1 Selective removal of chlorinated organic compounds from lindane wastes by combination of nonionic surfactant soil flushing and Fenton oxidation Carmen M. Dominguez*, Arturo Romero, Aurora Santos Dpto. Ingeniería Química y de Materiales, Facultad de Ciencias Químicas, Universidad Complutense Madrid. Ciudad Universitaria S/N. 28040, Madrid, Spain Paper submitted to “Chemical Engineering Journal” for consideration "Special issue on emerging advanced oxidation technologies and developing perspectives for water and wastewater treatment" 2 Abstract The extensive use of the organochlorine pesticide lindane in the second half of the 20th century generated large volumes of wastes over the world. Among these wastes, a dense non-aqueous phase liquid (DNAPL), mainly composed of chlorobenzenes, hexachlorocyclohexanes and heptachlorocyclohexanes, was dumped in insecure landfills remaining in the subsurface and contaminating the groundwater. A coupled process, combining soil flushing (with a nonionic surfactant) and Fenton oxidation, was proposed to deal with this problem. A commercial surfactant (E-Mulse 3 ®) was used to extract most of the residual DNAPL in soil at column conditions. The resulting surfactant flushing solution (SFS) presented high concentration of chlorinated organic compounds (COCs = 3693 mg L-1). In order to recover the surfactant and abate the COCs, the SFS was treated by Fenton process using three doses of hydrogen peroxide (200%, 100% and 50% of the theoretical stoichiometric amount for the complete mineralization of COCs; maintaining a molar ratio of H2O2:Fe = 32). Conversions of COCs above 80% were obtained when H2O2 doses of 100% and 200% of the stoichiometric amount were used at 144 h and 48 h, respectively. Non-aromatic compounds resulted to be less prone to oxidation by hydroxyl radicals than chlorobenzenes. The oxidation of the surfactant was significantly lower than that of the pollutants; therefore the surfactant capacity was maintained after the oxidation treatment and it could be reused in further flushing steps, improving the economy of the process. Keywords: Chlorinated organic compounds, Dense-non aqueous phase liquid, Lindane, Fenton process, Surfactant enhanced aquifer remediation, Soil flushing. 3 1. Introduction The organochlorine pesticide lindane (also known as γ-HCH or γ- hexachlorocychlohexane) was manufactured in many European countries, such as Czech Republic, Spain, France, Germany, United Kingdom, Italy and Poland, from the 1950s to the 1990s, while in others like Romania, the production continued until 2006, as recently reported by the Directorate General for Internal Policies (EU) [1]. In Spain there were four production sites dedicated to the production of this pesticide [1, 2] being the company INQUINOSA, located in Sabiñánigo (Huesca), the largest one. Due to the persistency in the environment, bioaccumulation, harmful impacts on human health and high potential for long range transport of HCH isomers [3-5], some of them (,  and -HCH) have been included in the list of persistent organic pollutants by the Stockholm Convention [6]. Apart from solid wastes (mixture of HCH isomers, called technical-HCH), liquid wastes from the chlorination failed reactions and distillation tails, were obtained. The liquid wastes were mainly composed by HCH isomers, chlorobenzenes and solvents [2, 7], and they were usually dumped in insecure landfills, generating a dense non-aqueous phase liquid (DNAPL) that migrated through the subsurface causing both soil and groundwater contamination. Due to the chlorinated nature of the compounds that make up this phase, the DNAPL is denser than water (1.5 g mL-1) and immiscible, so tends to sink until reach impermeable bedrock. When the DNAPL is in contact with groundwater, a plume of contamination is generated. The treatment of the groundwater pumped from Sabiñanigo landfills (supplied by the Aragon Government) has been recently addressed by different oxidation treatments with successful results [8-10]. However, as long as there is DNAPL in the subsoil, it will continue to contaminate the groundwater and therefore, its abatement should be a priority. 4 The injection and further extraction of a surfactant solution to enhance the removal of NAPLs in the subsurface, known as SEAR process (Surfactant Enhanced Aquifer Remediation) has received increasing attention in the last years [11-15]. Surfactants are a kind of natural or synthetic compounds that promote the wetting, solubilization and emulsification of organic and inorganic pollutants [16]. Surfactants are amphiphilic substances with hydrophilic and hydrophobic parts that can reduce the surface tension of water (from 74 to approximately 25 ± 5 mN m-1) [16], making them ideal for solubilization of hydrophobic compounds. Nonionic surfactants are particularly used for soil flushing (in situ process where extracting agents are used in order to improve the mobility of NAPLs by reducing the interfacial tension (IT) between NAPL and groundwater [15]) due to their high solubilization capacity and low toxicity [17]. Moreover, they usually present lower critical micelle concentrations (CMC, the minimal concentration of surfactant at which the micelle formation occurs) than ionic surfactants [13], meaning that nonionic surfactant maintain their surfactant capacity when working at relatively low concentrations. Although it has been proved that the soil flushing process using surfactants as extracting agent is an efficient process to remove residual DNAPLs, the pollutants are not destroyed and further treatments are required for this scope. Therefore, the development of an efficient post-treatment for the extracted surfactant flushing solution (SFS) is one of the main requirements for the application of a SEAR treatment. Besides, the recovery of the surfactant is a key factor from an economical point of view [15]. There are relatively few reported studies concerning the coupling of soil flushing with surfactants and a post treatment of the resulting polluted solution [12, 18-24]. In most of these works synthetic organic phases were used as source of contamination or only a single organic compound (in relatively low concentration) was analyzed in the SFS. In the 5 present work we apply the Fenton oxidation to the treatment of a SFS containing a complex mixture of COCs (28 chlorinated compounds present in a DNAPL extracted from Sabiñanigo´s landfill) at high concentration. The selective oxidation of COCs and the eventual recovery of the surfactant capacity after the Fenton treatment will be studied. Fenton reagent has been chosen as oxidation post-treatment due to its high efficiency in the degradation of organic matter, operational-simplicity and low cost [25, 26]. The ferrous iron (Fe2+) initiates and catalyzes the decomposition of hydrogen peroxide, resulting in the generation of hydroxyl radicals, with high oxidizing power [27]. H2O2 + Fe2+ → •OH + OH– + Fe3+ (1) Fe3+ + H2O2 → Fe2+ + HO2 • + H+ (2) From the best of our knowledge this is the first time that the Fenton treatment has been successfully applied to the selective abatement of such a complex mixture of COCs, obtained by flushing a soil contaminated with a real DNAPL (coming from lindane wastes) using a nonionic surfactant. 2. Materials and Methods 2.1 Reagents Working standard solutions of the HCH-isomers and chlorobenzenes (Sigma-Aldrich) were prepared for their calibration by gas chromatography. Other reagents used in the present work (all of them of analytical grade) were: iron (II) sulfate (Fe2SO4), hydrogen peroxide, titanium oxisulfate, methanol, butyl cyclohexane (C10H20) and tetrachloroethane (C2H2Cl4) (Sigma-Aldrich). The solutions were prepared using high- 6 purity water obtained from a Millipore Direct-Q system (resistivity >18 MΩ cm at 25 ºC). 2.2 DNAPL from the Bailin’s landfill The DNAPL used in the present work was extracted from Bailin´s landfill (Sabiñánigo) and it was kindly supplied by the Aragon Government. The exact location of the extraction well, as well as a photograph of the DNAPL can be found elsewhere [7]. The DNAPL consists in a viscous black-brown liquid, with an unpleasant odor and a density of 1.5 g mL-1. Once extracted, the DNAPL was immediately stored and preserved in hermetic glass bottles at 4 °C until its use. 2.3 Surfactant The surfactant selected for the soil flushing step was E-Mulse 3 ® from EthicalChem ®. It is slightly viscous, with yellowish color and citrus odor. E-Mulse 3 ® is highly stable under normal conditions and soluble in water. It is composed of nonionic surfactants (60-90%) combined with citrus terpenes (10-40%), more specifically, limonene (CAS: 94266-47-4). The combined use of surfactants with other additives, such as organic solvents, chelating agents, salts or ligand ions has demonstrated a stronger capability to remove soil pollutants than the original surfactants [13, 16]. E-Mulse 3 ® has been selected because of its biodegradable and non-toxic nature. More information about this product can be found in www.ethicalchem.com. 2.4 Soil flushing experiment A methacrylate column (37 cm length x 3.7 cm internal diameter) was filled with 306.8 g of sand (Sigma-Aldrich), 60 g of marls from the polluted site and 7.31 g of DNAPL http://www.ethicalchem.com/ 7 without further preparation (a porosity value Ɛ = 0.3 was estimated). Firstly, the column was flushed with tap water (Q = 0.2 mL min-1) during 24 h in order to reach pore water saturation. For that purpose a positive displacement pump (SIMDOS ®) was used. The obtained eluate, named water flushing solution (WFS) was collected and analyzed. The surfactant solution was stored in a glass bottle and pumped upwards through the column at Q = 0.2 mL min-1 during 350 h. Firstly, with the purpose of maximizing the solubilization of pollutants, an adequate initial concentration of surfactant was selected (15 g L-1) since it has been demonstrated that the pollutants solubility increases with the concentration of the surfactant [17, 28]. This value is in the range commonly found in the literature for soil flushing experiments using nonionic surfactants [29-31]. Secondly, in order to ensure that the contact time between the surfactant solution and the DNAPL was enough, the soil flushing was performed at a linear velocity of 7 cm day-1. The column eluate, named soil flushing solution (SFS) was collected in successive fractions and periodically analyzed. The total volume of SFS was 4.2 L; it was stored at 4 ºC until its treatment by Fenton process. 2.5 Fenton experiments The oxidation runs were carried out at ambient conditions (25 ºC, 1 atm) in cylindrical glass reactors (2 mL) with rotatory agitation (80 rpm). The reagents (H2O2 and Fe2SO4) were simultaneously added at the beginning of each experiment at the appropriate concentration. The initial pH of the reaction medium was not adjusted (pH0  6.5). In order to follow the evolution of the reaction, several reactors were prepared and placed simultaneously and sacrificed for each reaction time. 8 The theoretical stoichiometric dose of hydrogen peroxide can be calculated by using the proposed oxidation reactions (Eqs. 3 to 11). Doses of hydrogen peroxide of 50%, 100% and 200% of the theoretical stoichiometric amount for the complete mineralization of the COCs (information of these compounds is collected in Table 1) to carbon dioxide, water and chloride, were tested, which correspond to 5, 10 and 20 g L-1, respectively. As can be seen in Eqs 3 to 11, the oxidation valence of chloride does not change after pollutant oxidation whereas the carbon skeleton of the organic molecule is oxidized to carbon dioxide and water. 2 C6H5Cl + 29 H2O2 → 12 CO2 + 34 H2O + 2 Cl– (Eq. 3) C6H4Cl2 + 14 H2O2 → 6 CO2 + 16 H2O + 2 Cl– (Eq. 4) 2 C6H3Cl3 + 27 H2O2 → 12 CO2 + 30 H2O + 6 Cl– (Eq. 5) C6H2Cl4 + 13 H2O2 → 6 CO2 + 14 H2O + 4 Cl– (Eq. 6) 2 C6HCl5 + 25 H2O2 → 12 CO2 + 26 H2O + 10 Cl– (Eq. 7) 2 C6H5Cl5 + 29 H2O2 → 12 CO2 + 34 H2O + 10 Cl– (Eq. 8) C6H4Cl6 + 14 H2O2 → 6 CO2 + 16 H2O + 6 Cl– (Eq. 9) 2 C6H5Cl7 + 29 H2O2 → 12 CO2 + 34 H2O + 14 Cl– (Eq. 10) C6H6Cl6 + 15 H2O2 → 6 CO2 + 18 H2O + 6 Cl– (Eq. 11) The oxidation experiments were performed by duplicate, the standard deviation being lower than 5% in all cases. 2.6. Analytical methods 9 The identification and quantification of COCs was carried out by gas chromatography analysis coupled with flame ionization and electron capture detectors (GC-FID/GC- ECD). The exit of the column (HP-5-MS, 30 m × 0.25 mm i.d., 5% phenol methyl siloxane) was split to both detectors, FID and ECD, and simultaneously measured. Helium was used as carrier gas (2.8 mL min-1). The GC injection port temperature was set at 180 °C (the injection volume was 2 µL). The program temperature started at 80 °C followed by a temperature ramp (10 °C min-1) to 310 °C and then held constant for 1 min. Samples were diluted with methanol before the analysis. In order to minimize experimental error in COCs quantification, butyl cyclohexane and tetrachloroethane were used as internal standard compounds (ISTDs) for the FID and ECD analyses, respectively. Gas chromatography coupled with a mass spectrometric detector (GC-MS) was used for the identification of the oxidation intermediates generated during Fenton experiments, both coming from the COCs and the surfactant. For this purpose, a gas chromatograph (Agilent 6890N) with a mass spectrometric detector (HP5973) was used. The chromatograph was also equipped with a HP-5-MS column and a CTC CombyPAL sampler. The concentration of hydrogen peroxide was determined by colorimetric titration with a BOECO S-20 UV-VIS spectrophotometer at 410 nm [32] (Eisenberg, 1943), while the pH evolution was measured using a Basic 20-CRISON pH electrode. Carboxylic acids were measured by ionic chromatography (Metrohm 761 Compact IC) with anionic chemical suppression and a conductivity detector. A Metrosep A SUPP5 5-250 column (25 cm length, 4 mm diameter) was used as stationary phase and an aqueous solution (0.7 mL min-1) of Na2CO3 (3.2 mM) and NaHCO3 (1 mM) as mobile phase. The 10 injection system (injection volume = 250 L) was coupled to an online filtering system (0.45 m). Chlorides in solution were also measured by IC analysis. The interfacial tension (IT) and the critical micelle concentration (CMC) of the samples were measured using a Krüss tensiometer. The parameter defined in this work as the equivalent surfactant concentration (ESC) was determined by using the same experimental device and applying a novel method based on progressive dilutions of the surfactant-containing samples until obtain a value of IT slightly higher than the corresponding one to the CMC of the surfactant (IT = 34 mN m-1 at concentrations  85 mg L-1 of E-Mulse 3 ®). This value (multiplied by the dilution factor applied) indicates the surfactant capacity of the sample, including the contribution of the E-Mulse 3 ®, as well as the possible surfactant oxidation products generated during Fenton treatment that also have surfactant capacity. Moreover, the concentration of limonene, the co- solvent included in the commercial surfactant E-Mulse 3 ®, representing around 15% of the total commercial surfactant mass, was followed by GC-FID. 3. Results 3.1 Soil flushing experiment A total volume of 4.2 L of SFS was collected in 12 fractions. Taking into account the COCs concentration and the volume collected of each fraction, a total of 6160 mg of COCs were extracted from the column, which represents more than 85% of the DNAPL mass initially introduced. This fact highlights the great ability of the surfactant to solubilize the DNAPL. A SFS fraction with a representative concentration of COCs (the obtained one between 50 and 125 h of the soil flushing experiment) was selected to 11 carry out the second step of the process: the degradation of COCs and the recovery of the surfactant by Fenton oxidation. The GC-FID chromatogram obtained for the selected SFS is shown in Fig. 1 (the corresponding one to t = 0 h). As can be observed, 28 COCs have been identified, including chlorobenzene (CB), three dichlorobenzenes (DiCBs), three trichlorobenzenes (TriCBs), three tetrachlorobenzenes (TetraCBs), pentachlorobenzene (PentaCB), five pentachlorocyclohexenes (PentaCXs), six hexachlorocyclohexenes (HexaCXs), three heptachlorocyclohexanes (HeptaCHs) and five hexachlorocyclohexanes (HCH). The total concentration of COCs is 3639 mg L-1. Information regarding these compounds (the chemical formula, the acronyms used and the concentration values of each COC, expressed in mg L-1) is collected in Table 1. The same compounds were detected when the soil flushing was performed with tap water (WFS). The concentration of these compounds in WFS has been also included in Table 1 for comparative purposes. The total concentration of COCs in WFS was 68.2 mg L-1, meaning that the addition of the surfactant greatly increased the solubility of the DNAPL (by more than 50 times). When the surfactant molecules are present in the water-soil heterogeneous system (at an appropriate concentration, higher than the CMC) ellipsoidal or spheroidal micelles are formed [13]. These micelles, with hydrophilic surfaces and lipophilic cores, disperse the contaminants and dramatically improve their solubility in water phase. On the other hand, the distribution of COCs was also highly affected by the presence of the surfactant, as can be seen in Fig. 2, where the percentage of each group of COCs is represented both in WFS and SFS. The concentration of COCs in WFS is determined by the solubility of these compounds in water and the composition of the organic phase. The chlorinated compounds constituting the DNAPL are hydrophobic and therefore, 12 very poorly soluble in water. Their hydrophobic character generally increases with the chlorine content of the molecule. Thus, chlorobenzene was the compound showing the highest concentration (practically 50% of the total concentration of COCs) in WFS. However, the distribution of COCs in a surfactant-containing solution was mainly determined by the composition of the original DNAPL. In this sense, the distribution of COCs obtained in SFS agrees quite well with the composition of the DNAPL, which was detailed characterized in a previous work by its complete dissolution in acetone [7]. The corresponding results have been also included in Fig 2. Moreover, as a consequence of the flushing process, part of the surfactant was adsorbed onto the column (the ESC of the SFS was reduced by 25% compared with the fresh surfactant solution, from 15 to 11 g L-1), which is pretty common in flushing processes [13, 28]. Moreover, it has been demonstrated that the sorption of surfactants is also favored in the presence of organic pollutants [28]. 3.2 Fenton system activity The time-evolution of hydrogen peroxide and COCs conversion under the selected operating conditions is shown in Figs. 3a and 3b, respectively. As the oxidizable organic matter in the reaction medium is comprised of COCs as well as the surfactant, the doses of hydrogen peroxide commonly used in these systems are relatively high [15, 33]. Therefore, considering the possible unproductive consumption of hydroxyl radicals by the surfactant, a dose of hydrogen peroxide of 20 g L-1, which corresponds to twice the theoretical stoichiometric amount for the complete mineralization of the chlorinated pollutants (Eqs 3-11), was selected. A control experiment was firstly undertaken to verify the stability of COCs in SFS during the experimental time (Blank-SFS). When H2O2 was added to the reaction medium, in the absence of catalyst (Blank-SFS-H2O2), 13 there was no consumption of this reagent during the experimental time and therefore, the degradation of COCs was negligible. Similar results were obtained when only Fe2+ was added (Blank-SFS-Fe). In this case, a small decrease in the pH was observed (from 6.5 in the starting solution to pH 4 after 144 h), but this change did not affect the concentration or distribution of COCs in solution. However, the H2O2-Fe system (Fenton’s reagent) led to a gradual decomposition of H2O2 and consequently, to an important COCs degradation. As observed in Fig. 1, after 5 and 29 h reaction time, the chromatographic peaks corresponding to the identified COCs were significantly reduced, achieving a global conversion of COCs above 80% at 144 h reaction time (Fig. 3b). These results demonstrated the high efficiency of Fenton oxidation in the degradation of such recalcitrant compounds even in the presence of high concentrations of surfactant. It has been demonstrated that the presence of high concentrations of surfactant in the reaction medium highly decreases the efficiency of the process [34, 35] since i) surfactant micelles act as a protective environment strongly reducing the availability of COCs towards oxidant species [36] and ii) surfactant compete with target pollutants consuming hydroxyl radicals [15, 34]. Moreover, it should to be also taken into account that the treatment of a real soil flushing solution needs to last longer to reach the same efficiency than when synthetic soil flushing solutions are treated [15]. In short, the results here obtained are very promising, especially considering that the classical Fenton process was considered not enough efficient for the treatment of soil washing/soil flushing solutions containing high concentrations of surfactants [15], as is the present case. In addition, it must be taken into account that these are preliminary results and once the main variables of the process have been optimized, the degree of elimination of COCs is expected to increase significantly. Besides, the selective recovery of the surfactant could allow the recycle of the solution after the Fenton 14 treatment. Nevertheless, if the conversion of pollutants were lower than desired, an additional treatment should be coupled after the Fenton process. The increasing concentration of chlorides in solution with reaction time (data not shown) indicates the release of this anion as a consequence of the oxidation of COCs. The concentration detected is lower than expected due to the conversion of COCs and the stoichiometry of the reactions, probably due to the presence of surfactant, which can interfere in the analysis. As a result of the oxidation of COCs (and very probably also to the oxidation of part of the surfactant, as it can be seen in section 3.5) and the consequent generation of carboxylic acids as oxidation intermediates [37], the pH of the system rapidly decreased to acid values (pH = 2.5 from 4 h reaction time). Therefore, the Fenton oxidation of real SFSs can be successfully performed without pH adjustment, which is in agree with the results found in the literature [15]. Moreover, an additional experiment adjusting the initial pH to 3 (H2SO4, 0.1 M) was performed, (data not shown) obtaining the same COCs conversion profiles from 2-4 h reaction time. This fact reveals that the initial pH has a slight influence on the initial rate of COCs degradation, but when the pH of the reaction medium reaches a similar value (pH  3, due to the generation of the oxidation by-products), the rate of COCs degradation also does so. The solution acquired a brown tone from the beginning of the reaction (homogeneous and stable solution), which may be related with the oxidation of by-products generated both from COCs and surfactant oxidation and/or the oxidation of ferrous iron to ferric iron (Eq. 2). For the sake of discrimination, the consumption of H2O2 when no organic compounds were in the reaction medium (in the absence of COCs and surfactant, Blank-H2O2 + Fe experiment) has been also included. As observed, the elimination of H2O2 is much faster in the absence of organics than in the presence of these compounds. The complete 15 conversion of H2O2 was achieved at 24 h reaction time, while 100 h were required when the SFS was treated. This fact highlights the high extension of parasitic reactions in the blank experiment. The decomposition of H2O2, catalyzed by iron in solution, leads to high concentrations of hydroxyl radicals that, when there are not organic matter in their surroundings, favors the production of parasitic scavenging reactions [38]. Firstly H2O2 can act as a hydroxyl radical scavenger giving rise hydroperoxyl radicals and secondly, ferric ions may catalyze H2O2 decomposition into water and oxygen (unreactive at these operating conditions: room temperature and ambient pressure). Thus, it can be concluded that the presence of organic pollutants, and to a greater extent, that of the surfactant seems to protect hydrogen peroxide molecules and leads to a more progressive consumption of this reagent. 3.3 Effect of hydrogen peroxide concentration on COCs degradation Once the efficiency of Fenton process on the degradation of COCs in SFS has been demonstrated, a systematic study varying the concentration of hydrogen peroxide was performed with the aim of optimize the consumption of this reagent, which usually represents the main cost of the Fenton process [39]. Thus, two concentrations of hydrogen peroxide (5, 10 g L-1), which correspond to 50% and 100% of the theoretical amount for the complete mineralization of COCs, were tested and the results here obtained were compared to those corresponding to the 200% of H2O2. In these experiments, the molar ratio H2O2:Fe of 32 was maintained. The conversion of H2O2 with reaction time in these experiments is depicted in Fig 4. The consumption of this reagent followed first-order reaction kinetics and no significant differences have been found working with 100% and 200% of H2O2 doses. However, when using a concentration of H2O2 corresponding to 50% of the stoichiometric 16 amount, the decomposition of the oxidant slowed down, even though the pollutant/hydroxyl radical ratio increased. This fact could be related to an insufficient concentration of iron in solution, whose behavior can be also highly affected by the presence of the surfactant (constant in all cases). The initial concentration of hydrogen peroxide (and iron in solution) had an important effect on the degradation kinetics of the pollutants. To facilitate the interpretation of the results, the obtained data have been grouped by families and divided into two types: aromatic compounds, ie, chorobenzenes, which include CB, DiCBs, TiCBs, TetraCBs and PentaCB, and non-aromatic compounds, including PentaCXs, HexaCXs, HeptaCHs and HCHs (Fig 5). Attending to the results obtained when using the highest dose of hydrogen peroxide (200%), it can be concluded that non-aromatic compounds are less prone to oxidation by hydroxyl radicals than chlorobenzenes due to the saturated structure and the absence of double bonds in the formers [40]. Moreover, the reactivity order of these compounds seems to be related with the chlorine content of the molecule, obtaining the following degradation ranking: CB > DiCBs > TriCBs > TetraCBs = PentaCB, PentaCXs, HexaCXs, HeptaCHs, HCHs). A plateau in the conversion profile of these pollutants has been observed from 48 h reaction time, which is associated to the depletion of hydrogen peroxide. The generation of chlorinated compounds different from the starting ones at significant amount during Fenton oxidation was discarded by GC-ECD analyses. Moreover, no other aromatic organic compounds were detected by GC-MS analyses. Based on the results obtained in a previous work [9] it is expected that the oxidation of COCs takes place through a complex parallel-series reaction pathway, in which some contaminants may also appear as intermediate compounds from the oxidation of other contaminants 17 via abstraction (chlorine and hydrogen) and hydroxylation (•OH addition) reactions. These compounds are further oxidized to generate carboxylic acids, which are slowly transformed into carbon dioxide, chloride and water. As can be seen in Fig 5, the pollutants degradation greatly improved with the initial concentration of H2O2 for all the COCs. Taking CB as an example, the complete conversion of this compound was achieved at 30 h reaction time when working with the highest dose of H2O2 (200%), whereas conversions of 95% and 60% were achieved at the end of the treatment (144 h, complete consumption of H2O2) when using the stoichiometric and sub-stoichiometric doses of this reagent, respectively. Attending to the degradation of COCs, similar results were obtained when working with doses of 100% and 200% (for the complete consumption of H2O2). In the former case, longer reaction times were required (144 h vs 48 h). However, the use of substoichiometric doses of hydrogen peroxide (viz. 50%) is highly discouraged since the conversion of COCs achieved at these conditions was too low. 3.4 Hydrogen peroxide efficiency In order to quantify the efficiency of hydrogen peroxide during Fenton treatment, the amount of COCs degraded (in mg) per gram of oxidant consumed has been calculated for the three doses of H2O2 tested along reaction time and represented vs the amount of H2O2 consumed in each case (Fig 6). The lower the amount of H2O2 consumed the greater efficiency of this reagent. This fact could indicate that the hydroxyl radicals initially generated are very selective to the oxidation of COCs. As the amount of H2O2 consumed increased, the efficiency of the treatment decreased, highlighting the extension of scavenging reactions. The evolution 18 of the efficiency was not affected by the initial concentration of this reagent, obtaining values of this parameter even higher in the case of 200% of the stoichiometric dose. Therefore, attending to the efficiency in the consumption of this reagent, it seems interesting to work with high concentrations of hydrogen peroxide. 3.5 Surfactant stability The stability of the surfactant during the oxidation treatment is crucial since the reutilization of the SFS (for a new cycle of soil flushing) after pollutant abatement would be very advantageous for the economy of the process [15, 18, 23, 41]. Surfactant can be oxidized by hydroxyl radicals and therefore, competes with COCs during the Fenton process [15]. To know more about the unproductive consumption of the oxidant due to the presence of surfactant, additional experiments varying its initial concentration (3, 15, 30 and 45 g L-1) in the absence of COCs and using the same concentration of hydrogen peroxide (20 g L-1) and iron (0.625 g L-1), were performed. The consumption of H2O2 resulted to be independent of the initial concentration of surfactant (data not shown), that supports the hypothesis previously raised. Surfactant molecules (above a minimum concentration value) act as a protective agent of hydrogen peroxide, allowing the progressive decomposition of this reagent and minimizing the production of radical recombination reactions. Thus, working with high concentrations of surfactant is not associated with wasting of oxidant in this case. In order to evaluate the stability of the surfactant, several parameters have been considered: the evolution of limonene, the additive used in E-mulse 3 ® surfactant (which represents around 15% of the commercial product), during the Fenton reactions (Fig 7a) and the ESC (Fig 7b) and IT values at the end of the reactions (144 h). 19 As expected, the concentration of limonene remained invariable in the blank experiments but this compound was highly reactive in the Fenton system. The degradation rate of limonene increased with the initial concentration of hydrogen peroxide (Fig 7a). It was completely degraded at 24 h reaction time when using the highest concentration of H2O2 and conversions above 95% and 80% were obtained at the end of the treatment with doses corresponding to 100% and 50% of the stoichiometric amount, respectively. Limonene was attacked by hydroxyl radicals giving rise to the hydroxylated compound carvenol and carvone (both of them identified by GC-MS analysis). The degradation of this compound does not have to be significant from a practical point of view since its oxidation products (carvenol and carvone) provably present a similar behavior, enhancing also the performance of the surfactant in COCs solubilization. On the other hand, as observed in Fig 7b the oxidation of limonene is not directly associated with a loss in the surfactant capacity of the SFSs. The surfactant remained stable after blank experiments and it was partially oxidized by hydroxyl radicals during the Fenton treatments. Formic and malonic acids were detected as the only surfactant oxidation by-products, being the mineralization of the surfactant very scarce. The degradation of the surfactant (calculated as the ESC reduction) was proportional to the initial concentration of H2O2 (50%, 40% and 20% of reduction when working with doses of H2O2 of 200%, 100% and 50% of the stoichiometric amount, respectively). In any case, the oxidation treatments did not altered the IT of the effluents (34 mN m-1), showing all of them an ESC value much higher than the CMC (85 mg L-1 at 25 ºC). Thus, the SFSs apparently maintained their surfactant capacity and could be reused for further soil flushing cycles. In order to verify this point, an additional test was carried out. The reaction effluent corresponding to 200% H2O2 was put in contact with a known 20 mass of DNAPL, achieving the complete solubilization of the pollutants. The concentration of COCs measured in this solution was equivalent ( 3500 mg L-1) to the initial SFS, which indicates that the solubilization capacity of the surfactant is maintained after the oxidation treatment and the surfactant solution can be reused. Conclusions The results obtained in the present work reveal that the combination of soil flushing with nonionic surfactant and the subsequent selective oxidation of chlorinated organic compounds by using the Fenton´s reagent is a useful alternative for the remediation of soils polluted with chlorinated DNAPLs. Most of DNAPL was extracted from the soil by using 15 g L-1 of E-Mulse 3 ® solution in the soil flushing step. The resulting SFS, containing a complex mixture of COCs (28 chlorinated pollutants, COCs = 3639 mg L-1) was subsequently treated by Fenton oxidation. Chlorobenzenes were more prone to oxidation by hydroxyl radicals than the non-aromatic compounds (pentachlorocyclohexenes, hexachlorocyclohexenes, heptachlorocyclohexanes and hexachlorocyclohexanes). The initial concentration of hydrogen peroxide improved the oxidation rate and therefore, the degree of COCs abatement. This treatment allowed the selective oxidation of COCs (above 80%) and the recovery of the surfactant (all the Fenton effluents showed an IT value of 34 mN m-1 and surfactant concentrations much higher than the CMC). The application of this technology to field scale will require for additional studies to select the optimal conditions for each case. Acknowledgments 21 The authors acknowledge financial support from the Comunidad Autónoma of Madrid (Project S2013-MAE-2739 CARESOIL-CM) and from the Spanish MINECO (Project CTM2016-77151-C2-1-R) and the Aragon Government and EMGRISA Company for the supply of samples. Carmen M. Dominguez acknowledges the Spanish MINECO for the “Juan de la Cierva” post-doctoral contract (FJCI-2016-28462). Table and Figure camptions Table 1. Information about the COCs identified and quantified in the soil flushing solutions with tap water (WFS) and with 15 g L-1 of surfactant (SFS) using a Q = 0.2 mL min-1. Figure 1. GC-FID chromatograms of the SFS before (0 h) and after 5 and 29 h of Fenton treatment (COCs,0 = 3639 mg L-1, H2O2,0 = 20 g L-1, Fe = 0.65 g L-1, pH0 ≈ 6.5). Figure 2. Distribution of COCs (grouped by families) in WFS, SFS and DNAPL solved in acetone (data obtained from [7]). Figure 3. Evolution of hydrogen peroxide (a) and global COCs conversion (b) with reaction time in different experimental systems under the following conditions (when applicable): COCs,0 = 3639 mg L-1, H2O2,0 = 20 g L-1 (200% of the stoichiometric amount), Fe = 0.65 g L-1, pH0 ≈ 6.5. Figure 4. Evolution of hydrogen peroxide conversion at different hydrogen peroxide doses (50%, 100% and 200% of the stoichiometric amount) during reaction time under the following experimental conditions: COCs,0 = 3639 mg L-1, H2O2:Fe = 32, pH0 ≈ 6.5. Figure 5. Influence of hydrogen peroxide concentration (50%, 100% and 200% of the stoichiometric amount) on COCs conversion (grouped by families) during reaction time 22 under the following experimental conditions: COCs,0 = 3639 mg L-1, H2O2:Fe = 32, pH0 ≈ 6.5. Figure 6. Evolution of hydrogen peroxide efficiency with oxidant consumption at different hydrogen peroxide concentrations (50%, 100% and 200% of the stoichiometric amount) under the following experimental conditions: COCs,0 = 3639 mg L-1, H2O2:Fe = 32, pH0 ≈ 6.5. Figure 7. Evolution of limonene with reaction time (a) and equivalent surfactant concentration (ESC) at the end of the experiments (144 h) (b) in different experimental systems under the following conditions: COCs,0 = 3639 mg L-1, H2O2:Fe = 32, pH0 ≈ 6.5. References 1. M. Vega, D. Romano, E. Uotila, Lindane (persistent organic pollutant) in the EU, Directorate General for Internal Policies. Policy Department C: Citizens’ Rights and Constitutional Affairs. Petitions (PETI). PE 571 (2016). 2. J. Fernández, M. Arjol, C. Cacho, POP-contaminated sites from HCH production in Sabiñánigo, Spain, Environ. Sci. Pollut. Res. 20 (2013) 1937-1950. 3. A. Dorsey, Toxicological Profile for Alpha-, Beta-, Gamma, and Delta- hexachlorocyclohexane., Agency for Toxic Substances and Disease Registry (2005). 4. J. Vijgen, The legacy of lindane HCH isomer production, Main report. IHPA, January (2006). 23 5. S. Wacławek, V. Antoš, P. Hrabák, M. Černík, Remediation of hexachlorocyclohexanes by cobalt-mediated activation of peroxymonosulfate, Desal. Water Treat. 57 (2016) 26274-26279. 6. J. Vijgen, P. Abhilash, Y.F. Li, R. Lal, M. Forter, J. Torres, N. Singh, M. Yunus, C. Tian, A. Schäffer, Hexachlorocyclohexane (HCH) as new Stockholm Convention POPs—a global perspective on the management of Lindane and its waste isomers, Environ. Sci. Pollut. Res. 18 (2011) 152-162. 7. A. Santos, J. Fernandez, J. Guadaño, D. Lorenzo, A. Romero, Chlorinated organic compounds in liquid wastes (DNAPL) from lindane production dumped in landfills in Sabiñanigo (Spain), Environ. Pollut. 242 B (2018) 1616-1624. 8. A. Santos, J. Fernandez, S. Rodriguez, C.M. Dominguez, M.A. Lominchar, D. Lorenzo, A. Romero, Abatement of chlorinated compounds in groundwater contaminated by HCH wastes using ISCO with alkali activated persulfate, Sci. Total Environ. 615 (2018) 1070-1077. 9. C.M. Dominguez, N. Oturan, A. Romero, A. Santos, M.A. Oturan, Removal of organochlorine pesticides from lindane production wastes by electrochemical oxidation, Environ. Sci. Pollut. Res. (2018) 1-10. 10. C.M. Dominguez, N. Oturan, A. Romero, A. Santos, M.A. Oturan, Removal of lindane wastes by advanced electrochemical oxidation, Chemosphere 202 (2018) 400- 409. 11. C.N. Mulligan, R. Yong, B. Gibbs, Surfactant-enhanced remediation of contaminated soil: a review, Eng. Geol. 60 (2001) 371-380. 24 12. R.D. Villa, A.G. Trovó, R.F.P. Nogueira, Soil remediation using a coupled process: soil washing with surfactant followed by photo-Fenton oxidation, J. Hazard. Mater. 174 (2010) 770-775. 13. X. Mao, R. Jiang, W. Xiao, J. Yu, Use of surfactants for the remediation of contaminated soils: a review, J. Hazard. Mater. 285 (2015) 419-435. 14. A.T. Besha, D.N. Bekele, R. Naidu, S. Chadalavada, Recent advances in surfactant- enhanced In-Situ Chemical Oxidation for the remediation of non-aqueous phase liquid contaminated soils and aquifers, Environ. Technol. Innovation 9 (2017) 303-322. 15. C. Trellu, E. Mousset, Y. Pechaud, D. Huguenot, E.D. Van Hullebusch, G. Esposito, M.A. Oturan, Removal of hydrophobic organic pollutants from soil washing/flushing solutions: a critical review, J. Hazard. Mater. 306 (2016) 149-174. 16. S. Wang, C.N. Mulligan, An evaluation of surfactant foam technology in remediation of contaminated soil, Chemosphere 57 (2004) 1079-1089. 17. G. Zheng, A. Selvam, J.W. Wong, Enhanced solubilization and desorption of organochlorine pesticides (OCPs) from soil by oil-swollen micelles formed with a nonionic surfactant, Environ. Sci. Technol. 46 (2012) 12062-12068. 18. J. Gómez, M. Alcántara, M. Pazos, M. Sanromán, Remediation of polluted soil by a two-stage treatment system: desorption of phenanthrene in soil and electrochemical treatment to recover the extraction agent, J. Hazard. Mater. 173 (2010) 794-798. 19. E. Mousset, N. Oturan, E.D. Van Hullebusch, G. Guibaud, G. Esposito, M.A. Oturan, Influence of solubilizing agents (cyclodextrin or surfactant) on phenanthrene degradation by electro-Fenton process–study of soil washing recycling possibilities and environmental impact, Water Res. 48 (2014) 306-316. 25 20. D. Huguenot, E. Mousset, E.D. Van Hullebusch, M.A. Oturan, Combination of surfactant enhanced soil washing and electro-Fenton process for the treatment of soils contaminated by petroleum hydrocarbons, J. Environ. Manage. 153 (2015) 40-47. 21. M. Muñoz-Morales, M. Braojos, C. Sáez, P. Cañizares, M. Rodrigo, Remediation of soils polluted with lindane using surfactant-aided soil washing and electrochemical oxidation, J. Hazard. Mater. 339 (2017) 232-238. 22. S. Satyro, M. Race, R. Marotta, M. Dezotti, D. Spasiano, G. Mancini, M. Fabbricino, Simulated solar photocatalytic processes for the simultaneous removal of EDDS, Cu (II), Fe (III) and Zn (II) in synthetic and real contaminated soil washing solutions, J. Environ. Chem. Eng. 2 (2014) 1969-1979. 23. E. Mousset, N. Oturan, E.D. Van Hullebusch, G. Guibaud, G. Esposito, M.A. Oturan, Treatment of synthetic soil washing solutions containing phenanthrene and cyclodextrin by electro-oxidation. Influence of anode materials on toxicity removal and biodegradability enhancement, Appl. Catal., B: Environ. 160 (2014) 666-675. 24. E.R. Bandala, Y. Velasco, L.G. Torres, Decontamination of soil washing wastewater using solar driven advanced oxidation processes, J. Hazard. Mater. 160 (2008) 402-407. 25. S. Esplugas, J. Gimenez, S. Contreras, E. Pascual, M. Rodrı́guez, Comparison of different advanced oxidation processes for phenol degradation, Water Res. 36 (2002) 1034-1042. 26. J.J. Pignatello, E. Oliveros, A. MacKay, Advanced oxidation processes for organic contaminant destruction based on the Fenton reaction and related chemistry, Crit. Rev. Environ. Sci. Technol. 36 (2006) 1-84. 26 27. H. Fenton, LXXIII.—Oxidation of tartaric acid in presence of iron, J. Chem. Soc., Trans. 65 (1894) 899-910. 28. W. Chu, W. Choy, J. Hunt, Effects of nonaqueous phase liquids on the washing of soil in the presence of nonionic surfactants, Water Res. 39 (2005) 340-348. 29. C.N. Mulligan, F. Eftekhari, Remediation with surfactant foam of PCP- contaminated soil, Eng. Geol. 70 (2003) 269-279. 30. L.G. Torres, F. Ramos, M.A. Avila, I. Ortiz, Removal of methyl parathion by surfactant-assisted soil washing and subsequent wastewater biological treatment, J. Pestic. Sci. 37 (2012) 240-246. 31. L. Rios, M. David, J. Vazquez‐Arenas, W. Anderson, Use of surfactants and blends to remove DDT from contaminated soils, Can. J. Chem. Eng. 91 (2013) 238-244. 32. G.M. Eisenberg, Ind. Eng. Chem., Anal. Ed. 15 (1943) 327-328. 33. J. Quiroga, A. Riaza, M. Manzano, Chemical degradation of PCB in the contaminated soils slurry: Direct Fenton oxidation and desorption combined with the photo-Fenton process, J. Environ. Sci. Health Part A 44 (2009) 1120-1126. 34. J.K. Saxe, H.E. Allen, G.R. Nicol, Fenton oxidation of polycyclic aromatic hydrocarbons after surfactant-enhanced soil washing, Environ. Eng. Sci. 17 (2000) 233- 244. 35. C. Yang, D. Wang, Effect of anionic surfactants on the process of Fenton degradation of methyl orange, Water Sci. Technol. 60 (2009) 2803-2807. 36. C. Trellu, N. Oturan, Y. Pechaud, E.D. van Hullebusch, G. Esposito, M.A. Oturan, Anodic oxidation of surfactants and organic compounds entrapped in micelles–Selective degradation mechanisms and soil washing solution reuse, Water Res. 118 (2017) 1-11. 27 37. J. Rosas, F. Vicente, A. Santos, A. Romero, Soil remediation using soil washing followed by Fenton oxidation, Chem. Eng. J. 220 (2013) 125-132. 38. E. Neyens, J. Baeyens, A review of classic Fenton’s peroxidation as an advanced oxidation technique, J. Hazard. Mater. 98 (2003) 33-50. 39. R. Munter, M. Trapido, Y. Veressinina, A. Goi, Cost effectiveness of ozonation and AOPs for aromatic compound removal from water: A preliminary study, Ozone: Sci. Eng. 28 (2006) 287-293. 40. D.D. Dionysiou, A.P. Khodadoust, A.M. Kern, M.T. Suidan, I. Baudin, J. Laîné, Continuous-mode photocatalytic degradation of chlorinated phenols and pesticides in water using a bench-scale TiO2 rotating disk reactor, Appl. Catal., B: Environ. 24 (2000) 139-155. 41. C. Trellu, O. Ganzenko, S. Papirio, Y. Pechaud, N. Oturan, D. Huguenot, E.D. Van Hullebusch, G. Esposito, M.A. Oturan, Combination of anodic oxidation and biological treatment for the removal of phenanthrene and Tween 80 from soil washing solution, Chem. Eng. J. 306 (2016) 588-596. 1 Table 1 COC Chemical Acronym Concentration Concentration formula in WFS (mg L-1) in SFS (mg L-1) chlorobenzene C6H5Cl CB 32.2 358 1,3-dichlorobenzene C6H4Cl2 1,3 DiCB < 0.1 21 1,4-dichlorobenzene C6H4Cl2 1,4 DiCB 2.6 91 1,2-dichlorobenzene C6H4Cl2 1,2 DiCB 2.4 66 1,3,5-trichlorobenzene C6H3Cl3 1,3,5 TriCB < 0.1 40 1,2,4-trichlorobenzene C6H3Cl3 1,2,4 TriCB 2.8 207 1,2,3-trichlorobenzene C6H3Cl3 1,2,3 TriCB < 0.1 41 1,2,3,4+1,2,4,5-tetrachlorobenzene C6H2Cl4 1,2,4,5 TetraCB 0.3 79 1,2,3,5-tetrachlorobenzene C6H2Cl4 1,2,3,5 TetraCB 0.4 117 pentachlorocyclohexene-1 C6H5Cl5 PentaCX-1 2.8 200 pentachlorobenzene C6HCl5 PentaCB 0.1 14 pentachlorocyclohexene-2 C6H5Cl5 PentaCX-2 2.9 109 pentachlorocyclohexene-3 C6H5Cl5 PentaCX-3 0.6 36 hexachlorocyclohexene-1 C6H4Cl6 HexaCX-1 0.1 40 pentachlorocyclohexene-4 C6H5Cl5 PentaCX-4 0.6 0 pentachlorocyclohexene-5 C6H5Cl5 PentaCX-5 1.5 104 hexachlorocyclohexene-2 C6H4Cl6 HexaCX-2 0.3 49 hexachlorocyclohexene-3 C6H4Cl6 HexaCX-3 0.1 53 alpha-hexachlorocyclohexane C6H6Cl6 -HCH 1.2 153 hexachlorocyclohexene-4 C6H4Cl6 HexaCX-4 0.1 45 beta-hexachlorocyclohexane C6H6Cl6 -HCH < 0.1 4 gamma-hexachlorocyclohexane C6H6Cl6 -HCH 4.8 528 heptachlorocyclohexane-1 C6H5Cl7 HeptaCH-1 2.1 541 delta-hexachlorocyclohexane C6H6Cl6 -HCH 8.2 400 epsilon-hexachlorocyclohexane C6H6Cl6 -HCH 1.4 69 heptachlorocyclohexane-2 C6H5Cl7 HeptaCH-2 < 0.1 142 heptachlorocyclohexane-3 C6H5Cl7 HeptaCH-3 0.5 132 HCHs 15.8 1154 COCs 68.2 3639 Figure 1 1 ,2 ,4 T ri C B 1 ,2 ,3 T ri C B 1 ,2 ,4 ,5 T e tr aC B 1 ,2 ,3 ,5 T e tr aC B P en ta C X -1 ISTD a - H C H b -H C H d - H C H H ex aC X -1 H ep ta C H -2 1 ,3 ,5 T ri C B P en ta C B C B 1 ,3 D iC B 1 ,4 D iC B 1 ,2 D iC B P en ta C X -2 P en ta C X -3 P en ta C X -4 P C X -5 H ex aC X -2 H ex aC X -3 H ex aC X -4 g- H C H H ep ta C H -1 e- H C H H ep ta C H -3 Limonene t = 0 h t = 5 h t = 29 h Figure 2 0 10 20 30 40 50 WFS (COCs=68.2 mg L -1 ) SFS (COCs=3639 mg L -1 ) DNAPL (acetone) H C H s C O C s ( % ) H ep ta C H s H ex aC X s P en ta C X s P en ta C B T et ra C B s T riC B s D iC B s C B Figure 3 0 20 40 60 80 100 120 140 0.0 0.2 0.4 0.6 0.8 1.0 b) 1 -X C O C s t (h) 0 20 40 60 80 100 120 140 0.0 0.2 0.4 0.6 0.8 1.0 t (h) Blank-SWSS Blank-SWSS + H 2 O 2 Blank-SWSS + Fe Blank-H 2 O 2 + Fe SWSS + H 2 O 2 + Fe 1 -X H 2 O 2 a) Figure 4 0 20 40 60 80 100 120 140 0.0 0.2 0.4 0.6 0.8 1.0 t (h) 200% H 2 O 2 100% H 2 O 2 50% H 2 O 2 1 -X H 2 O 2 Figure 5 0 20 40 60 80 100 120 140 0 20 40 60 80 100 X t (h) 50% H 2 O 2 0 20 40 60 80 100 Chlorobenzenes CB DiCBs TriCBs TetraCBs PentaCBs X 200% H 2 O 2 20 40 60 80 100% H 2 O 2 X 0 20 40 60 80 100 120 140 0 20 40 60 80 100 50% H 2 O 2 t (h) X 0 20 40 60 80 100 X non-aromatic COCs PentaCX HexaCX HeptaCH HCH 200% H 2 O 2 20 40 60 80 X 100% H 2 O 2 Figure 6 0 4 8 12 16 20 0 100 200 300 400 500 3000 4000 200 % H 2 O 2 100 % H 2 O 2 50 % H 2 O 2 H 2 O 2 consumed (g) m g C O C /g H 2 O 2 Figure 7 0 3 6 9 Blanks F e H 2 O 2 ( 2 0 0 % ) H 2 O 2 ( 5 0 % ) + F e H 2 O 2 ( 1 0 0 % ) + F e H 2 O 2 ( 2 0 0 % ) + F e S W S S E S C (g L -1) b) Fenton effluents0 20 40 60 80 100 120 140 0.0 0.2 0.4 0.6 0.8 1.0 t (h) Blank-SWSS Blank-SWSS + H 2 O 2 Blank-SWSS + Fe SWSS + H 2 O 2 (200%) + Fe SWSS + H 2 O 2 (100%) + Fe SWSS + H 2 O 2 (50%) + Fe 1 -X lim o n e n e a) Graphical Abstract 20 40 60 80 100 144 h H C H X C O C s ( % ) 50% H2O2 100% H2O2 200% H2O2 22 h 0 20 40 60 80 100 P en ta C B T et ra C B H ex aC X C B T riC B D iC B     H ep ta C H P en ta C X X C O C s ( % ) DNAPL DNAPL Surfactant solution (E-Mulse 3 ®) Surfactant flushing solution (SFS) 3639 mg L-1 COCs Fe3+ Fe2+ H2O2 H2O2 ·OH + OH- ·OOH + H+ Surfactant Recovery Highlights • Nonionic surfactant E-Mulse 3 ® was used for soil flusing experiments • Efficient extraction of DNAPL by surfactant soil flushing (COCs=3693 mg L-1) • Selective oxidation of COCs by Fenton process (X>80%, 144 h) • COCs degradation greatly improved with the initial concentration of H2O2 • Surfactant solution can be reused after Fenton treatment (34 mN m-1) Revised Manuscript Selective removal of chlorinated organic compounds from lindane wastes by combination of nonionic surfactant soil flushing and Fenton oxidation Abstract 1. Introduction 2. Materials and Methods 3. Results Conclusions Acknowledgments The authors acknowledge financial support from the Comunidad Autónoma of Madrid (Project S2013-MAE-2739 CARESOIL-CM) and from the Spanish MINECO (Project CTM2016-77151-C2-1-R) and the Aragon Government and EMGRISA Company for the supply of samples. Ca... Table and Figure camptions References Table 1 Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Graphical Abstract Highlights