Application of state-of-the-art computational methods to gain mechanistic insights into industrial processes development work for the manufacture of active pharmaceutical substances

Abstract

DFT studies were conducted on reactions of pharmaceutical interest to evaluate the feasibility of QM methodology to predict the course of chemical reactions on the syntheses of APIs and to provide mechanistic rationales on abnormal side reactions. Herein, energy calculations for gas phase, at the B3LYP/6-31G(d) level of theory, provided scientific explanation for several questions addressed in this thesis. The unsuccessful synthesis of doxycycline acetate through the route using doxycycline monohydrate and acetic acid starting materials was explained by an energy barrier (EB) more than three times higher (16 kJ/mol) than that of doxycycline oxalate (5 kJ/mol) synthesized with oxalic acid. Overall energy balance (OEB) calculations for an alternative synthetic route starting from doxycycline hyclate and sodium acetate showed this to be energetically more favorable (exergonic OEB=-14 kJ/mol) than the only plausible side reaction (endergonic OEB=37 kJ/mol), leading to a successful synthesis of doxycycline acetate. High levels of DTG impurity observed on the synthesis of umeclidinium bromide precursor were due to favorable OEB on the hydrolyses side reactions of N-depropylaclidinium (OEB=-403 kJ/mol) and MDTG (OEB=-401 kJ/mol), and to a lower EB of MDTG hydrolysis (25 kJ/mol) than that of N-depropylaclidinium synthesis (EB=47 KJ/mol). On the synthesis of sancycline, the study of the mechanisms showed that hydrogenolysis is the preferred pathway because it presents a lower EB (16 kJ/mol) than hydrogenation (78 kJ/mol), and a more exergonic OEB (-76 kJ/mol vs -52 kJ/mol for the hydrogenation). Anhydro sancycline impurity high levels can be explained by the slightly higher EB (25 kJ/mol) than that of the hydrogenolysis. The high levels of the impurity dimer observed during the synthesis of a precursor of umeclidinium bromide are explained by the lower EB for the side reactions leading to the dimer formation (105 kJ/mol, via main product and via secondary product) when compared with those of the synthesis of the main product (113 kJ/mol) and of the secondary product formation (118 kJ/mol). The in-silico data obtained under computational conditions simulating experimental procedures, like temperature, pressure, and the reaction solvent, confirmed the gas phase results.