Predictive, Data-Driven Design of Red-Light Photoredox Catalysts for C─Heteroatom Bond Formation

Photocatalysis is a powerful tool for the synthesis of organic molecules, yet its widespread application is hindered by the dependence on high-energy light sources and expensive metal-based catalysts, which can limit scalability and environmental sustainability. In this study, we present a modular design strategy for organic dyes engineered for efficient red-light absorption, enabling photocatalytic reactions under low-energy irradiation. Our findings establish a clear relationship between the oxidation potential of the photocatalyst and the nature of its donor moiety, as well as between the reduction potential and the electronic characteristics of its core structure. Moreover, we demonstrate that the E_0-0 energy of a photocatalyst can be predicted via multivariate linear regression using the donor's oxidation potential and the core's reduction potential as descriptors. Utilizing this strategy, we synthesized red-light-absorbing photocatalysts that efficiently promote C─heteroatom cross-coupling reactions under mild conditions. This approach overcomes the limitations of blue-light photocatalysis by offering broad substrate compatibility, including π-conjugated aryl bromides and photolabile functional groups, while minimizing undesirable hydrodehalogenation. By reducing reliance on precious metals and improving energy efficiency, our approach provides a scalable alternative to traditional photocatalysis and advances the development of metal-free photocatalysts for sustainable chemistry.