A Novel Combination of Surfactant Addition and Persulfate-assisted Electrokinetic Oxidation for Remediation of Pyrene-Contaminated Soil

M. Abtahi,a S. Jorfi,b,* N. Mehrabi,c R. Saeedi,d R. Darvishi Cheshmeh Soltani,e and G. Barzegarf a Department of Environmental Health Engineering, School of Public Health, Shahid Beheshti University of Medical Sciences, Tehran, Iran bEnvironmental Technologies Research Center, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran cDepartment of Environmental Health Engineering, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran dDepartment of Health Sciences, School of Health, Safety and Environment, Shahid Beheshti University of Medical Sciences, Tehran, Iran eDepartment of Environmental Health Engineering, School of Health, Arak University of Medical Sciences, Arak, Iran fDepartment of Environmental Health Engineering, Behbahan Faculty of Medical Sciences, Behbahan, Iran


Introduction
Pyrene is a polycyclic aromatic hydrocarbon (PAH) which may be produced commercially to be utilized as an intermediate in the production of pigments or it can be found in emissions of incomplete combustion of kerosene and diesel.The crystalline structure of pyrene (C 16 H 10 ) is a four-ring PAHs that is solid at room temperature due to different neutrons 1,2 .The long-term and continuous emission of PAHs into the environment can pose a serious threat to human health and environment through air pollution or soil contamination pathways, e.g., through industrial discharges and some agricultural tradi-tions and improper practices of waste disposal 3,4 .Thus, the quick and effective removal of PAHs from the environment as well as reducing their adverse effects seems to be necessary.Three-and four-ring PAHs are in particular associated with soil particles, and due to their low aqueous solubility and high sorption properties, their removal from subsurface environments using traditional technologies, such as soil washing, conventional chemical oxidation or bioremediation, is inefficient 5 .In fact, their removal from fine soil particles is still a major challenge for geo-environments, because of the high variation in their hydraulic conductivity, a lot of specific contaminants, environmental and dynamic conditions, reversible geochemical processes, and the interactions between soil particles and contaminants 6,7 .
Until now, different physicochemical, thermal, and biological methods have been applied for decontamination of polluted soils.However, these methods are often costly and energy-intensive, and are almost ineffective and could, in turn, remain secondary toxic materials, which are potentially harmful to the environment [8][9][10][11] .Thus, electrochemical removal has been considered as a reliable technology for the remediation of contaminated soils 12,13 .The main advantages of electrokinetic remediation (EK remediation) of soil contaminated with petroleum hydrocarbons is that the procedure can be performed as an in-situ or ex-situ technology, in soils with heterogeneous texture and low permeability, or in saturated and unsaturated soils.It is also possible to use electrochemical oxidation for simultaneous removal of heavy metals, radionuclides, and organic pollutants.Flexibility of this technology, high energy efficiency, and cost-effectiveness are also merited 14,15 .EK remediation includes the flow of electrical current in the soil, liquid pole electro-osmotic migration, electrical migration of ions, charged particles and colloids, electrolysis of water in the soil in the vicinity of the electrodes, migration of hydrogen and hydroxide ions into the soil resulting in temporary changes in soil pH, gas production at the electrodes, development of the non-uniform electric field, as well as the production of electro-osmotic flow 16,17 .In EK decomposition of water reactions, oxygen and hydrogen ions (H + ) are produced due to oxidation at the anode, while hydrogen gas and hydroxide ions (OH -) are made because of the reduction in cathode 18,19 .One of the problems in EK remediation of soil contaminated with organic compounds is that PAHs are hydrophobic, thus reducing the efficiency of the process.One of the approaches in order to enhance the solubility of PAHs is the application of intermediate facilitating factors, such as surfactants 20,21 .Addition of these factors leads to creation of some changes in the fluid properties, such as dielectric constant, pH, fluid viscosity, as well as changes in surface properties of soil particles 22,23 .Among the surfactants, non-ionic materials have been widely used for transferring non-polar organic substances by electro-osmotic flow.Indeed, surfactants have lower dielectric constant and higher viscosity than water.Consequently, the addition of these facilitators not only reduces the electro-osmotic flow, but also increases the solubility of PAHs 24 .Another problem concerning EK remediation is the great amount of time required to remove contaminants from the soil.This problem could be solved in the EK remediation process using other oxidations.The application of persulfate as an oxidizing agent can improve the efficiency and significantly reduce the required time for soil decontamination 25,26 .The process of oxidation by means of activated persulfate is a promising approach, due to stability, reactivity, solubility, oxidizing properties, and high amount of organic matter degradation 27 .Persulfate (S 2 O 8

2-
), with standard oxidation potential of E 0 = 2.01 V, is a novel approach applied in chemical oxidation, and has been considered a promising remediation technique of recalcitrant organics.Moreover, S 2 O 8 2-activation leads to production of sulfate radical (SO 4 -• , E 0 = 2.7 V), as a strong oxidant for the degradation of organic matter, even at neutral pH values, and acts more selectively for destruction of compounds with carboncarbon double bonds and benzene rings 28,29 .Persulfate activation would be carried out via different approaches, such as heat, metal and base transition (elevated pH>11), UV-radiation, electron-beam and ultrasonication 30 .For soil remediation, UV irradiation is not capable of penetrating the soil matrix, and alkaline activation requires high quantities of alkaline substances.Furthermore, transition of metal-catalyzed activation (e.g., Fe 2+ ) exhibits best results in strict pH value range of 2-4, and metal can act as a radical scavenger 31 .Thermally activated persulfate is known as a green source of sulfate radicals and has been previously applied to remove organic pollutants 32 .Thermal activation of persulfate also has many benefits, like the enhancement of solubility of slightly soluble PAHs, increase in efficiency of reaction and mineralization, and the consumption of fewer chemical substances 33 .Because of the heat generation through EK remediation, persulfate activation can be accomplished at the same time.
Integrated EK remediation with persulfate can be used to overcome limitations of conventional EK remediation and facilitate the distribution of persulfate in soils with low permeability.At the same time, it can directly cause oxidation-reducting reactions in the soil 34 .This study is intended to remediate pyrene-contaminated soil using persulfate-assisted EK remediation, and to use its results for the unwashed real hydrocarbon-polluted soils.

Chemicals
All chemicals applied in the present research, including pyrene, sulfuric acid, soda, methanol, n-hexane, acetone, Tween 80, acetonitrile and sodium persulfate were analytical-grade purchased from Merck, Germany, and all solutions were prepared with deionized water and kept at 4 °C, prior to experiments.

Soil sample
All soils were obtained from an oil manufacturing district in southern Iran.The soil samples were collected from the upper layers (0-40 cm) by soil cores, air-dried, and passed through a 2-mm sieve.The process of homogenization of soil grains was then performed by a shaker, followed by washing thrice with acetone, autoclaving, and storing in nonreactive plastic bags at 4 °C.Since the washed soil still contained pyrene, a determined amount of pyrene was dissolved in n-hexane and added to the soil to achieve a final soil concentration of 200 mg kg -1 in all samples.It was mixed vigorously and left under hood for 24 h until the n-hexane evaporated.The pyrene stock solution was added to a shaker for 1 minute before spiking 35 .Also, an unwashed soil sample of the same origin and without washing was used as naturally contaminated soil.The X-ray diffraction (XRD) patterns of the soil sample were recorded by X-ray diffractometer (Model: GNR-MPD3000) using Cu anode at λ = 0.15060 nm, voltage = at 40 kV, and current intensity = 30 mA.X-ray Fluorescence (XRF) was applied for finding the chemical substances of soil particles.The pyrene and its degradation metabolites were also determined by GC-MS (Model: Agilent 7890, USA) with HP-5MS capillary column (30 m × 0.25 mm × 0.25 µm film thickness, 5 % Phenyl -95 % Methyl Siloxane phase).The carrier gas (helium) was fed with a fixed 1 mL min -1 flow rate.The oven temperature was fixed to 40 °C for 1 min, followed by an enhancement to 300 °C at 5 °C min -1 and kept constant for 3 min.Ultimately, the sample was poured in instrument with 10:1 splitting ratio.Mass spectra were obtained by electron impact (EI) at 70 eV with the full scan mode.

Lab-scale experimental setup
The reactor was a cube Plexiglas chamber (Supplementary file S1) consisting of 10 houses, each with a length of 30 cm, width of 10 cm, and height of 20 cm, divided into 10 completely equal sections using nine separating walls in the longitudinal sector.Other facilities in the reactor included 2 pairs of graphite electrodes with a thickness of 4 mm, width of 4 cm, and length of 15 cm for each section, and a DC power supply (Model: Zx 200).The voltage was adjustable between 0-100 V cm -1 .

Experimental procedure
All the experiments were carried out in batch mode operation.In fact, for determination of the optimum conditions of the experimental factors, such as voltage (0.3 -1.5 V cm -1 ), moisture content (10 -100 %), effect of Tween 80 as a chemical surfactant (0 -40 mL kg -1 ), effect of sodium persulfate (0 -100 mg kg -1 ) and reaction time (0-120 h) on pyrene removal, were investigated consecutively based on one factor at a time experimental design.Additionally, an unwashed soil sample was subjected to persulfate-assisted EK remediation based on the selected conditions.

Analytical methods
The remaining concentrations of pyrene in the soil samples were extracted using an ultrasonic instrument based on EPA method 3550B 36 .Briefly, the sample was put in a special tube containing n-hexane and acetone with 1:1 v/v ratio.With completion of extraction, the process of centrifugation of sample was carried out for 10 min at 6000 rpm, followed by filtration using polytetrafluoroethylene (PTFE).A specific amount of filtered solution was separated in order to perform the analysis.The pyrene concentration was obtained using gas chromatography (GC) system (Chrompack CP 9001) equipped with a flame ionization detector (FID) using (HP-5) capillary column (30 m, 0.32 mm i.d. and 0.2 µm film thickness).Nitrogen was used as carrier gas at a rate of 2 mL min -1 .The temperature program was as follows: column temperature was held at 120 °C for 1 min, and then raised to 240 °C at a rate of 20 °C min -1 , then held for 1 min.The temperatures of injector and detector were fixed at 250 and 300 °C, respectively.Removal was analyzed through quantifying initial and final pyrene concentration according to Eq. ( 1): (1) where, C 0 and C t are the initial and residual pyrene concentrations (mg L -1 ).Pyrene recovery rate was about 85 % immediately after spiking.Considering the value of 15 % as a non-extracted portion of pyrene from the soil, we multiplied all raw data by a factor of 1.15.Soil moisture, EC, and pH were analyzed based on ASTM (D4972-13) 37 .The soil texture was determined according to the method described by ASTM (D6913M-17) 38 .

Soil analysis
Some selected chemical and physical characteristics of the soil are presented in Table 1.Accordingly, the studied soil with 32 % clay was of the clay-loam.SiO 2 and CaO also accounted for most of the mineral parts of the soil.LOI of 22 % made a significant portion of the soil mass.This was because the soil was contaminated with different types of petroleum compounds.Furthermore, the pres-ence of 5.7 wt % of chloride ions indicated severe salinity of the studied soil.According to the results of XRD analysis (results are not presented and can be found in supplementary files), the dominant crystalline phases of soil samples included calcite (CaCO 3 ), quartz (SiO 2 ), halite (NaCl), dolomite (CaMg (CO 3 ) 2 , gypsum (CaSO 4 ⋅ 2H 2 O) and clays.

EK remediation
The results indicated that any changes in the applied voltage directly affected the electro-osmotic flow in a way that, with an increase in voltage, the electro-osmotic flow increased.Voltage changes, however, have the maximum effect on the soil geochemical characteristics that critically affect the electro-osmotic flow.Moreover, the voltage of 1 V cm -1 promotes the development and maintenance of the electro-osmotic flow during the reaction 39,40 .Effect of voltage in the range of 3.0 to 1.5 V cm -1 on the efficiency of the EK remediation process in removing pyrene from soil was examined.All the other variables, including natural soil pH equal to 7 ± 0.3, reaction time of 120 h, initial pyrene concentration of 200 mg kg -1 , humidity of 50 %, and zero concentrations of Tween 80 and persulfate were kept constant.Pyrene removal efficiency increased as the voltage increased, but no significant difference was observed between the voltages 1 and 1.5 V cm -1 (p value >0.05).Pyrene removal efficiencies in 0.3, 0.5, 0.7, 1, and 1.5 V cm -1 were 28, 31, 40, 46 and 47.5 %, respectively (Fig. 1a).
At the beginning of the reaction, due to the movement of the electro-osmotic flow from the anode towards the cathode, ionic migration of the contaminants exhibited an increasing trend 41,42 .In fact, due to the stronger electric field and the resulting higher current density through the electrodes, most of the processes associated with the removal of contaminants, such as electro-osmosis and electrical migration, occur more efficiently and thus more contaminants are eliminated 43 .
In effect, the high natural EC in loamy soils contributed to the high electro-osmotic flow at the beginning of the reaction.After the maximum streaming, access of water pole and moving ions (H + and OH -), due to the electrolysis of water molecules and deposition of ions in the area of the cathode, decreased and resulted in reduced electric current and constant amount of pollutant removal 44 .In this study, due to lower power consumption and non-significant difference (p value > 0.05) in removal efficiency of voltages 1 and 1.5 V cm -1 , we alternatively selected voltage of 1 V cm -1 for the rest of the experiments.In addition, pseudo-first-order degradation rate constants were 0.0045 and 0.0046 h -1 for voltages of 1 and 1.5 V cm -1 , respectively, indicating a non-significant difference between them (Fig. 1b).

Effect of moisture content
In the selected conditions including voltage of 1 V cm -1 , and the other variables, such as pyrene concentration of 200 mg kg -1 , reaction time of 120 h, natural soil pH equal to 7±0.3, zero concentration of Tween 80 and sodium persulfate, the effect of moisture content in the range of 10 %, 25 %, 50 % and 100 % on the efficiency of the EK remediation for pyrene removal were studied.Pyrene removal efficiency for moisture contents of 10, 25, 50 and 100 % were 15, 29, 45, and 64 %, respectively (Fig. 1c).The higher the moisture content, the greater is the electrical current established, and thus the more H + ions are produced 45 .When an electric field is established in moist soils, electrolysis of water molecules occurs, which reduces the moisture content of the soil.Changes in moisture content took place in the cathode, due to the electro-osmotic flow of water molecules from the anode to the cathode.Although the electro-osmotic streaming reduces with increasing pH and ion concentration in the liquid, the higher the current, the more efficient is the contaminants removal process 44 .Rather than removal efficiency, which was the highest in moisture content of 100 %, the reaction rate constant was much higher for saturation condition (0.0087 h -1 ) compared to moisture content of 50 % (0.0043 h -1 ) (Fig. 1d).Therefore, the moisture content of 100 % was selected for the rest of the experiments.Ln C 0 /C

Characteristic
Effect of Tween 80 addition In the selected experimental conditions as mentioned previously, the effect of adding Tween 80 concentrations of 20 and 40 mL kg -1 in pyrene removal using EK remediation was investigated.Pyrene removal efficiency for Tween 80 concentrations of 20 and 40 mL kg -1 were 76 and 78 %, respectively (Fig. 1e), compared to 64 % removal without surfactant addition.Surfactant used in the EK process was affected by chemical oxidation and reduction, and thus decomposed with pyrene and became harmless or low-risk material.In general, surfactant enhanced the water solubility of pyrene.Higher concentration of surfactant than the critical concentration of its micelle increases the mobility of the polycyclic aromatic hydrocarbons within the aqueous phase, and therefore increases the pyrene removal efficiency 46 .Results of several studies have shown that without addition of surfactant, contaminant decomposition was limited 47 .Adding a small amount of Tween 80 can facilitate the decomposition of endosulfan contamination by using bacteria.In effect, increasing the concentration of Tween 80 will not make a significant difference in removal efficiency 48 .Since an insignificant difference was observed between the removal efficiency as well as the reaction rate constant in Tween 80 concentrations of 20 mL kg -1 and 40 mL kg -1 (Fig. 1f), the value of 20 mL kg -1 was selected due to less usage of chemical.
EK remediation is based on various mutual mechanisms, but the principal electron transfer reactions that occur at electrodes during the process is the electrolysis of water upon electric field application.Water decomposition (electrolysis reactions) occurs at the electrodes.The electrolysis reactions Persulfate: 100 mg kg -1 g produce oxygen gas and hydrogen ions (H + ) derived from oxidation at the anode, and produce hydrogen gas and hydroxyl (OH − ) ions because of the reduction at the cathode, as illustrated in Eqs. 2 and 3: At anode (oxidation): (2) (3) Effect of persulfate addition In the selected experimental conditions including voltage of 1V cm -1 , saturated moisture, Tween 80 concentration of 20 mL kg -1 , pyrene concentration of 200 mg kg -1 , reaction time of 120 h, natural soil pH equal to 7 ± 0.3, and the effect of sodium persulfate in two concentrations of 50 and 100 mg kg -1 on EK remediation of pyrene was studied.Pyrene removal efficiencies for sodium persulfate concentration of 50 mg kg -1 and 100 mg kg -1 were 100 % that were obtained after the reaction time of 96 and 24 h, respectively (Fig. 1g and h).Firstly, persulfate moves to the soil by the electrical or electro-osmotic migration, and then it is activated by temperature and pH 49,50 .For activation of persulfate, it is necessary to make sure that the temperature is 45 o C.These conditions can be achieved in an electric field.As a result of high-voltage gradient, soil heats and acid is produced in the anode, which causes soil acidification and, thus, higher removal efficiency is obtained 51 .
In an electric field, the incomplete solubility of salt deposits, and the mobility of sodium persulfate ions leads to the production of higher amounts of ion concentration in water.On the other hand, the amount of electrical current is reduced over time, due to the electrical migration of ions towards the electrodes.However, with addition of sodium persulfate, the electrical current significantly increased, as a result of increasing interstitial fluid concentration 45,52,53.Persulfates have a high potential without activation, as shown in Eq. ( 4): Persulfate provides more oxidation potential when it is activated via an activator like heat or Fe ions.Since the studied soil contained Fe 2 O 3 iron oxides, based on XRF analysis, a possible pathway of persulfate activation is iron-based activation.The mechanism of persulfate activation through iron is as follows (Eq.( 5)), which yields sulfate radicals with higher oxidation potential, in comparison to persulfate.
Thermal activation of persulfate in EK oxidation is also a possible mechanism, due to increasing temperature.The mechanism of persulfate oxidation via heat catalysts, which caused sulfate radicals generation to remove pyrene, can be expressed via Eqs.(6-10) as follows 54 : Based on Eq. 6, persulfate can be catalyzed by heat which caused production of free radicals of SO 4 −• 55 .The production of sulfate radicals in solution could cause the production of the hydroxyl radical (HO • , E o = + 2.7 V) through radical inter-conversion reactions 56 .According to Eqs. 9 and 10, both SO 4 −• and HO • radicals are responsible for oxidation of recalcitrant organic matter in the heat-activated persulfate oxidation process.

Variations of soil pH
The soil pH of the lab-scale EK setup varied as expected.The soil pH near the anode was 3.82, while the pH near the cathode was about 10.22.The produced protons (H + ) and hydroxyl (OH − ) ions via electrolysis reactions (Eqs. 1 and 2) transported to the electrodes with different charges.In general, the mobility of protons is two times greater than hydroxyl ions, so the protons dominate the system and an acid front moves across the soil until it meets the hydroxyl front in a district close to cathode, where the ions may recombine to generate water.Therefore, the soil sample is separated in two different parts having various pH values, namely, high and low pH zones close to cathode and anode, respectively.The certain amounts of soil pH are highly dependent on the intensity of protons and hydroxyl transport, as well as soil geochemical features 57 .

Kinetic study
Pseudo-first-order and pseudo-second-order models were applied for kinetic studies of pyrene degradation via plotting −ln (C 0 /C t ) and 1/C t versus time, respectively (Fig. 2a and b), in the selected experimental conditions such as pH value of 7 ± 0.3, persulfate concentration of 100 mg kg -1 , pyrene con-centration of 200 mg kg -1 , Tween 80 concentration of 20 mL kg -1 , reaction time of 24 h, voltage of 1 V cm -1 , and moisture content of 100 %.The equations of each kinetic model (Eq.( 11), pseudo first-order and Eq. ( 12), pseudo second-order) are as follows 58,59 : where, C 0 and C t are the initial and residual pyrene concentrations (mg kg -1 ), respectively, t is the reaction time (min), and k n is the corresponding rate constants (n = 1 and 2).The reaction kinetics of pyrene removal through the persulfate-assisted EK remediation of pyrene-contaminated soil was best fitted for pseudo first-order reaction with the regression coefficient of 0.968 and rate constant of 0.191 min −1 .

Mineralization and intermediates of pyrene degradation
Intermediates of persulfate-assisted EK remediation of synthetically pyrene-contaminated soil were studied in selected experimental conditions and time intervals 4 and 24 h.A GC-MS analysis was performed exactly 4 h after beginning of the reaction (Fig. 3b).The main intermediate metabolites at this time were obtained as benzene (C 6 H 6 ), o-toluic acid (C 8 H 8 O 2 ), acetic acid (C 2 H 4 O 2 ), cyclopenta cycloheptene (azulene) (C 10 H 8 ), 1,3-cyclopentadiene (C 5 H 6 ), 2-ethyl-pyridine (C 7 H 9 N), naphthalene (C 10 H 8 ), decanoic acid (C 10 H 20 O 2 ), 1-chloro pyrene (C 16 H 9 Cl) and pyrene (C 16 H 10 ).It is obvious that pyrene degradation had been preceded and simpler compounds and aliphatic hydrocarbons had formed.In addition, the presence of 1-chloropyrene can be attributed to the binding of Cl -produced from electrolysis reaction of saline soil and pyrene.It was found from Fig. 3c that the main final products of pyrene degradation through persulfate-assisted EK remediation were benzoic acid (C 7 H 6 O 2 ), dimethyl phosphinic azide (C 2 H 6 N 3 OP), methoxyacetic acid (C 3 H 6 O 3 ), and hexadecane (C 16 H 34 ).Similar to the quantitative analysis, pyrene was fully removed after 24 h and converted to aliphatic and simpler components.Results indicated high mineralization rate, based on TPH analysis using Eq.13: where, P 0 is initial pyrene concentration and TPH represents the residual total petroleum hydrocarbons after oxidation.Based on the experimental results, TPH content of remediated soil sample was 92 mg kg -1 , corresponding to mineralization rate of 61 % in selected experimental conditions (Fig. 4).

Energy consumption
One of the most important issues in EK processes is the energy consumption, which affects the scale-up and full scale applications of this technology, since the high cost of energy is a limiting factor.Upon selecting the operating conditions in the labscale study, the energy consumption (P, kWh kg -1 ) was calculated using Eq. ( 14): where, V is the cell voltage (V cm -1 ), I is the average cell current (A), t is the electrolysis time (h) and M is the remediated soil mass [60][61][62] .Energy consumption was calculated in the selected experimental conditions, such as initial pH value of 7 ± 0.  concentration of 200 mg kg -1 , persulfate concentration of 100 mg kg -1 , Tween 80 content of 20 mL kg -1 , reaction time of 24 h, soil mass of 0.5 kg, moisture content of 100 %.According to the obtained results, an energy consumption of 0.11 kWh kg -1 was calculated for remediation of the studied soil to obtain a complete pyrene removal.This finding is in line with literature 63 .

Synergetic and scavenging effect
The synergetic effect of persulfate-assisted EK system was determined at selected experimental conditions using synergy index (Eq.( 15)): where, SI is the synergy index, K is the pseudo-first-order kinetic constant of pyrene removal in all experimented processes (PTEK: Persulfate-assisted EK+ addition of Tween 80, EK: electrokinetic remediation, PEK: electrokinetic remediation + persulfate addition, TEK: electrokinetic remedia-tion+ addition of Tween 80, P: oxidation through persulfate).The removal efficiency of various studied processes are presented in Fig. 5a.It is obvious that persulfate-assisted EK remediation was significantly more efficient compared to individual processes, which can be attributed to synergetic effect of all processes.This can be verified by calculation of synergy index (Fig. 5b), in which the highest value of 1.99 was obtained for persulfate-assisted EK remediation process.This significant synergy index probably can be associated with the decomposition of pyrene molecules due to solubility enhancement effect of Tween 80, oxidation influence of persulfate, as well as the oxidation function of EK reactions.These findings are in accordance with literature 64 .
In persulfate-assisted EK system, it is believed that the production of H 2 O 2 , SO 4 •− , Cl 2 and heat are responsible for pyrene destruction.Scavenging effect of non-target organics on the function of produced oxidants was investigated through addition of tert-butyl alcohol (t-BuOH), benzoquinone (BQ), sodium thiosulfate (ST) and iso-propanol (IPA) as scavenger and effective quenching agent for oxidants (Cl 2 , H 2 O 2 ) and SO 4 •− radicals 65 .It was observed that the degradation efficiency of pyrene reduced from 100 % to 63.6 %, 70.28 %, 52.3 % and 82.59 %, respectively, in the presence of IPA, t-BuOH, ST and BQ agents (Fig. 6).Results of scavenging experiments demonstrated that pyrene degradation was in direct relation with all the oxidation agents.

EK remediation of unwashed soil
Reliability of the persulfate-assisted EK remediation was also evaluated in the remediation of a real unwashed contaminated-soil, based on TPH analysis.In addition, the types of different hydrocarbons were also determined via GC-MS analysis.Initial TPH content of 912 mg kg -1 decreased to 560 mg kg -1 (38 % removal) in the selected conditions like those presented in this work.The unwashed soil sample was mainly contaminated with octadecane, hexadecane, nonadecane, eicosane, heneicosane and similar compounds (Table 2).The production of various aliphatic intermediates, such as comphene, docecane, tetradecane, etc., are clear from GC-MS analysis (Fig. 7 a and b).The high cumulative concentration of different hydrocarbons is also the main reason for lower removal efficiency of the persulfateassisted EK remediation of the unwashed soil sample.These results are in good agreement with our previous study on soil remediation with the modified Fenton oxidation 66 as well as with the literature on remediation of hydrocarbon-contaminated soil 67,68 .

Conclusions
Synthetically pyrene-spiked and naturally hydrocarbon-contaminated soil samples were subjected to persulfate-assisted EK remediation in the labscale experimental setup.A total pyrene removal of 76 % was obtained for initial pyrene concentration of 200 mg kg -1 , voltage of 1 V cm -1 , natural pH, and saturated conditions in a reaction time of 120 h.The addition of persulfate (100 mg kg -1 ) significantly enhanced the removal efficiency, and a complete removal was obtained in a much shorter reaction time of 24 h.In addition, moisture was also a key factor in the enhancement of pyrene degradation, and desirable results were found to be in higher moisture contents.A removal efficiency of 29 % in a low moisture content of 25 % and a reaction time of 120 h indicated the desired ability of the studied process to be operated in soils with low moisture contents.The initial TPH concentration of 912 mg kg -1 decreased to 560 mg kg -1 in the same conditions, as determined for pyrene removal in the synthetically contaminated soil, for a naturally hydrocarbon-contaminated soil sample.Although the removal (38 %) was lower, the total mass of removed hydrocarbon was considerable.Based on the results of this research, it is obvious that persulfate-assisted EK remediation is a viable, reliable, and efficient alternative for remediation of hydrocarbon-contaminated soil in lab-scale conditions.However, further experiments are needed to optimize the operating parameters for full-scale applications.