Date of Award

2019

Document Type

Open Access Dissertation

Degree Name

Doctor of Philosophy in Chemistry (PhD)

Administrative Home Department

Department of Chemistry

Advisor 1

Lynn R. Mazzoleni

Committee Member 1

Sarah A. Green

Committee Member 2

Kathryn A. Perrine

Committee Member 3

Evan S. Kane

Committee Member 4

David C. Podgorski

Abstract

Organic aerosol affects human health and climate. These effects are largely determined by the composition of the organic aerosol, which is a complex mixture of species. Understanding the complexity of organic aerosol is critical to determining its effect on human health and climate. In this study, long range transported organic aerosol collected at the Pico Mountain Observatory was analyzed using ultrahigh resolution mass spectrometry. Organic aerosol transported in the free troposphere had an overall lower extent of oxidation than aerosol transported in the boundary layer. It was hypothesized that the lower oxidation was related to a more viscous phase state of the aerosol during transport. The results suggest that biomass burning organic aerosol injected into the free troposphere are more persistent than organic aerosol in the boundary layer. A sample was also analyzed using tandem FT-ICR MS/MS fragmentation, providing information about the functional group composition in the aerosol sample. This was done using a segmented scan approach, which revealed an unprecedented molecular complexity of unfragmented precursor ions. In addition to the expected CO2 and H2O neutral losses, neutral losses corresponding to carbonyl functional groups (C2H4O, CO) were observed. The abundance of carbonyl functional groups suggests a slower rate of aging in the atmosphere. Analysis of nitrogen and sulfur containing neutral losses highlighted a surprising abundance of reduced nitrogen and sulfur loss (NH3 and SH2). This further supports the hypothesis of slower aging in the free troposphere. Additional research was done to develop an R software package (MFAssignR) to perform molecular formula assignment with improved decision-making transparency, noise estimation, isotope identification, and mass recalibration. MFAssignR was found to assign the same molecular formula as other molecular formula assignment methods for the majority (97-99%) of mass peaks that were assigned a molecular formula by the compared methods. Additionally, MFAssignR was more effective at assigning molecular formulas to low intensity peaks relative to the other methods tested, leading to more overall molecular formula assignments. MFAssignR is available via GitHub and is the first open source package to contain a full pipeline of functions for data preparation and analysis for ultrahigh resolution mass spectrometry.

Download

Share

COinS