A smog chamber and modeling study of the gas phase NOx–air photooxidation of toluene and the cresols

An experimental and modeling study of irradiated toluene–NO_x–air, toluene–benzaldehyde–NO_x–air, and cresol–NO_x–air mixtures at part-per-million concentrations has been carried out. These mixtures were irradiated at 303 ± 1 K in a 5800-liter Teflon-lined, evacuable environmental chamber, with temperature, humidity, light intensity, spectral distribution, and the concentrations of O₃, NO, NO₂, toluene, PAN, formaldehyde, benzaldehyde, o-cresol, m-nitrotoluene, and methyl nitrate beingmonitored as a function of time. For the toluene and toluene–benzaldehyde–NO_x–air runs a variety of initial reactant concentrations were investigated. Cresol–NO_x–air runs were observed to be much less reactive in terms of O₃ formation and NO to NO₂ conversion rates than toluene–NO_x–air runs, with the relative reactivity of the cresol isomers being in the order meta » ortho > para. The addition of benzaldehyde to toluene–NO_x–air mixtures decreased the reactivity, in agreement with previous studies. Alternative mechanistic pathways for the NO_x photooxidations of aromaticsystems in general are discussed, and the effects of varying these mechanistic alternatives on the model predictions for the toluene and o-cresol–NO_x–air systems are examined. Fits of the calculations to most of the experimental concentration–time profiles could be obtained to within the experimental uncertainty for two of the mechanistic options considered. In both cases it is assumed that (1) O₂ adds to the OH–toluene adduct ～75% of the time forming, after a further addition of O₂, a C₇ bicyclic peroxy radical, and (2) this C₇ bicyclic peroxy radical reacts with NO ～75% of the time to ultimately form α-dicarbonyls and conjugated γ-dicarbonyls (e.g., methylglyoxal + 2-butene-1,4-dial) and ～25% of the time to form organic nitrates. The major uncertainties in the mechanisms concern (1) the structure of the bicyclicperoxy intermediate, and (2) the γ-dicarbonyl photooxidation mechanism. Good fits to the o-cresol concentration–time profiles in the toluene–NO_x runs are obtained if it is assumed that o7-cresol reacts rapidly with NO₃ radicals. However, it is observed that the model underpredicts nitrotoluene yields by a factor of ～10, but this is in any case a minor product. It is concluded that further experimental work will be required toadequately validate the assumptions incorporated in the aromatic photooxidation mechanisms presented here.