Ab initio and transition state theory study of the OH + HO₂ → H₂O + O₂(³Σ_g^-)/O₂(¹Δ_g) reactions: Yield and role of O₂(¹Δ_g) in H₂O₂ decomposition and in combustion of H₂

Reactions of hydroxyl (OH) and hydroperoxyl (HO₂) are important for governing the reactivity of combustion systems. We performed post-CCSD(T) ab initio calculations at the W3X-L//CCSD = FC/cc-pVTZ level to explore the triplet ground-state and singlet excited-state potential energy surfaces of the OH + HO₂ → H₂O + O₂(³Σ_g^-)/O₂(¹Δ_g) reactions. Using microcanonical and multistructural canonical transition state theories, we calculated the rate constant for the triplet and singlet channels over the temperature range 200-2500 K, represented by k(T) = 3.08 × 10¹²T^0.07exp(1151/RT) + 8.00 × 10¹²T^0.32exp(-6896/RT) and k(T) = 2.14 × 10⁶T^1.65exp(-2180/RT) in cm³ mol^-1 s^-1, respectively. The branching ratios show that the yield of singlet excited oxygen is small (<0.5% below 1000 K). To ascertain the importance of singlet oxygen channel, our new kinetic information was implemented into the kinetic model for hydrogen combustion recently updated by Konnov (Combust. Flame, 2015, 162, 3755-3772). The updated kinetic model was used to perform H₂O₂ thermal decomposition simulations for comparison against shock tube experiments performed by Hong et al. (Proc. Combust. Inst., 2013, 34, 565-571), and to estimate flame speeds and ignition delay times in H₂ mixtures. The simulation predicted a larger amount of O₂(¹Δ_g) in H₂O₂ decomposition than that predicted by Konnov's original model. These differences in the O₂(¹Δ_g) yield are due to the use of a higher ab initio level and a more sophisticated methodology to compute the rate constant than those used in previous studies, thereby predicting a significantly larger rate constant. No effect was observed on the rate of the H₂O₂ decomposition and on the flame speeds and ignition delay times of different H₂-oxidizer mixtures. However, if the oxidizer is seeded with O₃, small differences appear in the flame speed. Given that O₂(¹Δ_g) is much more reactive than O₂(³Σ_g^-), we do not preclude an effect of the singlet channel of the titled reaction in other combustion systems, especially in systems where excited oxygen plays an important role.