Chemosphere 233 (2019) 17e24

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Enhanced electrokinetic remediation and simulation of cadmium-

contaminated soil by superimposed electric field
Zicheng Sun a, b, Bo Wu a, Penghong Guo c, Sa Wang a, Shuhai Guo a, *
a
Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang, 110016, China
b
University of Chinese Academy of Sciences, Beijing, 100049, China
c
Arizona State University, Tempe, AZ, 85281, USA

h i g h l i g h t s g r a p h i c a l a b s t r a c t

The superimposed electric field was

developed by switching two sets of
electrodes.

A superimposed region was formed,
which accelerated the move of the
acid front.

The ‘focusing’ region of Cd was
compressed by the migration of the
acid front.

The transport of Cd under the
superimposed electric field was
simulated by a model.

The removal efficiency of Cd was
increased in the simulated optimal
situations.

a r t i c l e i n f o a b s t r a c t

Article history: The ‘focusing’ effect has become a limiting factor for the removal of heavy metals from soils by elec-
Received 5 March 2019 trokinetic (EK) remediation. A superimposed electric field EK (SE-EK) method is proposed to address this
Received in revised form problem. Two sets of fixed electrodes placed at different positions were switched to move the ‘focusing’
23 May 2019
region of Cd to the cathode by controlling the location of the pH jumping front. Moreover, a model was
Accepted 27 May 2019
Available online 27 May 2019
established to simulate and optimize the process of Cd transport in soil under the superimposed electric
field. Results showed that, after 35 d of SE-EK remediation, Cd was mainly accumulated in the soil section
Handling Editor: X. Cao near the cathode (S5), where the acid and alkaline fronts converged. The removal rate of Cd in the soil
sections from S1 to S4 reached 87.60%, which was 6.13 times that in conventional EK remediation.
Keywords: Meanwhile, the energy utilization efficiency in SE-EK was 6.38 times that in conventional EK. The pH
Electrokinetic remediation changes and Cd distribution during the SE-EK experiment were simulated well, with good agreement
Cadmium between the modeled and experimental data. The removal of Cd in SE-EK remediation could therefore be
Superimposed electric field optimized through simulating the distribution of Cd in five situations with differences in switching time
‘Focusing’ effect
and electrode position. This research provides valuable technical support for effective EK remediation of
Simulation
heavy metal contaminated soil.
© 2019 Published by Elsevier Ltd.

1. Introduction

Electrokinetic (EK) remediation is an effective technology for

* Corresponding author.
removing heavy metal contaminants from soil (Kim et al., 2011;
E-mail address: shuhaiguo@iae.ac.cn (S. Guo). Virkutyte et al., 2002). It applies a low potential gradient to

https://doi.org/10.1016/j.chemosphere.2019.05.233
0045-6535/© 2019 Published by Elsevier Ltd.
18 Z. Sun et al. / Chemosphere 233 (2019) 17e24

transport mobile heavy metals mainly by electromigration and 2.2. EK experiment design
electroosmosis (Acar and Alshawabkeh, 1993). Previous studies
have reported that a ‘focusing’ effect, caused by the accumulation of The apparatus consisted of a Perspex soil chamber
heavy metals in the region of pH jumping, occurs in the process of (32 cm  17 cm  10 cm), a DC power supply system, a monitoring
EK remediation (Giannis and Gidarakos, 2005; Probstein and Hicks, system, and cylindrical stainless-steel electrodes (10 cm in length
1993). The ‘focusing’ effect has become a limiting factor for the and 1 cm in diameter) (Fig. 1a). The DC power supply system
removal of heavy metals from soil through EK remediation (Li et al., included two DC power supplies and a time switch. The monitoring
2012a). system could monitor and record current every 30 min with a
Many enhancement methods have been reported to control the precision of 0.1 mA.
‘focusing’ effect by improving the mobility of heavy metals. In Two experiments were conducted: a conventional EK treatment
addition to adding chelating/complexing agents (Giannis et al., and a SE-EK treatment. The conventional EK excluded the switching
2009; Yuan et al., 2016), acidic solutions (Villen-Guzman et al., of electrodes in order to evaluate the distribution of Cd during the
2014) and using ionic exchange membranes (Kim et al., 2005), experiment. The distance between the anodes and cathodes was
the commonly used method is approaching anodes EK (AAs-EK) 25 cm. In SE-EK, two sets of electrodes were inserted into the soil
remediation. This method has been applied to move the ‘focusing’ (Fig. 1a). The superimposed electric field was developed in two
region of heavy metals by providing progressive soil acidification steps:
(Li et al., 2012a, 2012b; Shen et al., 2007; Wei et al., 2016; Zhang
Step 1. the primary anode and cathode were operated the same
et al., 2016). Wei et al. (2016) and Zhang et al. (2016) reported
way as in the conventional experiment.
that the ‘focusing’ region was compressed near the cathode and the
contaminants were 2.34- and 3.5-fold enriched. Shen et al. (2007) Step 2. the secondary anode and cathode were operated. The
found enrichment of more than 99% of total Cd at the cathode af- objective was to increase the acid-front area and cause the region of
ter remediation with AAs-EK and various enhancement methods. Cd accumulation to migrate towards the cathode.
Modelling is a necessary tool to simulate the transport of con- The primary anode and cathode were 25 cm apart. The switch-
taminants, especially in cases of complex operation, such as when ing time and position of the secondary anodes were determined by
applying the AAs-EK method. Several numerical models have been the distribution of Cd during the conventional EK experiment. The
developed to simulate the transport of heavy metals secondary anodes needed to be placed on the left-hand side (near
(Asadollahfardi et al., 2016; Kim et al., 2003; Mascia et al., 2007; the primary anodes) of the accumulated-Cd region to accelerate the
Masi et al., 2016; Rezaeea et al., 2019). However, most of them are transport of Cd. The position of the secondary cathodes was
based on a conventional electric field with fixed electrodes and determined by the migration velocity of the acid and alkaline front
constant initial conditions. Few studies have been conducted with a estimated from the changes of soil pH in conventional EK. The
model that simulates the transport of heavy metals by including the purpose was to control the location of the pH jumping front.
process of switching electrodes because of the complexity involved Five and 6 kg of wet soil (moisture content: 25% w/w) were
with the changes in the initial conditions. stacked into the chambers for conventional EK and SE-EK, respec-
In this paper, we propose a novel superimposed electric field EK tively. The height of the soil was approximately 8 cm. A constant
(SE-EK) method for the removal of heavy metals from soil. A electrical potential (1 V cm1) was applied for 35 d for each
piecewise model was established to simulate the transport of Cd experiment. The average moisture content was maintained artifi-
under the superimposed electric field. The superimposed electric cially with distilled water during the experiments.
field was developed by switching two sets of fixed electrodes The soil was divided into five sections from anode to cathode
placed at different positions. The specific aims of this experiment (S1eS5) (Fig. 1b). The two experiments were carried out in tripli-
were (1) to evaluate the removal efficiency of Cd by SE-EK reme- cate. In conventional EK, the soil samples were collected every 5 d.
diation; and (2) to simulate the distribution of Cd in soil under the In SE-EK, they were collected after Step 1 and Step 2. At each
superimposed electric field. The findings are helpful for effective sampling point, a soil column of 1 cm in diameter and 1 cm in depth
decontamination of heavy metals from soil using the EK technique. was dug using a stainless-steel spoon. The five samples from each
sampling line were mixed uniformly. The pit formed by sampling
2. Materials and methods was backfilled with the soil in the same sampling line to reduce the
influence of sampling on the transport of Cd.
2.1. Materials

The Cd-contaminated soil used in this study was artificially 2.3. Model
prepared. Uncontaminated soil was collected from the topsoil layer
(0e30 cm) in the Shenbei Area, Shenyang, China. The physical and A numerical model was established to simulate the transport of
chemical properties of the collected natural soil were as follows: Cd under the electric field. The soil used in this study was unsat-
the clay proportion was 38.6%; the soil pH was 6.11; the cation urated. The water content of soil has an influence on the tortuosity
exchange capacity was 12.3 cmol kg1; the organic matter content factor and affects the ionic mobility (Chou et al., 2012). Therefore,
was 1.06%; the zeta potential was 14.1 mV; the conductivity was the soil was set to be unsaturated to make the simulation more
270.1 mS cm1; and the carbonate content was 1.51%. accurate. The following assumptions were made in this model: (1)
The soil sample was air-dried and ground to pass through a 20- the soil is isotropic and isothermal; (2) soil particles are non-
mesh sieve. Thirty kilograms of uncontaminated soil were divided conductive and their surface conductivity is negligible; (3) elec-
into twelve equal parts. Every part of the soil was spiked with 1 L trophoresis can be ignored, because colloid migration is hindered
(700 mg L1) Cd(NO3)2$4H2O solution. The nitrate solution and the by the immobile phase of the porous medium (Yu and Neretnieks,
soil were mixed thoroughly and maintained at room temperature 1997); (4) the hydraulically driven advection flow is negligible
until dry. The dry soil was ground and passed through the 20-mesh because of its lower order of magnitude compared to electroos-
sieve before being mixed together. All the Cd-contaminated soil motic flow (5) the pore geometry characteristics (e.g. porosity,
was equilibrated for 7 d and left at room temperature. The final Cd tortuosity); do not change over time (because of the current lack of
concentration in the contaminated soil was 98.31 mg kg1. understanding in terms of formalizing them as a function); and (6)
 Z. Sun et al. / Chemosphere 233 (2019) 17e24 19

Fig. 1. Schematic diagram of the experiment setup (a) and sampling sites (b).

the precipitate of Cd(OH)2 can be completely dissolved when the 

soil pH is sufficiently low. 0; if q  jb=aj
t¼ 2 (4)
Under these assumptions, the partial differential equation (PDE) aq þ bq; if q > jb=aj
that describes the transport of Cd2þ in porous media of soil during
Step k (k ¼ 1, 2, …, n) of the SE-EK was evaluated using the mass where a and b are empirical constants.
conservation law: The effective electroosmosis permeability coefficient was
expressed by the saturated electroosmotic permeability coefficient
keo (m2 V1 s1) (Lo

pez-Vizcaínoa et al., 2017), which can be related
vck 1 Jk to zeta potential z (V) and the dielectric constant ε (F m1) and
¼  þ Rk (1) viscosity of fluid m (N s m2) according to
vtk q vx
HemlholtzeSmoluchowski theory (Eykholt and Daniel, 1994), and
where ck (mol m3) is the concentration of the Cd2þ in the aqueous the tortuosity was introduced to include the porous matrix effects
phase, Jk (mol m2 s1) is the total mass flux of Cd2þ in porous in the original HemlholtzeSmoluchowski formulation (Masi et al.,
media, q (m3 m3) is the volumetric moisture content, and Rk is the 2016):
production rate of the Cd2þ due to chemical reactions. The model  3  3
was disassembled in terms of the SE-EK experiments. q q εz
keeo ¼ keo ¼ nt (5)
The initial conditions of Step n are the result of Step n1. Every n n m
step of the model included the transport of Hþ, OH, and Cd2þ, as
The PDE that describes the transport of mass in porous media of
well as the main reactions in the soil solution.
soil under the constant electric field was evaluated as follows:
The total mass flux under an electric field in porous media of soil
Ji (mol m2 s1) (Alshawabkeh and Acar, 1992) can be expressed as:
 vci vF
vci v2 c
q ¼ D*i 2i þ u*i þ keo þ qRi (6)
vt vx vx vx


Ji ¼  D*i Vci  ci u*i þ keeo VF (2) where Ri is the production rate of the i-th species due to chemical
reactions.
where Di (m2 s1) is the effective diffusion coefficient of the i-th Several main reactions were included in the models among the
species, ci (mol m3) is the concentration of the i-th species in the wide set of possible reactions: (1) the electrolysis of water at the
aqueous phase, F (V) is the electric potential, keeo (m2 V1 s1) is the electrodes; (2) the adsorption/desorption of Hþ; (3) the reaction of
effective electroosmosis permeability coefficient in soil, and u*i (m2 Hþ and OH; (4) the adsorption/desorption of Cd; and (5) the
V1 s1) is the effective ionic mobility of the i-th species, which is precipitation/dissolution of Cd.
estimated by the NernsteEinstein relation (Alshawabkeh and Acar, Without considering other ions in the soil, the buffering ca-
1992) and expressed by the respective diffusion coefficient Di (m2 pacity of soil was determined by the adsorption/desorption of Hþ.
s1) in the free solution (Mattson et al., 2002): We assumed that the process of adsorption/desorption of Hþ on soil
particles reached equilibrium instantaneously. The following rela-
tion was used (Mascia et al., 2007):
zi F zF
u*i ¼ D*i ¼ qtDi i (3)
RT RT r vsi r vsi vic
Rad ¼  ¼ (7)
q vt q vic vt
where R (8.314 J K1 mol1) is the universal gas constant, T (K) is the
absolute temperature, F (96485 C mol1) is Faraday's constant, and where Rad is the production rate of the i-th species due to
t is the tortuosity factor. The diffusion coefficients of Hþ, OH and adsorption/desorption, Si (mol kg1) is the adsorbed concentration,
Cd2þ were taken from the literature (Shackelford and Daniel, 1991) and r (kg m3) is the bulk dry density of the soil. Combining Eqs (6)
and the corresponding effective ionic mobilities were calculated. and (7) gives:
The tortuosity factor was increased with the increase in soil
water content. Chou et al. (2012) evaluated the existing models  
r vsi vci v2 c
 vci vF
used to describe the tortuosity factor. They showed that the tor- q 1þ ¼ D*i 2i þ u*i þ keo  qiR (8)
q vci vt vx vx vx
tuosity factor can be best described by the following equation,
proposed by Mullins and Sommers, 1986:
20 Z. Sun et al. / Chemosphere 233 (2019) 17e24

2.4. Analytical method

r vsH
Rd ¼ 1 þ (9)
q vcH Soil pH was measured using a PHS-3C pH meter (Rex, China) by
slurries with a soil to water ratio of 1:2.5. Electrical conductivity
where Rd is the retardation factor. was measured from a 1:5 soil weight to water volume ratio extract
The adsorption of Cd was described by an algebraic equation using a DDS-11A conductivity meter (Inesa, China). For Cd content
that was developed based on the linear and pH-dependent analysis, soil samples were air-dried and passed through a 100-
adsorption (Kim et al., 2003): mesh screen. The 0.500 g samples, in triplicate, were digested
with HNO3þHClO4þHF in 50-mL Teflon beakers. The Cd content in
8 the extract was determined by a Varian AA240 atomic absorbance
< 0; if rad  0
rad ¼ apH þ b; if 0 < rad < 1 (10) spectrometer (Varian, USA) (Sun et al., 2001).
:
1; if rad  1
2.5. Data analysis
where rad is the adsorption rate, and a and b are empirical
constants. The numerical model was implemented by Matlab 7.0 (Math-
The precipitation reaction of Cd2þ was described by: Works, USA) with a finite difference method. All statistical analyses
were performed in SPSS 21.0 (IBM, USA). Means and standard de-
CdðOHÞ2 4 Cd2þ þ 2OH (11) viations of experimental data were calculated and plotted using
SigmaPlot 10.0 (Systat Software, USA).

CdðOHÞ2
c2OH cCd  K sp ¼ 5:0  1015 (12) 3. Results and discussion

CdðOHÞ
where K sp 2
is the solubility product of the reaction. 3.1. Transport of Cd
The boundary conditions were expressed as:
The migration of Cd under an electric field is hindered by the
  increasing pH, and thus the ‘focusing’ effect of Cd occurs. According
dc 1 I
x ¼ 0; H ¼  JH to the variation of pH and the distribution of Cd during conven-
dt d F
tional EK (Fig. 2a and b), the positions of the secondary electrodes
1014 (13) were determined.
cOH ¼
cH Hþ and OH were produced at the anodes and cathodes,
dcCd 1 respectively. The acid and alkaline fronts moved in opposite di-
¼  JCd rections. After 15 d, the soil in S1 and S2 was acidified (pH < 2.55).
dt d
The Cd decreased to a low level (3.81e6.22 mg kg1) in S1eS2, and
accumulated in S3. Therefore, the secondary anode was positioned
1014 10 cm from the primary anode and the switching time was set to
x ¼ l; cH ¼
cOH 15 d in SE-EK. The acid and alkaline fronts converged at S4 after
  20 d, where the Cd accumulated. It could be inferred that the
dcOH 1 I (14) migration velocity of the acid front was about twice as fast as that of
¼  JOH
dt d F the alkaline front in this experiment. In order to ensure the acid
dcCd 1 front and alkaline front met near the cathode (S5) in Step 2, the
¼ JCd  Rp distance between the two cathodes was set to 5 cm, a quarter of the
dt d
distance between the secondary anode and cathode.
where I (A m2) is the electric current density, d (m) is the diameter The soil pH changes and the Cd distribution after Step 1 and Step
of the electrode, and Rp is the production rate of Cd due to 2 of SE-EK remediation are presented in Fig. 3. After 15 d, the soil
precipitation. pH in the area between the two anodes (S1eS2) decreased to below

Fig. 2. The change in soil pH (a) and Cd distribution (b) during the conventional EK remediation experiment, where C0 is the initial concentration of Cd and C is the residual
concentration of Cd.
 Z. Sun et al. / Chemosphere 233 (2019) 17e24 21

Fig. 3. The pH profile (a) and Cd distribution (b) in soil after Step 1 and Step 2 of the SE-EK remediation, where C0 is the initial concentration of Cd and C is the residual con-
centration of Cd.

2.56, and increased to 9.30 in S5. From 15 to 35 d, more soil (S3eS4) electroosmosis permeability constant to be equal to the initial
was acidified and the high soil pH region was narrowed. The pH electroosmosis permeability, like many previous studies
jumping front occurred in S5. These results showed that SE-EK is an (Asadollahfardi et al., 2016; Rezaeea et al., 2019). The migration of
effective method for controlling soil pH during EK remediation. Hþ, OH, and Cd2þ in the soil under the superimposed electric field
After 35 d of remediation, 87.60% of Cd was removed in S1eS4, was modeled with these parameters.
which was 6.13 times that in the conventional EK (14.3%). The
distribution of the residual Cd was correlated with the variation in 3.3. Numerical simulation results
soil pH. The Cd mainly accumulated in S5, where the acid and
alkaline fronts converged. The concentration of Cd reached The soil pH and Cd concentration were simulated and compared
432.33 mg kg1 in the ‘focusing’ region, 4.40 times the initial con- with the experimental data (Fig. 4). After 15 d, a pH jumping front
centration. The compression of the ‘focusing’ region can also be occurred around 10 cm from the anode. The migration of Cd was
achieved by the AAs-EK method (Shen et al., 2007; Wei et al., 2016; accumulated at the region of pH jumping. From 15 to 35 d, the
Zhang et al., 2016). The present results indicate that the SE-EK accumulated-Cd region moved to the cathode and stabilized at
method is an alternative and efficient technique for compressing around 20e25 cm from the anode where the acid and alkaline
the ‘focusing’ region. fronts converged.
The processes of Cd adsorption/desorption and precipitation/
3.2. Model parameters dissolution are affected by the soil pH. Previous research has shown
that desorption of Cd can increase linearly with a decrease in pH
The parameters adopted in the numerical model are summa- below 4.8 (Yuan et al., 2007). Low pH leads to the dissolution of Cd
rized in Table 1. Porosity and volumetric moisture content were into the pore solution as Cd2þ migrates towards the cathode.
experimentally determined. The parameters a and b in Eq. (4) were Meanwhile, the high redox potential in low pH conditions can
derived from the literature (Chou et al., 2012). The retardation cause the set off of Cd from soil particules (Shen et al., 2007). Cd
factor was adjusted to fit the soil pH data during conventional EK. could form insoluble Cd(OH)2 in near-neutral to alkaline condi-
The parameters a and b in Eq. (10) were adjusted to fit the exper- tions, hindering the migration of Cd. Therefore, as the EK remedi-
imental sorption data, the results of which are reported in Fig. S1. ation progresses, Cd will accumulate at the pH jumping fronts
The saturated electroosmosis permeability is mainly dependent on through the processes of absorption and precipitation.
the zeta potential of the soilepore fluid interface, which is pH Good agreement was achieved between the simulated and
dependent (Masi et al., 2016). In this study, we assumed the experimental data for both pH and Cd (R2 > 0.9). These findings
indicate that the model-simulated data explained well the change
in pH and the distribution of Cd during the SE-EK experiment.
Therefore, the removal efficiency of SE-EK remediation can be
Table 1
Model parameters.
optimized by modelling.

Parameters Description Unit Value

3.4. Optimization of removal efficiency
n Porosity e 0.46
q Volumetric moisture m3 m3 0.408
content
To optimize the removal efficiency, five situations (M1eM5)
a Parameter Eq. (4) e 0.2038 with different switching times (from 10 to 20 d) were simulated.
b Parameter Eq. (4) e 0.0244 The simulated pH changes and Cd distributions are shown in Fig. 5.
t Tortuosity factor e 0.024 In all five situations, the Cd decreased to a low level in the acidic
Rd Retardation factor e 11.5
region and accumulation occurred at the end of the acid front after
a Parameter Eq. (10) e 0.2265
b Parameter Eq. (10) e 0.4503 Step 1. Thus, the distance between the two anodes (d1) was set to
keo Saturated electroosmosis m2 1.18  1010 6.5, 8, 10, 11.5, and 13 cm in M1eM5, respectively. The distance
permeability V1 S1 between the two cathodes (d2) was set to a quarter of the distance
keeo Effective electroosmosis m2 8.23  1011 between the secondary anode and cathode, because the migrated
permeability V1 S1
velocity of Hþ was almost twice that of OH.
22 Z. Sun et al. / Chemosphere 233 (2019) 17e24

Fig. 4. Simulated and experimental profiles of pH and Cd after 15 d and 35 d during the SE-EK experiments.

In M1-M5, the area of the low pH region increased markedly the first 15 d, which was consistent with the trend of the electric
after 35 d. The Cd ‘focusing’ region in M3 was closer to the cathode current (Fig. 6b). After 15 d, the electric current stabilized at about
than that in the other four situations. This indicates that the 5 mA in conventional EK. In SE-EK, the current suddenly increased
removal efficiency of Cd in M3 was higher. Therefore, the switching from 5.5 to 21.7 mA when the electrodes were switched. The
time and electrode distance in M3 (switching time ¼ 15 d and increased electric current may have been because the utilization
electrode distances of d1 ¼10 cm and d2 ¼ 5 cm) were optimal efficiency of ions improved. The switching of electrodes also led to a
among the five situations. higher conductivity at S1eS3 in SE-EK than that in the conventional
The switching time influenced the location of the pH jumping EK.
front, which affected the location of the ‘focusing’ effect. When the The energy consumption per unit volume (Ev, KWh m3) in EK
acid and alkaline fronts formed by the primary cathode converged, remediation was calculated by the voltage between the electrodes,
the alkaline front was consumed by the acid front. The acid front the electric current, and the remediation time.
and the alkaline front produced by the secondary cathode Fig. 6c shows the energy consumption in the two experiments.
converged, which stabilized the pH. If the switching time was too The accumulated energy consumption increased monotonically
short, the area of high pH would have increased and the location during the two experiments. Although there was no difference in
where the acid and base met would have become more distant energy consumption between the conventional EK (60.0 KWh m3)
from the cathode, such as in M1 and M2. If the switching time was and SE-EK (57.4 KWh m3), the parameter value of energy utiliza-
too long, more Hþ would have been needed to be produced by tion efficiency, a relative ratio of the removal rate of heavy metals
water electrolysis near the secondary anode to neutralize the OH and the energy consumption, proposed by Fu et al.(2017), reached
produced by the primary cathode. In M4 and M5, 35 d was insuf- 1.53 in SE-EK, which was 6.38 times that in conventional EK (0.24).
ficient to stabilize the pH. Therefore, the locations of the Cd Compared with other enhanced EK methods for Cd-contaminated
‘focusing’ region were further away from the cathode in M1, M2, soil, such as using ion exchange membranes (0.14) (Kim et al.,
M4, and M5 than in M3. 2005), a pulsed power supply (0.15) (Ryu et al., 2010), catholyte
circulation (0.84) (Gao et al., 2013), or AAS-EK (1.14) (Zhang et al.,
3.5. Electric current and energy consumption 2016), the energy utilization efficiency in SE-EK was much higher.

Conductivity is an index of water-soluble ion content in soil. As 4. Conclusion

the time elapsed, the conductivity dropped because of the loss of
mobile ions, as shown in Fig. 6a. This decrease mainly occurred in A new method (the SE-EK method) to increase the removal of Cd

Fig. 5. Simulated profiles of pH (a) and Cd concentration (b) after Step 1 (dashed line) and Step 2 (solid line) of SE-EK remediation. The switching time was 10, 12.5, 15, 17.5, and 20 d
in M1eM5, respectively.
 Z. Sun et al. / Chemosphere 233 (2019) 17e24 23

Appendix A. Supplementary data

Supplementary data to this article can be found online at

https://doi.org/10.1016/j.chemosphere.2019.05.233.

References

Alshawabkeh, A.N., Acar, Y.B., 1992. Removal of contaminants from soils by elec-
trokinetics: a theoretical treatise. J. Environ. Sci. Heal. Part A: Environ. Sci.Eng.
Toxicol. 27, 1835e1861.
Acar, Y.B., Alshawabkeh, A.N., 1993. Principles of electrokinetic remediation. Envi-
ron. Sci. Technol. 27, 2638e2647.
Asadollahfardi, G., Rezaee, M., Mehrjardi, G.T., 2016. Simulation of unenhanced
electrokinetic process for lead removal from kaolinite clay. Int. J. Civ. Eng. 14,
263e270.
Chou, H., Wu, L., Zeng, L., Chang, A., 2012. Evaluation of solute diffusion tortuosity
factor models for variously saturated soils. Water Resour. Res. 48, 10539.
Eykholt, G.R., Daniel, D.E., 1994. Impact of system chemistry on electroosmosis in
contaminated soil. J. Geotech. Engrg. 120, 797e815.
Fu, R., Wen, D., Xia, X., Zhang, W., Gu, Y., 2017. Electrokinetic remediation of chro-
mium (Cr)-contaminated soil with citric acid (CA) and polyaspartic acid (PASP)
as electrolytes. Chem. Eng. J. 316, 601e608.
Gao, J., Luo, Q., Zhang, C., Li, B., Meng, L., 2013. Enhanced electrokinetic removal of
cadmium from sludge using a coupled catholyte circulation system with
multilayer of anion exchange resin. Chem. Eng. J. 234, 1e8.
Giannis, A., Gidarakos, E., 2005. Washing enhanced electrokinetic remediation for
removal cadmium from real contaminated soil. J. Hazard Mater. B123, 165e175.
Giannis, A., Nikolaou, A., Pentari, D., Gidarakos, E., 2009. Chelating agent-assisted
electrokinetic removal of cadmium, lead and copper from contaminated soils.
Environ. Pollut. 157, 3379e3386.
Kim, S.O., Kim, J.J., Yun, S.T., Kim, K.W., 2003. Numerical and experimental studies
on cadmium (II) transport in kaolinite clay under electrical fields. Water Air Soil
Pollut. 150, 135e162.
Kim, W.S., Kim, S.O., Kim, K.W., 2005. Enhanced electrokinetic extraction of heavy
metals from soils assisted by ion exchange membranes. J. Hazard Mater. 118,
93e102.
Kim, K.J., Kim, D.H., Yoo, J.C., Baek, K., 2011. Electrokinetic extraction of heavy metals
from dredged marine sediment. Separ. Purif. Technol. 79, 164e169.
Li, G., Guo, S., Li, S., Zhang, L., Wang, S., 2012a. Comparison of approaching and fixed
anodes for avoiding the ‘focusing’ effect during electrokinetic remediation of
chromium-contaminated soil. Chem. Eng. J. 203, 231e238.
Li, S., Li, T., Li, G., Li, F., Guo, S., 2012b. Enhanced electrokinetic remediation of
chromium-contaminated soil using approaching anodes. Front. Environ. Sci.
Eng. 6, 869e874.
Fig. 6. The variation in conductivity (a), electrical current (b), and accumulated energy

pez-Vizcaínoa, R., Yustresb, A., Leo
Lo
nb, M.J., Saezc, C., Can
~ izaresc, P., Rodrigoc, M.A.,
consumption (c) in the conventional EK and SE-EK experiments.
Navarrob, V., 2017. Multiphysics implementation of electrokinetic remediation
models for natural soils and porewaters. Electrochim. Acta 225, 93e104.
Mullins, G.L., Sommers, L.E., 1986. Characterization of cadmiumand zinc in four soils
treated with sewage-sludge. J. Environ. Qual. 15, 382e387.
by more effectively moving the ‘focusing’ region of Cd to the Mattson, E.D., Bowman, R.S., Lindgren, E.R., 2002. Electrokinetic ion transport
cathode has been presented in this paper. After 35 d of SE-EK through unsaturated soil: 1. Theory, model development, and testing. J. Contam.
Hydrol. 54, 99e120.
remediation, the level of Cd accumulated near the cathode (S5),
Mascia, M., Palmas, S., Polcaro, A.M., Vacca, A., Muntoni, A., 2007. Experimental
where the acid and alkaline fronts converged, was 4.40 times the study and mathematical model on remediation of cd spiked kaolinite by elec-
initial concentration. The removal rate of Cd in soil sections from S1 trokinetics. Electrochim. Acta 52, 3360e3365.
to S4 reached 87.6%, which was 6.13 times that under conventional Masi, M., Ceccarini, A., Iannelli, R., 2016. Multispecies reactive transport modelling
of electrokinetic remediation of harbour sediments. J. Hazard Mater. 326, 187.
EK remediation (14.3%), because of the expansion of the acidified Probstein, R.F., Hicks, R.E., 1993. Removal of contaminants from soils by electric
soil area. Meanwhile, the energy utilization efficiency in SE-EK was fields. Science 260, 498e503.
6.38 times that in conventional EK. A model was established to Rezaeea, M., Asadollahfardia, G., Gomez-Lahozb, C., Villen-Guzmanb, M., Paz-
Garciab, J.M., 2019. Modeling of electrokinetic remediation of Cd- and Pb-
describe the transport of Cd in soil under the superimposed electric contaminated kaolinite. J. Hazard Mater. 366, 630e635.
field. Good agreement was achieved between the modeled and Ryu, B.G., Yang, J.S., Kim, D.H., Baek, K., 2010. Pulsed electrokinetic removal of Cd
experimental data for both pH (R2 ¼ 0.999) and Cd (R2 ¼ 0.996). The and Zn from fine-grained soil. J. Appl. Electrochem. 40, 1039e1047.
Shackelford, C.D., Daniel, D.E., 1991. Diffusion in saturated soil. I: background.
removal efficiency of SE-EK remediation was optimized by the J. Geotech. Engrg. 117, 467e484.
simulation of the Cd distribution in five situations with switching Shen, Z., Chen, X., Jia, J., Qu, L., Wang, W., 2007. Comparison of electrokinetic soil
times from 10 to 20 d. This research is helpful for the effective remediation methods using one fixed anode and approaching anodes. Environ.
Pollut. 150, 193e199.
decontamination of heavy metals from soil using the EK technique. Sun, Y.C., Chi, P.H., Shiue, M.Y., 2001. Comparison of different digestion methods for
total decomposition of siliceous and organic environmental samples. Anal. Sci.
17, 1395.
Villen-Guzman, M., Paz-Garcia, J.M., Rodriguez-Maroto, J.M., Gomez-Lahoz, C.,
Garcia-Herruzo, F., 2014. Acid enhanced electrokinetic remediation of a
contaminated soil using constant current density: strong vs. weak acid. Separ.
Purif. Technol. 49, 1461e1468.
Acknowledgments
Virkutyte, J., Sillanpaa, M., Latostenmaa, P., 2002. Electrokinetic soil remediation-
critical overview. Sci. Total Environ. 289, 97e121.
This work was supported by the Key Deployment Projects of the Wei, X., Guo, S., Wu, B., Li, F., Li, G., 2016. Effects of reducing agent and approaching
Chinese Academy of Sciences (grant no. KFZD-SW-311) and the anodes on chromium removal in electrokinetic soil remediation. Front. Environ.
Sci. Eng. 10, 1e9.
National Natural Science Foundation of China (grant nos. 21707150, Yuan, L., Xu, X., Li, H., Wang, N., Guo, N., Yu, H., 2016. Development of novel assisting
21507144 and 21677150). agents for the electrokinetic remediation of heavy metal-contaminated kaolin.
24 Z. Sun et al. / Chemosphere 233 (2019) 17e24

Electrochim. Acta 218, 140e148. decontamination of soils. J. Contam. Hydrol. 26, 291e299.
Yuan, S., Xi, Z., Jiang, Y., Wan, J., Wu, C., Zheng, Z., Lu, X., 2007. Desorption of copper Zhang, P., Jin, C., Sun, Z., Huang, G., She, Z., 2016. Assessment of acid enhancement
and cadmium from soils enhanced by organic acids. Chemosphere 68, schemes for electrokinetic remediation of cd/pb contaminated soil. Water Air
1289e1297. Soil Pollut. 227, 1e12.
Yu, W., Neretnieks, I., 1997. Theoretical evaluation of a technique for electrokinetic