Conclusions and Further Work

Abstract

The complementary application of experimental and computational approaches has helped in improving methods for discovery of solid forms and providing a more complete and accurate interpretation of the solid-state structure and diversity. By combining multi-well plate for high throughput crystallisation, Raman spectroscopy for data collection and chemometric approach for data analysis, an efficient methodology for solid form screening was developed. The developed methodology has proved to be an effective tool for discovering solid forms by utilizing only mgs of material. Future work would include further refinement of the methodology by incorporating stirring element and temperature control unit to have better control over crystallisation conditions, by combining Raman with another analytical techniques and applying this methodology under variety of crystallization conditions. A random forest model has for the first time enabled the prediction of the crystallisability (crystals vs. no crystals) of organic molecules with ~70 % accuracy. Future work would include incorporating information about crystallisation conditions (solvent, rate of solvent evaporation, RH, and temperature) and extending this technique to multi-component systems such as salts and co-crystals. Complementary application of experimental and computational tools has provided a better understanding of the factors underpinning the solid-state structure and diversity in two groups each comprising of two structurally related molecules. Future work include to investigate if any of the experimentally observed crystal structure of these molecules can act as a template for the predicted structure of structurally related molecules on the crystal energy landscape. Other computational tools such as electrostatic potential surfaces calculations and Hirshfeld surface analysis can be tested to study the influence of overall molecular shape on distinct solid-state behaviour.