Automated and continuous synthesis of drug substances

A continuous synthesis line was developed integrating different common reaction steps namely nitration, substitution and hydrogenation. Mesalazine as model drug substance was produced from 2-chlorobenzoic acid in continuous flow mode. A multi-instrument PAT strategy was implemented to equip the full line with in-line monitoring capability for critical quality attributes (CQAs) at several points. While most applications of PAT in research and manufacturing of the drug substance are limited to in-process control of quality, in this study real-time process information was generated to support automated process control actions. Process models were developed for all steps of the synthesis route either mechanistic based on reaction kinetics, data driven based on process data or depicting the residence time distribution (RTD). These models facilitated the development and implementation of an automation concept. Feedback loops were established using model predictive controllers for each unit operation, with the objective to maximize yield and minimize side products. The results indicate the high potential of an advanced control concept based on process models and real-time process data.