Conférence SigmaAldrich avec Prof Corina Schindler (University of British Columbia)

Titre: Des cycloadditions [2+2] et l'optimisation électrochimique in situ basée sur les données analytiques

Résumé : Data-driven reaction optimization is limited by the scarcity of descriptors that are simultaneously information-rich and experimentally scalable. Electrochemical measurements offer a direct, physically grounded probe of reaction environments, but their integration into machine-learning workflows has been hindered by the lack of high-throughput hardware and automated analysis pipelines. Here, we introduce an in situ electrochemical optimization framework that combines custom parallelized hardware, automated data processing, and machine learning to enable high-throughput acquisition and encoding of electrochemical descriptors for reaction modeling. The platform centers on a miniaturized, parallel three-electrode array that performs non-destructive electroanalytical measurements prior to electrolysis within the same experimental configuration, ensuring that each reaction is represented by descriptors intrinsically tied to its operative electrochemical environment. When coupled with computer-vision–based feature extraction and Bayesian optimization, the framework enables efficient, data-efficient exploration of large reaction design spaces under tightly controlled conditions. Together, these elements establish a generalizable electrochemical machine-learning platform for scalable reaction optimization and predictive modeling.