Much of the recent focus of CCDC research has related to the control and prediction of the organic solid state. The field of crystal engineering is one in which the CCDC has played a leading role in establishing. In common with other areas, this has involved a great deal of collaboration, in this case that collaboration has most recently been with fellow members of the Pfizer Institute for Pharmaceutical Materials Science (PIPMS).

Work over the last decade in particular has shown how analysis of the relative likelihoods of interaction patterns in the Cambridge Structural Database (CSD) can be highly applicable to understanding solid forms. Lately, significant research effort, in collaboration with PIPMS, has been focussed on the quantification of the likelihood that a hydrogen bond interaction will occur. The culmination of this research, a tool which can predict likely and unlikely hydrogen bond outcomes, is now being used within our consortium of industrial development chemists in the pharmaceutical and agrochemical businesses. Application of the tool allows solid state chemists to assess the possibility of polymorphism and understand the implications for stability of the observed hydrogen bonding. Recent developments have tailored the methodology such that it can be used to aid cocrystal screen design. Coformers are run against the target molecule to establish if there are strong hetero- interactions between the components. This tool will help chemists design cocrystal screens rationally and investigate possible solvated forms.

Research Plans

Strong intermolecular interactions, either mediated by hydrogen bonds or dipolar interactions have been fundamental to the development of the field of crystal engineering. However, reliance only on strong intermolecular interactions as a tool for crystal design does not result in reliably engineered structures in all cases. There are numerous examples of polymorphic systems where either the hydrogen bonding pattern between the polymorphs is the same, or there is no hydrogen bonding present. This indicates that for a greater understanding of polymorphism and how molecular building blocks interact, we should be looking towards the less well defined, weaker, non-H-bond interactions. A project will be undertaken where we extract and classify non-H-bond interactions to examine whether they can be incorporated into a prediction strategy alongside, stronger, more traditional interactions.

There is a strong desire to understand, predict and control the morphology of a crystal. Awkward morphologies, such as needles or plates, can cause significant problems when large scale, bulk processing is required, for example in the pharmaceutical industry. The CSD is a rich source of knowledge on intermolecular interactions and we are now combining this information with crystal morphology data. We hope to develop heuristic rules for morphology prediction, identifying molecular interactions which have a high chance of resulting in a particular morphology.

Recent work has shown that molecular docking packages like CCDC's GOLD can be successfully applied to the crystal engineering problem of predicting multi-component crystal forms. Future plans in this area include re-parameterising a current scoring function for use in small molecule structures, tests on cross-docking of guests between different frameworks and blind tests on pharmaceutical co-crystal systems.