Generating and interrogating crystallographic data to predict solid-state properties

  • Lygia Silva de Moraes

Student thesis: Doctoral Thesis

Abstract

A bespoke database of crystalline structures of salt forms of active pharmaceutical ingredients has been expanded to 300 structures, by the addition of 75 new crystal structures of salt forms of four bases that were all derivatives of phenylethylamine. These structures, with the addition of salt forms obtained from the CCDC database were analysed in detail according to features of the crystal structure such as cation conformation, presence of multiple crystallographically independent cations in the asymmetric unit and crystal packing similarity. General trends of hydrate formation were observed for these bases. Halides, aliphatic sulfonates, benzoates and monocarboxylate counterions can be classified as predominantly anhydrous while dicarboxylates, inorganic salts with tetrahedral anions and aryl-sulfonates are likely to be hydrates, depending on synthesis and crystallisation conditions. This work also studied the physicochemical properties of these salt forms, such as hardness, solubility and melting point. A new method for analysis of nanoindentation data was presented and used in this work showing a linear relationship between hardness and Young's Modulus. This linearity is characterised by the gradient of the trendline, the elasticity index. It was observed that salt forms of racemic methylephedrine will have higher elasticity index compared to the other salts. There is also a relationship between the hardness / Young's Modulus of the compounds and the size of the anion, clearly observed for halides but also observed in benzoates with different substitutions on the aromatic ring. There is a linear relationship between solubility and melting point when comparing salt forms according to the type of active pharmaceutical ingredient used. Five counterions had interesting results when plotting solubility versus melting point, and these counterion groups can behave in three different ways according to the analysis of the trendlines: (1) when solubility tends to zero the value of x axis tends to the molecular weight of the free acid; (2) when solubility tends to zero the x axis, when corrected by the density of the free acid, also tends to the molecular weight of the free acids; and (3) the samples have solubility increasing with melting point - the inverse behaviour to normal expectation.
Date of Award16 Aug 2019
Original languageEnglish
Awarding Institution
  • University Of Strathclyde
SupervisorAlan Kennedy (Supervisor) & Blair Johnston (Supervisor)

Cite this

'