3.1 Temporal distribution of PM2.5 and PM2.5-PAHs
A total of 504 air samples were collected at the monitoring site in Wuxi, and Table 1 summarizes the annual concentrations of PM2.5 and 16 PAHs. The median PM2.5 concentration decreased annually from 2016 (64.3 µg/m3) to 2021 (34.0 µg/m3), accompanied by a decrease in ∑PAHs levels. According to the PRC National Standard, the average 24-hour ambient PM2.5 concentration limit is 75 µg/m3. The violation rates were 33.3% (2016), 19.0% (2017), 31.0% (2018), 15.5% (2019), 14.3% (2020), and 10.7% (2021). Statistics showed that the deaths resulted from outdoor air pollution are conservatively projected between 350,000 and 500,000 in China annually. The economic losses caused by relevant health cost has accounted for 1.16–3.8% of the gross domestic product (GDP) (Ma et al.,2022). A 1% increase of population density caused a 0.214% increase of daily PM2.5. Meanwhile, industrialization, energy consumption, and traffic pollution also significantly increased PM2.5 accumulation rate (Lou et al.,2016). The median concentration of PM2.5 in Wuxi was 64.3 µg/m3 in 2016. This was 6.4 times higher compared to the WHO air quality guideline stipulating that PM2.5 exposure limit is 10 µg/m3. Furthermore, 33.3% of the samples were higher than the PRC National Standard. Notably, a series of stringent emission control measures imposed in Wuxi are remarkable. An annual decrease in the concentration of PM2.5, with a median of 34.0 µg/m3 in 2021, reaching the annual mean concentration limit of 35 µg/m3. The PM2.5-bound PAHs also showed a decreasing trend from 2016–2021. With the improvements and amendments of Air Pollution Prevention and Control Law of the People's Republic of China in 2018, a series of related regulations were put into practice in Wuxi. Industrial restructuring and energy structure reformation were implementing over time. Total industrial coal consumption is decreasing year by year, and new energy vehicles achieve greater penetration. Especially, the substantial air quality optimization in 2020 and 2021 could attribute to the COVID-19 pandemic.
Table 1
Annual concentrations of 16 PAHs in PM2.5 at sampling site
| 2016 | 2017 | 2018 | 2019 | 2020 | 2021 |
PM2.5(µg/m3) | 64.3 (34.9–86.2) | 50.6 (33.2–70.4) | 57.9 (39.3–80.8) | 44.4 (31.7–62.8) | 38.9 (27.5–50.8) | 34.0 (21.3–54.7) |
Nap(ng/m3) | 0.13 (0.13–0.13) | 0.13 (0.13–0.13) | 0.13 (0.13–0.13) | 0.13 (0.13–0.13) | 0.13 (0.13–0.13) | 0.13 (0.13–0.13) |
Acy(ng/m3) | 0.08 (0.08–0.08) | 0.08 (0.08–0.08) | 0.08 (0.08–0.08) | 0.08 (0.08–0.08) | 0.08 (0.08–0.08) | 0.08 (0.08–0.08) |
Flu(ng/m3) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) |
Ace(ng/m3) | 0.23 (0.05–0.41) | 0.22 (0.17–0.35) | 0.17 (0.05–0.31) | 0.05 (0.05–0.13) | 0.05 (0.05–0.05) | 0.05 (0.05–0.05) |
Phe(ng/m3) | 0.06 (0.06–1.49) | 0.88 (0.17–1.28) | 1.00 (0.67–1.28) | 0.42 (0.06–1.10) | 0.06 (0.06–0.31) | 0.84 (0.44–1.51) |
Ant(ng/m3) | 0.05 (0.05–0.23) | 0.05 (0.05–0.39) | 0.05 (0.05–0.51) | 0.05 (0.05–0.05) | 0.09 (0.05–0.55) | 0.24 (0.07–0.43) |
Fla(ng/m3) | 0.05 (0.05–0.05) | 0.05 (0.05–0.05) | 0.05 (0.05–0.05) | 0.05 (0.05–0.05) | 0.05 (0.05–0.05) | 0.05 (0.05-05) |
Pyr(ng/m3) | 0.03 (0.03–1.59) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–1.02) |
Chr(ng/m3) | 0.04 (0.04–0.97) | 0.70 (0.04–1.42) | 0.49 (0.04–0.78) | 0.04 (0.04–0.09) | 0.04 (0.04–0.09) | 0.04 (0.04–0.04) |
Baa(ng/m3) | 0.03 (0.03–1.67) | 1.15 (0.03–2.04) | 0.46 (0.03–0.98) | 0.03 (0.03–0.36) | 0.03 (0.03–0.22) | 0.03 (0.03–0.03) |
Bbf(ng/m3) | 0.03 (0.03–0.03) | 0.03 (0.03–0.64) | 0.03 (0.03–1.37) | 1.11 (0.69–2.03) | 0.36 (0.18–1.59) | 0.28 (0.14–0.72) |
Bkf(ng/m3) | 3.35 (3.35–4.62) | 2.27 (0.89–2.99) | 2.21 (1.45–2.80) | 1.33 (0.82–3.07) | 0.13 (0.07–1.26) | 0.10 (0.06–0.26) |
BaP(ng/m3) | 0.03 (0.03–0.73) | 0.37 (0.03–1.69) | 0.27 (0.21–0.65) | 0.16 (0.09–0.30) | 0.15 (0.07–0.29) | 0.08 (0.03–0.29) |
Dah(ng/m3) | 0.03 (0.03–0.35) | 0.03 (0.03–0.73) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.03) | 0.03 (0.03–0.05) |
Bghi(ng/m3) | 1.27 (1.27–3.13) | 1.63 (0.13–3.03) | 1.23 (0.83–1.58) | 0.54 (0.31-1.00) | 0.36 (0.21–0.84) | 0.58 (0.30–1.45) |
Ind(ng/m3) | 0.04 (0.04–0.04) | 0.04 (0.04–2.34) | 0.04 (0.04–0.04) | 0.04 (0.04–0.04) | 0.04 (0.04–0.04) | 0.04 (0.04–0.04) |
∑PAHs(ng/m3) | 5.27 (5.27–14.4) | 8.39 (0.43–17.8) | 7.06 (5.90-12.43) | 5.05 (3.33–9.18) | 3.14 (1.70–5.26) | 4.22 (2.70–5.12) |
Moreover, the BaP concentration in 2017 was found to exceed the recommended health-based standard of the EU at 1 ng/m3, whereas the annual concentration was in line with China's national standards at 1 ng/m3 specified by the Ambient air quality standards GB3095-2012. BaP is a common PAHs which is used as a surrogate to estimate the toxicity of overall PAH levels. The carcinogenic and genotoxic potential of BaP metabolites was confirmed. Upon enzymatic conversion by cytochrome P450, metabolites of BaP can form adducts with guanine, resulting in disruption of normal DNA replication (Moorthy et al.,2015).
Among the 16 PAHs, Nap, Acy, Flu, Fla, Pyr, Dah, and Ind were seldom detected, whereas Phe, Baa, Bkf, BaP, and Bghi accounted for a large proportion of the PAHs composition. Bkf (five rings) had the highest concentration between 2016 and 2019. Bbf (five rings) has increased since 2019, as shown in Fig. 2. Moreover, Bghi (six rings) was found to be a steady component, accounting for 11–19% during the observation period. Due to semi-volatility, PAHs possessing 2–3 aromatic rings are predominant in the gaseous phase, while PAHs with four or more aromatic rings are dominant in the particulate phase (Wang et al.,2014). The dominant PAHs were 5–6 ring compounds, indicating a prominent contribution of vehicle exhaust. The 2–3 ring PAHs are generated by pyrolysis of unburned fossil fuels, and four-ring PAHs are features of coal/biomass combustion (Feng et., 2021). Recently, the composition of PM2.5-bound PAHs has demonstrated a trend of diversification, suggesting that pollution sources are also increasing in Wuxi.
Furthermore, the seasonal distribution variation of PM2.5 and 16 PAHs was notable, as shown in Fig. 3. PM2.5 concentrations in spring (63.3 ± 24.9 µg/m3) and winter (69.1 ± 34.9 µg/m3) were significantly higher than summer (43.3 ± 28.1 µg/m3) and fall (39.4 ± 24.2 µg/m3), and total PM2.5-bound 16 PAHs was higher in winter (15.4 ± 10.5 ng/m3) than other seasons (5.85 ± 3.95 ng/m3, 5.54 ± 4.05 ng/m3, 5.97 ± 3.90 ng/m3 for spring, summer, and fall, respectively). China has the highest PAH emissions globally and higher atmospheric PAH pollution is observed in winter than in other seasons (Shen et al.,2013) due to its coal-dominant energy structure. The high PAHs levels in winter were attributed to the massive coal combustion emissions for heating and meteorology parameter, such as temperature and humidity (Sun et al.,2022). The energy consumption structure, which has given priority to coal, will not markedly change in the future, and severe PAHs contamination in winter will continue to sour.
3.2 Source apportionment
To reduce PAH emissions, it is necessary to specify the primary emission sources. MDRs have been widely applied for PM2.5-bound PAHs source apportionment. Typically, the diagnostic ratios of Ant/(Ant + Phe), Fla/(Fla + Pyr), Baa/(Baa + Chr), Ind/(Ind + Bghi), and BaP/Bghi are used to confirm the contributions of various atmospheric PAHs sources. The details of the MDRs in the different years and seasons are listed in Table 2.
Table 2
Details of MDRs values in temporal PM2.5
| Ant/(Ant + Phe) | Fla/(Fla + Pyr) | Baa/(Baa + Chr) | Ind/(Ind + Bghi) | BaP/Bghi |
2016 | 0.23 ± 0.12 | 0.66 ± 0.18 | 0.62 ± 0.10 | 0.27 ± 0.14 | 0.45 ± 0.27 |
2017 | 0.32 ± 0.08 | 0.60 ± 0.02 | 0.57 ± 0.11 | 0.53 ± 0.14 | 0.75 ± 0.92 |
2018 | 0.35 ± 0.05 | 0.58 ± 0.08 | 0.55 ± 0.10 | 0.68 ± 0.05 | 0.51 ± 0.63 |
2019 | 0.36 ± 0.06 | 0.50 ± 0.01 | 0.75 ± 0.07 | 0.46 ± 0.18 | 0.33 ± 0.17 |
2020 | 0.27 ± 0.08 | 0.66 ± 0.12 | 0.69 ± 0.13 | 0.44 ± 0.08 | 0.47 ± 0.12 |
2021 | 0.31 ± 0.16 | 0.46 ± 0.08 | 0.79 ± 0.09 | 0.36 ± 0.06 | 0.20 ± 0.07 |
Spring | 0.34 ± 0.12 | 0.74 ± 0.16 | 0.60 ± 0.11 | 0.70 ± 0.07 | 0.38 ± 0.43 |
Summer | 0.33 ± 0.13 | 0.56 ± 0.20 | 0.51 ± 0.11 | 0.21 ± 0.13 | 0.26 ± 0.14 |
Fall | 0.31 ± 0.12 | 0.62 ± 0.18 | 0.58 ± 0.10 | 0.42 ± 0.13 | 0.33 ± 0.49 |
Winter | 0.29 ± 0.17 | 0.54 ± 0.09 | 0.66 ± 0.12 | 0.40 ± 0.15 | 0.75 ± 0.96 |
| < 0.1 = petrogenic > 0.1 = pyrogenic | < 0.4 = petrogenic/unburned petroleum 0.4–0.5 = fossil fuel combustion >0.5 = biomass and coal combustion | < 0.2 = petrogenic 0.2–0.35 = petroleum combustion >0.35 = biomass and coal combustion | < 0.2 = petrogenic 0.2–0.5 = petroleum combustion >0.5 = biomass and coal combustion | < 0.6 = nontraffic > 0.6 = traffic |
Considering the detection rate of PAHs, samples above the detection limit were selected for MDR calculation. The Ant/(Ant + Phe) ratio is indicative of pyrogenic and petrogenic sources, and the Ant/(Ant + Phe) ratio was higher than 0.1 from 2016–2021, indicating that pyrogenic sources were predominant. The Fla/(Fla + Pyr) ratio is used as an indicator of petrogenic/unburned petroleum, fossil fuel, biomass, and coal combustion. In the present study, the Fla/(Fla + Pyr) ratio was mostly greater than 0.5, except for 2021, which suggests emission from biomass and coal combustion. Moreover, the Baa/(Baa + Chr) ratio was higher than 0.35 which indicated emission from biomass and coal combustion. For the Ind/(Ind + Bghi) ratio, majority of the samples were between 0.2 and 0.5, which represent petroleum combustion emission. However, these ratios were higher than 0.5, in 2018 (0.68 ± 0.05) and in spring (0.70 ± 0.07) when they were grouped by season. The BaP/Bghi ratios used to distinguish traffic sources indicated less traffic pollution emissions in this area.
The PMF species profiles are shown in Fig. 4. Factor 1 is dominated by Nap and Bkf. Considering that Nap is more likely to be detected in gaseous phases than PM2.5, and Bkf is thought to originate from diesel combustion (Yancheshmeh et al.,2014, Shen et al.,2019). Factor 1 was defined as the diesel combustion. Factor 2 is dominated by Acy, Flu, Ace, Phe, Ant, and Chr. Low molecular weight PAHs play a leading role. They are mainly derived from low temperature combustion (Wang et al., 2010), such as coal combustion. And they are likely to decrease with decreasing PMs diameter (Wang et al.,2018). Therefore, Factor 2 is defined as coal combustion. Factor 3 was dominated by Bghi and Ind, which are generated from the incomplete gasoline combustion and are markers of higher traffic flow in cities (Zaragoza-Ojeda et al., 2022). Therefore, Factor 3 was defined as gasoline emissions. Factor 4 is dominated by Fla, Pyr, Chr, Baa, and BaP; Chr and Baa are good tracers of biomass burning. The detectable BaP is often derived from the combustion of organic matter and the natural sources, such as forest fires and volcanic eruptions (Ofori et al.,2020, Wu et al.,2022). Factor 4 is defined as biomass burning. Factor 5 is dominated by Bbf and Dah. Dah was seldom detected in the PM2.5 samples. A remarkably high Bbf/Bkf ratio was observed in snack-street boiling, which was attributed to cooking fumes (Liu et al.,2018). Furthermore, Bbf was selected by the EU Scientific Committee on Food as one of the most suitable indicators of carcinogenic PAHs in food (EU 2011). Hence, factor 5 was defined as cooking. Cooking emissions, known as the cooking-like organic aerosol factor, are an important source of organic aerosols. Cooking emissions were found to account for 90% of PAHs on average, nine times higher than that of traffic (10%) (Lin et al.,2022).
3.3 Spatial distribution variations of PM2.5 and PAHs levels
To understand the spatial distribution differences of the EPA priority 16 PAHs, we conducted a literature review. Studies that reported data on 16 PAHs in atmospheric samples were included. The exclusion criteria were as follows:1) sampling dates earlier than 2016 or later than 2021; 2) sampling sites that were contaminated areas; 3) mean, standard deviation, and sample sizes that were not found in the literature; and 4) the 16 PAH types that were not the same as those in the current study. Eleven studies met the inclusion criteria described above and the results are shown in Fig. 5. Then, individual t-tests were performed between relevant data and the present study according to the sampling time for PM2.5 and ∑PAHs. PM2.5 concentration was found significantly different from Jamshedpur City and Ranchi of India (more heavily polluted than Wuxi), Shanghai of China (superior to Wuxi), and six provinces of Thailand (superior to Wuxi). As for ∑PAHs distribution, Seoul, Korea, Arequipa of Peru, and six provinces of Thailand were not significantly different from those in the present study. The ∑PAHs concentrations in other areas were higher than those in the present study, except for Shanghai, China (p < 0.01). The details of the PM2.5-bound PAHs are listed in Table S3. It is worth noting that PAHs levels were not always associated with PM2.5, and no significant differences were found between Wuxi and Eastern India, Changchun, Tehran, and Dalian; however, the distributions of PM2.5, were distinctly different. PAHs are also regulated by natural meteorological processes, such as size distribution of particles, particle diameter and particle density (Wang et al.,2019).
3.4 Health risk assessment of membrane-captured PAHs
Based on the health risk assessment model and toxicity parameters, the incremental lifetime carcinogenic risk of children, teenagers, and adults exposed to PAHs (PM2.5) in Wuxi city was calculated using Crystal Ball. The median TEQ was 0.70 for a total of 15 PAHs and TEQ conformed to the lognormal distribution, as shown in Fig.S2(a). Fig.S2(b) shows the individual TEQ of each PAH. The TEQ of BaP (0.178) was the highest, followed by that of Bkf (0.090), Dah (0.048), Ind (0.034), and Baa (0.014).
The ILCR calculated using Eqs. (2) was analyzed using a dynamic simulation. Monte Carlo Analysis has been widely applied in uncertainty model. Through 10000 simulations, the medians of the ILCR for long-term exposure to PAHs were 2.74E-8 (max:2.00E-6), 1.98E-8 (max:3.64E-6), and 1.71E-7 (max:2.21E-5) for children, teenagers, and adults, respectively. According to the previous experience, ILCR can be divided into very low (ILCR < 1.0E-6), low (1.0E-6 < ILCR < 1.0E-4), medium (1.0E-4 < ILCR < 1.0E-3) and high carcinogenic risk (ILCR > 1.0E-3), indicating the low carcinogenic risk of PAHs pollution in air in this area except that the maximum of ILCR for adults exceed the EPA recommend limit in winter. Other studies reported that adult ILCR exceeded 1.0E-6 in industrial area of northern China, and excess cancer risks exceeded EPA recommend limit largely in Jinan, located in the northern China (Jiang et al., 2022, Yu et al., 2022). In India, the TEQ ranged between 0.24 and 94.13 while the ILCR ranged between 1.0E-5 and 1.0E-3, representing relative high cancer risk (Ekka et al.,2021). Besides, PAHs were also reported in soils and water which could accumulated and pose threat to grain, fish and vegetables (Zhang et al.,2022, Sonego et al.,2022). Considering the multiple exposure pathways, the risk of exposure of adults to PAHs cannot be neglected.
Furthermore, the Spider chart and Tornado tool in Crystal Ball were employed for sensitivity analysis to identify prominent contaminants, as shown in Fig. 6. The Spider chart describes sensitivity by the slope of all variables, where the greater slope means more remarkable impact on the prediction. While the variables at the upper part have a greater impact on the prediction, on the contrary, the bottom variables generate a smaller oscillation amplitude to predict in the tornado tool. It was found that body weight had a greater negative effect on prediction when ingestion rate and exposure duration had a positive effect. Among the 15 PAHs, BaP, Bkf, and Dah significantly contributed to carcinogenic toxicity. The consequences for children and teenagers were consistent with those for adults, which can be found in the Supplementary Information. BaP is a well-known highly toxic and carcinogenic compound. Besides, recent studies reported that Bkf was also the most important contributor to the aryl hydrocarbon receptor (AhR) activation which is participated in the mechanism of PM2.5-induced cardiovascular diseases, followed by Bbf and Dah (Ho et al.,2022, Ma et al.,2022). Also, Bkf was found to negatively affect cell viability and necrosis in the placental cell line, indicating the potential reproductive toxicity (Jo et al.,2022). Bkf was found abundant in several environmental media, even in Arctic seas (Lakhmanov et al.,2022, Zonkpoedjre et al.,2022). Therefore, it is worthwhile to focus on Bkf pollution.
To accurately evaluate the associations between the parameters and ILCR prediction, a scatter plot of the sensitivity analysis was used to obtain the correlation coefficient, as shown in Fig. 7. The exposure duration, ingestion rate, BaP, Bkf, and Dah were positively correlated, with R values of 0.6312, 0.1353, 0.4573, 0.3043, and 0.1600, respectively. In contrast, body weight was inversely associated with ILCR prediction, with an R-value of -0.1359. Scatter plots of the children and teenagers are provided in the Supplementary Information.