3.1. Levels and spatial distributions of pollutants
All the contents of PCBs and HCBD were presented on a dry weight (dw) basis. At a national-scale, the PCBs contents ranged from < LOD to 241.22 ng/g with the mean contents of 62.33 ng/g in greenhouse soils, and ranged from < LOD to 192.89 ng/g with the mean contents of 51.09 ng/g in open-field soils (Table S5 and Fig. 1). Furthermore, the detection frequencies of PCBs were 84.31% in greenhouse soils and 76.47% in open-field soils. In contrast, the contamination of PCBs in greenhouse soils was more serious than in open-field soils. Different from their contamination levels and detection frequencies, the compositions of total PCBs (ΣPCBs) were similar in greenhouse and open-field soils. The major PCB homologue was tetra-CBs, followed by penta-CBs and hexa-CBs, which took over > 80% of ΣPCBs in greenhouse and open-field soils (Figure S1). The similar compositions indicated the possible homology of greenhouse and open-field soil’s pollution from human activities. Totally, the ΣPCBs levels in this study was at the same levels as previous reports in China, such as in agriculture soils of south Jiangsu and Taiyuan (Zhang et al. 2017; Sun et al. 2019).
In order to further identify the distribution characteristics, all the soil samples were classified as seven regions in China, including northeastern (HLJ, JL and LN), northwestern (XJ, GS and QH), northern (BJ, IM, SX and HB), central (HN, JX), eastern (SD, JS, ZJ and SH), southern (GD), and southwestern (YN, TB and SC). As shown in Table S6, the north and south China showed higher PCBs levels in open-field soils, especially in IM, BJ and HB, while the north, southwest and south China presented higher PCBs levels in greenhouse soils. But it should be noted that the PCBs levels in greenhouse soils were higher than in open-field soils in most regions, which may be attributed to the limited soil-air exchange in greenhouse (Gao et al. 2008; Hu et al. 2021). Usually, the polluted sources of soil PCBs are mainly from wastewater discharge, solid waste leakage, waste incineration, atmospheric deposition, etc. (Cetin et al. 2017). And the source/sink tendency of soil PCBs depended on their volatility. For example, the low-chlorinated PCBs tended to volatilize from soil, and high-chlorinated PCBs tended to sink into soil, until reached the balanced state in open-field soils (Ali et al. 2015; Xu et al. 2019). Therefore, the volatility was inhibited by the limited soil-air exchange due to the physical isolation in greenhouse.
For all the soil samples, the HCBD contents ranged from 0.85 to 24.18 ng/g with the mean contents of 8.19 ng/g in greenhouse soils, and ranged from < LOD to 20.19 ng/g in greenhouse and open-field soils with the mean contents of 6.52 ng/g in open-field soils (Table S5 and Fig. 1). Compared with PCBs, though HCBD showed lower contamination levels, the detection frequencies of HCBD were higher both in greenhouse (100%) and open-field soils (96.08%). Furthermore, the contamination of HCBD in greenhouse soils were also higher than in open-field soils in most regions, which also may be attributed to the limited soil-air exchange in greenhouse (Ali et al. 2015; Xu et al. 2019) (Table S6). Totally, the HCBD contents in this work were at comparable levels to that in agricultural soil of east China (Tang et al. 2014), the Yangtze River Delta (Sun et al. 2018c) and southwest China (Tang et al. 2016). According to the investigation results of soil PCBs and HCBD, it could be easily found that both of them presented their unique characteristics in spatial distributions, such as more serious pollution in greenhouse of north China. The reasons may be attributed to their pollution sources and regional environmental factors (including other pollutants and soil physicochemical properties) (Nieuwoudt et al. 2009; Habibullah-Al-Mamun et al. 2019; Niu et al. 2022).
3.2. Correlation between pollutants and environmental factors
The linkages between pollutants and environmental factors in greenhouse and open-field soils were obtained through redundancy analysis (Fig. 2). At a national-scale, the positive correlations were only obtained between OCPs and PCBs (r = 0.298, p༜0.05) in greenhouse soils (Table S7), indicating that they were likely to share the same pollution paths, such as irrigation (Meng et al. 2017). The positive correlations between Pb and PCBs (r = 0.311, p < 0.05) were also obtained in open-field soils (Table S8), suggesting that they were likely to share same pollution sources, such as industrial metal smelting (Diop et al. 2017). But there was no any environmental factor correlated with HCBD in greenhouse and open-field soils (Table S7, S8). In this study, no correlation between soil PCBs and properties can be attributed to the variety of pollution levels and region distributions across China (Niu et al. 2022). Meanwhile, there were no correlation between soil HCBD and other pollutants/soil properties.
To further study the relationships between pollutants and environmental factors, the correlation analysis between PCBs homologues and soil properties were also carried out in some economically developed regions, such as the north China and the east China. In north China, ΣPCBs were correlated with tetra-CBs (r = 0.907, p < 0.01 in greenhouse soils; r = 0.830, p < 0.01 in open-field soils) and hexa-CBs (r = 0.609, p < 0.05 in greenhouse soils; r = 0.881, p < 0.01 in open-field soils) (Table S9, S10). In east China, the total PCBs were correlated with tetra-CBs (r = 0.921, p < 0.01 in greenhouse soils; r = 0.904, p < 0.01 in open-field soils) and penta-CBs (r = 0.860, p < 0.01 in greenhouse soils; r = 0.569, p < 0.05 in open-field soils) (Table S11, S12). The correlations between ΣPCBs and homologues verified the same sources/paths for greenhouse and open-field soils pollution in these regions (Hu et al. 2021).
3.3. Regional source analysis of pollutants
Traceability analysis on soil PCBs in north and east China has been performed to identify their mainly emission sources using principal component analysis (PCA). The compositions of PCBs homologues were also compared with the Aroclor series products, including Aroclor 1016, 1221, 1232, 1242, 1248, 1254, 1260, 1262. Factors with eigenvalues greater than 1.0 were extracted and 3 principal components (> 80% variance) were acquired. In north China, the compositions of PCBs were less consistent with the Aroclor series products (Fig. 3a, 3b). In both greenhouse and open-field soils, hexa-CBs and hepta-CBs presenting larger loading in PC1 (> 0.53), were likely attributed to the recycling and disposal of electrical equipment containing PCBs (Li et al. 2019). Tetra-CBs and tri-CBs showing larger loading in PC2 (> 0.48), corresponded to the domestic coal, wood burning emissions, non-ferrous metal smelting and regeneration, and high temperature incineration of industrial as well as municipal waste (Lee et al. 2005; Nieuwoudt et al. 2009; Habibullah-Al-Mamun et al. 2019). Penta-CBs with loading of > 0.83 in PC3 were mainly from oil additives and metallurgy industry (Ba et al. 2009). This result revealed the silimar pollution sources of greenhouse soils and open-field soils in north China.
In east China, the compositions of PCBs were similar to the Aroclor series products in partial sample sites (Fig. 3c, 3d). For example, in both greenhouse and open-field soils of Shanghai, PCBs (G1, O1, O2) possibly derived from Aroclor 1242, which was identical with the result of Jiang et al. (2010). In rest greenhouse soils, tri-CBs and tetra-CBs showed larger loading in PC1 (> 0.53), indicated the mainly sources of the domestic coal, wood burning emissions, high temperature incineration of industrial and municipal waste, etc. (Lee et al. 2005; Nieuwoudt et al. 2009; Habibullah-Al-Mamun et al. 2019). And hepta-CBs and hexa-CBs showed larger loading (> 0.61) in PC2 and PC3, indicating the source from the recycling and disposal of electrical equipment containing PCBs (Li et al. 2019). In open-field soils, hexa-CBs and hepta-CBs showed larger loading in PC1 (> 0.42) in open-field soils, and hexa-CBs and tri-CBs presented larger loading (> 0.53) in PC2 and PC3. This indicated that the recycling and disposal of electrical equipment containing PCBs was the main source. Previous reports have indicated that the e-waste dismantling areas distributed in east China, such as Taizhou, were one of the most important sources for soil PCBs (Sun et al. 2018b; Liu et al. 2020).
3.4. Human health risk assessments
The health risks of pollutants in greenhouse and open-field soils were estimated via soil ingestion, dermal contact and inhalation exposure pathways (Niu et al. 2013; Li et al. 2021b). As shown in Fig. 4a, the non-cancer exposure risks of PCBs to children and adults in open-field soils were both higher than in greenhouse soil. Meanwhile, the non-cancer exposure risks of PCBs to children were higher than to adults both in greenhouse and open-field soils. The average HI values of PCBs were 0.022 for adults and 0.222 for children in greenhouse soils, and 0.030 for adults and 0.274 for children in open-field soils, respectively. In all these samples, lower HI values (< 1.0) for adults indicated the negligible non-cancer risks to adults. But in some provinces, such as BJ, XJ, SD and GD, higher HI values (> 1.0) for children both in greenhouse and open-field soils indicated the potential non-cancer risks of PCBs to children were not negligible. Among PCBs homologues, the average HI value of PCB-126 was the highest in greenhouse soils (HI = 0.010 for adults, HI = 0.121 for children) and in open-field soils (HI = 0.016 for adults, HI = 0.146 for children), followed by PCB-81 (Figure S2). For HCBD, all the HI values were below 1.0 for adults and children, implying the negligible non-cancer risks in soil samples (Fig. 4c).
The carcinogenic risks of PCBs to children were higher than that to adults (Fig. 4b). In greenhouse soils, the average values of PCBs to adults and to children were 8.37 × 10− 7 and 1.38 × 10− 6, respectively. Compared with in greenhouse soils, the carcinogenic risks of PCBs to children and adults in open-field soils presented higher risks (Fig. 4b). In open-field soils, the average values of PCBs to adults and children were 1.03 × 10− 6 and 2.13 × 10− 6 respectively. And the values in large number of samples (25.5% to adults and 37.3% to children) were between 10− 6 and 10− 4, implying relatively low carcinogenic risks to human beings. Among these PCB homologues, the most significant carcinogenic risks were also attributed to PCB-126 (Figure S1). In contrast, the carcinogenic risks of HCBD to adults and children were both below 10− 6 in greenhouse and open-field soils (Fig. 4d), corresponding to the negligible carcinogenic risks.
In this study, the non-cancer risks and carcinogenic risks of PCBs were similar to the Yellow River irrigation area and Lanzhou soils (Ding et al. 2018; Li et al. 2021b). However, the different patterns were found between the health risks and PCBs levels. For example, samples from IM presented the higher PCBs levels and the lower health risks in greenhouse soils. The reason could be explained by different toxic equivalency factor (TEF) of measured PCBs, which order was PCB-126 > PCB-169 > PCB-81 > PCB-77 > other PCBs (Van den et al. 2006; Niu et al. 2022). Compared with that in open-field, low-chlorinated (≤ 4) PCBs presented increasing ratios in ΣPCBs in greenhouse soils, corresponding to their lower risks. This might be attributed to the faster dechlorination of high-chlorinated PCBs in greenhouse, as greenhouse environment factors (moisture and temperature) were more favorable for microbial growth (Dou et al. 2020; Tao et al. 2022).