Mechanistic photodecarboxylation of pyruvic acid: excited-state proton transfer and three-state intersection

Photodissociation dynamics of pyruvic acid experimentally differs from that of commonly known ketones. We have employed the complete active space self-consistent field and its multi-state second-order perturbation methods to study its photodissociation mechanism in the S0, T1, and S1 states. We have uncovered four nonadiabatic photodecarboxylation paths. (i) The S1 system relaxes via an excited-state intramolecular proton transfer (ESIPT) to a hydrogen-transferred tautomer, near which an S1/S0 conical intersection funnels the S1 to S0 state. Then, some trajectories continue completing the decarboxylation reaction in the S0 state; the remaining trajectories via a reverse hydrogen transfer return to the S0 minimum, from which a thermal decarboxylation reaction occurs. (ii) Due to a small S1 -T1 energy gap and a large S1/T1 spin-orbit coupling, an efficient S1 → T1 intersystem crossing process happens again near this S1/S0 conical intersection. When decaying to T1 state, a direct photodecarboxylation proceeds. (iii) Prior to ESIPT, the S1 system first decays to the T1 state via an S1 → T1 intersystem crossing; then, the T1 system evolves to a hydrogen-transferred tautomer. Therefrom, an adiabatic T1 decarboxylation takes place due to a small barrier of 7.7 kcal/mol. (iv) Besides the aforementioned T1 ESIPT process, there also exists a comparable Norrish type I reaction in the T1 state, which forms the ground-state products of CH3CO and COOH. Finally, we have found that ESIPT plays an important role. It closes the S1-T1 and S1-S0 energy gaps, effecting an S1/T1/S0 three-state intersection region, and mediating nonadiabatic photodecarboxylation reactions of pyruvic acid.