Assessment of an NDL-PCBs Sequestration Strategy in Soil Using Contrasted Carbonaceous Materials through In Vitro and Cucurbita pepo Assays

Featured Application: A sequestration-based remediation strategy using several types of carbonaceous amendments and its potential to reduce transfer to plants. Abstract: The present study aims to assess the respective efﬁciency of Biochars (BCs) and activated carbons (ACs) to limit PCB 101, 138, 153 and 180 transfer to plants. A set of 6 high carbon materials comprising 3 BCs and 3 ACs was tested and used to amend a soil at 2% rate. Then, the two most efﬁcient carbonaceous materials were used as an amendment of an historically contaminated soil sampled in the St Cyprien vicinity (Loire, France). An environmental availability assessment was performed using the ISO/DIS 16751 Part A assay (n = 3). For the in vivo part, Cucurbita pepo were grown for 12 weeks. Signiﬁcant decreases of transfer were found for both assays notably for powdered ACs (up to 98%). By contrast, signiﬁcantly lower levels of transfer reduction were observed when BCs amendments were performed, ranging from 27 to 80% for environmental availability assessment and 0 to 36% for C. pepo . Reduction factors above 90% for the 2 selected materials were found from amended historically contaminated soils. Present results led to consider such a sequestering strategy as valuable to ensure plant production on non-dioxin-like polychlorobiphenyls (NDL-PCBs) contaminated soils.


Introduction
For a long time, the transfer of highly hydrophobic organic compounds in plants has been neglected, until that many publications claimed that plants are able to take up and accumulate those. Among these plants, species belonging to Cucurbitaceae family such as zucchini and pumpkin are able to accumulate polychlorobiphenyls [1,2], dioxins [3] and organochloride pesticides such as chlordecone [4] in roots with high translocation from roots to shoots for varieties of Cucurbita pepo [5]. The interception and uptake capacity depend primarily on the physico-chemical characteristics of organic pollutants (including solubility, hydrophobicity and molecular mass), but also on root traits of the species that may influence these criteria [6]. Conditions at the soil-root interface (i.e., rhizosphere) can modify the availability of organic pollutants via the exudation of organic acids as described in C. pepo [7] to favor root interception. The subsequent accumulation of organic pollutants in the root also depends on root characteristics, particularly the composition of membrane and parietal lipids [8] that in relation with hydrophobicity drives the retention in roots. As persistent organic pollutants (POPs) have very marked hydrophobicity characteristics (log Kow greater than 4), in the majority of plant species, these compounds are absorbed and accumulated in the roots without transfer to the vegetative aerial parts and fruits. Nevertheless, some varieties of Cucurbitaceae are able to accumulate large amounts of POPs in the aerial parts depending on the number and the position of the non-dioxin-like polychlorobiphenyls (NDL-PCBs) chlorine atoms [1,5].
Thus, acting directly on this transfer to plants would prevent the contamination of food-products. A possible strategy to reduce this halogenated POP transfer to vegetal is to trap them in the soil by making them non-available. In this perspective, several amendment materials have already been evaluated and proven to be effective to limit the POP transfer, as assessed on root vegetables [9,10] or through an in vitro assay [11,12]. To limit transfer to plant, compost amendment of chlordecone (CLD) contaminated soils displayed a reduction of transfer to root vegetables by a factor of 3 to 10 compared to non-amended soils [13]. However, such data is lacking for other POPs, and also for plants known to accumulate halogenated POPs inside their aerial parts like Cucurbitaceae family. Activated carbons (ACs) amendment showed also a reduction up to 90% in non-physiological based in vitro environmental availability assays [12,14]. Some studies reported the ability of activated carbons and biochars to trap and reduce the transfer of pesticides, PCBs and other organic pollutants in cucurbits [15][16][17][18][19].
Thus, the subject of the present study is to assess the efficiency of a trapping strategy when using different high carbon matrices as amendment material to ensure plant production on polluted agricultural land. In this frame we aimed to focus on Non-Dioxin-like Polychlorobiphenyls (NDL-PCBs).

Experimental Design
The experimental design aimed at assessing an NDL-PCBs sequestration strategy using AC or BC amendment on historically contaminated soils. As several distinct matrices were produced, a preliminary step was to select the most efficient sequestering materials before validation on a soil originating from a known historically contaminated site. Thus, the first step was carried out on an Organization for Economic Co-operation and Development (OECD) soil and the second on a soil originating from St Cyprien (Loire, France), a well-described PCBs contaminated site. In the 22 August 2008 a fire broke on this former facility storing large quantities of contaminated oil. This fire event produced a contamination of the agricultural areas nearby resulting in PCB impregnation of soil and subsequently led to the destruction of various locally produced foodstuffs, including vegetables. Recent contamination data of the site are provided elsewhere [20] and several studies were performed on various soil samples collected from this site during the past decade [20][21][22]. For each step, the environmental availability and the NDL-PCB transfer to aerial parts (leaves, stems, fruits, flowers) of Cucurbita pepo were assessed. Finally, the ability of the quick and cheap in vitro assay to predict the in vivo assay was discussed.

Production and Acquisition of Condensed Materials
A set of 6 highly carbonaceous materials was obtained from two distinct sources: (i) 3 ACs were obtained from a furnisher (ROTH Sochiel E.U.R.L., Lauterbourg, Grand Est, France) and (ii) 3 biochars were produced by CARBOFRANCE (Montier-sur-Saulx, Lorraine, France). Biochars were produced by pyrolysis using the industrial-scale ovens of CARBOFRANCE (Montiers-sur-Saulx, Lorraine, France). Oak stems originated from various locations of the East part of France, to ensure comparability with the raw material of this industry. Japanese knotweeds (Reynoutria japonica) were collected in Laxou (Lorraine, France) using an apparatus designed by Lorexam (Ludres, Lorraine, France).
Ovens were set to perform a pyrolysis at 500 • C or 700 • C. After the pyrolysis process, biochars samples were ground and sieved until <500 µm before their use. Brief description of the condensed materials is provided in Supplemental Information (Table S2) and characterization details of the media are provided elsewhere [11,23].

Soils Preparation
Two distinct types of soils (OECD soil and St Cyprien soil) were used for this two-step study. Their preparation was carried out at GISFI (Groupement d'Intérêt Scientifique sur les Friches Industrielles, Homécourt-France) and aging was carried out at the Bio-DA (Bioavailability-Bioactivity) platform.
For the first step on OECD soil, two distinct contaminated sets of OECD soils were produced: The first was contaminated by a mix of PCBs 101, 138, 153 and 180 whereas the second remained uncontaminated. An amount of 5% of the total mass of sand was sub-sampled and spiked to achieve the chosen NDL-PCBs concentrations in OECD soil. The spiking methodology used is fully described elsewhere [23]. All constituents of this soil were added and mixed accordingly to the OECD guideline 207 (Table 1) [24]. Concerning the sampled soil from St Cyprien contaminated site (Saint Cyprien, Loire, France), soil was air dried and sieved until <2 mm prior utilization. Then, OECD soil was left to age for 1 month prior to sub-sampling it to 7 aliquots. Finally, 6 out the 7 soil-aliquots were amended by 2% of one of the highly condensed matrices, thoroughly mixed and allowed to age for three additional months. A similar sub-sampling procedure was applied to St Cyprien soil to obtain 3 aliquots. A similar proportion of the two most efficient matrices from the first step was amended. Composition of each soil set and the NDL-PCBs concentrations are presented in Table 1.

Environmental Availability Assays
Using a balance, a 4 g of test sample of fresh soil was placed in a 50 mL falcon tube. Then a 40 mL volume of a 100 mmol·L −1 Hydroxypropyl-β-cyclodextrin (Kleptose HP oral grade, Roquette, France) extraction solution was added to the tube and shaken horizontally for 20 h (180 rpm, 20 ± 2 • C). Samples were centrifuged at (3000× g for 15 min) to separate the solid phase from the liquid phase (clear supernatant). Supernatant corresponding to the available fraction was extracted 3 times using 20 mL of petroleum ether. The collected petroleum ether was then evaporated until dry and 150 µL of nonane was added to recover analytes. Extracts were quantified as described in [11]. For each soil, 4 replicates were performed.

Plants Cultivation of C. pepo
Seeds of C. pepo (Diamant cultivar.) were sown in cups of potting soil and then transplanted at the 1st-leaf stage into 15-cm-diameter pots containing either OECD or historically contaminated soils corresponding to the different tested modalities. Each pot contained between 400 and 500 g of soil and one plant. Plants were grown in a greenhouse under these following controlled environmental conditions cooling above 20 • C. All pots were adjusted each week to 80% of the water holding capacity. After 12 weeks, all aerial parts were harvested, weighted and frozen at −20 • C before analysis.

Analytical Processes
All chemicals used during the analytical processes were of Pesticide Residues grade and glassware equipment was cleaned with an alkaline detergent (RBS T 105, Carl Roth GmbH, Karlsruhe, Germany).

Extraction Processes of Plant Samples
Biological matrices were extracted by Pressurized-Liquid Extraction (PLE) using a Büchi E-916 Speed-extractor with pure toluene, a temperature of 50 • C, a pressure of 135 Bars and 3 extraction cycles of 5 min.

Purification
Step Potential chromatographic interferences were removed from extracts using Multi-layer silica column (28397-U, Merck KGaA, Darmstadt, Germany). The silver nitrate treated layer removes sulphur-containing compounds; whilst two sulphuric acid treated layers oxidise sample lipids and remove any basic analytes. The potassium hydroxide treated layer removes any acidic sample components. NDL-PCB pass through the silica column unretained. Each extract was evaporated until dryness and dissolved in nonane.

Quantitation Processes
Quantitation of NDL-PCBs was performed on the extracts of the soil environmental availability and plants using a Gas Chromatography (Agilent 6890N, Santa Clara, CA, USA)-High Resolution Mass Spectrometry (Quattro Ultima, Waters, Milford, CT, USA) device in Laboratoire d'Analyse des Sols (Arras, Hauts de France, France). Quantitation was performed by selecting the two most abundant molecular ions for each organic compound, by isotopic dilution and at a resolution close to 10,000. The determination was performed in SIR (Single Ion Recording) mode and ionization by electron impact in low energy positive mode (35.5 eV) and with a trap current between 400 and 600 µA. Program of the column (DB-5MS 60 m, 250 µm internal diameter, 0.25 µm active phase thickness, Agilent J&W, Santa Clara, CA, USA) consisted of successive temperature gradients: 100 • C (1 min); 40 • C·min −1 until 200 • C (6 min); 3 • C·min −1 until 235 • C (10 min); 8 • C·min −1 unti 315 • C (9 min). Temperature of the transfer line was set to 250 • C.

Quality Control of the Analytical Processes
Quantitation was performed using isotopic dilution by adding 13 C internal standards (EC9605-SS, Wellington Laboratories, Guelph, ON, Canada) as described in Environment Canada Method 1/RM/31. Ratio of the molecular ions area of the parent compounds over internal standards of the same chlorine number allowed the quantitation after calibration (Regression coefficient R 2 > 0.99). In addition, perfluorokerosene (lock mass) was added allowing mass calibration of the mass spectrometer during the whole analysis. Recoveries of the whole analytical steps above 70% were obtained throughout analytical runs using recovery standards (EC9605-RS, Wellington Laboratories, Guelph, ON, Canada). In each run, chromatographic and method blanks were introduced showing no cross-contamination (all < LOQs). A quality control using an independent standard solution (ECPCS3 at 200 µg·L −1 , Wellington Laboratories, Guelph, ON, Canada) was introduced showing a <20% deviation. Limits of quantitation for 2 g of soil or plants are 0.2 µg·kg −1 for each NDL-PCB. Each step of the analytical process was performed as documented in the COFRAC quality accreditation of LAS laboratory for the NDL-PCBs quantitation of soil samples.

Statistical Analyses
All analyses were carried out using R software (version 4.0.2, R Foundation for Statistical Computing, Vienna, Austria). In order to assess the impact of BCs and ACs on pollutant concentrations in both assays from both steps, a Dunnett test (using the package multcomp, Hothorn 2020, version 2) was used in order to compare reductions of NDL-PCBs concentrations against their respective non-amended control (SS) (n = 3 for environmental availability and n = 5 for plant assay). Then, the ANOVA procedure followed by the Tukey-Kramer post-hoc test (using the package agricolae, Mendiburu 2020, version 1.3.3) were used in order to compare the efficiency of matrices on the groups found significant in the previous test. Differences were considered significant at p < 0.05.

Calculations of Relative Environmental Availability and Bioavailability Factors
For each pollutant, relative environmental availability and relative bioavailability factors were calculated as the ratio of each pollutant's concentration from tested amended soil compared to their respective non-amended SS soil as described previously [25].

Test on OECD Soils
A first set of experiments was performed on OECD artificial soil in order to select the most efficient matrices upon their ability to sequestrate NDL-PCBs.

Environmental Availability
The in vitro assay performed on artificial soils demonstrated the efficiency of the amendment strategy to reduce the NDL-PCBs leaching from soil. Indeed, for PCB 101, 138, 153 and 180, environmental availabilities concentrations from amended soils were significantly lower than those from the non-amended control group (Figure 1), excluding biochar-amended groups for PCB 180. PCB 101 appeared the most susceptible compound to this strategy. Indeed, contrarily to other congeners, even biochar amended soils presented over 50% reduction (62 to 80%, mean) of environmental availability concentrations ( Figure 1). Such amendment led to significant but lower reduction for the three most lipophilic compounds: 35 to 42% (mean) for PCB 138, 37 to 42% for PCB 153 and 27 to 33% for PCB 180. As expected, NDL-PCBs were far more sequestrated when AC was amended but contrasted results were obtained from granular AC1 compared to powdered AC2, AC3. Indeed, AC1 displayed the lowest reduction potential among ACs as no significant distinction was found between this AC and BC amendment (p > 0.10). In contrast, AC2 and AC3 amendment presented the most intense sequestration for all congeners as environmental availability levels were similar to the control levels, i.e., the analytical quantitation limit for each congener. Due to these analytical limits, only minimal reduction factors could be derived from this assay: 79% for PCB 101, 91% for PCB 138, 88% for PCB 153 and 45% for PCB 180. Thus, from this environmental availability assay, the following efficiency hierarchy (reduction of transfer potential) was found between the used sequestration materials: BC1 = BC2 ≤ BC3 = AC1 << AC2 = AC3.

Cucurbita pepo Assay
The in vivo assay on C. pepo was performed on the same soils from the in vitro assay. Due to technical difficulties inherent to the experimental design (weak soil quantity in each pot), some repetition did not grow properly. Thus, 4 repetitions were obtained from BC2 and BC3 groups, whereas only 1 repetition was obtained from AC3 group instead of the 5 repetitions sown for each modality. As the characteristics of the AC amendment could also strongly influence nutrient availability and water retention, they can interfere with plant development. Considering transfer, and similarly to previous environmental

Cucurbita pepo Assay
The in vivo assay on C. pepo was performed on the same soils from the in vitro assay. Due to technical difficulties inherent to the experimental design (weak soil quantity in each pot), some repetition did not grow properly. Thus, 4 repetitions were obtained from BC2 and BC3 groups, whereas only 1 repetition was obtained from AC3 group instead of the 5 repetitions sown for each modality. As the characteristics of the AC amendment could also strongly influence nutrient availability and water retention, they can interfere with plant development. Considering transfer, and similarly to previous environmental availability results, BCs and AC1 were the less effective materials. Results were even more contrasted as for all studied NDL-PCBs no significant reductions were found for these materials compared to the SS control group (Figure 2). At the opposite, AC2 and AC3 were also the most effective matrices and interestingly, levels of reduction found were greater than previous measurements as the analytical quantitation limits were far lower than concentrations found in the non-amended control group. Thus, the minimal reduction factors which could be derived from this assay were the following: 93% for PCB 101, 95% for PCB 138, 95% for PCB 153 and 92% for PCB 180. Thus, from this in vivo assay, the following efficiency hierarchy was found between the used sequestration materials: BC1 = BC2 = BC3 = AC1 << AC2 = AC3.
Appl. Sci. 2022, 12, x FOR PEER REVIEW 7 of 1 availability results, BCs and AC1 were the less effective materials. Results were even more contrasted as for all studied NDL-PCBs no significant reductions were found for these materials compared to the SS control group (Figure 2). At the opposite, AC2 and AC3 were also the most effective matrices and interestingly, levels of reduction found were greater than previous measurements as the analytical quantitation limits were far lowe than concentrations found in the non-amended control group. Thus, the minimal reduc tion factors which could be derived from this assay were the following: 93% for PCB 101 95% for PCB 138, 95% for PCB 153 and 92% for PCB 180. Thus, from this in vivo assay, the following efficiency hierarchy was found between the used sequestration materials: BC1 = BC2 = BC3 = AC1 << AC2 = AC3.

Selection of the Sequestering Materials
Similar efficiency trends were found between both assays showing that AC2 and AC3 were the two most efficient media. These two media were used as amendment sources fo the St Cyprien contaminated soil. However, as AC3 may also present a negative effect on

Selection of the Sequestering Materials
Similar efficiency trends were found between both assays showing that AC2 and AC3 were the two most efficient media. These two media were used as amendment sources for the St Cyprien contaminated soil. However, as AC3 may also present a negative effect on the C. pepo growth, AC2 only was assessed in vivo to validate this amendment strategy on an historically contaminated soil.

Environmental Availability
For the first experiment on an historically contaminated soil, the environmental availability in vitro assay confirmed an important sequestration efficiency for NDL-PCBs 138, 153 and 180 ( Figure 3). As all results concerning PCB 101 were below LOQs, including non-amended soil (SS), no sequestration potential could be derived for this specific compound. For the three most lipophilic compounds, intense reductions were obtained as all environmental values except one (PCB 138, AC2 amendment) were below LOQs. Thus, high levels of reduction were obtained and a minimum reductions factor of (i) 83 to 90% (mean) for PCB 138, (ii) 91% for PCB 153 and (iii) 27 to 44% for PCB 180 were obtained ( Table 2). These results provided evidence that even in a known historic contamination context (over a decade between contamination context and sampling of the soil) such a strategy could efficiently limit the environmental availability of these pollutants.
Appl. Sci. 2022, 12, x FOR PEER REVIEW 8 of 1 the C. pepo growth, AC2 only was assessed in vivo to validate this amendment strategy on an historically contaminated soil.

Environmental Availability
For the first experiment on an historically contaminated soil, the environmental avail ability in vitro assay confirmed an important sequestration efficiency for NDL-PCBs 138 153 and 180 ( Figure 3). As all results concerning PCB 101 were below LOQs, including non-amended soil (SS), no sequestration potential could be derived for this specific com pound. For the three most lipophilic compounds, intense reductions were obtained as al environmental values except one (PCB 138, AC2 amendment) were below LOQs. Thus high levels of reduction were obtained and a minimum reductions factor of (i) 83 to 90% (mean) for PCB 138, (ii) 91% for PCB 153 and (iii) 27 to 44% for PCB 180 were obtained ( Table 2). These results provided evidence that even in a known historic contamination context (over a decade between contamination context and sampling of the soil) such a strategy could efficiently limit the environmental availability of these pollutants.

Cucurbita pepo Assay
The second set of experiments on AC2 amended soil was performed on Cucurbita pepo plantations (n = 4) (Figure 4). Similarly, to environmental availability results, high reduction of NDL-PCBs transfer was observed for amended soil, except for PCB 101 for which all results, even for non-amended soil, were below LOQs. Minimal reduction levels derived from this assay were the following: (i) 94% (mean) for PCB 138, (ii) 96% for PCB 153 and (iii) 27 to 90% for PCB 180, confirming the reduction levels found on OECD soils.

In Vitro Assessment of Interactions between Pollutants and Carbonaceous Materials
The environmental availability test was used to express the portion of the pollutant that is desorbed from the soil and available in the soil solution for interception and uptake by plant roots. Considering this test, overall, BCs were less effective than ACs in trapping PCBs due to their limited microporous and mesoporous surfaces. The efficiency was of the same order in both soils, suggesting that the physicochemical properties of the soils did not influence this efficiency and that the OECD artificial soil realistically mimics the behavior of natural soil. There was no impact of biochar type regardless of the PCB considered, suggesting that the nature of the biomass and pyrolysis temperature did not influence sequestration potential even though BC1 had a 10-fold lower specific surface area than the other two.
For AC materials, AC2 and AC3 were the most effective in reducing PCB concentrations given their larger specific surface area and mesoporous surface, while AC1 was less effective due to its granular shape. This result confirms that surface properties and textural characteristics impact the sorption properties of carbonaceous materials in contaminated soils [11,12,14,26,27]. Regarding the type of pollutants, no difference could be shown in terms of sequestration efficiency by the different carbonaceous materials probably because the physicochemical properties related to adsorption in soil such as lipophilicity are of the same order (about log Kow of 7).

In Vitro Assessment of Interactions between Pollutants and Carbonaceous Materials
The environmental availability test was used to express the portion of the pollutant that is desorbed from the soil and available in the soil solution for interception and uptake by plant roots. Considering this test, overall, BCs were less effective than ACs in trapping PCBs due to their limited microporous and mesoporous surfaces. The efficiency was of the same order in both soils, suggesting that the physicochemical properties of the soils did not influence this efficiency and that the OECD artificial soil realistically mimics the behavior of natural soil. There was no impact of biochar type regardless of the PCB considered, suggesting that the nature of the biomass and pyrolysis temperature did not influence sequestration potential even though BC1 had a 10-fold lower specific surface area than the other two.
For AC materials, AC2 and AC3 were the most effective in reducing PCB concentrations given their larger specific surface area and mesoporous surface, while AC1 was less effective due to its granular shape. This result confirms that surface properties and textural characteristics impact the sorption properties of carbonaceous materials in contaminated soils [11,12,14,26,27]. Regarding the type of pollutants, no difference could be shown in terms of sequestration efficiency by the different carbonaceous materials probably because the physicochemical properties related to adsorption in soil such as lipophilicity are of the same order (about log Kow of 7).

Effect of Plant on the Interactions between Pollutants and Carbonaceous Materials
The former environmental availability assay reflects the ability of BC and AC to increase pollutant sorption on these materials, but it did not consider plant physiological processes that may influence this interaction. Among plants, it is well known that the Cucurbita family accumulate considerable amounts of POPs in their shoots [28]. For this reason, C. pepo was chosen in this study because it highly extracts and accumulates POPs including PCB from contaminated soils [2,[29][30][31][32][33]. Plant is a major driver of pollutants availability through a variety of root functions such as uptake, exudation that can alter biogeochemical parameters of the soil (e.g., pH and redox potential, concentrations of pollutants, complexing or chelating compounds) in the vicinity of the roots [34]. Roots release protons, responsible for a substantial decrease of rhizosphere pH, that could in turn change the sorption of PCB congeners on soil substrate that is maximum between pH 6.5 and 7.5 [35]. Root exudates like organic acids also play a great role on the sorption/desorption of pollutants from soil components and organo-mineral complexes and subsequently on their availability for plant uptake ( Figure 5). The amendment of soil with citric acid that is strongly exuded in Cucurbita pepo ssp. pepo allowed a significant increase of shoot PCB concentration [36] probably due to PCB desorption [32] and/or to root absorption of PCB via complexation [7]. POPs are transported from roots to shoots via the xylem sap [5,37,38] and other transfer pathways to shoots are negligible such as direct soil contamination, atmospheric deposition and volatilization from soil and subsequent redeposition on shoots [39]. Our results showed that only the addition of activated carbons to the soil decreases significantly POPs transfer to shoot and thus their extraction from soil by C. pepo roots. Previous studies also showed a reduction of PCB concentration in C. pepo shoot tissue: 63% of PCB (Aroclor 1254) with 12.5% AC amendment [18] and 54% of PCB with 11.1% BC amendment [16]. Furthermore, the addition of AC (0.1 and 0.23%; w/w) reduced from 73% to 98% of heptachlor epoxide in the shoot of winter squash (Cucurbit maxima) [40]. AC amendment (0.02 to 0.08%) also reduced up to 66% of the dieldrin concentration in cucumbers (Cucumis sativus L.) [17].

Effect of Plant on the Interactions between Pollutants and Carbonaceous Materials
The former environmental availability assay reflects the ability of BC and AC to increase pollutant sorption on these materials, but it did not consider plant physiological processes that may influence this interaction. Among plants, it is well known that the Cucurbita family accumulate considerable amounts of POPs in their shoots [28]. For this reason, C. pepo was chosen in this study because it highly extracts and accumulates POPs including PCB from contaminated soils [2,[29][30][31][32][33]. Plant is a major driver of pollutants availability through a variety of root functions such as uptake, exudation that can alter biogeochemical parameters of the soil (e.g., pH and redox potential, concentrations of pollutants, complexing or chelating compounds) in the vicinity of the roots [34]. Roots release protons, responsible for a substantial decrease of rhizosphere pH, that could in turn change the sorption of PCB congeners on soil substrate that is maximum between pH 6.5 and 7.5 [35]. Root exudates like organic acids also play a great role on the sorption/desorption of pollutants from soil components and organo-mineral complexes and subsequently on their availability for plant uptake ( Figure 5). The amendment of soil with citric acid that is strongly exuded in Cucurbita pepo ssp. pepo allowed a significant increase of shoot PCB concentration [36] probably due to PCB desorption [32] and/or to root absorption of PCB via complexation [7]. POPs are transported from roots to shoots via the xylem sap [5,37,38] and other transfer pathways to shoots are negligible such as direct soil contamination, atmospheric deposition and volatilization from soil and subsequent redeposition on shoots [39]. Our results showed that only the addition of activated carbons to the soil decreases significantly POPs transfer to shoot and thus their extraction from soil by C. pepo roots. Previous studies also showed a reduction of PCB concentration in C. pepo shoot tissue: 63% of PCB (Aroclor 1254) with 12.5% AC amendment [18] and 54% of PCB with 11.1% BC amendment [16]. Furthermore, the addition of AC (0.1 and 0.23%; w/w) reduced from 73% to 98% of heptachlor epoxide in the shoot of winter squash (Cucurbit maxima) [40]. AC amendment (0.02 to 0.08%) also reduced up to 66% of the dieldrin concentration in cucumbers (Cucumis sativus L.) [17]. It may be hypothesized that reduction of PCB concentration in shoot tissue is due to the sorption of PCB into the surface of activated carbon, which prevent these molecules from being available in the soil pore water and absorbed into the root, and thus translocated from the root to shoot tissue. This assay involves the use of a sorbent phase acting as an infinite sink for pollutants. Concerning the in vivo assays, the present study used C. pepo as a model plant which extracts pollutants from the water solution and accumulates them in aerial parts.

Conclusions
Sequestration strategy involving high carbonized material appeared to significantly reduce the transfer of NDL-PCBs to Curcubita pepo aerial parts, appearing promising to ensure vegetable cultivation. However, no significant reduction of transfer was observed when BC was used, few when granular AC was amended, showing that only powdered AC may be suitable to that purpose. In that particular case, reduction levels above 90% have been observed from these particularly micro-and mesoporous media. Regardless of the type of soil, in vivo and in vitro experiments showed a noticeable similarity of responses leading to consider that (i) the congener effect (ii) the nature of the sequestering material were the main parameters playing on the efficiency of such an amendment strategy. The convenient in vitro assay showed an interesting correlation needing to be further assessed to predict in vivo transfer to Cucurbita pepo plants. Such an assay appears as a promising tool in the frame of testing before the in situ application of such materials.