Quantum Mechanical Methods for Thermal Hazard Risk Assessment in Early Phase Pharmaceutical Development

Abstract

Quantum mechanical (QM) applications to predict heat of reaction (ΔHr) and thermal stability of strained aminocarbocylic salts, diazo compounds, and nitroalkanes for early phase thermal hazard risk assessment is presented. We provide examples on the use of explicit solvation to predict accurate ΔHr. Based on the QM calculations, the criticality class of a copper-catalyzed C–N coupling reaction is determined according to Stoessel’s reaction criticality class and the predictions are consistent with RC1 calorimetric experiment. We emphasize that to predict accurate ΔHr, it is important to consider the roles of reagents and solvents in QM calculations rather than simply considering the bond formation and bond breaking steps involved with reactants and products. Further, the use of predicted ΔHr for salt formation is applied to predict the thermal stability of bromoacetylene azetidine compounds to establish structure–stability relationship which would be useful to identify stable salt intermediates for safe reaction design. A strong correlation between ΔHr and the left limit of the DSC onset temperature (Tinit, °C) of the of exothermic peak is identified (Tinit = −2.85 ΔHr − 99.5). We propose that this model can be used as a prediction tool for novel azetidine salts to provide an estimate of thermal stability before synthesis. In this paper, for the first time we report molecular electrostatic potential (MESP) descriptor for the prediction of Tinit of diazo compounds. The deepest MESP minimum (Vmin) on the diazo group is considered as a probe to quantify the variation in structural effects. A strong correlation between Vmin and Tinit is found which would provide a new way of interpreting the thermal stability of novel diazo molecules just based on chemical structure. Further, the applicability of Vmin is verified on another set of compounds (nitroalkanes) and a good correlation is obtained. The structure–stability relationships that involves Vmin can be a useful QM descriptor for thermal stability prediction of a variety of molecules.