Stable carbon isotopes and levoglucosan for PM2.5 elemental carbon source apportionments in the largest city of Northwest China

Stable carbon isotopes provide information on aerosol sources, but no extensive long-term studies of these isotopes have been conducted in China, and they have mainly been used for qualitative rather than quantitative purposes. Here, 24 h PM2.5 samples (n = 58) were collected from July 2008 to June 2009 at Xi'an, China. The concentrations of organic and elemental carbon (OC and EC), water-soluble OC, and the stable carbon isotope abundances of OC and EC were determined. In spring, summer, autumn and winter, the mean stable carbon isotope in OC (δ13COC) were −26.4 ± 0.6, −25.8 ± 0.7, −25.0 ± 0.6 and −24.4 ± 0.8‰, respectively, and the corresponding δ13CEC values were −25.5 ± 0.4, −25.5 ± 0.8, −25.2 ± 0.7 and −23.7 ± 0.6‰. Large δ13CEC and δ13COC values in winter can be linked to the burning coal for residential heating. Less biomass is burned during spring and summer than winter or fall (manifested in the levels of levoglucosan, i.e., 178, 85, 370, 935 ng m−3 in spring, summer, autumn, and winter), and the more negative δ13COC in the warmer months can be explained by the formation of secondary organic aerosols. A levoglucosan tracer method combined with an isotope mass balance analysis indicated that biomass burning accounted for 1.6-29.0% of the EC, and the mean value in winter (14.9 ± 7.5%) was 7 times higher than summer (2.1 ± 0.4%), with intermediate values of 6.1 ± 5.6 and 4.5 ± 2.4% in autumn and spring. Coal combustion accounted for 45.9 ± 23.1% of the EC overall, and the percentages were 63.0, 37.2, 36.7, and 33.7% in winter, autumn, summer and spring respectively. Motor vehicles accounted for 46.6 ± 26.5% of the annual EC, and these contributed over half (56.7-61.8%) of the EC in all seasons except winter. Correlations between motor vehicle-EC and coal combustion-EC with established source indicators (B(ghi)P and As) support the source apportionment results. This paper describes a simple and accurate method for apportioning the sources of EC, and the results may be beneficial for developing model simulations as well as controlling strategies in future.