Polycyclic aromatic hydrocarbons (PAHs), a series of hydrophobic aromatic compounds composed of two or more benzene rings, are carcinogenic, teratogenic and mutagenic (Cvancarova et al. 2013). Sixteen PAHs have been listed as priority pollutants by the U.S. Environmental Protection Agency (EPA) (Zelinkova and Wenzl 2015). Generally, PAHs are the products of pyrolysis or rearrangement of molecules during incomplete combustion of organic matter under oxygen-deficient conditions (Dat and Chang 2017). In addition to some natural sources such as biological releases, natural fires and volcanic eruptions (Tsibart and Gennadiev 2013), many anthropogenic sources including the chemical industry, transportation, and domestic pollution have resulted in the current worldwide distribution of PAHs (Dat and Chang 2017; Duran and Cravo-Laureau 2016; Gao et al. 2018; Wilcke 2007). The produced PAHs are semivolatile, lipophilic, and persistent and easily accumulate in the environment, posing great risks to organisms including human beings.

Soil is a major sink of PAHs in terrestrial ecosystems and serves as an important medium for the volatilization, deposition and degradation of PAHs in the environment. The average concentration of PAHs in topsoil in China was found to be approximately 730 ng g−1 (Zhang and Chen 2017). The sources of PAHs in the soil are mainly from coal combustion, vehicular emissions, sewage disposal, and petroleum spills (Han et al. 2014). Therefore, PAHs are frequently detectable in soils located in areas with high urbanization and industrialization (Zhang et al. 2017). For example, the PAH content detected in 40 soil sampling points collected from a coke factory in Shandong Province, China, exceeded the heavy contamination standard (Han et al. 2014). Idowu et al. (2020) reported that the content of PAHs in the soils of a renowned industrial heritage city in Australia ranged from 2.51 to 392.93 mg kg−1, hundreds of times higher than the local standards. In addition, the total PAH contents in the soil near a cement factory in Beijing reached 1134.3 μg kg−1, 3.37 times higher than the local background value (Wang et al. 2018), leading to high health risks for children and adults near the cement factory. Moreover, the PAHs in soil can be revolatilized into the atmosphere (Jia et al. 2019), transported by the soil–plant system (Zhang et al. 2015), and leached into the deep soil profile or even groundwater (Widdowson et al. 2003), leading to serious ecological and environmental issues. Therefore, the contamination of PAHs in the environment is of great concern.

Once organic contaminants enter the soil, they gradually exhibit diverse speciation, such as dissolved in soil solution, adsorbed on soil organic matter and minerals, or bound to the soil solid phase to form unextractable residues (Sabate et al. 2006). Since not all forms of contaminants pose threats to organisms, it is necessary to take into account the bioaccessibility of contaminants, which is described as the process of association and dissociation between the bound contaminant and released contaminant, and both of them transport to a biological membrane and finally across it (Semple et al. 2004). Considering the bioaccessibility of pollutants is conducive to predicting more realistic exposure limits and providing more justifiable and prospective risk assessment (Ortega-Calvo et al. 2015). The evaluation methods and influencing factors of PAH bioaccessibility in soil have received extensive attention, among which chemical extraction is still the most commonly used indirect in vitro method to determine the bioaccessibility of PAHs in soils (Gao et al. 2019; Tao et al. 2006). It is worth mentioning that modified cyclodextrin with a cyclic structure has been broadly utilized to evaluate PAH bioaccessibility (Humel et al. 2020; Leech et al. 2020; Qin et al. 2020). Reid et al. (2000) showed that the optimized hydroxypropyl-β-cyclodextrin (HPCD) extraction method could best predict the available microbial concentrations of soil-associated phenanthrene. A better correlation between HPCD-extracted PAHs and earthworm-accumulated PAHs than other chemical extraction methods in soil was also observed by Zhang et al. (2017). Moreover, studies have revealed that PAH bioaccessibility is affected by PAH properties, soil properties and environmental factors (Duan et al. 2014), such as the molecular weight of PAHs (Cheng et al. 2021; Li et al. 2013), cocontaminants (Lin et al. 2008), soil composition (Portet-Koltalo et al. 2020), and soil organic matter (SOM) (Moeckel et al. 2014). However, the influence and degree of influence of various factors have been inconsistent in diverse studies, and the change in PAH bioaccessibility in different soil profiles of contaminated sites is still lacking sufficient research.

In the present study, we intended to investigate the influencing factors that affect the bioaccessibility of PAHs in contaminated sites. Soil samples, including different profile depths, were collected from two industrial sites in China. HPCD extraction was chosen as the PAH bioaccessibility assessment method. The relationship between PAH bioaccessibility and soil properties was analyzed. Therefore, this study sheds light on the risk assessment of organic contaminants in soil.

Materials and Methods

The PAH-contaminated soils were sampled from a coal gasification plant site in Nanjing (S1) and a steel plant site in Chongqing (S2), China. According to the size of the contaminated sites, 5 and 15 sampling points in S1 and S2 were sampled by Geoprobe (TecopSA TEC15DP, AMK, Spain) and a self-contained pump drilling rig (GK200-1, YiXin Machinery Corp., China), respectively. The soils in S1 were sampled at a depth of 6 m, while the soils in S2 were sampled at a depth of 11 m with the horizon as a base depth. Finally, a total of 34 and 53 soil samples were collected at S1 and S2, respectively. All the soil samples were air-dried at room temperature. After removing stones and root residues, all soil samples were ground evenly to pass through a 0.25 mm sieve, and their physicochemical properties were characterized, as shown in the supplemental material (Table S1 and Fig. S1).

In this study, 16 priority PAHs proposed by the U.S. EPA were determined through the accelerated solvent extraction method (ASE 200, Dionex, USA). The ASE program was conducted under the following conditions (Zhang et al. 2016): hexane/acetone (1:1, v/v) as the solvent, a temperature and pressure of 100°C and 10.34 MPa, respectively, and static extraction for 5 min. Then, the extract obtained from ASE was concentrated to 2 mL by a rotary evaporator, transferred to a silica gel column packed with Florisil, silica gel and anhydrous Na2SO4, and eluted with 15 mL of n-hexane/dichloromethane solution (9:1, v/v). Then, the eluate was concentrated and dissolved in 2 mL of acetonitrile, and the mixture was passed through a 0.22 μm polytetrafluoroethylene filter membrane. The final extract was determined by gas chromatography–mass spectrometry (Agilent 7890A/5975C, Santa Clara, CA) (Ni et al. 2018). The instrument limit of detection (LOD) and of quantification (LOQ) was 0.03–0.15 μg kg−1 and 0.10–0.55 μg kg−1, respectively.

The specific steps of HPCD extraction are as follows. One gram of soil sample was placed into a 50 mL glass centrifuge tube, and 20 mL of 50 mmol L−1 HPCD solution was added. The extraction was carried out with oscillation (200 r min−1) for 20 h at room temperature. Then, the mixture was centrifuged at high speed for 10 min (2000 r min−1) after static placement for 30 min. Next, the supernatant was discarded, and 20 mL of deionized water was added to the bottom soil, which was centrifuged again at high speed. The deposited soil was retained, and an appropriate amount of diatomite was added to the bottom soil sample, which was evenly stirred and loaded into the ASE extraction column. The sample was then extracted and determined following the same steps as for total PAH extraction. Finally, the PAH content extracted by HPCD was calculated by the subtraction method and regarded as the bioaccessible PAH content.

The potential toxic effects of PAHs were evaluated based on toxic equivalent (TEQ) analysis. In this study, the TEQ was computed according to the method from Richter-Brockmann and Achten (2018). The specific calculation is shown in Eq. (1).

$$ {\text{TEQ}} = \sum {\text{C}}_{{\text{(bio)i}}} \times {\text{TEF}}_{{\text{i}}} $$
(1)

where C(bio)i is the (bioaccessible) concentration of individual PAHs and TEFi is the corresponding toxic equivalent factor, representing the potential toxicity relative to that of benzo[a] pyrene (Nisbet and Lagoy 1992). The TEF values of 16 PAHs are presented in Table S2. Furthermore, to compare the toxicity intensity of PAH in soils of different sites, the toxic concentration ratio (Rt/c) based on the TEQ was calculated as Eq. (2).

$$ \left( {\text{Bio- }} \right){\text{R}}_{{{\text{t}}/{\text{c}}}} = {\text{C}}_{{\left( {\text{bio- }} \right){\text{TEQ}}}} /{\text{C}}_{{{\text{total}}}} $$
(2)

where C(bio-) TEQ is the (bioaccessible) TEQ concentration of PAHs in soil and Ctotal is the total PAH concentration in soil.

Stepwise linear regression (LR) analysis with soil properties as independent variables and PAH bioaccessibility as the dependent variable was carried out using the software Statistical Product and Service Solutions (SPSS). The F test was used for the linear regression model, and the t test was used for the regression coefficient significance test. The Kruskal–Wallis test and Mann–Whitney test were used to assess the difference between different sites. The content data of each PAH in the soil at different depths are expressed as the mean ± SD. Differences at p < 0.05 were regarded as statistically significant.

Results and Discussion

The residue concentrations of PAHs in soils are shown in Fig. 1. A total of fifteen kinds of PAHs were detected at site S1. The content of individual PAHs ranged from 0.41 to 668.85 mg kg−1, and the mean value of individual PAHs was 3.08–47.81 mg kg−1. Twelve kinds of PAHs were detected at site S2, with the concentrations ranging from 0.25 to 21.4 mg kg−1, and the mean content of individual PAHs ranged from 0.77 to 3.24 mg kg−1. In addition, PAHs were barely detectable in soils deeper than 6 m at site S2. Overall, the PAH contents at site S1 were significantly higher than those at site S2 (p < 0.05). As shown in Fig. 1c, 4-ring PAHs were the dominant fraction at both sites (33.6%–60.9%), followed by 3-ring PAHs at site S1 and 5-ring PAHs at site S2, regardless of soil profile depth. Two-ring PAHs showed the lowest proportion and were almost undetectable at site S2. The high-ring PAHs (4-, 5-, and 6-ring PAHs) at the two sites accounted for approximately 80% of all PAH contents.

Fig. 1
figure 1

The contents of polycyclic aromatic hydrocarbons (PAHs) at sites S1 and S2 (a) and the distribution of PAHs with different ring numbers in the soil profile (b)

Full size image

Compared with those found in existing studies, the total PAH contents at sites S1 and S2 were all relatively high (Wang et al. 2017), indicating that the risks of PAH contamination in the coal gasification plant and steel plant were high. PAHs naturally exist in coal due to diagenesis. The total amount of PAHs extracted from coal may range from tens to several thousand milligrams per kilogram (Hindersmann and Achten 2018); hence, the combustion of coal and coke has been an important source of PAHs in the environment and at industrial sites. Moreover, one study showed that PAHs in coal-rich soil were deficient in microbial degradation (Achten et al. 2011).

The composition of PAHs with different ring numbers in soils differed substantially (Hindersmann and Achten 2018; Karaca 2016; Siemering and Thiboldeaux 2021). Studies have shown that 4-ring PAHs are dominant in heavy traffic and industrial areas (Karaca 2016). High-ring (5- and 6-ring) and 2-ring PAHs are distributed in the soil near petrochemical complexes (Nadal et al. 2004), while 3- and 4-ring PAHs are mainly concentrated near chemical companies. In this study, The S1 and S2 sampling sites were located at the former sites of the coal gasification plant and steel plant, respectively. After the plant stopped operating, the plant-derived LMW PAHs in the soil were relatively quickly degraded due to their comparably higher water solubility, while the HMW PAHs persisted in the soil for a long time (Li et al. 2012). In addition, our diagnostic ratio results further corroborate that the dominant source of PAH in two sites was the combustion of coal and other fuels (Fig. S3), and there was source of traffic emission because the two sites were located in urban areas (Wu et al. 2021).

The bioaccessibility of PAHs (bioacc-PAHs) with different ring numbers is shown in Fig. 2. The bioaccessibility of PAHs extracted by HPCD at site S1 ranged from 1.18% to 92.93%, with an average of 52.02%, while those at site S2 ranged from 0.84% to 77.86%, with an average of 29.28%. In general, the bioacc-PAHs at site S1 were significantly higher than that at site S2, and the average bioacc-PAHs for 3-ring PAHs were the highest in soils at both sites. The bioacc-PAHs of 5- and 6-ring at site S1 were significantly lower than that of the 3- and 4-ring PAHs. Although there was no significant difference among the bioacc-PAHs with different ring numbers at site S2, the average bioacc-PAHs for 6-ring (20.81%) at site S2 was apparently lower than that of 3–5 ring (29.21%–31.96%). This result is consistent with that reported by other studies. For example, Crampon et al. (2016) and Oleszczuk et al. (2017) both demonstrated a decline in bioacc-PAHs with increasing number of aromatic rings in soils. The amounts of PAHs retained by humin in soils increased with increasing number of aromatic rings, leading to decreased bioaccessibility (Han et al. 2020). Three-ring PAHs had the highest bioaccessibility for grains, i.e., the highest likelihood to be taken up, whereas the 5- and 6-ring PAHs had the lowest likelihood (Tian et al. 2018). Such a tendency could be explained by the increased Kow of PAHs with high molecular weight (Oleszczuk et al. 2017); a higher Kow value represents stronger hydrophobicity and lipophilicity (Yu et al. 2018), leading to lower contents of water-soluble PAHs and poor migration ability of PAHs from soil particles to the soil solution. Therefore, PAHs with more aromatic rings and larger Kow values were more difficult to extract and more recalcitrant in soils than those with fewer aromatic rings (Cheng et al. 2021).

Fig. 2
figure 2

The bioaccessibility of polycyclic aromatic hydrocarbons with different ring numbers. Boxes sharing the same letter are not significantly different based on the Kruskal–Wallis test

Full size image

As shown in Fig. S2, the bioacc-PAHs varied greatly at different soil profile depths. A slight increasing trend with increasing depth was found in the bioacc-PAHs of the 3-, 4- and 6-ring PAHs at site S1, while at site S2, higher bioacc-PAHs were found in the mid-deep soil layers (1.51–2.25 m, 3.76–4.50 m) than in the deep layers (5.26–6.00 m) and upper soil layers (0–0.75 m), which could be attributed to the changes in soil properties. In general, the soil organic matter content reaches its highest value in the soil surface layer and decreases significantly with increasing depth, which affects the adsorption and desorption processes in the soil to a great extent (Gao et al. 2013). Several studies have also obtained similar results that the sorption of organic pollutants in deep soil is lower than that in surface soil (Knight et al. 2019; Gao et al. 2013; Si et al. 2009), which could explain the high bioaccessibility of contaminants in deep soil layers.

The organic matter content and mineral content of soils with different textures are very distinct, greatly influencing the morphological distribution and toxicity of organic pollutants in soil (Princz et al. 2018). There were four soil textures, i.e., sandy loam, loamy sand, silt loam and loam, in our studied sites (Fig. 3). In both sites, there was almost no significant difference in bioacc-PAHs among different soil textures. However, it is worth noting that the bioacc-PAH in silt loam soil in site S1 was significantly lower than that of loamy sand soil, and the bioacc-PAH of loam soil in site S2 was significantly lower than that of sand loamy soil, which may be due to the fact that silt loam and loam are exactly the two of the four studied soils with higher silt particle (2–0.02 mm) proportion. Wilcke et al. (1996) reported that PAHs in soil were adsorbed the most on silt particles and the least on sand particles, resulting in the low bioaccessibility of PAHs in silt soils. Ritore et al. (2022) observed a decline trend in the removal efficiency of hydrophobic organic contaminants (HOCs) in the soil with the transformation of soil texture from sandy to silt loam and clay soil. Therefore, the abundant aromatic ring structures in the organic matter of silt particles leads to the adsorption and binding of HOCs in soil, thus decreasing their bioaccessibility (Wilcke et al. 1996).

Fig. 3
figure 3

The bioaccessibility of polycyclic aromatic hydrocarbons in different soil textures. Boxes sharing the same letter are not significantly different based on the Kruskal–Wallis test

Full size image

The correlations between the bioaccessibility of PAHs and soil properties were analyzed. As shown in Table S4, the total bioacc-PAHs had a positive LR coefficient (0.478) with soil dissolved organic carbon (DOC) and a negative coefficient (− 0.247) with total organic carbon (TOC). Compared to those on insoluble organic matter, PAHs adsorbed on dissolved organic matter (DOM) have higher mobility and bioaccessibility. The former matter can immobilize PAHs, while the latter can solubilize PAHs in the soil, which has been suggested by numerous studies (Li et al. 2022; Tao et al. 2006; Yu et al. 2018). Therefore, the decreasing DOC value in the soil samples found in this study led to a reduction in the bioaccessibility of PAHs.

There were significant positive LR coefficients between 3, 4, and 6-ring bioacc-PAHs and pH. Studies have shown that pH affects bioacc-PAHs by influencing the content or morphological changes of DOM and SOM in soil (Yu et al. 2018). Under different pH conditions, SOM exists in diverse physical forms and has different adsorption capacities for PAHs (Yu et al. 2018). In addition, Yu et al. (2016) concluded that higher pH values promoted an increased DOM content in soils, thereby inhibiting the proportion of PAHs sequestered by the soil solid phase and improving their bioaccessibility. The cation exchange capacity (CEC) had a significant influence only on 6-ring PAHs. It was reported that an increased CEC in soil would deepen the sequestration degree of PAHs (Chung and Alexander 2002). Soil salinity showed subtractive LR coefficients with 3-ring and total bioacc-PAHs, which was consistent with the results obtained by Anderson (2013) that saline soil could retain PAHs in soil to a certain extent.

The total potassium, total iron and total manganese contents all had significant negative effects on the bioacc-PAHs (Table S4). Previous studies have shown that the presence of metal ions can enhance the adsorption capacity of soil for PAHs with poor desorption reversibility, thereby reducing its bioavailability (Luo et al. 2010; Saeedi et al. 2018). On the one hand, metal ions interact with solid organic matter in the soil, such as by forming complexes with polar functional groups of organic matter, modifying its structure and improving its adsorption capacity for PAHs (Saison et al. 2004). The higher the soil organic carbon content is, the stronger the enhancement of the adsorption affinity by metal cation modification (Liang et al. 2016). On the other hand, quantum mechanics have proven that the cationic–π interaction between metal ions and PAHs may improve the adsorption of PAHs in SOC (Liang et al. 2016).

The proportion of soil silt and clay had no significant effect on the total bioacc-PAHs but had different effects on the 2-ring, 4-ring and 5-ring PAHs (Table S4). Numerous studies have indicated the negative effects of the clay proportion in soil on the extractability of PAHs. However, its impact was limited, such as decreasing only the low molecular weight bioacc-PAHs (Portet-Koltalo et al. 2020) or having only a negative correlation in the soil in particular research areas (Pu et al. 2004). Theoretically, clay is rich in pores with a size less than 20 nm or even smaller and has a large surface area, which is difficult to invade by soil (micro)organisms or plant roots, making it an excellent place for pollutant storage (Whittaker et al. 2019). Moreover, with high surface areas, enhanced sorption to contaminants through weak physical interactions was verified in clay soil (Pu et al. 2004).

The TEQ concentration based on total and bioaccessible PAH contents in 78 soil samples from the two sites was calculated (Table S3). The total TEQ concentration ranged from 0.0004 to 263.24 mg kg−1 at site S1 and 0.0002 to 10.11 mg kg−1 at site S2, with mean values that were 7.5 times higher at site S1 than at S2. The bioaccessible TEQ concentration ranged from 0 to 90.22 mg kg−1 at site S1 and 0 to 13.34 mg kg−1 at site S2, with mean values that were 9.2 times higher at site S1 than at S2. Due to the high PAH contents at site S1 (Fig. 1a), the total TEQ and bioaccessible TEQ were all higher at site S1 than at site S2. Therefore, a toxic concentration ratio (Rt/c) was calculated to normalize the PAH concentrations. It was shown that the Rt/c (CTEQ/Ctotal) at site S2 was significantly higher than that at site S1 (p < 0.01) (Fig. 4), which can be attributed to the fact that the proportion of HMW PAHs with high toxicity at site S2 was about 10% higher than that at site S1(Fig. 1b). This result indicates that higher PAH toxicity would be achieved at site S2 rather than S1 when the contaminated sites show the similar total PAH contents. However, the Bio-Rt/c (Cbio-TEQ/Ctotal) based on the bioaccessible TEQ was significantly higher at site S1 than at site S2 (p < 0.01). This opposite result means that higher bioaccessible toxicity would be present at site S1 than S2 under the condition of similar total PAH contents, even though the proportion of HMW PAHs was larger at site S2 than S1. Therefore, the proportion of HMW PAHs and the bioaccessibility of PAHs both dominate the toxicity of PAHs in contaminated soils. Since PAHs are likely sequestered by soil organic matter and minerals in soil (Umeh et al. 2019), blocking contact with organisms, we suggested that to avoid overestimating the toxicity of PAH in soil, it is necessary to take bioaccessibility into account.

Fig. 4
figure 4

The ratio of toxic equivalent concentration to total polycyclic aromatic hydrocarbons (PAHs) concentration (Rt/c), and the ratio of bioaccessiable toxic equivalent concentration to total PAHs concentration (Bio-Rt/c) in site S1 and S2

Full size image