J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirect

www. jou rna l s . e l sev i e r . com/ jou rna l -o f - env i r onmenta l - sc i ences

An assessment of the effectiveness and impact ofelectrokinetic remediation for pyrene-contaminated soil

Sujuan Xu1,2, Shuhai Guo1,⁎, Bo Wu1, Fengmei Li1, Tingting Li1

1. Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang 110016, China2. Graduate School of Chinese Academy of Sciences, Beijing 100049, China

A R T I C L E I N F O

⁎ Corresponding author.E-mail addresses: xusujuan919@yeah.net

http://dx.doi.org/10.1016/j.jes.2014.09.0141001-0742/© 2014 The Research Center for Ec

A B S T R A C T

Article history:Received 16 December 2013Revised 3 April 2014Accepted 30 April 2014Available online 23 September 2014

The effectiveness of electrokinetic remediation for pyrene-contaminated soil was investigatedby an anode–cathode separated system using a salt bridge. The applied constant voltage was24 V and the electrode gap was 24 cm. Two types of soil (sandy soil and loam soil) wereselected because of their different conductive capabilities. The initial concentrations of pyrenein these soil samples were 261.3 mg/kg sandy soil and 259.8 mg/kg loam soil. After treatmentof the sandy soil and loam soil for seven days, 56.8% and 20.1% of the pyrene had beenremoved respectively. Under the same power supply voltage, the removal of the pollutantfrom the sandy soil was greater than that from the loam soil, due to the higher current andlower pH. Further analysis revealed that the effectiveness of electrokinetic remediation wasaffected by the energy expenditure, and was associated with changes in soil properties.© 2014 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences.

Published by Elsevier B.V.

Keywords:Electrokinetic remediationAnode–cathode separated systemElectrochemical oxidationSoilPyrene

Introduction

Electrokinetic remediation is an effective technology for theremediation of organic contaminated soil (Alcántara et al., 2008,2012; Pazos et al., 2010; Reddy et al., 2011). Electrochemicaloxidation, an important component process of electrokineticremediation, has been suggested as an effective pathway toremove organic pollutants from soil (Rahner et al., 2002;Andreottola and Ferrarese, 2008; Istrate et al., 2011). Theapplication of an electric field creates favorable conditions forredox reactions to occur in the soil between the electrodes (Jinand Fallgren, 2010). Thus, electrochemical reactions not onlyhappen on the electrode surfaces, but also within the entire soilmatrix (Torres et al., 2003; Alshawabkeh et al., 2004; Sanromán etal., 2005; Jin and Fallgren, 2010). It is surmised that reactiveradicals are generated, such as hydroxyl radicals (OHU) and

(S. Xu), shuhaiguo@iae.a

o-Environmental Science

oxygen free radicals (OU), which are strong oxidizing agents andcan oxidize organic compounds (Rahner et al., 2002; Sanromán etal., 2005; Pazos et al., 2010).

The effectiveness of electrokinetic remediation dependson many factors (Page and Page, 2002; Virkutyte et al., 2002).According to the Nernst equation, there is a negativecorrelation between pH and the redox potential, and a highredox potential is good for oxidation reactions; thus, a low pHis conducive to electrochemical oxidation reactions. Indeed,Sanromán et al. (2005) reported that acidic conditions canenhance the electrochemical oxidation reaction. The strengthof the electric current affects the electrochemical reaction rateand the overall performance of electrokinetic remediation(Baek et al., 2012; Mouli et al., 2004), and the electrochemicaloxidation rate increases with current density (Bouya et al.,2012). In addition, the electric current itself is affected by

c.cn (S. Guo).

s, Chinese Academy of Sciences. Published by Elsevier B.V.

pH0 2 4 6 8 10 12

Vol

ume

of 1

mol

/L H

NO

3/NaO

H a

dded

(μL)

-400

-300

-200

-100

0

100

200

300

400

Sandy soilLoam soil

Fig. 1 – Acid/base buffering capacities of the soils used in thisstudy.

2291J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

the conductive capability, moisture content, and voltage(Al-Hamdan and Reddy, 2008; Park et al., 2005). In previousstudies of electrokinetic remediation, the anode and cathodehave generally been located in the same electrokineticchamber. As a result of electroosmosis and electromigration,organic pollutants and H+ ions migrate within that electroki-netic cell (Acar et al., 1995; Lima et al., 2011; Cang et al., 2013),inducing changes in the redox environment. Therefore, thisconventional apparatus setup neither identifies the factorsinfluencing the electrochemical oxidation process, nor as-sesses the electrochemical oxidation ability of the anode andcathode.

In this study, an anode–cathode separated electrokineticremediation system was employed. An agar–KCl saturatedsalt bridge was applied to connect the anode and cathoderegions to form a complete electrokinetic remediation system.Due to the separation of the anode and cathode, theelectroosmosis process and the migration of H+ ions betweenthe anode and cathode chambers were impeded. Therefore,our expectation was that the results would be able to reflectthe independent effects of the anode and cathode onelectrochemical oxidation. Sandy soil and loam soil wereselected to determine the effects of different conductivecapabilities and redox conditions on the removal of thechosen organic pollutant (pyrene) from the soil. The efficiencyof the treatment was evaluated based on the amount ofpyrene removed and the impact of the technology on soilproperties (pH, conductivity, moisture content). At the sametime, the energy expenditure was also estimated.

1. Materials and methods

1.1. Soil

In order to investigate the effect of soil properties onelectrochemical oxidation, sandy soil and loam soil wereused, samples of which were taken from Shenyang City,Liaoning Province, China. The soil was air-dried at roomtemperature after removal of debris and plant roots, and thenpassed through a 20-mesh sieve. Table 1 presents thecharacteristics of the soils used in this study. The concentra-tions of iron and manganese in the sandy soil were bothhigher than those in the loam soil, indicating that redox

Table 1 –Main characteristics of the tested soils.

Parameter Soil 1 Soil 2

Texture Sandy LoamParticle size analysis (%)>20 μm 98.7 40.220–2 μm 1.1 48.7<2 μm 0.2 11.1

pH 6.18 6.44EC (μS/cm) 23.3 72.5CEC (cmol/kg) 17.88 21.65Organic matter (%) 0.20 1.34Iron (mg/kg) 36,460.4 21,512.5Manganese (mg/kg) 844.8 601.9

reactions in the sandy soil may occur at a relatively higherrate than those in the loam soil (Jin and Fallgren, 2010). Fig. 1displays the buffer capacities of the two soils at different pHvalues. The volume of 1 mol/L HNO3 added to the soil wastaken to be negative to be consistent with the definition ofbuffer capacity. It can be seen that the loam soil had a higheracid/base buffer capacity than the sandy soil at different pHvalues. Therefore, the pH of the sandy soil was expected tochange more easily than the loam soil due to the electrolysisreaction.

Pyrene, a four-ring polycyclic aromatic hydrocarbon (PAH),was used as the target compound. It is a rigid, carcinogenic,unreactive and non-electrooxidizable compound (Barathi andKumar, 2013), and has been used as a target compound formany researches (Sun and Yan, 2007; Barathi and Kumar,2013; Cang et al., 2013). The soils were artificially contaminat-ed with a pyrene–acetone solution, and the soil–acetone–pyrene mixture was then placed in a ventilation hood for oneweek until the acetone had completely evaporated and thecontaminated soil was dry. During the drying period, occa-sional stirring was necessary to increase the rate of drying andfurther ensure uniform pyrene distribution. The soil wasspiked with pyrene at a target concentration of 300 mg/kg(mass of pyrene/mass of dry soil).

1.2. Experimental setup and procedure

Fig. 2 shows the electrokinetic remediation setup usedthroughout this study and the sampling sites. The setupconsisted of two rectangular Perspex boxes (15 cm in length,10 cm in width, 4 cm in height), two pairs of columnargraphite electrodes (1 cm in diameter, 10 cm in length), twoagar–KCl-saturated salt bridges (1 cm in diameter, 30 cm inlength), an ammeter, and a stabilized DC power supply (24 V).The electrode gap was 24 cm. The reaction cell was closedwith a Perspex cover to prevent evaporation of water.

As summarized in Table 2, four experiments weredesigned to investigate the pyrene removal rate underdifferent treatment conditions. Control 1 and Control 2 werecontrol experiments with no electric field applied; while Test 1and Test 2 were treated with a constant electric potential(24 V). Approximately 700 g of pyrene-spiked soil with the

Sampling spot Salt bridge

6 cm 6 cm 6 cm 6 cm

Fig. 2 – Schematic representation of the experimental setup and sampling positions. (1) electrode; (2) salt bridge; (3) soil; (4)sampling spot; (5) power supply; (6) relay; (7) timer; (8) amperemeter.

2292 J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

initial moisture content (about 20%, W/W) was carefullystacked into each Perspex box. The soil samples were left toequilibrate for one day before applying the voltage gradient. Afraction of the soil in each Perspex box was obtained todetermine the initial soil pH, moisture content, and concen-tration of pyrene. Sampling and monitoring events wereperformed after 0, 1, 3, 5, and 7 days. For each samplinground, the samples at the same distance from the threesampling lines were thoroughly mixed together to form onecomposite sample for analysis of the moisture content, pH,and pyrene concentration. The current intensity in theammeter was recorded periodically. The initial concentra-tions of pyrene were 261.3 mg/kg sandy soil and 259.8 mg/kgloam soil.

1.3. Chemicals and analytical methods

The analytical methods used were performed according to Lu(2000). Electrical conductivity (EC) and pH were measuredfrom a 1:2.5 soil weight to water volume ratio extract using aconductivity meter (DDS-11A, INESA, China) and pH meter(PHS-3E, INESA, China), respectively. Moisture content wasestablished by heating the samples at 105°C until a constantweight was achieved. The total levels of Fe and Mn in the soilwere determined through the atomic absorption spectrum(AAS) after digestion with HF–HNO3–HClO4. The acid/basebuffering capacity was determined using the methods de-scribed by Yeung et al. (1997).

Pyrene was extracted using dichloromethane/acetone (5:5,V/V) (Huang et al., 2012), and then its concentration wasanalyzed using high performance liquid chromatography

Table 2 – Summary of experimental conditions applied.

Run Soil(700 g/chamber)

Appliedvoltage (V)

Sampling time(day)

Control 1 Sandy soil 0 0, 1, 3, 5, 7Control 2 Loam soil 0 0, 1, 3, 5, 7Test 1 Sandy soil 24 0, 1, 3, 5, 7Test 2 Loam soil 24 0, 1, 3, 5, 7

(HPLC, Waters, USA) equipped with a variable wavelengthfluorescence detector (FLD, model 2475, Waters, USA) and aPAH column (dimensions: 250 mm × 4.6 mm; particle size:5 μm; Waters, USA). Prior to injection, the pyrene extract wasfiltered through a 0.22-μm Teflon filter. The column temper-ature was 25°C and the injection volume was 10 μL. Themobile phase was a mixture of acetonitrile and water (90:10,V/V). The excitation and emission wavelengths of pyrenewere 332 and 378 nm, respectively.

1.4. Quality control

The concentration of pyrene was determined by externalstandard method. The linearity of the calibration curve waschecked, giving an average regression coefficient (R2) of 0.999.Duplicate standard samples and sample blanks were com-monly injected to ensure a uniform response and that thesystem remained uncontaminated. All analytical determina-tions were performed in triplicate, and the results werepresented herein as means.

2. Results and discussion

2.1. Variation of the electric current

The electric current was recorded at regular time intervalsduring the experiment and the results are shown in Fig. 3.During the progress of the experiments, the current displayedthe same tendency in both Test 1 and Test 2; that is, toincrease up to a maximum value, and then decrease untilreaching a steady state. This was similar to the findingsreported in some previous studies (Reddy et al., 2006;Al-Hamdan and Reddy, 2008; Tsai et al., 2010; Lu et al., 2012).In Test 1, the electric current increased during the first 72 hrafter the voltage had been applied, decreased between 72 and96 hr, and then plateaued after 96 hr (3.4–8.9 mA). In Test 2,the current increased rapidly from 13.5 to 15.9 mA during thefirst 24 hr, then decreased to 0.06 mA. The current in Test 1was higher than that in Test 2, except during the first 24 hr.

Time (hr)0 24 48 72 96 120 144 168

Elec

tric

curr

ent (

mA

)

0

20

40

60

80

100

120

140

Test 1 (sandy soil)Test 2 (loam soil)

Time (hr)0 24 48 72 96 120 144 168

Elec

tric

curr

ent (

mA

)

0

4

8

12

16

Initial moisture content: 21%

Fig. 3 – Electric current measured during the experiment. (Inset: current in Test 2).

2293J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

The naturally high EC of loam soil was responsible for theinitially higher current in Test 2. However, the maximumcurrent in Test 1 was approximately eight times higher thanthat in Test 2, and the total amount of electricity passingthrough the sandy soil was approximately six times higherthan that through the loam soil. These findings indicated thatthe effect of electrochemical oxidation in the sandy soil wasstronger than that in the loam soil under the same electricvoltage power supply. After the maximum current had beenreached, the availability of pore water and mobile ionsdecreased gradually, caused by the electrolysis of water andthe precipitation of ions in the cathode region, resulting in adecrease in the electric current.

2.2. Decomposition of pyrene under electrokinetic treatment

The total removal rates of pyrene at the end of all the tests areshown in Fig. 4. The amount of pyrene in Control 1 did notdecrease (data not shown), whereas the total removal rate ofpyrene in Control 2 was 1.0%, indicating that the naturalpyrene removal efficiency of the loam soil was higher thanthat of the sandy soil, despite both being low. The totalremoval rate of pyrene in Test 1 was 56.8%, which was 2.8

Total Anode chamber Cathode chamber

Pyre

ne re

mov

al (%

)

0

10

20

30

40

50

60

70

80

90

100Control 2 Test 1 Test 2

Fig. 4 – Pyrene removal rates after 7 days of treatment.

times higher than that in Test 2, and the removal rates ofpyrene in the anode and cathode chambers in Test 1 wereboth higher than those in Test 2. This was because the currentin Test 1 was higher than that in Test 2, thereby increasingthe electrochemical oxidation rate in the former (Bouya et al.,2012). The generation of reactive radicals near the electrodesand their subsequent diffusion into the soil matrix wouldallow them to attack the active site of pyrene or oxidize it in astepwise manner (ECP, 2003; Sanromán et al., 2005;Andreottola and Ferrarese, 2008), resulting in the removal ofpyrene. These oxidation reactions would have occurred atall of the interfaces between the soil and the pore water,consistent with the microconductor principle (Rahner et al.,2002). In addition, pollutants adsorb more readily onto smallgrains and soil organic matter (Luo et al., 2005a). The grainsize of the loam soil was smaller than that of the sandy soil,and the organic matter content of the former was higher thanthat of the latter (Table 1). The combined effects may haveresulted in the removal of pyrene being more difficult in Test2. Fig. 4 also reveals that the pyrene removal rate in the anodechamber was higher than that in the cathode chamber forboth soils, indicating that anode electrochemical oxidation isthe more effective pathway for pyrene removal.

Fig. 5 presents the pyrene removal rates from the anode tothe cathode at different points of times throughout Test 1 andTest 2. The concentrations of pyrene in Test 1 and Test 2decreased over time; in general, the main changes occurredduring the first five days, and the increase in the pyreneremoval rate in Test 1 was greater than that in Test 2. Thevariations in the removal rates were related to changes in thecurrent (Figs. 3 and 5). A rapid rate of removal of pyreneoccurred just at the time when the current increased;correspondingly, as the current decreased in the later stagesof treatment, the removal rate exhibited a slower increase.However, the increase in the pyrene removal rate variedthroughout the soil medium. The pyrene removal rate in theanode chamber was higher than that in the cathode chamber,especially in Test 1. With the extension of treatment time, thedifference in the removal rates between the anode chamberand cathode chamber became increasingly significant. In the

Anode chamber

Distance from anode (cm)0

Pyre

ne re

mov

al (%

)

0

20

40

60

80

100

120Cathode chamber

Distance from cathode (cm)036912

Pyre

ne re

mov

al (%

)

0

20

40

60

80

100

120Anode chamber

Distance from anode (cm)

Pyre

ne re

mov

al (%

)

0

20

40

60

80

100

120Cathode chamber

Distance from cathode (cm)036912

Pyre

ne re

mov

al (%

)

0

20

40

60

80

100

120

0 day 1 day 3 days 5 days 7 days

2tseT1tseT

21963021963

Fig. 5 – Changes in pyrene removal rate under different modes of electrokinetic operation.

2294 J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

anode chamber of Test 1, the removal rates of pyrenedecreased with increasing distance from the anode duringthe first three days. In the later stages, the removal rate in themiddle of the anode chamber was lower than that anywhereelse, indicating that pyrene concentrated there. The removalrates were not significantly different in the cathode chamberof Test 1. In Test 2, the pyrene removal rate decreased withincreasing distance from the anode in the anode chamber, butincreased with increasing distance from the cathode in thecathode chamber. This indicated that, in Test 2, the pyrenewas mobilized towards the cathode region via pore-liquidtransport for electroosmosis (Lima et al., 2011).

2.3. Effects on soil properties

2.3.1. Variation of soil moisture contentWhen an electric field is applied to a wet soil mass, theelectrolysis of water occurs, which induces a decrease in soilmoisture content (Acar et al., 1995; Page and Page, 2002). Inthis study, the soil moisture content at the beginning of theexperiment was about 21.0%. It can be seen from Fig. 6 thatthe soil did not dry out completely, even though no fluid wasprovided at the electrodes or in other regions. The moisturecontent decreased over time in both Test 1 and Test 2, withthe main changes occurring in the first five days after thebeginning of the experiment. The reason for this was that the

Anode chamber

Distance from anode (cm)

Moi

stur

e co

nten

t (%

)

0

5

10

15

20

25Cathode chamber

Distance from cathode (cm)036912

Moi

stur

e co

nten

t (%

)

0

5

10

15

20

25

1tseT

0 day 1 day 3 days

219630

Fig. 6 – Changes in soil moi

main changes in current also occurred during this period(Fig. 3), which induced electrochemical reactions to consumesoil water. After 7 days of treatment, the moisture content inthe anode chamber of Test 1 had decreased to 11.0%, 12.2%,and 11.5% at 0, 6, and 12 cm from the anode, respectively; inthe cathode chamber, the moisture content had decreased to19.7%, 20.3%, and 17.5% at 0, 6, and 12 cm from the cathode,respectively. The moisture content was high in the middle ofeach soil chamber. In Test 2, the moisture content increasedwith decreasing distance to the cathode in the cathodechamber. Obviously, the moisture content at the cathodewas higher than the initial value (2.1%) and a strongermoisture content gradient was visible in the cathode chamberof Test 2. This was because of the effect of electroosmosis,whereby the water flowed from anode to cathode (Park et al.,2005). Furthermore, electroosmosis decreases with increasingacidity and ion concentration in the pore fluid (Yeung, 2006).The current in Test 1 was higher than that in Test 2, and thusmore H+ ions were produced in the former than in the latter.This higher content of H+ ions would have resulted in greaternet positive charge on the surfaces of soil particles, decreasingthe electroosmotic flow to the cathode (Sanromán et al., 2005).In addition, the loam soil was denser than the sandy soil, andelectroosmosis function is more obvious in denser soil(Probstein and Hicks, 1993). Thus, for the reasons mentionedabove, the occurrence of electroosmosis as a result of changes

Anode chamber

Distance from anode (cm)

Moi

stur

e co

nten

t (%

)

0

5

10

15

20

25 Cathode chamber

Distance from cathode (cm)036912

Moi

stur

e co

nten

t (%

)

0

5

10

15

20

25

2tseT

5 days 7 days

219630

sture content over time.

Anode chamber

Distance from anode (cm)

pH

0

2

4

6

8

10Cathode chamber

Distance from cathode (cm)036912

pH

0

2

4

6

8

10 Anode chamber

Distance from anode (cm)

pH

0

2

4

6

8

10Cathode chamber

Distance from cathode (cm)036912

pH

0

2

4

6

8

10

2tseT1tseT

0 day 1 day 3 days 5 days 7 days

0 3 6 9 12 0 3 6 9 12

Fig. 7 – Variation in pH across the soil sections during the experiment.

2295J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

in moisture content was hindered in Test 1. As a whole, theamount of soil pore fluid consumed in Test 1 was higher thanthat in Test 2, because the electrolysis reaction occurred at arelatively higher rate in Test 1 due to the higher current. Thefaster electrolysis reaction indicated that the electrochemicaloxidation was more intense in Test 1, and that the removalrate of the pollutant was higher.

2.3.2. Variation of soil pHElectrolysis of water generates H+ ions at the anode and OH−

ions at the cathode (Acar et al., 1995; Zhou et al., 2003), whichinduce the changes in soil pH. During the experimentalprocess, soil pH was measured in sections from the anode tothe cathode after each sampling. As a result of the electrolysisof water, the soil pH showed significant changes during theexperimental period (Fig. 7). Due to the separation of theanode and cathode by the salt bridge, the anode chamber andcathode chamber were separately affected by H+ and OH− ions.The pH in the cathode chamber increased over time, while inthe anode chamber it decreased over time during the first5 days after the device had been energized. The pH thenremained steady after 5 days in both Test 1 and Test 2,ranging from 2.57 to 2.83 in the anode chamber and from 9.32

Anode chamber

Distance from anode (cm)

Elec

trica

l con

duct

ivity

(μS/

cm)

0

300

600

900

1200

1500

1800Cathode chamber

Distance from cathode (cm)036912

Elec

trica

l con

duct

ivity

(μS/

cm)

0

300

600

900

1200

1500

1800

1tseT

219630

0 day 1 day 3 days

Fig. 8 – Variation in EC across the soil se

to 9.34 in the cathode chamber at the end of Test 1, and from3.69 to 4.83 in the anode chamber and from 8.23 to 8.76 in thecathode chamber at the end of Test 2. Thus, it can be seenthat, using this experimental setup, the soil pH can rapidly(within five days) turn into a strong acid or strong base. Thevariations of pH in Test 2 were smaller than those in Test 1,perhaps because the current in Test 2 was lower, meaningthat fewer H+ and OH− ions were generated, resulting in lesschange of pH in Test 2. In addition, the high buffer capacity ofthe loam soil could also have contributed to this phenomenon(Luo et al., 2005a).

2.3.3. Variation of electrical conductivity (EC)Fig. 8 presents the change in EC across the soil over timefor the different treatments. After electrokinetic treatment,the EC of each type of soil increased compared to the initialvalue. The EC of the anode chamber was higher than thatof the cathode chamber; and in the cathode chamber, theEC decreased with increasing distance from the cathode.This indicated that the electrolysis reaction and acidifica-tion of the soil resulted in a large number of ions releasedinto the soil, in turn increasing the EC (Fan et al., 2011;Alshawabkeh et al., 2004). In general, the change in EC in Test

Elec

trica

l con

duct

ivity

(μS/

cm)

Elec

trica

l con

duct

ivity

(μS/

cm)

Anode chamber

Distance from anode (cm)

0

300

600

900

1200

1500

1800Cathode chamber

Distance from cathode (cm)036912

0

300

600

900

1200

1500

1800

2tseT

219630

5 days 7 days

ctions in the different experiments.

Table 3 – Relationship between pyrene removal rate andsoil properties (n = 24).

Moisturecontent

pH EC Pyreneremoval rate

Moisturecontent

1 0.712 ⁎⁎ −0.700 ⁎⁎ −0.535 ⁎⁎

pH 0.712 ⁎⁎ 1 −0.846 ⁎⁎ −0.928 ⁎⁎

EC −0.700 ⁎⁎ −0.846 ⁎⁎ 1 0.829 ⁎⁎

Pyreneremoval rate

−0.535 ⁎⁎ −0.928 ⁎⁎ 0.829 ⁎⁎ 1

⁎⁎ Correlation is significant at the 0.01 level (two-tailed test).

2296 J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

1 was higher than that in Test 2, because of the more violentelectrolysis reaction in Test 1 due to the higher current, in otherwords, the ion content in Test 1 was higher than that in Test 2.

2.3.4. Relationships between pyrene removal rate and soilpropertiesThe above results indicated that anode electrochemicaloxidation was the main pathway for pyrene removal, whichwas affected by the soil pH, soil moisture content, and EC. Thedistribution change profiles of these factors all had a relativetendency that changed during 0–5 days, and then stabilized.

Pearson correlation analysis was performed to further studythe relationships among the moisture content, pH, EC, andpyrene removal rate, and the results are shown in Table 3. Asignificant negative correlation was found between the pH andpyrene removal rate (R = −0.928, p < 0.01), while a significantpositive correlation was found between the EC and the pyreneremoval rate (R = 0.829, p < 0.01). The electrokinetic processresulted in the electrolysis of water, producing H+ andOH− ions,which then induced changes in the soil pH and EC (Grundl andMichalski, 1996; Reddy and Cameselle, 2009), thereby affectingthe oxidation environment and current, which in turn influ-enced the removal of pyrene. Therefore, the main factorinfluencing the removal of this pollutant was the change insoil pH induced by the electrolysis reaction during electroki-netic remediation, which is a similar conclusion to that madein other studies (Yeung et al., 1997; Luo et al., 2005a).

In addition, a significant negative correlation was foundbetween the moisture content and the pyrene removal rate(R = −0.535, p < 0.01). Here, the electrolysis reaction induced thereduction in moisture content; and the faster the electrolysisreaction, the more intense was the electrochemical oxidationreaction. Furthermore, electrochemical oxidation resulted in theremoval of pyrene. Therefore, the declining moisture contentillustrated that the strength of electrochemical oxidation wasgradually decreasing. It is surmised that if the soil moisturecontent was maintained at a certain level, the conditions wouldbemore conducive to electrochemical oxidation, and the organicpollutant would be removed more effectively.

Table 4 – Energy consumption of the electrokinetic process afte

Test 1

Elapsed time (day) 1 3 5Pyrene removal (%) 12.54 44.58 53.59Energy consumption (10−3 kWh) 4.80 87.72 129.60

2.4. Energy expenditure

We calculated the energy expenditure according to thefollowing equation (Luo et al., 2005b; Tsai et al., 2010):

P ¼Z

VIdt

where, P (kWh) is the energy expenditure; V (V) is the voltage; I(A) is the current; and t (hr) is the time. In this study, aconstant electric potential (24 V) was applied. So, the energyexpenditure was related directly to the time integral of thecurrent across the reaction system. Table 4 lists the energyexpenditures of the various experiments.

By the end of the experiment, the calculated energyconsumptions in Test 1 and Test 2 were 138.09 × 10−3 and21.86 × 10−3 kWh, respectively. The relatively higher energyconsumption for Test 1 occurred because of the high currentthat developed during the experiment (Fig. 3). Under thecondition of constant voltage, the energy expenditure isproportional to the current. However, at the same voltageand the same treatment time, the removal efficiency ofpyrene in Test 1 was higher than that in Test 2. A significantpositive correlation existed between pyrene removal rate andthe energy expenditure (R = 0.994; p < 0.01). The high energyexpenditure enhanced the electrochemical reaction, furtherincreasing the removal rate of pyrene.

3. Conclusions

During this research, an anode–cathode separated electroki-netic remediation system was set up and run on thelaboratory scale to assess the effectiveness of electrokineticremediation for the removal of pyrene from two types of soil.The results demonstrated that anode electrochemical oxida-tion was the main effective pathway for the removal ofpyrene. We also found that the removal efficiency wasaffected by energy expenditure, and pH was the mostimportant factor affecting pyrene removal. Overall, thisstudy suggests that if an organic contaminated soilpossesses the appropriate level of conductive capability,and suitably low pH conditions, using electrokinetic remedi-ation to restore the soil may be easier and more successful.

Acknowledgments

This work was supported by the Knowledge InnovationProject Key-Direction Project Sub-project of the ChineseAcademy of Sciences (No. KZCX2-EW-407), the NationalNatural Science Foundation of China (Nos. 21047006,21107119) and the Key Project of Science and Technology ofChina (No. 2013ZX07202-007).

r seven days of treatment.

Test 2

7 1 3 5 756.85 11.09 15.98 18.41 20.11

138.09 8.58 18.25 21.20 21.86

2297J O U R N A L O F E N V I R O N M E N T A L S C I E N C E S 2 6 ( 2 0 1 4 ) 2 2 9 0 – 2 2 9 7

R E F E R E N C E S

Acar, Y.B., Gale, R.J., Alshawabkeh, A.N., Marks, R.E., Puppala, S.,Bricka, M., et al., 1995. Electrokinetic remediation: basics andtechnology status. J. Hazard. Mater. 40 (2), 117–137.

Alcántara, T., Pazos, M., Cameselle, C., Sanroman, M., 2008.Electrochemical remediation of phenanthrene fromcontaminated kaolinite. Environ. Geochem. Health 30 (2),89–94.

Alcántara, M.T., Gómez, J., Pazos, M., Sanromán, M.A., 2012.Electrokinetic remediation of lead and phenanthrene pollutedsoils. Geoderma 173–174, 128–133.

Al-Hamdan, A.Z., Reddy, K.R., 2008. Transient behavior of heavymetals in soils during electrokinetic remediation.Chemosphere 71 (5), 860–871.

Alshawabkeh, A.N., Sheahan, T.C., Wu, X., 2004. Coupling ofelectrochemical and mechanical processes in soils under DCfields. Mech. Mater. 36 (5), 453–465.

Andreottola, G., Ferrarese, E., 2008. Application of advancedoxidation processes and electrooxidation for the remediationof river sediments contaminated by PAHs. J. Environ. Sci.Health A 43 (12), 1361–1372.

Baek, K., Mao, X., Ciblak, A., Alshawabkeh, A.N., 2012. Greenremediation of soil and groundwater by electrochemicalmethods. GeoCongress 2012: State of the Art and Practice inGeotechnical Engineering, pp. 4348–4357.

Barathi, P., Kumar, A.S., 2013. Electrochemical conversion ofunreactive pyrene to highly redox-active 1,2-quinonederivatives on a carbon nanotube-modified gold electrodesurface and its selective hydrogen peroxide sensing. Langmuir29 (34), 10617–10623.

Bouya, H., Errami, M., Salghi, R., Zarrouk, A., Assouag, A., Zarrok,H., et al., 2012. Electrooxidation of cypermethrin pesticide: acomparative study of SnO2 and boron doped diamond anodes.JCPRC5 4 (7), 3468–3477.

Cang, L., Fan, G.P., Zhou, D.M., Wang, Q.Y., 2013.Enhanced-electrokinetic remediation of copper-pyreneco-contaminated soil with different oxidants and pH control.Chemosphere 90 (8), 2326–2331.

ECP, 2003. White Paper: Electro Chemical Geo Oxidation(ECGO)—A Synthesis. In: EPA (Ed.), Electrochemical Process,Wayne, PA (USA).

Fan, G.P., Cang, L., Zhou, D.M., Zhou, L.X., 2011. Effect of differentsoil types on the remediation of copper–pyrene compoundcontaminated soils by EK-oxidation process. Environ. Sci. 32(11), 3435–3439.

Grundl, T., Michalski, P., 1996. Electroosmotically driven waterflow in sediments. Water Res. 30 (4), 811–818.

Huang, D.N., Guo, S.H., Li, T.T., Wu, B., 2012. Coupling interactionsbetween electrokinetics and bioremediation for pyreneremoval from soil under polarity reversal conditions.Clean-Soil. Air. Water. 41 (4), 383–389.

Istrate, I., Badea, A., Cocârţă, D., Bulmău, C., 2011. Electrochemicaltreatment of PAH contaminated sediments and human healthrisk assessment. Present. Environ. Sustain. Dev. 5, 101–112.

Jin, S., Fallgren, P.H., 2010. Electrically induced reduction oftrichloroethene in clay. J. Hazard. Mater. 173 (1–3), 200–204.

Lima, A.T., Kleingeld, P.J., Heister, K., Loch, J.P.G., 2011. Removal ofPAHs from contaminated clayey soil by means ofelectro-osmosis. Sep. Purif. Technol. 79 (2), 221–229.

Lu, R.K., 2000. Soil Agricultural Chemical Analysis Method. ChinaAgriculture Press, Beijing, China.

Lu, P., Feng, Q., Meng, Q.J., Yuan, T., 2012. Electrokineticremediation of chromium- and cadmium-contaminated soilfrom abandoned industrial site. Sep. Purif. Technol. 98,216–220.

Luo, Q., Zhang, X., Wang, H., Qian, Y., 2005a. Mobilization ofphenol and dichlorophenol in unsaturated soils bynon-uniform electrokinetics. Chemosphere 59 (9), 1289–1298.

Luo, Q., Zhang, X., Wang, H., Qian, Y., 2005b. The use ofnon-uniform electrokinetics to enhance in situ bioremediationof phenol-contaminated soil. J. Hazard. Mater. 121 (1), 187–194.

Mouli, P.C., Mohanb, S.V., Reddy, S.J., 2004. Electrochemical processesfor the remediation of wastewater and contaminated soil:emerging technology. J. Sci. Ind. Res. India 63, 11–19.

Page, M.M., Page, C.L., 2002. Electroremediation of contaminatedsoils. J. Environ. Eng. 128 (3), 208–219.

Park, J.Y., Kim, S.J., Lee, Y.J., Baek, K., Yang, J.W., 2005. EK-Fentonprocess for removal of phenanthrene in a two-dimensionalsoil system. Eng. Geol. 77 (3–4), 217–224.

Pazos, M., Rosales, E., Alcántara, T., Gómez, J., Sanromán, M., 2010.Decontamination of soils containing PAHs byelectroremediation: a review. J. Hazard. Mater. 177 (1–3), 1–11.

Probstein, R.F., Hicks, R.E., 1993. Removal of contaminants fromsoils by electric fields. Science 260 (5107), 498–503.

Rahner, D., Ludwig, G., Röhrs, J., 2002. Electrochemically inducedreactions in soils — a new approach to the in-situ remediationof contaminated soils? Part 1. The microconductor principle.Electrochim. Acta 47 (9), 1395–1403.

Reddy, K.R., Cameselle, C., 2009. Electrochemical RemediationTechnologies for Polluted Soils, Sediments and Groundwater.John Wiley & Sons, Hoboken, New Jersey.

Reddy, K.R., Ala, P.R., Sharma, S., Kumar, S.N., 2006. Enhancedelectrokinetic remediation of contaminated manufactured gasplant soil. Eng. Geol. 85 (1), 132–146.

Reddy, K.R., Darko-Kagya, K., Al-Hamdan, A.Z., 2011.Electrokinetic remediation of pentachlorophenolcontaminated clay soil. Water Air Soil Pollut. 221 (1–4), 35–44.

Sanromán, M., Pazos, M., Ricart, M., Cameselle, C., 2005.Decolourisation of textile indigo dye by DC electric current.Eng. Geol. 77 (3–4), 253–261.

Sun, H.W., Yan, Q.S., 2007. Influence of Fenton oxidation on soilorganic matter and its sorption and desorption of pyrene. J.Hazard. Mater. 144 (1), 164–170.

Torres, R.A., Torres,W., Peringer, P., Pulgarin, C., 2003. Electrochemicaldegradation of p-substituted phenols of industrial interest on Ptelectrodes. Attempt of a structure–reactivity relationshipassessment. Chemosphere 50 (1), 97–104.

Tsai, T.T., Sah, J., Kao, C.M., 2010. Application of iron electrodecorrosion enhanced electrokinetic-Fenton oxidation toremediate diesel contaminated soils: a laboratory feasibilitystudy. J. Hydrol. 380 (1–2), 4–13.

Virkutyte, J., Sillanpää, M., Latostenmaa, P., 2002. Electrokineticsoil remediation—critical overview. Sci. Total Environ. 289 (1),97–121.

Yeung, A.T., 2006. Contaminant extractability by electrokinetics.Environ. Eng. Sci. 23 (1), 202–224.

Yeung, A.T., Hsu, C., Menon, R.M., 1997. Physicochemicalsoil–contaminant interactions during electrokineticextraction. J. Hazard. Mater. 55 (1–3), 221–237.

Zhou, D.M., Zorn, R., Czurda, K., 2003. Electrochemicalremediation of copper contaminated kaolinite by conditioninganolyte and catholyte pH simultaneously. J. Environ. Sci. 15 (3),396–400.