Identifying new co-crystal complexes from powder data

The likelihood and different methods of co-crystal formation (i.e. crystal structures composed of more than one component) are currently hot topics in crystallography and crystal engineering. This is particularly true in the case of Active Pharmaceutical Ingredients (APIs), where the pharmaceutical industry is starting to realize the huge potential of being able to modify the physical properties of their new API by co-crystallizing it with a so-called Inactive Ingredient. These Inactive Ingredients are compounds that the FDA in the US has licensed as being acceptable to add to any medicine without undertaking any additional trials on it. Co-crystals or crystalline molecular complexes with the same API will have strikingly different solubilities and dissolution rates depending on the nature of co-crystal formers. 

There is a popularly held view in crystallography that co-crystals are less likely to exhibit polymorphism than single component systems, although there is little evidence to support this view. It is true that there are few known co-crystal systems with documented polymorphism. But there is also an old adage in crystallography that "the number of polymorphs known is proportional to the time and money spent looking for them". As polymorphism in co-crystals is a relatively new area of research then perhaps it is not surprising that few cases have yet been documented. 

This study is part of an investigation of the effect of solvent-mediation on co-crystallization by looking at chlorobenzoic and fluorobenzoic acids co-crystallized with benzimidazole in four different solvents; methanol, ethanol, 1-propanol and acetone. The results presented here focus on 4-chlorobenzoic acid with benzimidazole. It is particularly important in these experiments to be able to get a quick snapshot of the results; these experiments can give a large number of powder samples to be analyzed in a very short time, and the ideal scenario is to be able to obtain a powder pattern on the RAPID II in around 5-10 minutes per sample to give a quick indication of which samples it is worth pursuing.

Each powder pattern was collected using a 0.5 mm capillary using Mo radiation for a total of 5 minutes per image whilst rotating through 5° in ω. As the unit cells for these compounds are likely to be of a reasonable size the resolution for the powder patterns is not particularly good; however, they are intended purely as fingerprints to give an initial indication of how much effort should be put in to the analysis. In this context, these results are very useful. None of the recrystallizations from any of the solvents greatly resembles either of the reference patterns, although the patterns from powders recrystallized from methanol and ethanol solvents resemble each other, as do the patterns from powders recrystallized from acetone and 1-propanol. The particular regions that suggest differences are highlighted below with black and red arrows.

Results courtesy of Andy Parkin, Lynne H Thomas and Chick C Wilson, Structural Chemistry Group,  University of Glasgow

Tags: WAXS, cellulose, wood, fibres, two-dimensional X-ray diffraction