Effect of Hydrophilic Organic Seed Aerosols on the Formation of Secondary Organic Aerosol from α-Pinene Ozonolysis in Dry and Humid Conditions

Our previous study (Song et al., 2007) showed that hydrophobic primary organic aerosols (POA) could not enhance secondary organic aerosol (SOA) mass yield from α-pinene ozonolysis, indicating that SOA species may not be well mixed with hydrophobic POA species. However, aerosol composition and properties of urban hydrophobic POA will change in the atmosphere. POA can be oxidized via heterogeneous uptake of oxidants such as hydroxyl radical (OH), ozone and nitrate radical (NO3). Some aerosol-phase organic monomers may undertake condensed-phase chemistry to form oligomers. As a result, urban POAs can gradually become hydrophilic over time. On the other hand, some atmospheric organic aerosols are already somewhat hydrophilic. For example, POA formed from biomass burning are already somewhat hydrophilic. These hydrophilic organic aerosols may act contrastingly to those hydrophobic POAs and could enhance the absorption of SOA species into the aerosol-phase by providing additional absorptive mass. In this study, we investigate the SOA formation from α-pinene ozonolysis in the presence of various surrogates of oxidized/hydrophilic POAs such as fulvic acid, levoglucosan and glycerol. Fulvic acid serves as surrogates for aged oligomeric or polymeric substances in ambient organic aerosols; levoglucosan is used to resemble hydrophilic POAs from biomass burning; glycerol serves as a model hydrophilic organic aerosol. Each seed aerosol will be tested in both dry or humid conditions. Humidity will also be varied to exame the extent of the effect of aerosol-phase water on SOA yield. Organic species in the gas phase are measured with a Proton Transfer Reaction-Mass Spectrometer (PTR-MS) while the growth of aerosols and their composition are analyzed using a Scanning Mobility Particle Sizer (SMPS) and an Aerosol Mass Spectrometer (AMS), respectively. The data will be interpreted with the gas-particle partitioning theory, and the implications of our results on modeling SOA formation will be discussed.