From crystal structure to crystal particle - a combined experimental and computational study of polymorphism in molecular crystals

  • Pietro Sacchi

Student thesis: Phd


As the majority of pharmaceuticals are marketed as crystalline products, the control of their solid form landscape during development, as well as during the manufacturing process, is of central interest to pharmaceutical industries. Most organic molecular compounds can adopt different solid-state arrangements when crystallising, resulting in individual phases each with its own properties. Although the methods for the computational prediction of this polymorphism are advancing rapidly, a connection between the latter and the experimental conditions needed to isolate the desired polymorphic phase selectively and accurately are still lacking. In fact, each polymorphic system constitutes its own island, and its understanding requires extensive experimental effort. This work explores the combined use of experimental and computational methods to study the polymorphism of molecular compounds. First, an assessment of the accuracy of structural comparison methods in identifying crystal forms in the Cambridge Structural Database (CSD) is presented. Our results show that while automated comparison methods are indeed a useful tool for phase identification, the assessment of an expert is still required for unambiguous classification. Then, the polymorphism of tolfenamic acid (TFA) is studied in two parts. The first reports the unexpected discovery of the ninth polymorph of TFA and its characterisation through experimental and computational techniques. In the second part, crystal growth kinetics of three polymorphs of TFA, including the new polymorph TFA-IX, were measured experimentally and compared with the assistance of in-house written Python programs with the objective of understanding why TFA-IX had not been discovered before. This is the first example of a comparison of experimental growth kinetics for three distinct polymorphs, and methods to compare overall crystal growth rates of polymorphs are critically analysed and discussed. Finally, the observation of a reversible polymorphic transformation between two polymorphs of ritonavir (RVR) during ball-mill grinding experiments was used as starting point for the production of a code to compute particle stabilities at the nanoscale. Computations for RVR showed that a stability switch at nanoscale size is possible even for polymorphs with relatively high lattice energy difference. These results show how computational tools could be used to direct polymorph discovery, especially through the versatile and effective technique of ball-mill grinding.
Date of Award1 Aug 2023
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorRobin Smith (Supervisor) & Aurora J. Cruz-Cabeza (Supervisor)


  • Crystal size and shape
  • Crystal growth
  • Computational methods
  • Polymorphism
  • Molecular crystals

Cite this