Development of Pd-immobilized porous polymer catalysts via Bayesian optimization

Abstract

In this study, a Pd-polymeric porous immobilized catalyst is prepared for the Suzuki–Miyaura coupling reactions by employing a Bayesian optimization method to optimize the catalyst. This research represents the first endeavor to utilize machine learning for the optimization of polymer-immobilized catalysts and provides a novel perspective on utilizing machine learning for the optimization of complex materials. This study presented the workflow of machine learning-guided optimization of Pd-immobilized porous polymer catalysts. Two independent variables (DVB and 1-decanol content) were involved in polymerization to maximize TOF as target variable in Suzuki–Miyaura coupling reaction. Bayesian optimization was applied for predictive modeling, and the optimized conditions were experimentally validated in subsequent iterations. By applying this workflow, the catalytic activity of immobilized polymer porous catalysts was successfully optimized using machine learning.