Source identification of water-soluble organic aerosols at a roadway site using a positive matrix factorization analysis

Abstract

Daily PM2.5 measurements were carried out at a local roadway every sixth day from May 2011 to August 2013 to obtain seasonal quantitative information on the primary and secondary sources of two water-soluble organic carbon (WSOC) fractions. Filter samples were analyzed for OC, elemental carbon (EC), WSOC, hydrophilic and hydrophobic WSOC fractions (WSOCHPI and WSOCHPO), and ionic species. An XAD solid phase extraction method and a total organic carbon analyzer were used to isolate the two WSOC fractions and determine their amounts, respectively.The WSOC/OC and WSOCHPI/WSOC ratios were 0.62±0.13 and 0.47±0.14, respectively. Similar seasonal profiles in EC, OC, and WSOC concentrations were observed, with higher concentrations occurring in the cold season and lower concentrations in the warm season. However, opposite results were obtained in WSOC/OC and WSOCHPI/WSOC ratios, with the higher in the warm season and the lower in the cold season. Correlation analyses indicated that two WSOC fractions in winter were likely attributed to secondary formation processes, biomass burning (BB), and traffic emissions, while WSOCHPI observed in other seasons were associated with secondary formation processes similar to those of oxalate and secondary inorganic species.A positive matrix factorization (PMF) model was employed to investigate the sources of two WSOC fractions. PMF indicated that concentrations of WSOC fractions were affected by five sources: secondary NO3− related, secondary SO42− and oxalate related, traffic emissions, BB emissions, and sea-salt. Throughout the study period, secondary organic aerosols were estimated to be the most dominant contributor of WSOC fractions, with higher contributions occurring in the warm seasons. The contribution of secondary aerosol formation processes (NO3− related+SO42− and oxalate related) to WSOCHPI and WSOCHPO was on an average 56.2% (45.0–73.8%) and 47.7% (39.6–52.1%), respectively. The seasonal average contribution of WSOCHPI and WSOCHPO attributed to BB was 19.0% (14.3–25.3%) and 14.8% (7.2–19.5%), respectively, with higher fractions occurring in the fall and winter. Traffic sources contributed to WSOCHPI and WSOCHPO from 4.2 to 21.0% (an average of 11.6%) and from 7.9 to 32.3% (an average of 19.9%), respectively, with higher fractions in the fall and winter compared with the other seasons. During the study period, for an episode associated with high local O3 level (~110ppbv) and high WSOCHPI/WSOC (0.80), secondary formation processes contributed 67.1% to WSOCHPI, and 72.6% to WSOCHPO, respectively. However, for an episode associated with local and severe regional haze pollutions, contributions of secondary formation processes to WSOC fractions were observed to be low (32.4–43.1%), while traffic and BB emissions contributed 16.8% and 24.3% to WSOCHPI, respectively, and 18.3% and 18.7% to WSOCHPO, respectively. The PMF results suggest that the contribution of traffic emissions to concentrations of two WSOC fractions cannot be neglected at the studied roadway site.