Concentration, source identification, and exposure risk assessment of PM2.5-bound parent PAHs and nitro-PAHs in atmosphere from typical Chinese cities

Sixteen parent PAHs and twelve nitro-PAHs were measured in PM2.5 samples collected over one year (2013–2014) at nine urban sites in China. During the sampling period, concentrations of individual nitro-PAHs were one or two orders of magnitude lower than their parent PAHs. Typical seasonal variations in parent PAH concentrations, which increased 10- to 80- fold in winter compared to summer, were observed in this study. Conversely, the mean atmospheric concentrations of nitro-PAHs were similar in all four seasons, with the exception of 9-nitroanthracene (9n-Ant). Compared to other nitro-PAHs which were secondary formation products, 9n-Ant had a higher concentration and made up a larger proportion of total nitro-PAHs. Positive matrix factorization results indicated that 9n-Ant sources included biomass burning (20%), vehicle exhaust emissions (43%), and secondary formation (30%). Overall, the elevated concentrations of parent PAHs observed in winter correlated with the contribution from coal combustion at all sites, especially in north China (>80%). The contribution of secondary formation products to total nitro-PAHs was measured during the summer, and was especially high in the larger cities such as Shanghai (84%), Beijing (76%), Guangzhou (60%), and Chengdu (64%), largely due to the summer concentrations of parent PAHs were markedly lower than in winter.

P23 Figure S3. 5-factor loadings by PMF analysis from parent PAHs and nitro-PAHs data of PM2.5 sample at nine urban sites across China.
P24 Figure S4. 4-factor loadings by PMF analysis from parent PAHs and nitro-PAHs data of PM2.5 sample at nine urban sites across China.
P26 Figure S5. 6-factor loadings by PMF analysis from parent PAHs and nitro-PAHs data of PM2.5 sample at nine urban sites across China.

S2.4 PMF receptor model
Detailed concepts and applications of PMF model for source apportionment were described in EPA PMF 5.0 Fundamentals and User Guide (http://www.epa.gov/heasd/research/pmf.html). In principle, the PMF model is based on the following equations: where Xij is the concentration of the jth congener in the ith sample of the original data sets; Aik is the contribution of the kth factor to the ith sample; Fkj is the fraction of the kth factor arising from congener j; Rij is the residual between the measured Xij and the predicted Xij using p principal components.
where Sij is the uncertainty of the jth congener in the ith sample of the original data sets containing m congeners and n samples. Q is the weighted sum of squares of differences between the PMF output and the original data sets. One of the objectives of PMF analysis is to minimize the Q value. Before the PMF analysis, both the concentration file and uncertainty file were inserted into the model. In this study, uncertainties of 20% for parent PAHs and 15% for nitro-PAHs were adopted based on the results from QA/QC, respectively.
During the PMF analysis, the model was run for 3-7 factors and was always with random seeds.
Finally, the five-factor solution was adopted for further discussion in this study. The 5-factor results adopted in this study were based on: 1) The correlation between the model estimated concentrations and the measured concentrations was highest for 5-factor result (See Figure S2).