An Improved Mechanism to Predict Secondary Organic Aerosols from Toluene for Air Quality Models

HARSHAL M. PARIKH (1), Annmarie G. Carlton (2), William Vizuete (1), Eric H. Chen (1), Richard M. Kamens (1)

(1) University of North Carolina - Chapel Hill, (2) U.S. Environmental Protection Agency

     Abstract Number: 162
     Last modified: April 27, 2010

Abstract
Predictions of Secondary Organic Aerosol (SOA) from precursors typically use an Odum-type approach in current air quality models (AQM). For toluene, AQMs employ two semi-volatile products for high-NOx pathway (RO2 + NO) and a single non-volatile product for low-NOx pathway (RO2 + HO2). Stoichiometric and partitioning coefficients for SOA products were determined using chamber studies under limited conditions of low relative humidity and high initial particle number concentration of (NH4)2SO4 seed. This SOA mechanism is evaluated against smog chamber data from 14 experiments conducted with different combinations of initial toluene, NOx and hydrocarbon mix injections, initial seed type and, humidity. The mean absolute error between observed and predicted final aerosol mass from all experiments is 16.15 ug/m3. SOA mass is severely overestimated under dry conditions for lower initial inorganic and aged aerosol mass in the presence of hydrocarbon mix. For humid conditions, SOA mass is underestimated earlier in the day. Based on these results, modifications that are feasible for future implementation in AQMs, were made to this mechanism, that include: 1) an additional pathway of SOA formation due to partitioning of polar product species into the aqueous-phase of the particle, 2) logarithmic dependence of SOA formation on available particle surface area from semi-volatile and non-volatile products partitioning into the organic-phase of the particle. The modified mechanism exhibited lower value of mean absolute error (3.48 ug/m3) and its overall performance for both wet and dry conditions was significantly improved over that of initial mechanism.