Extractable organic carbon and its differentiation by polarity in diesel exhaust, wood smoke, and urban particulate matter

Abstract

In order to determine whether characterization of organic carbon (OC) in particulate matter (PM) is limited due to the extraction or analysis, we have evaluated the OC extractability from three model PM matrices (wood smoke, diesel exhaust, and urban PM). We have compared hot pressurized water (HPW) fractionation and sequential organic solvent Soxhlet extraction. The evaluation of extraction efficiency was based on three methods: thermal optical transmittance (TOT), reflectance, and total organic carbon analyses. For all three PM samples, comparable OC yields were obtained with both extraction methods accounting for 45-60% of the total OC found in PM (based on TOT). These recoveries also suggest that a significant portion of OC in all PM materials remains unextractable. Further investigation of OC distribution in different polarity fractions using spectrometric techniques (including gas chromatography with mass spectrometry, carbon 1s near-edge X-ray absorption fine structure, and proton nuclear magnetic resonance spectroscopies) revealed that sequential extraction may be useful approach to differentiate the constituents attributed to primary emission sources and to secondary aerosol formation, respectively. In contrast to the common expectation that polarity of constituents corresponds to the polarity of the extracting solvent, distribution of OC species (among the different polarity fractions) also depended on their availability (i.e., analyte-matrix interactions). It appears that secondary OC species (e.g., diacids) are not deeply embedded into the matrix, and thus were recovered in the solvent of a corresponding polarity. By contrast, primary OC constituents are strongly bound to the matrix and thus required harsh conditions to be recovered.