Self-optimization of Claisen-Schmidt condensation in an automated microflow reaction system using machine learning

An efficient self-optimization system was developed for managing chemical reactions in a plug flow reactor, aiming to minimize reactant and intermediate concentrations while maximizing product yield. The approach was demonstrated using a Claisen–Schmidt condensation between 2-methoxybenzaldehyde and acetone. Kinetic parameters, including activation energy, pre-exponential factors, and reaction orders, were determined and integrated into mass balance equations to predict final reactant and product concentrations. The self-optimization system autonomously adjusted flow rates, achieving experimental results with a deviation of ±10% from theoretical predictions. Compared to classical methods, which first determine kinetic parameters in batch systems, this system significantly reduces the time required to reach optimal conditions. Additionally, it minimizes chemical consumption, enhancing both environmental sustainability and economic efficiency. This work highlights the potential of self-optimization in chemical reaction engineering, offering a faster and more resource-efficient alternative to conventional optimization approaches.