Characterising the Seasonal Sources of Urban Organic Aerosol Using the Non-Targeted High-Resolution Mass Spectrometry and Factor Analysis

Sri Hapsari Budisulistiorini, DANIEL J. BRYANT, Jonathan Taylor, Alfred Mayhew, David Topping, Jacqueline Hamilton, University of York

     Abstract Number: 39
     Working Group: Urban Aerosols

Abstract
Organic aerosol (OA) comprises a significant proportion of urban PM2.5. Using positive matrix factorisation (PMF) of bulk measurements of organic aerosol composition by aerosol mass spectrometry, the oxidised organic aerosol (OOA) is classified based on its volatility and oxidation state. This classification is useful but does not directly reveal the sources of OOA. Recently, molecular characterisation studies using high-resolution mass spectrometry (HRMS) techniques provided a deeper insight into the organic aerosol composition. The benefit of abundant organic molecules identified by the HRMS techniques comes with a drawback of difficulties in identifying their sources in the atmosphere.

This study aims to develop a new analytical approach to characterise the seasonal sources of OA. We applied PMF on HRMS data to classify organic aerosol sources in Beijing, China. PM2.5 samples were collected in Beijing, China, in winter 2016 and summer 2017. The organic species were identified using ultra-performance liquid chromatography coupled to a Q-Exactive Ultra-High Mass Range Hybrid Quadrupole-Orbitrap Mass Spectrometer with electrospray ionisation (UHPLC-ESI-Orbitrap MS). Datasets containing organic molecules identified over the observation periods were then analysed using the PMF, widely used to characterise the sources of aerosol measured by the online aerosol mass spectrometer.

The PMF analysis conducted in robust mode resulted in five and seven-factor solutions for winter and summer. The summer factor solution analysis revealed two biogenic secondary organic aerosol (SOA), two anthropogenic SOA, and three biomass burning OA sources. The contributions are ca. 22%, 15%, and 63% for biogenic, anthropogenic, and biomass burning sources, respectively. The large contribution of biomass burning is due to the long-range transport of biomass burning plumes from the western region.

This study demonstrates the capability of combining the non-targeted HRMS technique with factor analysis to characterise the sources of organic aerosol. Future studies will develop deep learning methods for source apportionment.