Modelling of secondary organic aerosol formation from isoprene photooxidation chamber studies using different approaches

Secondary organic aerosol (SOA) formation from the photooxidation of isoprene was simulated against smog chamber experiments with varied concentrations of isoprene, nitrogen oxides (NOx = NO + NO2) and ammonium sulfate seed aerosols. A semi-condensed gas-phase isoprene chemical mechanism (ISO-UNC) was coupled with different aerosol-phase modelling frameworks to simulate SOA formation, including: (1) the Odum two-product approach, (2) the 1-D volatility basis-set (VBS) approach and (3) a new condensed kinetic model based upon the gas-particle partitioning theory and reactive uptake processes. The first two approaches are based upon empirical parameterisations from previous studies. The kinetic model uses a gas-phase mechanism to explicitly predict the major intermediate precursors, namely the isoprene-derived epoxides, and hence simulate SOA formation. In general, they all tend to significantly over predict SOA formation when semivolatile concentrations are higher because more semivolatiles are forced to produce SOA in the models to maintain gas-particle equilibrium; yet the data indicate otherwise. Consequently, modified dynamic parameterised models, assuming non-equilibrium partitioning, were incorporated and could improve the model performance. In addition, the condensed kinetic model was expanded by including an uptake limitation representation so that reactive uptake processes slow down or even stop; this assumes reactive uptake reactions saturate seed aerosols. The results from this study suggest that isoprene SOA formation by reactive uptake of gas-phase precursors is likely limited by certain particle-phase features, and at high gas-phase epoxide levels, gas-particle equilibrium is not obtained. The real cause of the limitation needs further investigation; however, the modified kinetic model in this study could tentatively be incorporated in large-scale SOA models given its predictive ability.