Machine-learning-driven exploration of Suzuki–Miyaura cross-coupling with a polymer-supported Pd in continuous-flow system

Abstract In this study, we developed a hybrid methodology for the machine-learning-driven optimization of a continuous-flow reaction with a polymer-supported Pd catalyst, aiming to both boost productivity and elucidate the influencing factors under the reaction conditions. A porous polymer bearing phosphine ligand was prepared by using polymerization-induced phase separation and Pd was coordinated to the support to construct the flow reactor. Suzuki–Miyaura cross-coupling reactions were performed in the continuous-flow system. Combining Bayesian optimization and linear regression realized the optimization of continuous-flow conditions and analysis of key influencing factors, demonstrating the utility of the present machine-learning method. Indeed, the continuous-flow system with the monolith reactor was also applicable to a range of chloroarenes, which emphasized the importance of our catalytic system for fine chemical production.