Crystal structure generation

Gavezzotti, A., Flippini, G., Kroon, J., van Eijck, B. P. and Klewinghaus, P. (1997). The crystal polymorphism of tetrolic acid (CHsC CCOOH) a molecular dynamics study of precursors in solution, and in crystal structure generation. Chem. Fur. J., 3, 893-9. [183]... 

For some compounds, suitable crystals simply can not be obtained and all attempts are frustrated. The reasons are not well understood. In some cases, repeated attempts with varied conditions always produce multiple or twinned crystals. Trying a range of different solvents is the most obvious approach. It is not just that the crystal quality may vary with the solvent used incorporation of solvent in the crystal structure generates a different crystalline form (a solvate), which will probably have quite different crystal growth characteristics. In most cases the nature of the solvent is unimportant in the quest for a definitive crystal structure analysis. However, the study of different solvates (and different polymorphs) of the same compound can itself be of interest, with significant changes in molecular structure resulting from different intermolecular interactions. 

Figure 5.17 Topography of crystalline polyethylene. The chain folds of the crystal structure generate the rough appearance. (Reproduced with permission from B.D. Ratner and V.V. Tsukruk (eds), Scanning Probe Microscopy of Polymers, American Chemical Society, Washington D.C. 1998 American Chemical Society.)...

When crystal-structure generation procedures are attempted on high-nuclearity complexes, the limitations inherent in the description of the metal atoms become increasingly severe. With crystalline Fe3(CO)i2 and Ru3(CO)i2, the experimental crystal structures could be compared with other theoretical structures, although the PPE of the theoretical crystals were invariably higher than the range expected. A comparison of the observed crystal packing of Fe3(CO)i2 (constructed with only one of the two disordered molecules related by centers of inversion) and the best solution of the theoretical approach is shown in Fig. 2. 

One might object that the intrinsic structural flexibility of the molecule should discourage attempts at crystal-structure generation with a rigid molecular building... 

Once a molecular geometry, or ensemble of conformers, has been produced, the central computational step involved in CSP is the generation of trial crystal structures (Fig. 5.1). The goal of crystal structure generation methods is to produce aU possible, physically reasonable arrangements of molecules in a translationaUy repeating arrangement. 

Many types of energy calculations have been applied in CSP studies, starting from the simplest atom-atom descriptions of interactions. While fairly simple brute force (grid and random strncture generation) approaches are often suc-cessfnl at addressing the crystal structure generation problem, the assessment of relative energies requires the development of sophisticated methods to achieve the required accuracy for CSP. One approach has been the development of more elaborate, anisotropic (nonspherical) atom-atom models." Recently, the application of QM solid-state electronic strncture calculations has also been demonstrated to be very successful. " ... 

The success of crystal structure generation methods for rigid molecules has been quite high across the blind tests, with most of the participants locating the observed structure somewhere in their lists for most molecules in categories 1 and 2. The notable exceptions are molecules XI, IX, and xm (Table 1). 

The Prom procedure is encoded in a mature software that has been used over the years with considerable success [1,9,12]. As an example of the general landscape that a crystal structure predictor must face, an exercise in crystal structure generation by Prom will now be discussed. 

Table 14.6 Crystal structure generation by the Prom procedure for chlorocorannulene. Raw structures output of the first search first sorting clustered structures from raw structures (no optimization) final sorting number of optimized crystal structures best energy (kJ mol ) best lattice energy of final sorting. For some space gronps, results for the dicarbonyl derivative are also reported. Crystal data are collected in the Snpplementary material, file chlorcor.oeh and colcor.oeh...

Table 14.7 Crystal structure generation by the Prom procedure for coraimulene carboxylic acid. See Table 14.6 for the meaning of the headers. Complete crystal data are collected in the Supplementary material, file coohcor.oeh...

---