Ambient fine particulate matter (PM_2.5) has become a major concern of human society, as it is associated with numerous health risks and increased mortality rate. To evaluate the health effect of PM_2.5, oxidative potential (OP), the capability of PM_2.5 in catalyzing the consumption of antioxidants and cellular reductants or the generation of reactive oxygen species (ROS), has been proposed as a potential toxicity mechanism. The measurement protocols of OP have been gradually developed in the last two decades, while the association of OP with PM_2.5 mass and chemical composition remains highly unclear. In this dissertation, I targeted on advancing the understanding of this connection for the ambient PM_2.5 in the midwestern US.

The first objective of this dissertation is to find out the interactions among different redoxactive PM_2.5 chemical species in causing oxidative stress. We used DTT assay, the most widely used OP endpoint, as the OP indicator, and started with the standard solutions of known redoxactive substances, i.e., quinones (9,10-phenanthraquinone, 1,2-naphthoquinone, 1,4- naphthoquinone and 5-hydroxy-1,4-naphthoquinone) and metals [Cu (II), Mn (II) and Fe (II)]. Different interaction trends were observed for different metals:​ Fe had a strong synergistic interaction with all quinones in •OH generation, while Cu showed antagonistic interactions with quinones in both DTT consumption and •OH generation. Mn interacted synergistically with quinones in DTT oxidation, but antagonistically in •OH generation. We observed a similar pattern of the interaction between these metals (Fe, Mn, and Cu) and ambient humic-like substances (HULIS) compared to their interaction with quinones. The comparison of OP between metals + ambient HULIS and total water-soluble PM_2.5 showed that •OH generation in DTT involves a contribution (~50 %) from additional compounds besides known redox-active chemical species of PM_2.5 (i.e. Fe, Mn, Cu and HULIS), which indicated the existence of the interaction between these chemicals with other compounds in hydrophilic fractions of PM_2.5 extracts.

To better understand the mechanisms of OP and ease measurement protocols of OP, we combined five acellular OP endpoints, including consumption rate of dithiothreitol (OPDTT), ascorbic acid (OPAA-SLF) and glutathione (OPGSH-SLF), and the generation rate of •OH in DTT (OPOH-DTT) and in surrogate lung fluid (OPOH-SLF), and developed a semi-automated multi-endpoint ROS-activity analyzer (SAMERA). The system was evaluated for its performance using both ambient PM_2.5 extracts and positive controls. We observed a high analytical precision and a good agreement between automated system measurement and manual operation. SAMERA takes 3 hours to analyze one sample for all these OP endpoints, which is a substantial improvement over the manual analysis protocol.

SAMERA was deployed for analyzing the OP for five endpoints of both water-soluble and methanol-soluble PM_2.5 in the midwestern US. A large set of PM_2.5 samples (N = 241) were collected from five sites, setup in different environments, i.e., urban, rural, and roadside, in Illinois, Indiana, and Missouri from May 2018 to May 2019. PM_2.5 mass concentrations in the Midwest US as obtained from these samples were spatially homogeneously distributed, while most OP endpoints showed significant spatiotemporal heterogeneity. The correlations between OP and PM_2.5 mass were generally poor, further indicating the disparity between their trends. Moreover, the poor-to-moderate correlations among different OP endpoints demonstrated the rationale for analyzing multiple OP endpoints to achieve a better and comprehensive OP assessment.

In addition to OP analyses, we measured the common PM_2.5 chemical species, including carbonaceous species, inorganic ions, metals, and metalloids, on the same set of PM_2.5 samples from the midwestern US. A good mass balance was reached with excellent correlations between measured and reconstituted PM_2.5 mass at all sites. We observed some decent correlations between certain redox-active species (e.g., Cu, Fe, and WSOC) with various OP endpoints, which corroborated their sensitivity towards corresponding endpoints. Furthermore, we performed source apportionment analyses on OP and PM_2.5 mass using a positive matrix factorization (PMF) model, and observed a huge contrast in the major contributing sources between OP and mass of PM_2.5. Although secondary sources were the main contributors towards PM_2.5 mass, we found many anthropogenic sources, which also included some minor emission sources (< 5% of PM_2.5 mass), had substantial contributions towards various OP endpoints. Overall, our results indicate that the sources contributing substantially to PM_2.5 mass are not necessarily equally important in terms of their health effects and source apportionment analysis should include more health-relevant metric such as OP or toxicity in their designs for a more efficient air pollution management.