Introduction

Polycyclic aromatic hydrocarbons (PAHs) are by-products of incomplete combustion of organic materials and are ubiquitous in the environment (Wilcke 2007). Some of them are known or suspected carcinogens or mutagens (Duan et al. 2015). Soil is a mainly terrestrial reservoir of PAHs. Some of the high-molecular-weight PAHs have been found to persist in soil for several decades (Shuttleworth and Cerniglia 1995).

It is well known that dissipation of PAHs is more difficult in long-term field-contaminated soils than in laboratory-spiked soils (Ahn et al. 2005; Smith et al. 2011). The main difference between spiked and field-contaminated soils is the aging effect that decreases extractability and bioavailability of PAHs to soil microorganisms and other biological receptors with increasing PAHs–soil interaction time (Hatzinger and Alexander 1995; White and Alexander 1996). Several mechanisms have been suggested and explained for the aging of chemicals in soils, including partitioning into or onto humic substances (Chiou et al. 1979, 1983), or diffusion into three-dimensional micropores of soil particles (Steinberg et al. 1987). Many studies have reported that the partition to soil organic matter (SOM) was the chief mechanism for organic chemicals, particularly for hydrophobic organic chemicals (HOCs) to be sequestered in soil (Luthy et al. 1997). However, very little attention has been devoted to the physical protection of HOCs in soil aggregates. SOM is not only the most important sorbent for HOCs in soils but also the major agent of soil aggregate formation and stabilization in soils (Tisdall and Oades 1982; Balesdent et al. 2000). Soil aggregate formation can conversely protect SOM by developing physical barriers between microorganisms and enzymes and their substrates (Elliott and Coleman 1988). Then, the PAHs combined with SOM can also be protected within soil aggregates. Hence, the changes of soil structure might affect the bioavailability of PAHs in them. Steinberg et al. (1987) reported that the release of residual 1,2-dibromoethane in agricultural topsoils up to 19 years was greatly improved by pulverization of the soil, and Nam et al. (2003) reported that soil aggregation might be another important determine factor in the decreased biodegradation of aged phenanthrene (Phe) except for soil organic carbon (SOC) content. The apparent sequestration of HOCs by geosorbents is not well understood, because there is no direct observation of the HOCs’ molecular distributed within natural geosorbents (Luthy et al. 1997).

Particle–size fractionation of soils has been widely applied to separate SOM pools with different quality and turnover rates (Christensen 2001). For example, the SOM in sand–size fraction is largely comprised of fresh or slightly degraded plant material, the silt–size fraction contains partially decomposed residues, and the SOM in clay–size fraction is dominated by highly processed organic matter with aromatic and aliphatic structures (Kiem et al. 2002; Chen and Chiu 2003). Because SOM is the main adsorbent for PAHs in soils, many studies have also been carried on the distribution of PAHs in particle–size aggregates of field-contaminated soils (Müller et al. 2000; Doick et al. 2005; Ni et al. 2008) and the bioavailability of PAHs associated with different particle–size aggregates (Amellal et al. 2001). The results of Shor et al. (2003) showed that intraaggregate mass transport limitations, and not the inherent bacterial PAH utilization ability, were most important in controlling biodegradation rate of PAHs in sediments.

Phytoremediation is a green technology that uses plants to decontaminate soils of organic and inorganic compounds. Plants may promote PAH dissipation through a mixture of mechanisms, such as plant uptake and accumulation, increasing soil microbial activity, and alteration of soil physical and chemical properties in the rhizosphere (Harvey et al. 2002; Parrish et al. 2005; Wenzel 2009). In our previous study, we have compared the dissipation of PAHs in freshly spiked and two long-term field-contaminated soils in phytoremediation and found that PAHs are easily dissipated in freshly spiked soils than in field-contaminated soils (Wei et al. 2017). Moreover, the dissipations of PAHs in the two field-contaminated soils were also different and we explained that that was due to the different SOM contents in soils. In our another previous study, however, we found that the percentage removals of total PAHs from field-contaminated soils in phytoremediation had a positive correlation with the percentage distributions of total PAHs in coarse sand fraction (r = 0.813, p < 0.10) and had a significantly negative correlation with those in fine silt fraction (r = 0.898, p < 0.05) (Ni et al. 2013). Hence, we suppose that the different dissipation of PAHs in the two field-contaminated soils might be related with soil aggregate structure, not only determined by the content of SOM. To our knowledge, there is little information on the variation of soil aggregate–size distribution in phytoremediation and its relationship with the dissipation of PAHs in soils. In this study, therefore, we further reported the variations of soil aggregates in the two field-contaminated soils in phytoremediation with alfalfa (Medicago sativa L.) in our previous study (Wei et al. 2017) and analyzed the relationship between variation in soil aggregate–size distribution and PAH dissipation.

Materials and methods

Chemicals

PAH mixture standard (47940-U) and EPA 610 PAH Mix (4S8743) were purchased from Sigma-Aldrich. Acetonitrile (HPLC grade) was obtained from Fisher Scientific. Dichloromethane was bought from Sinopharm Chemical Reagent Co., Ltd. and redistilled before use.

Soils

Two PAH field-contaminated soil samples were used in this study. One is Anthrosols, the other is Phaeozems. According to the international soil classification system, the texture of Anthrosols and Phaeozems is silt sandy loam (39.5% sand, 48.6% silt, and 11.9% clay) and loam (51.1% sand, 36.7% silt, and 12.2% clay), respectively. All soil samples were air dried and sieved through a 2-mm mesh screen and thoroughly mixed. Selected soil properties and concentrations of 15 USEPA priority PAHs (except acenaphthylene) of the two soils are listed in Tables 1 and 2, respectively.

Table 1 Selected properties of the tested soils
Full size table
Table 2 Concentrations of PAHs in field-contaminated soils before and after the 10-month plant cultivation
Full size table

Pot experiments

The procedure of pot experiments was same as presented in our previous study (Wei et al. 2017). There were two treatments with three replicates, an unplanted control and soil planted with alfalfa. At the end of pot experiments, plants were removed and soil samples were freeze dried. Dried soil samples were sieved through a 2-mm mesh screen for aggregate–size fractionation and PAH analysis.

Soil aggregate–size fractionation

Soil samples of initial, unplanted, and planted treatments were fractionated into coarse sand (2000–105 μm), fine sand (105–53 μm), coarse silt (53–20 μm), fine silt (20–2 μm), and clay (< 2 μm) according to the method of Ni et al. (2008). All particle–size aggregates were freeze dried and weighed for further analysis.

Sample extraction

PAHs in bulk soils and aggregates were extracted with a Soxhlet extractor. The detailed procedure was presented in our previous study (Wei et al. 2017).

PAH analysis

The 15 USEPA priority PAHs were analyzed in this study, including naphthalene (Nap), acenaphthene (Ace), fluorene (Flu), Phe, anthracene (Ant), fluoranthene (FluA), pyrene (Pyr), benzo[a]anthracene (BaA), chrysene (Chry), benzo[b]fluoranthene (BbF), benzo[k]fluoranthene (BkF), benzo[a]pyrene (BaP), dibenz[a,h]anthracene (DBA), benzo[g,h,i]perylene (BghiP), and indenol[1,2,3-cd]pyrene (IP). Acenaphthylene was excluded due to its low fluorescence. The PAHs were analyzed using a Waters® ACQUITY UPLC system equipped with a fluorescence detector and a reversed phase column ACQUITY UPLC® BEH Shield RP18 (150 mm × 2.1 mm, 1.7 μm). The programs of gradient mobile phase and wavelength were listed in Tables S1 and S2, respectively.

Quality control

The detection limit of the 15 PAHs ranged from 0.03 to 1.53 μg L−1. Method blanks (solvent) and spiked samples (EPA 610 PAH Mix (4S8743), Sigma-Aldrich, USA, spiked into soil) were extracted and analyzed using the same procedure as the samples. The recoveries of PAHs ranged from 77.8 ± 3.1 to 97.2 ± 0.5%, except that the Nap recovery was 28.7 ± 5.7% (Table S3). Results of blanks extracted were below detection limits. The results of samples are presented without recovery ratio correction.

Statistical analysis

Statistical analysis of the data was performed in replicates by one-way analysis of variance (ANOVA) using Tukey’s test (p < 0.05).

Results

PAH dissipation in soils

Table 2 shows the concentrations of PAHs detected in Anthrosols and Phaeozems before and after a 10-month plant cultivation. For the Anthrosols, the total PAH concentration significantly decreased for both of the unplanted and planted soils compared with the initial soil (p < 0.05); the reduction was 10.2 and 15.4% in unplanted and planted soils, respectively. The largest reduction was BghiP, which dissipated about 50% for both of unplanted and planted treatments. For the Phaeozems, there were no significant changes of total PAH concentrations in both the unplanted and the planted soils relative to the initial soil.

Variation in soil aggregate–size distribution

Soil mass recoveries of the aggregate–size fractionation ranged from 96.7 to 99.5% for Anthrosols and from 96.1 to 98.4% for Phaeozems. As shown in Fig. 1, for Anthrosols, mass percentages of coarse sand and fine sand were significantly reduced while coarse silt and fine silt were significantly increased for planted soil when comparing with initial and unplanted soils (except for fine silt fraction) (p < 0.05) (Fig. 1a). Comparing with initial soil, mass percentages of coarse sand were significantly reduced and coarse silt was significantly increased in unplanted soil (p < 0.05) (Fig. 1a). For Phaeozems, there were no significant variation of aggregate–size distribution among different treatments except that the mass percentage of coarse silt was significantly decreased for planted and unplanted soil when comparing with initial soil (p < 0.05) (Fig. 1b).

Fig. 1
figure 1

Variation in soil aggregate–size distribution before and after the 10-month plant cultivation. Within a same aggregate size and different treatments for each soil, values followed by different letters are significantly different according to a Tukey’s test (p < 0.05)

Full size image

Distribution of PAHs in soil aggregates

Figure 2 is the distribution of PAHs in soil particle–size aggregates before and after the 10-month plant cultivation. The recoveries of PAHs in the sum of particle–size aggregates to total PAHs in bulk soils ranged from 88.2 to 107.9% for Anthrosols and from 99.7 to 112.9% for Phaeozems. For Anthrosols, the percentages of PAHs were significantly decreased in coarse sand while significantly increased in coarse silt in planted soil when comparing with initial and unplanted soils (p < 0.05) (Fig. 2a). There were no significantly differences of percentages of PAHs in particle–size aggregates between initial and unplanted soils (Fig. 2a). For Phaeozems, there were no significantly differences of percentages of PAHs in particle–size aggregates among different treatments (p < 0.05) (Fig. 2b).

Fig. 2
figure 2

Distribution of PAHs in soil aggregates before and after the 10-month plant cultivation. Within a same aggregate size and different treatments for each soil, values followed by different letters are significantly different according to a Tukey’s test (p < 0.05)

Full size image

Discussions

PAH dissipation in soils

For the Anthrosols, the total PAH concentration significantly decreased for both of the unplanted and planted soils compared with the initial soil (p < 0.05) and the largest reduction was BghiP (Table 2). The common understanding about the microbial degradation of high-molecular-weight PAHs is the cooxidation or cometabolism (Keck et al. 1989; Kanaly and Harayama 2000). Binet et al. (2000) also reported that ryegrass was able to accelerate the dissipation of a range of PAHs, including DBA and BghiP. For the Phaeozems, there were no significant changes of total PAH concentrations in both the unplanted and the planted soils compared with the initial soil, which might attribute to the high SOC content that limits the bioavailability of PAHs. It has been reported that SOM is the predominant sorbent for HOCs (Pignatello 1998). However, our results further indicated that the influences of SOM on the dissipation of PAHs are not only reflected in the SOM quantity but also in the stability of soil aggregates affected by SOM (See the following sections for details.).

Variation in soil aggregate–size distribution

SOM is one of the most principal adsorbent that can affect the bioavailability of HOCs in soils (Luthy et al. 1997). Moreover, Nam et al. (2003) reported that soil aggregation might be another important determining factor in the decreased biodegradation of aged Phe except for SOC content. Hence, variations in soil aggregate–size distribution before and after the 10-month phytoremediation were analyzed (Fig. 1). The different variations of particle–size aggregates in Anthrosols and Phaeozems might be due to the different SOM content, because that organic matter is believed one of the primary aggregating agents in soils (Tisdall and Oades 1982; Annabi et al. 2011). Some components of SOM such as polysaccharides, humic substances, root material, and fungal hyphae have a vital role in soil structural stabilization. Whitbread (1995) reported that the stability of soil macroaggregates increased at higher organic matter contents, probably due to organic bonding mechanisms. The SOC content in Phaeozems (8.51%) is far greater than that in Anthrosols (1.41%); the stability of macroagrregates in Phaeozems is therefore greater than that in Anthrosols.

Soil aggregation is affected by many factors such as changes in SOM, moisture content and microbial community, crop type, root growth, and tillage application (Denef et al. 2002; Verchot et al. 2011). Some of these factors can interact with and mutually affect each other. For example, soil water holding capacity (WHC) usually increases with increasing SOM content (Hudson 1994). In this study, variation of soil aggregate–size distribution was greater in Anthrosols than Phaeozems (Fig. 1) which also might be due to the different WHC of the two soils. WHC of Anthrosols (44.7%) is less than that of Phaeozems (63.9%) (Table 1), which might cause relatively more drying–wetting cycles in Anthrosols than Phaeozems, because the amount of water for watering unplanted and planted treatments was almost same at each time for both soils. Jager and Bruins (1975) reported that soils become structureless after repeated drying–wetting cycles, which is primarily ascribed to breakdown of air-dry aggregates by fast rewetting. With rapid wetting, air pressure develops inside aggregates, due to the water fast entering the soil pores, which will result in slaking of unstable aggregates. For Anthrosols (Fig. 1a), the significant difference of the distribution of soil particle–size aggregates between planted and unplanted treatments might be due to the root growth. In the short term, root growth may stimulate aggregate breakdown and soil disintegration by creating zones of collapse, thereby inducing soil loosening (Materechera et al. 1994). Reid et al. (1982) suggested that reduced stability of aggregates following the short-term growth of corn could be ascribed to a destruction of the linkages between organic matter, iron or aluminum, and mineral particles by the roots.

Relationship between variation in soil aggregates and PAH dissipation

Combining the variations of soil particle–size aggregates (Fig. 1) and the distribution of PAHs in particle–size aggregates (Fig. 2), it can be inferred that the different dissipation of PAHs between Anthrosols and Phaeozems in phytoremediation might be related to the variations of their particle–size aggregates. In Fig. 1a, mass percentages of coarse sand and fine sand fractions were significantly reduced while coarse silt and fine silt were significantly increased in planted and unplanted treatments of Anthrosols, which indicates that some of the soil macroaggregates (the coarse sand and fine sand fractions) were broken up into microaggregates (coarse silt and fine silt) in Anthrosols. As soil macroaggregates were broken up, some of PAHs trapped within aggregates became exposed to soil microorganisms and plants and were dissipated, and some were still trapped and distributed into smaller aggregates. The variations of the distribution of PAHs in particle–size aggregates of Anthrosols (Fig. 2a) also back up this point. The results of the Pearson correlation (one-tailed test) showed that the percentages of total PAHs dissipated in Anthrosols were significantly positively correlated with the decreased mass percentage of coarse sand (r = 0.786, p = 0.032, n = 6) and the increased mass percentage of coarse silt (r = 0.823, p = 0.022, n = 6). Nam et al. (2003) reported that soil aggregation might be another important determine factor in the reduced biodegradation of aged phenanthrene except for SOC content. Pee et al. (2015) also reported that ultrasonic irradiation increased the bioaccessibility of PAHs in sediment by breaking up large particles into small particles.

There was no significant change of particle–size aggregates in Phaeozems (Fig. 1b), and there was almost no reduction of soil PAH concentrations (Table 2), which was mainly due to the sequestration of PAHs in soil micropores and the inaccessibility of PAHs to microorganisms (LeBoeuf and Weber 1997; Xing and Pignatello 1997). SOM content is considered one of the most important factors that affects the bioavailability of PAHs in soils; however, Fismes et al. (2002) reported that PAH concentrations in vegetables increased with PAH concentrations in soils, although the SOM content was also proportionally increased with PAH concentrations in soils. Hence, the effect of SOM on the bioavailability of PAHs in soils is reflected not only in the quantity of SOM but also in other factors such as the soil aggregate stability influenced by SOM. Therefore, almost no reduction of PAHs in Phaeozems might be due to the higher stability of particle–size aggregates which caused the inaccessibility of PAHs sequestered in them. One may think that no reduction of total PAHs in Phaeozems might due to the lack of degradation microorganisms. However, the results of Shor et al. (2003) showed that intraaggregate mass transport limitations, and not the inherent bacterial PAH utilization ability, were most important in controlling biodegradation rate of PAHs in sediments. Hence, we think that the different dissipation of PAHs between Anthrosols and Phaeozems was mainly due to the different variations of soil aggregates.

It is well known that long-term tillage causes an alteration of soil dry aggregate–size distribution (Yang and Wander 1998; Eynard et al. 2004) and water-stable aggregation (Unger et al. 1998; Álvaro-Fuentes et al. 2008). Drying–wetting cycles and root growth can also have effects on aggregate formation in soils (Denef et al. 2002). Soil management systems that promote aggregate destruction, primarily of macroaggregates, can increase SOM degradation rates due to the exposure of the organic matter that was previously sequestered in the aggregates (Six et al. 1998; Lobe et al. 2011). SOM is the main adsorbent for PAHs in soils. Thus, the PAHs associated with SOM can be accessed by soil microorganisms in these processes. In this study, the percentages of total PAHs distributed in macroaggregates (coarse sand and fine sand) are 65.4 and 80.1% for the initial Anthrosols and Phaeozems, respectively. Other research results also showed that PAHs are mainly distributed in soil macroaggregates (Ni et al. 2008; Liao et al. 2013). Therefore, soil management systems, due to their effect on aggregate transformation rates, can also influence PAH degradation. Hence, proper soil managements such as traditional tillage and wetting and drying cycles can be chosen to enhance bioaccessibility of PAHs in long-term polluted soils in phytoremediation and natural attenuation by breaking soil macroaggregates into smaller aggregates.

Conclusions

In this study, the dissipation of PAHs was greater in Anthrosols than in Phaeozems, which was attributed to the lower SOC content of the Anthrosols. Moreover, the variation in soil aggregate–size distribution and the distributions of PAHs in particle–size aggregates of Anthrosols suggest that lower SOC contents leads to easier breaking up of soil macroaggregates into smaller size aggregates in Anthrosols, which made some trapped PAHs bioaccessible and easier to be bioremediated. Thus, proper soil management that breaks up macroaggregates (e.g., traditional tillage and wetting and drying cycles) can be chosen to enhance bioaccessibility of PAHs in bioremediation of long-term polluted soils. The results in this study also provide a new idea and method for the remediation of PAHs in field-contaminated soils in situ and ex situ.