Date of Award

Summer 7-2015

Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Chemical Engineering


Spyros Pandis


The development of effective atmospheric particulate matter (PM) mitigation strategies relies on understanding source contributions to the corresponding PM levels. Most previous work has focused on PM mass concentrations. However, recent health studies suggest that particle number may also contribute to the negative health effects of particles, and particle number concentrations indirectly affect the energy balance of our planet. In the first part of this work, the aerosol number-focused CTM PMCAMx-UF is used to investigate the sources of particle number concentrations during a photochemically-active period in the Eastern U.S. A new aerosol number emissions inventory is developed for the July 2001 period. With the new emissions as input, PMCAMx-UF reproduces particle number concentrations within 12% of observations in Pittsburgh. Nucleation is predicted to be the dominant source of total number concentrations (>90%). Gasoline vehicles are predicted to contribute 36% to primary particle number concentrations, followed by industrial sources (31%), non-road diesel (18%), on-road diesel (10%), biomass burning (1%), and long-range transport (4%). The effects of reductions in diesel PM emissions are investigated in the second part of this work. A 50% reduction of diesel particulate emissions results as expected in lower (23%) black carbon mass concentrations and similar changes both in magnitude and spatial pattern to the absorption coefficient (27-30%). Contrary to what is expected, an average 1.6% increase of the total particle number concentrations is predicted due to a decrease in the coagulation and condensation. sinks. At the same time, a 1.6% decrease in N100 (particles larger than 100 nm) particle concentrations is predicted. Both changes in number concentrations are significantly different from those expected assuming a proportional change. These results suggest that mitigation of large diesel particles and/or particle mass emissions will reduce absorption climate-relevant properties related to black carbon and have health benefits; however the changes could also have the unintended effect of increased ultrafine particle number concentrations. The changes in cloud condensation nuclei (CCN) are predicted to be significantly less than expected assuming a proportional reduction during this photochemically active period. Next, the mass-focused CTM PMCAMx was used to explore the contribution of biomass burning organic aerosol (bbOA) to organic aerosol (OA) mass concentrations in the continental U.S. for three representative months of the modeling year 2008. PMCAMx uses the volatility basis set (VBS) to track the evolution of semivolatile OA. The simulation-averaged predicted bbOA contribution to OA concentrations is 8% for April, 22% for July, and 10% for September. Locally and on days of maximum biomass burning, bbOA sources can be significant both downwind (21-27% of OA) and near fire (56-90%). Compared to the CTM CAMx, which assumes non-volatile OA, PMCAMx predicts 20% more bbOA on average and estimates that most of the bbOA is secondary (oxidized vapors that condensed) whereas CAMx predicts all primary (emitted directly in the particle phase). Generally PMCAMx predicts less bbOA near fires due to the evaporation of primary OA and more bbOA downwind of fires due to the formation of secondary OA. Finally, a trajectory model is used to quantify the changes in OA concentrations in a biomass burning plume as it travels and dilutes. The model suggests that in the first few hours the evaporation of primary OA dominates over the production of secondary products and that the OA levels normalized to CO levels decrease despite the secondary production. Sensitivity studies are performed for parameters that affect OA concentrations (e.g. OH concentration, aging rate), and the results are analyzed to determine which scenarios lead to net loss, increase, or no significant changes of OA concentrations.