RESUMEN
The [2 + 2] photocycloaddition provides a simple, single-step route to cyclobutane moieties that would otherwise be disfavored or impossible due to ring strain and/or steric interactions. We have used a combination of optical and X-ray transient absorption spectroscopies to elucidate the mechanism of the Cu(I)-catalyzed intermolecular photocycloaddition reaction using norbornene and cyclohexene as model substrates. We find that for norbornene the reaction proceeds through an initial metal-to-ligand charge transfer (MLCT) state that persists for 18 ns before the metal returns to the monovalent oxidation state. The Cu K-edge spectrum continues to evolve until â¼5 µs and then remains unchanged for the 50 µs duration of the measurement, reflecting product formation and ligand dissociation. We hypothesize that the MLCT transition and reverse electron transfer serve to sensitize the triplet excited state of one of the norbornene ligands, which then dimerizes with the other to give the product. For the case of cyclohexene, however, we do not observe a charge transfer state following photoexcitation and instead find evidence for an increase in the metal-ligand bond strength that persists for several ns before product formation occurs. This is consistent with a mechanism in which ligand photoisomerization is the initial step, which was first proposed by Salomon and Kochi in 1974 to explain the stereoselectivity of the reaction. Our investigation reveals how this photocatalytic reaction may be directed along strikingly disparate trajectories by only very minor changes to the structure of the substrate.