Calculating Equilibrium Phase Distribution during the Formation of Secondary Organic Aerosol Using COSMOtherm

Abstract

Challenges in the parametrization of compound distribution between the gas and particle phase contribute significantly to the uncertainty in the prediction of secondary organic aerosol (SOA) formation and are rooted in the complexity and variability of atmospheric condensed matter, which includes water, salts, and a multitude of organic oxidation products, often in two separated phases. Here, we explore the use of the commercial quantum-chemistry-based software COSMOtherm to predict equilibrium partitioning and Setchenow coefficients of a suite of oxidation products of α-pinene ozonolysis in an aerosol that is assumed to separate into an organic-enriched phase and an electrolyte-enriched aqueous phase. The predicted coefficients are used to estimate the phase distribution of the organic compounds, water and ammonium sulfate, the resulting phase composition, and the SOA yield. Four scenarios that differ in terms of organic loading, liquid water content, and chemical aging are compared. The organic compounds partition preferentially to the organic phase rather than the aqueous phase for the studied aerosol scenarios, partially due to the salting-out effect. Extremely low volatile organic compounds are predicted to be the dominant species in the organic aerosols at low loadings and an important component at higher loadings. The highest concentration of oxidation products in the condensed phase is predicted for a scenario assuming the presence of non-phase-separated cloud droplets. Partitioning into an organic aerosol phase composed of the oxidation products is predicted to be similar to partitioning into a phase composed of a single organic surrogate molecule, suggesting that the calculation procedure can be simplified without major loss of accuracy. COSMOtherm is shown to produce results that are comparable to those obtained using group contribution methods. COSMOtherm is likely to have a much larger application domain than those group contribution methods because it is based on fundamental principles with little calibration (Graph Presented).