Intermediate-Volatility Organic Compounds: Source Emission Profiles, Resolved Chemical Composition and Contribution to Secondary Organic Aerosol Formation

Atmospheric aerosols are tiny particles or droplets suspended in the air. Aerosols with diameters less than 2.5 microns are fine particulate matter, or PM_2.5. Human exposure to fine particulate matter has been associated with higher mortality, stroke, lung cancer and cardiovascular diseases. Organic aerosol (OA) contributes significantly to fine particulate matter mass globally. Intermediate-volatility organic compounds (IVOCs) are a group of organic compounds that has effective saturation concentration (C*) at 298K between 10³ and 10⁶ μg/m³ (roughly equivalent to C₁₂ to C₂₂ n-alkanes). They are important precursors to secondary organic aerosol (SOA) formation.

To investigate the contribution of IVOCs to ambient SOA formation, it is crucial to correctly model them with emission, volatility distribution and chemical composition from direct measurements. Traditional emission inventories such as the National Emission Inventory only account for gas-phase organic emissions of volatile organic compounds (VOCs). There are large gaps in IVOCs and semi-volatile organic compounds (SVOCs) between traditional emission profiles and recent experiment results. Also, despite recent efforts to include IVOCs and SVOCs in inventories, most of the model assumptions are overly simplistic and scaled based on limited experimental data. For example, if they are included, IVOCs are often represented using a single surrogate species. No emission profiles or simulations have been developed to incorporate the most recent experimental data that measured comprehensive organic species in a wide range of volatility spectrum.

This thesis consists of three main parts: compiling comprehensive organic emission profiles; implementing SOA parameterization in a chemical transport model (CTM) and run simulations; and developing a positive matrix factorization (PMF) technique to resolve the IVOC chemical composition from traditional gas chromatography-mass spectrometry (GC-MS) data.

Organic emission data on gasoline and diesel vehicles and aircraft emissions from previously published papers are compiled to create comprehensive emission profiles for use in CTMs. The organic emissions from all three source categories show tri-modal volatility distributions, and the volatility distributions are consistent across sources using the same fuel type. The traditional profiles dramatically underestimate IVOCs and SVOCs, which are important classes of SOA precursors. Accounting for IVOCs in gasoline exhaust almost doubles the predicted SOA production compared to the traditional profile. For gas-turbine and diesel sources, IVOCs and SVOC vapors combining contribute factors of 13 (gas-turbine) and 44 (diesel) more SOA than VOCs alone.

We developed a new parameterization to model the SOA formation from mobile source IVOC emissions designed for implementation in CTMs. The parameterization has six lumped IVOC species: two aromatics and four aliphatics, to account for the volatility and chemical composition of the IVOC emissions. Simulation results show that mobile sources contribute 2.7 μg m^-3 of IVOCs at Pasadena site, which is 43% of measured concentrations of hydrocarbon IVOCs. They also contribute ~1 μg m^-3 in daily peak SOA concentration, a 67% increase compared to the base case without IVOC emissions. Therefore, it is crucial to include mobile-source IVOC emissions in simulations. Results from the exploratory model runs suggest that additional 12% to 26.8% of non-mobile organic emissions are likely IVOCs.

We also developed a PMF-based technique to resolve the chemical composition of IVOCs in traditional GC-MS data. Evaluation on multiple datasets shows this technique can recover much of the chemical information compared to more sophisticated instruments. Source apportionment analysis on tunnel samples shows major contributions from diesel-

source IVOCs. PMF analysis on ambient samples in Pasadena, California shows more than 70% IVOCs are oxygenates indicating complex atmospheric oxidation. SOA modeling of gasoline vehicle emissions shows an 80% SOA yield increase under low-NOx conditions, highlights the need to include the IVOC aromatics in resolved chemical composition and CTM studies.

This thesis presents systematic efforts to better understand IVOCs and incorporate them into the model, from emission to chemical composition and SOA formation. To include IVOCs in CTM simulations, we compiled model-ready emission profiles for mobile-sources from direct measurements. To simulate the SOA formation from IVOCs and account for chemical composition, we developed a PMF technique to extract necessary information from traditional GC-MS data. And to evaluate the contribution of IVOCs to ambient OA, we implemented SOA parameterization and perform CTM simulations. This pipeline could be reused in the future study of other sources with IVOC emissions.