Predicting glass transition temperatures of unstable drugs using co-amorphous mixtures

Abstract The model compounds benzocaine (BENZ) and lidocaine (LIDO) exhibited amorphous phase instability, evident by the absence of measurable glass transition temperatures ( T g ) during differential scanning calorimetric (DSC) analysis; this was associated with their high propensity to crystallise upon cooling. Through formation of co-amorphous mixtures (CAMs) with each other and ketoprofen (KETO), this instability was minimised with some instances of glass solution formation. Stability enhancement by these means enabled T gCAM measurements, followed by modelling neat drug T g through rearranged and modified forms of the Fox and Gordon-Taylor (GT) equations. Despite its simplicity, the rearranged form of the Fox equation (R-Fox) led to predicted T g of − 30.3 and − 68.5 °C for BENZ and LIDO, respectively, which were comparable to literature values of − 31.0 and − 63.4 °C. The more complex modified GT equation (M-GT) led to similar T g predictions of − 34.3 and − 64.8 °C and allowed for T g interpolation with respect to non-ideal mixing effects. It was then possible to determine the optimal ratio of BENZ:LIDO as 5:5 mol/mol, with a GT model deviation of + 8.2 °C and + 7.5 °C for the R-Fox and M-GT predicted neat T g values, respectively. This outcome affirmed these models as successful T g predictors of poor glass forming compounds. Observed deviations from literature were contextualised by recognising the role of intermolecular interactions and their relationship with p K a , rotatable bonds, molecular mass, aromatic rings, and number of hydrogen bond donors and acceptors.