertsin and Kitaigorodsky. In recent work this subjective method has been replaced by more or less well-defined algorithms, always based on energies calculated by an empirical force field. It turns out that lists obtained in that way can contain hundreds of possible structures. Two strategies are used random search versus systematic search. The... 

Some of the motivation for refinement of the methods has been a search for an understanding of what determines the crystal structure of individual metals. The first serious attempt at predicting structures was directed at Na, Mg, and A1 (Harrison, 1964), and it succeeded completely for these three metals. Subsequent studies (reviewed by Harrison, 1966a, p. 189ff) suggested that these metals were a fortuitous choice, however. More recent, careful studies seem to have confirmed that second-order pseudopotential theory is hot adequate to determine the structures correctly. One contrary example is the work of Hafner and Nowotny (1972), which correctly gave nine out of nine structures for pure metals. An extension of... 

---