Machine Learning-Guided Photocatalytic Cross-Coupling of Phenols and Heteroaryl Halides

Developing sustainable methods for C(sp2)–C(sp2) bond formation that avoid transition-metals and prefunctionalized substrates remains a central goal in synthetic chemistry. Phenols and N-heteroarenes (azines) are abundant feedstocks, yet their cross-coupling is hindered by mismatched redox properties and competing pathways. Herein, we report a photochemical strategy that couples phenols with heteroaryl halides under redox-neutral conditions using an organic dye photocatalyst and base. Concurrent oxidation of the phenol component and reduction of the azine component generates complementary radicals that cross-couple efficiently, delivering moderate to high yields (up to 91%) with high functional group tolerance. Mechanistic experiments and density functional theory (DFT) studies elucidate the radical reaction pathways, while substrate clustering, high-throughput experimentation (HTE), and machine learning (ML) enable prediction of C–C versus SNAr reactivity across broad chemical space.