Source apportionment of primary and secondary organic aerosols using positive matrix factorization (PMF) of molecular markers

Monthly average ambient concentrations of more than eighty particle-phase organic compounds, as well as total organic carbon (OC) and elemental carbon (EC), were measured from March 2004 through February 2005 in five cities in the Midwestern United States. A multi-variant source apportionment receptor model, positive matrix factorization (PMF), was applied to explore the average source contributions to the five sampling sites using molecular markers for primary and secondary organic aerosols (POA, SOA). Using the molecular makers in the model, POA and SOA were estimated for each month at each site. Three POA factors were derived, which were dominated by primary molecular markers such as EC, hopanes, steranes, and polycyclic aromatic hydrocarbons (PAHs), and which represented the following POA sources: urban primary sources, mobile sources, and other combustion sources. The three POA sources accounted for 57% of total average ambient OC. Three factors, characterized by the presence of reaction products of isoprene, alpha-pinene and beta-caryophyllene, and displaying distinct seasonal trends, were consistent with the characteristics of SOA. The SOA factors made up 43% of the total average measured OC. The PMF-derived results are in good agreement with estimated SOA concentrations obtained from SOA to tracer yield estimates obtained from smog chamber experiments. A linear regression comparing the smog chamber yield estimates and the PMF SOA contributions had a regression slope of 1.01 +/- 0.07 and an intercept of 0.19 +/- 0.10 mu g OC m(-3) (adjusted R-2 of 0.763, n = 58). (C) 2009 Elsevier Ltd. All rights reserved.