Molecular view modeling of atmospheric organic particulate matter: Incorporating molecular structure and co-condensation of water

Abstract

Abstract Most urban and regional models used to predict levels of organic particulate matter (OPM) are based on fundamental equations for gas/particle partitioning, but make the highly simplifying, anonymized-view (AV) assumptions that OPM levels are not affected by either: a) the molecular characteristics of the condensing organic compounds (other than simple volatility); or b) co-condensation of water as driven by non-zero relative humidity (RH) values. The simplifying assumptions have allowed parameterized chamber results for formation of secondary organic aerosol (SOA) (e.g., “two-product” (2p) coefficients) to be incorporated in chemical transport models. However, a return towards a less simplistic (and more computationally demanding) molecular view (MV) is needed that acknowledges that atmospheric OPM is a mixture of organic compounds with differing polarities, water, and in some cases dissolved salts. The higher computational cost of MV modeling results from a need for iterative calculations of the composition-dependent gas/particle partition coefficient values. MV modeling of OPM that considered water uptake (but not dissolved salts) was carried out for the southeast United States for the period August 29 through September 7, 2006. Three model variants were used at three universities: CMAQ-RH-2p (at PSU), UCD/CIT-RH-2p (at UCD), and CMAQ-RH-MCM (at TAMU). With the first two, MV structural characteristics (carbon number and numbers of functional groups) were assigned to each of the 2p products used in CMAQv.4.7.1 such that resulting predicted Kp,i values matched those in CMAQv.4.7.1. When water uptake was allowed, most runs assumed that uptake occurred only into the SOA portion, and imposed immiscibility of SOA with primary organic aerosol (POA). (POA is often viewed as rather non-polar, while SOA is commonly viewed as moderately-to-rather polar. Some runs with UCD/CIT-RH-2p were used to investigate the effects of POA/SOA miscibility.) CMAQ-RH-MCM used MCM to generate oxidation products, and assumed miscibility of SOA and POA. In a ∼500 km wide band from Louisiana through to at least North Carolina, CMAQ-RH-2p and UCD/CIT-RH-2p predicted that water uptake can increase SOA levels by as much as 50–100% (from a range of ∼1–2 μg m−3 to a range of ∼1–4 μg m−3). CMAQ-RH-MCM predicted much lower effects of water uptake on SOA levels (