Simultaneous Screening and Reaction Optimization in an Automated Segmented-Flow Microreactor System

High-throughput technologies greatly enhance the speed at which medicinal chemists screen through libraries of reagents, catalysts, and ligands to synthesize a target molecule. Often overlooked during a screen, however, is the response of the reaction to both the discrete variables of interest and to continuous variables such as temperature, reaction time, and reagent concentrations. The best catalyst can easily be overlooked, for instance, if its activation is dependent on one or more other discrete or continuous factors. As enumeration of all discrete and continuous variables in reaction development is both cost-prohibitive and time-consuming, we propose a deterministic mixed-integer nonlinear programming approach to optimizing a target synthesis. Our approach is demonstrated online using a fully automated segmented flow screening system capable of handling the harsh reagents and reaction conditions found in most organic syntheses. Reactions are studied on the 20-μL scale, with results that are scalable because of the rapid heat and mass transfer rates achievable in segmented flow microfluidic systems. Our approach is demonstrated for the identification of the optimal solvent and reaction conditions in a selective alkylation reaction and for the optimization of ligand selection in a palladium-catalyzed cross-coupling reaction.