(673a) Gas-Particle Partitioning of Primary Organic Aerosol Emissions From Gasoline and Diesel Vehicles

Gas-particle partitioning (i.e., phase equilibrium between the vapor and condensed phases) of primary organic aerosol (POA) emissions from motor vehicles has important implications toward understanding atmospheric organic aerosol concentrations. Source testing from a fleet of 45 in-use light-duty gasoline (LDGV) and two in-use light-duty diesel vehicles (LDDV) was performed at the California Air Resources Board Haagen-Smit Laboratory. Three in-use heavy-duty diesel vehicles (HDDV) were tested at the California Air Resources Board Heavy-Duty Engine Testing Laboratory. Four different methods were used to characterize the semi-volatile nature of the POA emissions: 1) artifact-corrected POA emission factors from a constant volume sampler (CVS); 2) isothermal dilution from the CVS into an environmental chamber; 3) evaporation induced by heating in a thermodenuder; and 4) inference of volatility distributions from thermal desorption/gas chromatography/mass spectrometry (TD-GC-MS) analysis. All four methods indicate that the POA emissions are semi-volatile. Measurements from the CVS showed that 30-80 % of POA emissions collected on bare-quartz filters reside in the particle phase; the remainder is semi-volatile vapors. Dilution into the chamber reduced the fraction of organic material in the particle phase by a factor of two or more. Nearly all LDGV POA emissions evaporated when heated to 100 ^oC in a thermodenuder. However, only a portion (~40%) of the LDDV and HDDV POA emissions evaporated when heated to 100 ^oC; this may be caused by a different gas-particle partitioning mechanism, such as adsorption instead of absorption. The median volatility distribution for the light-duty gasoline and diesel vehicles determined from the TD-GC-MS method was used to model experimental data with partitioning theory with excellent performance, especially for dilution and thermodenuder data. It is recommended that future studies apply this volatility distribution in order to estimate POA concentrations atmospherically-relevant conditions.