Computer modeling of smog chamber data: Progress in validation of a detailed mechanism for the photooxidation of propene and n‐butane in photochemical smog

Abstract A detailed mechanism is presented for reactions occurring during irradiation of part‐per‐million concentrations of propene and/or n ‐butane and oxides of nitrogen in air. Data from an extensive series of well‐characterized smog chamber experiments carried out in our 5800‐liter evacuable chamber–solar simulator facility designed for providing data suitable for quantitative model validation were used to elucidate several unknown or uncertain kinetic parameters and details of the reaction mechanism. The mechanism was then tested against the data base from the smog chamber runs. In general, most calculated concentration–time profiles agreed with experiments to within the experimental uncertainties. Fits were usually attained to within ∼±20% or better for ozone, NO, propene, and n ‐butane, to within ∼±30% or better for NO 2 , PAN, methyl ethyl ketone, 2‐butyl nitrate, butyraldehyde, and (in runs not containing propene) methyl nitrate, to within ⋐±50% or better for the minor products 1‐butyl nitrate and propene oxide, and to within a factor of 2 for methyl nitrate in propene‐containing runs. Propionaldehyde was consistently underpredicted in all runs; it is probably a chamber contaminant. For formaldehyde and acetaldehyde, the major products in both systems, fits to within ⋐±20% were often obtained, yet for a number of experiments, significantly greater discrepancies were observed, probably as a result of experimental and/or analytical problems. The good fits to experimental data were attained only after adjusting several rate constants or rate constant ratios related to uncertainties concerning chamber effects or the chemical mechanism. The largest uncertainty concerns the necessity to include in the mechanism a significant rate of radical input from unknown sources in the smog chamber. Other areas where fundamental kinetic and mechanistic data are most needed before a predictive, detailed propene + n ‐butane‐NO x ‐air smog model can be completely validated concern other chamber effects, the O 3 + propene mechanism, decomposition rates of substituted alkoxy radicals, primary quantum yields for radical production as a function of wavelength for aldehyde and ketone photolyses, and the mechanisms and rates of reactions of peroxy radicals with NO and NO 2 .