Data-Driven Workflow for the Development and Discovery of N-Oxyl Hydrogen Atom Transfer Catalysts

N-oxyl species are promising hydrogen atom transfer (HAT) catalysts to advance C–H bond activation reactions. However, because of the complex structure–activity relationship within the N-oxyl structure, catalyst optimization is a key challenge, particularly for simultaneous improvement across multiple parameters. This paper describes a data-driven approach to optimize N-oxyl hydrogen atom transfer catalysts. A focused library of 50 N-hydroxy compounds was synthesized and characterized by three parameters─oxidation peak potential, HAT reactivity, and stability─to generate a database. Statistical modeling of these activities described by their intrinsic physical organic parameters was used to build predictive models for catalyst discovery and to understand their structure–activity relationships. Virtual screening of 102 synthesizable candidates allowed for rapid identification of several ideal catalyst candidates. These statistical models clearly suggest that N-oxyl substructures bearing an adjacent heteroatom are more optimal HAT catalysts compared to the historical focus, phthalimide-N-oxyl, by striking the best balance among all three target experimental properties.

Machine learning models revealed that N-oxyl compounds bearing adjacent heteroatoms to carbonyls are more promising hydrogen atom transfer catalysts by striking a balance between multiple parameters.