Chemical Composition, Seasonal Variation And Sources Of Water-soluble Organic Components In PM_2.5 In Shanghai

With the rapid urbanization and economic development, air pollution in Chinais gradually changing from coal-smoke pollution into a complex pollution of bothcoal and petroleum combustion. Organic aerosols account for more than30%of thetotal mass of fine particles, among which a large part （30%to60%） is water soluble.As water-soluble organic matter is hygroscopic, it would have very important impacton air quality, climate and human health.Water-soluble organic compounds in fineparticleshas many sources, the two most important sources are secondary organicaerosol （SOA） and biomass burning. With the frequent occurrence of urban haze inrecent years in China, water-soluble organic compounds and SOA in the fineparticlesare drawing more and more concern. Analyzing the composition andsources of the water-soluble organic compounds could provide scientific basis forthe study of formation mechanism of air pollution and urban haze, and for themaking of pollution-control policies.The concentration and seasonal variations of water-soluble organic carbon（WSOC）in PM_2.5were studiedby analyzing theintegrated24h samples collected at asuburban site （Baoshan, BS） and an urban site （Xujiahui, XJH） in Shanghai duringJanuary2010and January2011. The concentrations and compositions of organictracers for the main biogenic and anthropogenic precursors of SOA, includingisoprene, α-pinene,β-caryophyllene and toluene were measured. Sourceapportionment of WSOC was conducted with positive matrix factorization （PMF）method. Contributionof secondary organic carbon （SOC） was also estimated withthe OC/EC ratio method, the WSOC-basedmethod, and the SOAtracer-based method.The SOC concentrations from the above mentioned methods were compared witheach other. The following conclusions were reached:1) The annual average concentrations of PM_2.5in Shanghai were62.5μg/m3（BS）and60.1μg/m3（XJH）, organic matter （OM） accounted for32%and31%of thetotal PM_2.5mass, respectively.The average OC/EC ratio was4.2and4.4at BS and XJH, respectively, showing that secondary pollutantshad considerablecontribution to the fine particulate matter in Shanghai.2) The WSOC average concentrations were4.9μg/m3and5.0μg/m3, respectivelyin BS and XJH.Due to the seasonal change in meteorological conditions, theWSOC concentrations had aseasonal trend of autumn>winter>spring>summer.WSOC occupied42%and45%of the organic carbon （OC） at BS and XJH,slightly higherthan other cities in China such as Beijing and Guangzhou. Inaddition, WSOC concentration at the BS and XJH sampling sites showed avery strong correlation （R2=0.93）, suggesting that the WSOC in atmosphericaerosols in Shanghai had regional characteristics.3) The average concentrations of the isoprene SOA tracers atBS and XJH were13.3ng/m3and10.3ng/m3, while the concentrations of α-pinene tracers were4.7and3.5ng/m3respectively. All the above tracers had highest concentrationsin summer, and lower concentrations in other threeseasons.β-caryophylleneoxidation product was not detected in our wintersamples, and most of the sampleshad concentrations of less than2ng/m3. Nosignificant seasonal trend was found in the concentration of β-caryophyllenetracer. The annual average concentration of toluene oxidation products was1.2ng/m3, with higher concentrations in summer and autumn, followed byspring and winter.4) Correlation analysis between the SOA tracers inPM_2.5in Shanghai showedthatsignificant mutual correlations were found between the isoprene SOAtracers （C5-alkenetriols and2-methyltetrols） and the α-pinene tracers （HDMGAand MBTCA）, suggesting that they were formed in similar way.Thecorrelations between2-methylglyceric acid and other isoprene tracers wereweaker. But2-methylglyceric acid,3-hydroxyglutaric acid, β-caryophyllinicacid and2,3-dihydroxy-4-oxopentanoic acidwere significantly correlated witheach other. Concentrations of isoprene tracers at BS were significantlycorrelated with that at XJH in all the sampling periods （R>0.7）, for the tracersof other three precursors, strong correlations between the two sites were also found in the summer, but the correlations were weak in other sampling periods,suggesting the impact of local emissions.5) Saccharide compounds in the PM_2.5samples in Shanghai were mainlycomprised of anhydro sugars （levoglucosan, mannosan, galactosan）,monosaccharides （fructose and glucose） and sugar alcohols （arabitol andmannitol）, where levoglucosan accounted for about60%of the total saccharideconcentration. Significant correlation was found between levoglucosan andWSOC, with R2of0.82（BS） and0.91（XJH） respectively, indicating thatbiomass burningwas a very important contributorto WSOC in PM_2.5inShanghai.6) Positive matrix factorization （PMF） method was applied to resolve the mainsources of WSOC in PM_2.5in Shanghai.Six factors （sources） were derived, andthey were:resuspended soil, SOA from isoprene, SOA from terpenes andtoluene, SOA associated with pinonic acid, biomass burning and SOAassociated with secondary sulfate and nitrate. Their contributions to WSOCwere3%,1%,13%,10%,46%, and27%, respectively. Comparison of the SOCcontribution estimated from different methods showed thattheSOCOC/EC-basedmethod would overestimate the contribution of SOC inautumn due to the existence of biomass burning, and the SOCtracer-basedmethodmight underestimate the contribution of SOC due to the uncertainties in thechamber test and in the chemical analysis of organic tracers. Based on theSOCPMFresults, the SOCWSOC-basedmethod couldgive reasonableestimation ofSOC contribution.