Solvent Dependency of Catalyst-Substrate Aggregation Through π-π Stacking in Photoredox Catalysis

Assemblies of photoredox catalysts and their target substrates prior to photoexcitation is a phenomenon naïvely overlooked by the majority of synthetic chemists, but can have profound influences on reactivity and selectivity in photocatalytic reactions. In this study, we determine the aggregation states of triarylamine radical cationic photocatalysts with various target arene substrates in different solvents by specifically parameterized polarizable molecular dynamics simulations. A π-stacking interaction previously implicated by more expensive, less-representative quantum calculations is confirmed. Critically, this study presents new insights on: i) the ability of solvents (MeCN vs DMF) to make or break a photocatalytic reaction by promoting (MeCN) or demoting (DMF) its catalyst-substrate assemblies, which is a determining factor for reactivity, ii) the average "lifetimes" of assemblies in solution from a dynamic simulation. We find that both in the ground state and the photoexcited state, the cationic radical assemblies remain intact for periods often higher than 60 ps, rendering them ideally suitable to undergo intra-assembly electron transfer reactions upon photoexcitation. Such aspects have not addressed by previous studies on synthetic photocatalytic reactions involving non-covalent assemblies.