Abstract View
A Coupled Volatility and Molecular Composition Based Source Apportionment of Atmospheric Organic Aerosol
PHILIP RUND, Ben H. Lee, Claudia Mohr, Daniel Jaffe, Noah Bernays, Qi Zhang, Ryan Farley, Tuukka Petäjä, Joel A. Thornton, University of Washington
Abstract Number: 462
Working Group: Source Apportionment
Abstract
Improvements in organic aerosol (OA) source apportionment are needed to improve our understanding of the processes and precursors that control its atmospheric abundance. We apply Non-Negative Matrix Factorisation (NNMF) to molecular-level particle composition and volatility measurements obtained using a Time-of-Flight Chemical Ionization Mass Spectrometer (ToF-CIMS) coupled to a custom Filter Inlet for Gases and AEROsols (FIGAERO) at a mixed temperature forest site during summer, a boreal forest site during spring, and a remote mountain top observatory during late summer. We conduct NNMF of resolved thermogram timeseries containing >350 molecular components, utilizing volatility and abundance information. We also conduct NNMF using only the molecular-level abundance information. Typically, 5 to 10 factors explain a significant fraction of the spectral variance at each site. Factors derived from the resolved thermograms show distinct volatility, molecular composition, and temporal variations that illustrate different biogenic and anthropogenic OA precursors and aging timescales. The median mass contributions of NNMF factors produced for the temperate forest site, without a priori information imposed, align well with categories determined by an independent study of the same data using a spectral basis set produced from multiple laboratory chamber experiments. A weakness of the NNMF approach is the lower dynamic range of separation, as particulate organic nitrates, which were observed to have distinct patterns of variability but low overall abundance at some sites, were not separable. The routine is shown to be more robust using resolved thermograms as input rather than concentration timeseries, likely due to the additional component variability in volatility space. The added layer of volatility information and molecular-level identification of OA composition provided by the FIGAERO-CIMS shows potential with the NNMF algorithm to reproduce atmospherically relevant sources from observations as well as providing a framework to further identify chemical processes that lead to volatility lowering.