Combined experimental and computational studies of the polymorphism of small organic molecules.

Polymorphism is the ability of a molecule to adopt more than one crystalline form and the control of polymorphism is of importance to the fine chemical industry. Complimentary computational crystal structure prediction and experimental crystallisation techniques have been used to investigate the polymorphism of four organic molecules, none of which were previously known to be polymorphic. For each molecule computational crystal structure prediction produced possible crystal structures, which could correspond to new polymorphs. Manual crystallisation techniques were employed in three instances, and an automated crystallisation platform was used in the fourth, to discover new polymorphs. The crystal structures of all new polymorphs and solvates were fully determined, where possible, by single crystal X-ray diffraction for comparison to the predicted structures. The 5-fluorouracil crystallisation screen produced one new polymorph which corresponded to a low energy predicted structure. For 5-fluorocytosine, where no anhydrous forms had previously been determined, two new polymorphs were discovered, one of which was predicted by the computational results. The study on 3-azabicyclo 3.3.1 nonane-2,4-dione aimed to find a new hydrogen bond dimer-based polymorph inspired by the results of earlier prediction studies. The crystallisation screen produced one new polymorph which was structurally related to the previously reported chain-based structure, along with a high temperature plastic phase. Four polymorphs of 4-hydroxycoumarin were discovered, of which two were fully characterised by single crystal X-ray diffraction and two were identified by powder X-ray diffraction. Many of the newly discovered solvates of these molecules had their hydrogen bonding rationalised in terms of the hydrogen bonded motifs found in the predicted structures of the parent molecule. The viability of computationally predicting monohydrate structures was investigated, using 5-azauracil monohydrate as a test system. This proved a success, with the known crystal structure found by the computational method to be energetically competitive with the other hypothetical structures