Simulation of Monoterpene SOA Formation by Multiphase Reactions Using Explicit Mechanisms

Abstract

The formation of terpene secondary organic aerosol (SOA) was simulated using the unified partitioning aerosol phase reaction model that predicted multiphase reactions of hydrocarbons in the presence of electrolytic inorganic aerosols. To predict oxygenated products from the atmospheric oxidation of terpenes, the master chemistry mechanism, an explicit gas kinetic mechanism, was implemented. The resulting products were then classified into 51 lumping groups using mass-based stoichiometric coefficients according to their volatility and aerosol phase reactivity. In the presence of wet inorganic aerosol, the SOA model was approached by liquid–liquid phase separation between the organic and inorganic phases due to the hydrophobicity of terpene products (oxygen to carbon ratios <0.6). The model streamlined three SOA formation pathways including partitioning of gaseous oxidized products onto both the organic aerosol and aqueous aerosol phases, oligomerization in the organic phase, and aqueous phase reactions (acid-catalyzed oligomerization and organosulfate formation). In the model, the peroxy radical autoxidation mechanism, which is a recently derived explicit mechanism to form highly oxygenated molecules, was also included to form less volatile products. The model simulation was demonstrated for SOA data that were produced through the photo-oxidation of three different monoterpenes (α-pinene, β-pinene, and d-limonene) under various experimental conditions in a large outdoor photochemical smog chamber. Terpene SOA growth was considerably accelerated in the aqueous phase anchored in acidic seeds but much weaker with neutral seeds. This tendency is quite different from that of isoprene SOA, which noticeably grows even in the neutral aqueous phase. Unlike hydrophilic isoprene products, terpene products are hydrophobic and weakly soluble in the aqueous phase, and thus, the neutral aqueous phase is insufficient to increase SOA mass. The model underestimated the production of polar functional groups, such as −OH, −COOH, and −ONO2, compared to the compositions measured using Fourier-transform infrared spectral data. In particular, the model underestimated carboxylic acids due to the knowledge gaps in the mechanisms to form carboxylic acid in both gas-phase oxidation and in-particle chemistry. Under the current emission trends in which SO2 and NOx have been decreasing, the model simulation suggested that the reduction of sulfate is more efficient to reduce SOA mass than the reduction of NOx.

Topics & Concepts