AI-driven, self-optimizing electrochemical flow platform for rapid synthesis of 2-amino substituted benzothiazole libraries

Abstract

Electrochemical synthesis remains challenging due to the complex interplay of redox, mass-transfer, and operational parameters. We report a data-driven electrochemical flow platform for the efficient synthesis of functionalized benzothiazoles via a previously unexplored isocyanide insertion pathway. The reaction proceeds under mild, catalyst- and oxidant-free conditions and tolerates a broad range of 2-aminothiophenols and isocyanides, including sterically hindered, aliphatic, aromatic, and heterocyclic substrates. Closed-loop Bayesian optimization identified key interdependencies between current, residence time, and electrolyte loading, providing reproducible conditions that deliver the model product in up to 92% isolated yield, a fourfold improvement over the batch yield of 24%. The platform was further applied to the automated generation of a 21-member benzothiazole library, illustrating the generality, scalability, and operational robustness of integrating electrochemical flow synthesis with machine-guided optimization for rapid heterocycle construction. The optimized flow conditions were readily translated to gram-scale synthesis, delivering the target benzothiazoles with maintained yield and operational stability under continuous operation.