Molecular Composition of Monoterpene Secondary Organic Aerosol at Low Mass Loading

Yuqian Gao (1) and MURRAY JOHNSTON (1)

(1) University of Delaware

     Abstract Number: 467
     Last modified: May 12, 2010

Abstract
The molecular composition of secondary organic aerosol (SOA) from the ozonolysis of monoterpenes (a-pinene and b-pinene) was studied by nano electrospray ionization (nanoESI) and high resolution Fourier transform ion cyclotron resonance (FTICR) mass spectrometry. SOA particles were generated in a flow tube reactor with a reaction time of 23 sec. A microsampling assembly in combination with nanoESI-FTICR analysis permitted SOA with a mass loading as low as 3.5 ug/m3 to be characterized with high accuracy and precision mass analysis. The molecular composition of monoterpene SOA is complex, even at low mass loading. Hundreds of product molecular formulas are identified. Most of the peaks in the dimer region are direct combinations of detected peaks in the monomer region. Similar molecular formulas are found at all mass loadings; however the relative intensities change significantly. In particular, a species with the (neutral molecule) formula C17H26O8 increases substantially in intensity relative to other products as the mass loading decreases. Tandem mass spectrometry (MSMS) data suggests that this species is a dimer of C9H14O4 and C8H12O4. Based on molecular formula and additional MSMS data, these monomers are consistent with pinic acid and terpenylic acid, respectively. This and other data to be presented suggest that particle nucleation occurs through formation of dimers and/or higher order oligomers. Finally, van Krevelen plots constructed from H:C vs. O:C ratios for individual molecular formulas are compared to bulk H:C and O:C ratios obtained by aerosol mass spectrometry at similar mass loadings. The limiting values of the bulk ratios as the mass loading approaches 1 ug/m3 do not match specific molecular compounds identified by nanoESI-FTICR analysis. Instead, the bulk ratios appear to be determined by the relative yields of families of compounds within three composition domains identified in our previous work (Heaton et al., ES&T 2009).