High-throughput experimentation and machine learning-promoted synthesis of α-phosphoryloxy ketones via Ru-catalyzed P(O)O-H insertion reactions of sulfoxonium ylides

Abstract

Herein, we report a novel and highly efficient method for the synthesis of α-phosphoryloxy carbonyl compounds via Ru-catalyzed P(O)O-H insertion reactions of sulfoxonium ylides and phosphinic acids, with the assistance of high-throughput experimentation (HTE) and machine learning (ML). A variety of P(O)O-H derivatives, including diarylphosphates, alkyl phosphates, and alkoxyphosphates, are competent candidates to react with sulfoxonium ylides in this transformation, and various α-phosphoryloxy carbonyls and propylene phosphates are directly constructed. This approach utilizes readily available sulfoxonium ylide as a carbene precursor, and features mild conditions, operational simplicity, and broad functional groups tolerance, and could be used for late-stage functionalization of structurally complex bioactive molecules. Moreover, a conducive exploration of the reaction space is also conducted (756 reactions) and a machine learning model for reaction yield prediction has been developed and applied, showcasing the practical application of this newly workflow (HTE-ML) in the field of synthetic chemistry.