Primary to secondary organic aerosol: evolution of organic emissions from mobile combustion sources
A series of smog chamber experiments were conducted to investigate the transformation of primary organic aerosol (POA) and formation of secondary organic aerosol (SOA) during the photooxidation of dilute exhaust from a fleet of gasoline and diesel motor vehicles and two gas-turbine engines. In experiments where POA was present in the chamber at the onset of photooxidation, positive matrix factorization (PMF) was used to determine separate POA and SOA factors from aerosol mass spectrometer data. A 2-factor solution, with one POA factor and one SOA factor, was sufficient to describe the organic aerosol for gasoline vehicles, diesel vehicles, and one of the gas-turbine engines. Experiments with the second gas-turbine engine required a 3-factor PMF solution with a POA factor and two SOA factors. Results from the PMF analysis were compared to the residual method for determining SOA and POA mass concentrations. The residual method apportioned a larger fraction of the organic aerosol mass as POA because it assumes that all mass at m / z 57 is associated with POA. The POA mass spectrum for the gasoline and diesel vehicles exhibited high abundances of the C nH 2n+1 series of ions ( m / z 43, 57, etc.) and was similar to the mass spectra of the hydrocarbon-like organic aerosol factor determined from ambient data sets with one exception, a diesel vehicle equipped with a diesel oxidation catalyst. POA mass spectra for the gas-turbine engines are enriched in the C nH 2n−1 series of ions ( m / z 41, 55, etc.), consistent with the composition of the lubricating oil used in these engines. The SOA formed from the three sources exhibits high abundances of m / z 44 and 43, indicative of mild oxidation. The SOA mass spectra are consistent with less-oxidized ambient SV-OOA (semivolatile oxygenated organic aerosols) and fall within the triangular region of f44 versus f43 defined by ambient measurements. However there is poor absolute agreement between the experimentally derived SOA mass spectra and ambient OOA factors, though this poor agreement should be expected based on the variability of ambient OOA factors. Van Krevelen analysis of the POA and SOA factors for gasoline and diesel experiments reveal slopes of −0.50 and −0.40, respectively. This suggests that the oxidation chemistry in these experiments is a combination of carboxylic acid and alcohol/peroxide formation, consistent with ambient oxidation chemistry.