Master equation calculations were carried out to simulate the production of hydroxyl free radicals initiated by the reaction of acetyl free radicals (CH₃(CO)˙) with molecular oxygen. In particular, the competition between the unimolecular reactions and bimolecular reactions of vibrationally excited intermediates was modeled by using a single master equation. The vibrationally excited intermediates (isomers of acetylperoxyl radicals) result from the initial reaction of acetyl free radical with O₂. The bimolecular reactions were modeled using a novel pseudo-first-order microcanonical rate constant approach. Stationary points on the multi-well, multi-channel potential energy surface (PES) were calculated at the DFT(B3LYP)/6-311G(2df,p) level of theory. Some additional calculations were carried out at the CASPT2(7,5)/6-31G(d) level of theory to investigate barrierless reactions and other features of the PES. The master equation simulations are in excellent agreement with the experimental OH yields measured in N₂ or He buffer gas near 300 K, but they do not explain a recent report that the OH yields are independent of pressure in nearly pure O₂ buffer gas.

Alkyl peroxyls form in the atmospheric oxidation of hydrocarbons and in their combustion. When NO concentration is low, they can appreciably react with themselves. This reaction has both propagation and termination channels. Multireference second-order perturbative energy calculations CAS(16,12)-PT2/6-311G(2df,p) have been carried out on the CAS(8,8)-MCSCF/6-311G(d,p) geometries pertaining to the reaction pathways explored. The tetroxide intermediate put forward first by Russell in 1957 is found as a stable energy minimum, but the calculations indicate that, as the system moves from atmospheric to combustion temperatures, its formation becomes problematic. A concerted synchronous transition structure, apt to connect it with the termination products, formaldehyde, methanol, and dioxygen, is not found. The concerted dissociation of the two external O–O bonds in the tetroxide leads to the (CH23O•)23⋯3O2 complex, with overall singlet spin multiplicity. Both termination via H transfer, to give H2CO, CH3OH, and O2, or dissociation to 2 CH3O•+O2 (possible propagation) are feasible. The former could occur in principle with production of either excited O21 or excited H23CO. However, if a sufficiently easy intersystem crossing (ISC) could take place in the complex, the process would end up with all ground-state molecules. The (possible) propagation channels are favored by higher temperatures, while lower temperatures favor the ISC mediated termination channel. A fairly good qualitative agreement with experimental T dependence of the relevant branching ratio is found. From the tetroxide over again, dissociation of a single external O–O bond leads to CH3O• and CH3O3•, or possibly to a (CH3O•⋯CH3O3•)1 complex, but further transformations along this line are not competitive.