Dynamic consideration of smog chamber experiments

Abstract. Recent studies of the α-pinene + ozone reaction that address particle nucleation show relatively high molar yields of highly oxidized multifunctional organic molecules with very low saturation concentrations that can form and grow new particles on their own. However, numerous smog-chamber experiments addressing secondary organic aerosol (SOA) mass yields, interpreted via equilibrium partitioning theory, suggest that the vast majority of SOA from α-pinene is semivolatile. We explore this paradox by employing a dynamic volatility basis set (VBS) model that reproduces the new-particle growth rates observed in the CLOUD experiment at CERN and then modeling SOA mass yield experiments conducted at Carnegie Mellon University (CMU). We find that the base-case simulations do overpredict observed SOA mass but by much less than an equilibrium analysis would suggest; this is because delayed condensation of vapors suppresses the apparent mass yields early in the chamber experiments. We further find that a second VBS model featuring substantial oligomerization of semivolatile monomers can match the CLOUD growth rates with substantially lower SOA mass yields; this is because the lighter monomers have a higher velocity and thus a higher condensation rate for a given mass concentration. The oligomerization simulations are a closer match to the CMU experiments than the base-case simulations, though they overpredict the observations somewhat. However, we also find that if the chemical conditions in CLOUD and the CMU chamber were identical, substantial nucleation would have occurred in the CMU experiments when in fact none occurred. This suggests that the chemical mechanisms differed in the two experiments, perhaps because the high oxidation rates in the SOA formation experiments led to rapid termination of peroxy radical chemistry.