Engineering Geology 85 (2006) 132 – 146

www.elsevier.com/locate/enggeo

Enhanced electrokinetic remediation of contaminated

manufactured gas plant soil
Krishna R. Reddy a,⁎, Prasanth R. Ala a , Saurabh Sharma a , Surendra N. Kumar b
a
University of Illinois at Chicago, Department of Civil and Materials Engineering, 2095 Engineering Research Facility, 842 West Taylor Street,
Chicago, Illinois 60607, USA
b
STAT Analysis Corporation, 2201 West Campbell Park Drive, Chicago, Illinois 60612, USA
Accepted 15 September 2005
Available online 17 April 2006

Abstract

This paper evaluates different flushing agents to enhance the efficiency of electrokinetic remediation of a manufactured gas
plant (MGP) soil contaminated with polycyclic aromatic hydrocarbons (PAHs) and heavy metals. Because of high concentrations,
PAHs were of environmental concern and required to be removed to acceptable levels. Four flushing agents, which included two
surfactants (3% Tween 80, and 5% Igepal CA-720), one cosolvent (20% n-Butylamine) and one cyclodextrin (10% hydroxypropyl-
β-cyclodextrin or HPCD), were examined to enhance the solubilization of PAHs in the soil. Four electrokinetic experiments were
conducted at 2.0 VDC/cm voltage gradient and 1.4 hydraulic gradient in order to assess the effectiveness of these flushing solutions
for the removal of PAHs. Variables measured during the application of electric potential were electric current, electroosmotic flow,
and contaminant removal from the soil. After the completion of each test, the soil was further examined for moisture content, pH,
redox potential, electrical conductivity, and residual contaminant distribution. It is found that cosolvent increased the soil pH, while
the surfactants and HPCD did not induce substantial change in the soil pH. The current densities fluctuated with time for all tests
and remained less than 1 mA/cm2. The current density for the test conducted with cosolvent was higher as compared to the tests
conducted with surfactants and HPCD. Electroosmotic flow was the maximum with the cosolvent, while the lowest flow was
observed with Tween 80 surfactant. Overall, Igepal CA-720 surfactant yielded the highest removal efficiency due to partial
solubilization of PAHs, causing some PAHs to migrate towards the cathode. Heavy metals are found to be strongly adsorbed/
precipitated and showed negligible migration behavior in all the tests. Based on the contaminant mass remaining in the soil, it is
apparent that further optimization of the electrokinetic system is required to improve PAH removal efficiency for the MGP soil.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Electrokinetics; Remediation; Soils; Polycyclic aromatic hydrocarbons; Heavy metals; Surfactants; Cosolvents; Cyclodextrins

1. Introduction 2000). The contaminants found at these sites include

There are over 3000 to 5000 former manufactured
gas plant (MGP) sites across the United States (USEPA,
⁎ Corresponding author. Tel.: +1 312 996 4755; fax: +1 312 996
2426.
E-mail address: kreddy@uic.edu (K.R. Reddy).
0013-7952/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.enggeo.2005.09.043
mainly polycyclic aromatic hydrocarbons (PAHs), but
small amounts of heavy metals are also often encoun-
tered. PAHs are compounds composed of two or more
fused aromatic rings, and PAHs with higher molecular
weights are proven carcinogenic and mutagenic.
Because of their low volatility, low solubility, and low
biodegradability, PAHs are difficult to treat (Williamson
 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146 133

Table 1 removed by EO flow. However, it is difficult to apply the

Properties of the contaminated manufactured gas plant soil electrokinetic remediation method to remove hydropho-
Property Test method Value bic and strongly adsorbed contaminants especially from
Specific gravity ASTM D854 2.54 the low permeability clayey soils. The use of solubilizing
Grain size distribution ASTM D422 % gravel = 1.8–15.4 agents, such as surfactants, cosolvent and cyclodextrins
% sand = 50.1–65.6 is considered to enhance the efficiency of removing these
% fines = 32.6–34.5
hydrophobic pollutants from the soils (Maturi, 2004).
Atterberg limits ASTM D4318 Non-plastic
Hydraulic conductivity ASTM D2434 2.1 × 10− 4 cm/s The purpose of the present study is to develop an
pH ASTM D4972 6.9 effective electrokinetic remediation system for the
Organic content ASTM D2974 2.69–3.75% removal of hydrophobic PAHs from the field soil
USCS classification ASTM D2488 SM obtained from actual MGP site. In particular, several
flushing solutions, specifically two different surfactants
(5% Igepal CA-720 and 3% Tween 80), a cosolvent
et al., 1998; Hatheway, 2002). Conventional ex situ (20% n-Butylamine), and a cyclodextrin (10% HP-β-
remediation methods, such as excavation, incineration, CD), were examined for their potential use in the
thermal desorption, soil washing, and bioremediation, removal of hydrophobic PAHs from the field soil. A
are found to be either expensive and/or ineffective at series of bench-scale electrokinetic experiments were
field scale application (Shosky, 1996). Therefore, in situ conducted using these different flushing solutions to
remediation of soils is preferred due to simplicity, less assess the extent of contaminant migration and removal.
site disturbance, and minimal public exposure. As a
result, a variety of in situ technologies have been 2. Experimental methodology
designed and developed, but they are found to be less
effective and costly for treatment of low permeability 2.1. Soil characterization
and heterogeneous soils (McGowan et al., 1996;
Chowdiah et al., 1998; Lee et al., 2001). Contaminated soil sample, selected for this study, was
Recently, attention has focused on developing in situ obtained from a former manufactured gas plant (MGP)
electrokinetic technique for the treatment of low site in Chicago, Illinois, USA. The received soil sample
permeable soils contaminated with heavy metals, radio- was thoroughly homogenized. The homogenized sample
nuclides, and selected organic pollutants. This technique was analyzed for different physical properties according
involves applying a low-level DC electric potential to the respective ASTM standard testing procedures and
through electrodes, which are placed into the contami- the results are presented in Table 1. Fig. 1 shows the grain
nated soil. If the contaminants are ionic compounds, they size distribution of the field soil. The homogenized soil
can be transported to the oppositely charged electrode by sample was also analyzed by standard EPA method SW
electromigration. In addition, electroosmotic flow (EO 6020 for metals (SW 7471A for mercury), EPA method
flow) provides a driving force for the movement of SW 8260B for volatile organic compounds (e.g., BTEX),
contaminants. Therefore, soluble contaminants may be and EPA method SW 8270C (Selective Ionic Mode) for

Fig. 1. Grain size distribution of contaminated manufactured gas plant soil.

134 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146

Table 2 was determined by titration analysis using 2 M nitric

Contaminants found in the manufactured gas plant soil acid as titrant solution. A soil slurry sample was
(a) Total metals (USEPA (b) Polycyclic aromatic hydrocarbons prepared by mixing 20 g of soil in 200 mL of water.
method SW6020/SW7471A) (USEPA method SW8270C(SIM)) The acid was added incrementally to the slurry while it
Chemical Concentration Chemical Concentration was being mixed with a magnetic stirrer. A deionized
(mg/kg) (mg/kg) water sample was used as a control sample. The
Aluminum 3800 2-Methylnaphthalene 230 equilibrium pH of the slurry was measured with a pH
Arsenic 11 Acenaphthene 25–40 meter (Thermo Orion model 720 A). The results showed
Barium 38 Acenaphthylene 84–120
that the buffering capacity of the aqueous MGP soil
Calcium 38000 Anthracene 69–92
Chromium 8.3 Benz(a)anthracene 66–82 slurry with a solids concentration of 8.5% is 3.7 eq/kg
Cobalt 4.9 Benzo(a)pyrene 59–62 (dry soil) at the inflection point of the titration curve (pH
Copper 13 Benzo(b) 31–33 6.2) (see Fig. 2). This indicates that the MGP soil
fluoranthene possesses high acid buffering capacity.
Iron 15000 Benzo(g,h,i) 4.8–33
perylene
Lead 25 Benzo(k) 23–30 2.2. Electrokinetic test setup
fluoranthene
Magnesium 15000 Chrysene 39–75 Fig. 3 shows the schematic of the electrokinetic test
Manganese 440 Dibenz(a,h) 9.1 setup used for this study and has been described in
anthracene
detail by Reddy and Parupudi (1997) and Reddy and
Nickel 14 Dibenzofuran 7.7
Potassium 700 Fluoranthene 92–130 Chinthamreddy (2003). The test setup mainly consist
Sodium 88 Fluorene 92 of an electrokinetic cell, two electrode compartments,
Thallium 1.8 Indeno(1,2,3-cd) 12–21 two electrode reservoirs, a power source, and a
pyrene multimeter. Plexiglas cell having inside diameter of
Vanadium 12 Naphthalene 600
6.3 cm and a total length of 19.1 cm was used as
Zinc 66 Phenanthrene 260–350
Pyrene 130–210 electrokinetic cell. Each electrode compartment includ-
ed a valve to control the flow into the cell, a slotted
Chemicals for which measured concentrations were below detection
limits are not listed.
graphite electrode, and a porous stone. Small holes in
the electrode compartment contained the electrode
pins, and filter paper was placed between the soil
PAHs (USEPA, 1986), and their respective concentrations sample and the electrode. The electrode reservoirs were
found in the MGP soil are presented in Table 2. made of 3.2 cm inner diameter. Plexiglas reservoirs
The presence of calcium carbonates or other were connected to the electrode compartments using
compounds such as magnesium carbonates or sodium Tygon tubing. Exit ports were created in the electrode
carbonates causes high buffering capacity of the soil. compartments, and the tubing was attached to these
Buffering capacity of soil refers to the capability of soil ports to allow the gases generated due to the
to neutralize acid. Buffering capacity of the MGP soil electrolysis of water to escape. The other end of

Fig. 2. Acid buffering capacity of contaminated manufactured gas plant soil.

 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146 135

Fig. 3. Electrokinetic test setup.

these gas tubes was connected to the reservoirs to (20% n-Butylamine), and a cyclodextrin (10% HP-β-
collect any liquid that was removed along with the CD) and these particular types of flushing solutions and
gases. A power source was used to apply a constant their concentrations were selected on the basis of results
voltage to the electrodes, and a multimeter was used to from several series of previous batch and electrokinetic
monitor the voltage and measure the current value experiments (Saichek and Reddy, 2004; Maturi, 2004).
through the soil sample during the test.
2.4. Testing procedure
2.3. Test variables
The contaminated field soil was placed in the
Table 3 shows the details of the four experiments electrokinetic cell in layers and compacted uniformly
conducted for this study. All of the experiments were
conducted at a constant voltage gradient of 2.0 VDC/ Table 3
cm. The hydraulic gradient that existed under these Electrokinetic testing program
experimental conditions was approximately 1.4, and is Test Voltage gradient Hydraulic Flushing
not significant enough to generate substantial hydraulic number (VDC/cm) gradient solution
flow because of the characteristic low permeability of 1 2.0 1.4 5% Igepal CA–720
the soil. Flushing solutions examined to enhance 2 2.0 1.4 3% Tween 80
solubilization of PAHs were: two different surfactants 3 2.0 1.4 20% n-Butylamine
4 2.0 1.4 10% HP-β-CD
(5% Igepal CA-720 and 3% Tween 80), a cosolvent
136 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146

Fig. 4. Measured current densities.

using a hand compactor. The electrode compartments
were then connected to the electrokinetic cell. In each
electrode compartment, filter papers were inserted
between the electrode and the porous stone as well as
between the porous stone and the soil. The electrode
compartments were connected to the anode and cathode
reservoirs using Tygon tubing. The anode reservoir was
filled with a selected flushing solution and the cathode
reservoir was filled with deionized water. The water level
in both reservoirs was monitored and adjusted carefully
throughout the tests in order to maintain a constant
hydraulic gradient across the specimen. The electroki-
netic cell was then connected to the power supply and a
constant voltage gradient of 2.0 VDC/cm was applied to
the soil sample. The flushing solution was circulated
using peristaltic pump in the anode reservoir and the
electroosmotic flow from the cathode reservoir was
collected periodically. Each test was terminated when the
current value, flow rate, or contaminant concentrations
in effluent was significantly reduced.
After the completion of each test, aqueous solutions
from the anode and cathode reservoirs and the
electrode assemblies were collected and the volumes
were measured. Then, the reservoirs and the electrode
assemblies were disconnected, and the soil specimen
was extruded from the cell using a mechanical
extruder. Each of the extruded soil specimen was
sectioned into three or five equal parts to determine
the final distribution of pH values across the soil
specimen. Each soil section was weighed and
preserved in a glass bottle. From each soil section,
10 g of soil was taken and mixed with 10 mL of a
0.01 M CaCl2 solution in a glass vial. The slurry was
shaken thoroughly by hand for several minutes and
the solids were allowed to settle for an hour. This
slurry was then used for measuring the soil pH, redox
potential and electrical conductivity. The pH, redox
potential and electrical conductivity of the aqueous
solutions from the electrodes were also measured. The
moisture content of each soil section was also

Fig. 5. Measured electroosmotic flow.

 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146 137

determined in accordance with ASTM D2216 (ASTM,
2004).
2.5. Chemical analyses
Representative samples of reservoir solutions, soil
sections, and the initial soil for each test were analyzed
for total metals and PAHs using the standard USEPA
methods (USEPA, 1986). The total metals in soil and
liquid samples were analyzed using the USEPA Method
SW6020 and the mercury was analyzed using
SW7471A for soil and 7470A for liquid samples. The
PAHs were analyzed using the USEPA Method
SW8270C (Selective Ionic Mode). The chemical
analyses were conducted with a stringent quality
control by the STAT Analysis Corporation, Chicago,

Fig. 6. (a) Removal of total metals from the soil. (b) Removal of toxic metals from the soil. (c). Removal of polycyclic aromatic hydrocarbon from the
soil.
138 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146

Fig. 7. (a) Moisture distribution in the soil. (b). Soil pH variation in the soil. (c). Redox potential variation in the soil. (d) Conductivity variation in the
soil.
 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146
Fig. 7 (continued).
Illinois, a certified laboratory. To ensure accuracy of the
test results, new electrodes, porous stones, and tubing
were used for each experiment, and the electrokinetic
cell and compartments were washed thoroughly and
then rinsed first with tap water and finally with
deionized water to avoid cross contamination between
the experiments.
3. Results and analysis
The results of the electrokinetic experiments were
analyzed to assess the electric current, electroosmotic
flow, and contaminant removal during the electric
potential application as well as the moisture content,
pH, redox potential, electrical conductivity, and residual
contaminant distribution in the soil after the experiments
were terminated.
3.1. Electric current density
The measured electric current densities for all the tests
are plotted against elapsed time in Fig. 4. The current
densities for each test were obtained by dividing current
values measured during the testing by the cross-sectional
area of the EK cell. The results showed that the current
density values fluctuated with time for all tests
conducted with surfactants, cosolvent and HPCD and
remain less than 1 mA/cm2. However, the current density
for the test conducted with 20% n-Butylamine was
higher as compared to the other tests using surfactants
and HPCD as flushing solutions. This behavior can be
explained by considering that when the flushing fluids
pass through the soil, the solubilization of the contam-
139
inant occurs and the ionic strength of pore fluid is
increased. Thus, when the voltage gradient is applied,
initially the current is low because it takes time for the
solution to migrate into the soil from the electrode
reservoirs and for the soil constituents/minerals and/or
contaminants to dissolve from the soil surface. After
some time (few hours), the initial current reaches its peak
value due to the strong ionic concentration of the pore
fluid and also due to the electromigration of contami-
nants towards their respective electrode. Then, current
value gradually decreases because of decrease in the
electromigration of the cations and anions in the pore
fluid. In addition, the products of the electrolysis
reactions or other chemical species may reduce the
current by neutralizing the migrating ions. For instance,
H + ions migrating towards the cathode could be
neutralized by OH− ions migrating towards the anode,
thereby forming water and diluting the number of ions in
solution. Change in soil pH due to electrolysis reactions
could also affect the current by causing changes such as
mineral dissolution, or chemical precipitation/dissolu-
tion. Unless flushing solutions, which introduce addi-
tional, non-reactive, ions as charge carriers, are used, the
current usually diminishes over time (Dzenitis, 1997).
Thus difference in the current data among the four tests
can be explained on the basis of their affinity towards the
hydrophobic contaminants. 20% n-Butylamine was
found to be more effective in solubilizing the con-
taminants from the field soil as compared to surfactants
and cyclodextrin (as discussed in Section 3.3). The trend
of current values of 5% Igepal 720 and 3% Tween 80
enhanced system was quite similar but slightly higher
current values were recorded with 5% Igepal enhanced
140 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146
system. This is due to the formation of more stable
micelles due to the higher concentration of the Igepal.
The lowest current values were recorded for HPCD
enhanced system. The higher initial values of electric
current are obviously due to the higher electrolyte
concentration in the pore water but soon it decreased and
showed a fluctuation in the later stages of operation. This
is due to the constant change in polarization of soil
particles due to the change in the double layer by the
action of HPCD solution.
3.2. Electroosmotic flow
Fig. 5 represents the electroosmotic flow data for all
tests. In all the tests, the electroosmotic flow at the
cathode increased with an increase in the operating
duration, i.e. elapsed time. It can be seen that electro-
osmotic flow behavior was dependent on the type of
flushing solutions. Maximum electroosmotic low was
observed within the cosolvent system, while the lowest
flow was observed with Tween 80 surfactant system. A
total of 7.2, 2.1, 10.7 and 7.5 pore volumes of flow were
measured in tests with 5% Igepal CA-720, 3% Tween 80,
20% n-Butylamine, and HPCD tests, respectively. The
electroosmotic flow variation is found to be consistent
with their respective trend as observed for the variation
in current densities in all the tested systems.
As seen in Fig. 4, the electric current varies
significantly with elapsed time, and was attributed to
the physico-chemical processes, such as the electromi-
gration of ionic species and the electrolysis reactions.
These processes affect the surface charge of the soil
particles (zeta potential) and the pore fluid properties,
such as dielectric constant and viscosity, with time, and
hence influence the electroosmotic flow. Initially, during
the beginning of the test, when the current is high
(electromigration is high), the transfer of momentum to
the surrounding fluid molecules may be substantial. This
often corresponds to a significant volume of electroos-
motic flow. A high ionic strength of the pore fluid can
also be detrimental for electroosmotic flow, because it
reduces the thickness of the diffuse double layer and,
thereby, constricts the electroosmotic flow. The charge
on the soil surface must also be considered, because
when the pH is below its ZPC, the soil particle surfaces
possess a positive zeta potential and the electroosmotic
flow occurs towards the anode, and when the pH is above
the ZPC, the soil particles have a negative zeta potential
and the electroosmotic flow occurs towards the cathode.
After few days, it was observed that the electroosmotic
flow sharply decreased with time. The reason for this is
that electroosmotic flow was inhibited by a decrease of
zeta potential of soil particles by excess H+ and heavy
metal precipitation by excess OH−.
3.3. Contaminant removal
Effluent samples collected at different time intervals
for all the tested systems were analyzed for metals and
PAHs. Fig. 6(a) shows the cumulative metal removal for
all of the metals shown in Table 2. These plots revealed
that 20% n-Butylamine has pronounced affinity as
compared to other flushing solutions for the cumulative
removal of metals from the soil under the constant
voltage gradient. Since toxic metals (all metals except
Al, Ca, Fe, Mg, K and Na in Table 2) are of prime
concern, the cumulative toxic metal removal is depicted
in Fig. 6(b). The removal of total PAHs with number of
pore volumes is shown in Fig. 6(c). These results show
that 20% n-Butylamine cosolvent has a maximum
affinity towards the removal of metals and PAHs as
compared to surfactants and HPCD. The affinity of all
the tested systems decreases in the following order 20%
n-Butylamine N 5% Igepal CA-720 N 10% HPCD N 3%
Tween 80 system. It is pertinent to mention here that a
very low concentration of the surfactants and HPCD
were employed during these investigations as compared
to n-Butylamine system. Therefore, the performance of
these systems may be increased using higher concentra-
tions of flushing solutions.
3.4. Moisture content, pH, redox potential and
electrical conductivity
The initial moisture content of the MGP soil was
15%, and the variation of moisture content with
normalized distance from the anode after the electroki-
netic treatment is shown in Fig. 7(a). The normalized
distance is defined as the distance to the specific
location from the anode divided by the total distance
from the anode to the cathode. In general, moisture
content of the soil near the anode increased slightly,
while the moisture content near the cathode decreased
slightly. This behavior can be seen in HPCD system,
where the moisture content clearly increases at anode
and decreases at cathode. In 20% n-Butylamine system
moisture content was high in the first three soil
specimens but then sharply decreases and again
increases at cathode end. This may be due to the
enhanced electroosmotic flow behavior of this system.
However interesting results were obtained for the
surfactant enhanced systems. It is pertinent to mention
here again that the concentration of the surfactants
employed in these investigations were low. It has been
 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146 141

observed that in 5% Igepal-system, the moisture content
of all the soil specimens was found to be higher than the
other tested systems. The moisture content decreases
from anode to cathode. But in 3% Tween 80-system the
moisture content at anode was found to be low at anode
end and was found to be just 15% (initial soil condition)
in the other soil specimens. Overall it is concluded that
the electrokinetic process in all the tested systems does
not significantly alter the moisture content. Slight
changes in moisture contents are evident which can be
attributed to the variations in the electroosmotic flow
that occurred as a result of the changes in parameters
such as the ionic strength, conductivity, and/or electrical
gradient. These results suggest that the electroosmotic
flow might not be uniform and there might be changes in
pore pressures (Eykholt, 1997). Nevertheless, it appears
that the soil moisture content remained fairly consistent
and comparable to the initial moisture content. It is
possible that regions where the electroosmotic flow was
high, a pressure gradient was created so that the solution
was pulled from regions where the electroosmotic flow
was lower. Since the solution was continuously trans-
ported through the soil, the moisture content did not
substantially deviate from the initial moisture content.
Fig. 7(b) shows the pH distribution across the soil
after the completion of the experiments. Considering
that the MGP soil had a pH of 6.9 before the experi-
ments, the soil pH after the experiment was analyzed for
each soil specimen of all the tests. Generally, the elec-
trolysis of water results in the formation of H+ ions (low
pH solution) at the anode and OH− ions (high pH
solution) at the cathode, and, primarily due to

Fig. 8. (a) Distribution of metals in soils using 5% Igepal CA-720 surfactant. (b). Distribution of PAHs in soils using 5% Igepal CA-720 surfactant.
142 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146

electromigration, these ions tend to migrate towards the
oppositely charged electrode(s). Because of high acid
buffering capacity of the MGP soil, the H+ ions are
neutralized and are not migrated through the soil.
However, OH− migrate through the soil towards the
anode. Thus, Fig. 7(b) illustrates that a weak acidic front
of solution was generated by the electrolysis reaction at
the anode and pH slightly decreased in the first section
near the anode in the surfactant and HPCD enhanced
systems. In the second section the difference in the pH
behavior becomes more pronounced as it increases for
surfactant enhanced systems while remains approxi-
mately the same as initial pH in the HPCD enhanced
systems. The third section, which is closest to the
cathode, high soil pH was observed in the surfactant
enhanced systems while remain constant to the initial
pH in the HPCD system. In the 20% n-Butylamine
system, the soil pH remains maximum for all the soil
section specimens. This cosolvent is highly alkaline in
nature and its migration into the soil by electroosmosis
increased pH throughout the soil. The transport of OH−
into the soil from the cathode also contributed to
increase in soil pH.
The redox potentials of the soil specimens for all the
tested systems are shown in Fig. 7(c) and reflect the
opposite trend to that observed for pH. Redox potentials
were low for the cosolvent test, while they were high for
the HPCD test. Electrical conductivity values, as shown
in Fig. 7(d), reveal that the test with cosolvent had
higher electrical conductivity and it decreased signifi-
cantly from the anode to the cathode. On the other hand,
the tests conducted with surfactants and HPCD had

Fig. 9. (a) Distribution of metals in soils using 3% Tween 80 surfactant. (b). Distribution of PAHs in soils using 3% Tween 80 surfactant.
 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146
lower electrical conductivity and the values increase
slightly from the anode to the cathode. The lowest
electrical conductivity was observed in the Tween 80
enhanced system.
3.5. Residual contaminant distribution
After the completion of experiments, the soil
samples were sectioned into three equal parts: S-1
(near anode), S-2 (middle), and S-3 (near cathode).
However, the soil sample of 20% n-Butylamine
enhanced system was sectioned into five equal parts:
S1 (near anode), S2, S3 (middle), S4, and S5 (near
cathode). The contaminant concentrations determined
for each of these sections are plotted together in order to
elucidate the migration behavior of the contaminants
through the soil.
Fig. 8(a) and (b) show the residual distribution of
total metals and PAHs concentrations respectively, in
Fig. 10. (a) Distribution of metals in soils using 20% n-Butylamine cosolvent. (b). Distribution of PAHs in soils using 20% n-Butylamine cosolvent.
143
the MGP soil treated with 5% Igepal CA-720. As seen
from Fig. 8(a) all the metals are found to be evenly
distributed throughout the soil sample even after the
completion of the test. Only cadmium concentration was
found to be higher at cathode i.e. section S3, while
copper and magnesium were migrated toward anode
(i.e. sections S1 and S2) from the cathode section S3.
This may be due to their respective ions and complexes
migration towards cathode and anode. This shows that
soluble metals are negligible in the field soil. This also
reflects that Igepal CA-720 is not suitable for the
removal of metals from the MGP soil under the
investigated concentration range. In contrast, Fig. 8(b)
showed that Igepal CA-720 has strong affinity to
remove a wide array of PAHs from the MGP soil.
This plot also indicates that all the PAHs from the MGP
soil were significantly removed near anode i.e. section
S1 and middle section i.e. section S2. Comparatively
higher concentration of PAHs is found to be
144 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146

accumulated at cathode end i.e. section S3. This reflects
the migration behavior of PAHs from anode towards
cathode. The results also confirm that hydrophobic
character of the PAHs increases with the number of
rings, as the concentration of higher ringed PAHs was
found high in S2 section, indicating that these PAHs
were strongly attached to the soil under the investigated
test conditions.
Figs. 9(a), 10(a) and 11(a) show the residual metal
concentrations in different sections for tests conducted
with 3% Tween 80, 20% n-Butylamine, and 10%
HPCD, respectively. These findings suggest that only
3% Tween 80 system influenced the migration of
mercury from the soil samples. Though the removal was
not so significant, but it was found comparably suitable
for the removal of mercury complexes from the soils
towards anode. Based on these results, it can be
formulated that there are no significant changes in the
metal concentrations of different sections for all these
tested systems. This implies that the metals are not
migrated towards the electrodes under the influence of
flushing solutions used. This indicates that flushing
solutions used were not effective for desorption and/or
dissolution of metals in the soils (Maturi, 2004). This
may be due to significant amount of organic matter that
strongly adsorbed metals. In addition, the high buffering
capacity of the soil may have caused metals to exist as
precipitates. Thus strong adsorption and precipitation of
metals results into low migration and for this reason
metals did not exist in pore water.

Fig. 11. (a) Distribution of metals in soils using 10% hydroxypropyle-β-cyclodextrin. (b). Distribution of PAHs in soils using 10% hydroxypropyle-
β-cyclodextrin.
 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146
The residual concentrations of PAHs in different
sections are plotted in Figs. 9(b), 10(b), and 11(b), for the
tests conducted with Tween 80, n-Butylamine, and
HPCD, respectively. Fig. 9(b) shows low concentration
of PAHs at anode indicating the migration of PAHs occurs
from cathode to anode. Further the concentration of all the
PAHs was found to be higher in the third section i.e. S3
section. This reflects that the flushing solution Tween 80
is also suitable to solubilize appreciable amount of PAHs
from the soil sample. It is pertinent to mention here that
the concentration of Tween 80 employed in this system
was just 3% and was comparatively low with respect to
the other flushing solutions including Igepal CA-720
system. Fig. 10(b) shows very low migration trend of
PAHs in 20% n-Butylamine test, while Fig. 11(b) shows
some migration of selected PAHs towards the cathode
(i.e., from section S1 towards sections S2 and S3) in
HPCD test.
In general the concentration of PAHs increased from
section S-1 to section S-3 in all the studied systems. This
shows that PAHs migrated towards the cathode. One
conclusion that can be drawn regarding this observation
is that the use of appropriate flushing solution enhances
desorption/solubilization of PAHs in soils. It is also
observed that surfactant enhanced removal of PAHs
depends upon the micelle formation at appreciable
CMC. The differences between the efficiency of 5%
Igepal 720 and 3% Tween 80 most likely resulted from
competitive behavior among various PAH compounds
for partitioning into the stable micelles as well as com-
petitive sorption of PAH compounds and surfactant to
the soil organic matter and soil particles. This study also
indicates that 10% HPCD system had contributed partial
solubilization of the PAHs resulting in their migration
towards the cathode. HPCD enhanced system was found
to be more effective for the solubilization of low polarity
PAHs. This partial solubilization of low-polarity PAHs
is attributed to the formation of inclusion complexes
within the relatively non-polar cavity of the HPCD. The
higher electroosmotic flow in 20% n-Butylamine test
resulted in higher contaminant removal as compared to
other tests; however, very low migration trend of PAHs
in this test show that 20% n-Butylamine did not effec-
tively solubilize/desorb PAHs in the soil.
The variation in the concentrations of PAHs in their
respective sections may also be contributed by hetero-
geneous distribution of the contaminants in the soil.
These results show that Igepal CA-720, Tween 80 and
HPCD systems are effective for solubilization of the
PAHs from the MGP soil (under investigated condi-
tions). These studies also elucidate that high buffering
capacity of the soil also impede the efficiency of the
145
contaminant removal. Metals are readily precipitated
under the tested conditions due to the high buffering
capacity of soil. PAHs were found to be efficiently
solubilized by the flushing solutions under the exper-
imental conditions. Substantial electroosmotic flow can
be induced in the MGP soils using different flushing
solutions resulting in the appreciable removal of
contaminants from the MGP soil. It is also believed
that longer durations and different applied voltage
gradient and higher concentrations of the flushing
solutions may also result in the better contaminant
removal efficiency.
4. Conclusions
Based on the experimental results, the following
conclusions may be drawn:
• The manufactured gas plant soil used in this study
was contaminated with both heavy metals and
polycyclic aromatic hydrocarbons. It is feasible to
enhance extraction of PAHs using surfactants,
cosolvents and cyclodextrin from this aged MGP
field soil. However, no significant removal of heavy
metals was observed in this study. This also indicates
that heavy metals are mostly present as precipitates
due to high pH and high acid buffering capacity of
the soil.
• Substantial electroosmotic flow can be induced in
the soil using different flushing solutions. Maxi-
mum electroosmotic flow was observed in the 20%
n-Butylamine enhanced system followed by HPCD
enhanced system. Comparatively low flow was
observed in surfactant enhanced systems.
• PAHs were solubilized in the surfactant and HPCD
enhanced systems more efficiently even at low
concentration as compared to cosolvent system
resulting in significant migration towards the cath-
ode. The solubilization of PAHs using surfactants
depends upon the stability and number of the
micelles formed during the test. The mechanism of
PAHs solubilization in HPCD enhanced system was
found to be partial solubilization. The migration of
PAHs in n-Butylamine enhanced system was attrib-
uted to desorption phenomenon.
• The partially solubilized PAHs migrated from anode
towards the cathode due to electroosmotic flow. The
soil pH remains high due to its high pH buffering
capacity and under such conditions heavy metals
remain strongly adsorbed/precipitated. Therefore, the
metals are not electromigrated and are not removed
from the soil under all the tested systems.
146 K.R. Reddy et al. / Engineering Geology 85 (2006) 132–146
Acknowledgements
The financial support for this project was received
from a Technology Challenge Grant provided by the
State of Illinois. The authors are grateful to Kranti
Maturi and Craig Chawla for their assistance in this
project.
References
American Society of Testing and Materials (ASTM), 2004. Annual
Book of Standards, vol. 04.08. Soil and Rock, West Consho-
hocken, PA.
Chowdiah, P., Misra, B.R., Kilbane II, J.J., Srivastava, V.J., Hayes,
T.D., 1998. Foam propagation through soils for enhanced in situ
remediation. J. Hazard. Mater. 62, 265–280.
Dzenitis, J.M., 1997. Steady state and limiting current in electro-
remediation of soil. J. Electrochem. Soc. 144, 1317–1322.
Eykholt, G.R., 1997. Development of pore pressures by nonuniform
electroosmosis in clays. J. Hazard. Mater. 55, 171–186.
Hatheway, A.W., 2002. Geoenvironmental protocol for site and waste
characterization of former manufactured gas plants; worldwide
remediation challenge in semi volatile organic wastes. Eng. Geol.
64, 317–338.
Lee, P.H., Ong, S.K., Golchin, J., Nelson, G.L., 2001. Use of solvents
to enhance PAH biodegradation of coal tar. Water Res. 35,
3941–3949.
Maturi, K., 2004. Enhanced electrokinetic remediation of soils
contaminated with co-existing PAHs and heavy metals. M.S.
Thesis, University of Illinois, Chicago, Illinois.
McGowan, T.F., Greer, B.A., Lawless, M., 1996. Thermal and
nonthermal technologies for remediation of manufactured gas
plant sites. Waste Manage. 16, 691–698.
Reddy, K.R., Chinthamreddy, S., 2003. Sequentially enhanced
electrokinetic remediation of heavy metals in low buffering clayey
soils. J. Geotech. Geoenviron. Eng., ASCE 129 (3), 263–277.
Reddy, K.R., Parupudi, U.S., 1997. Removal of chromium, nickel, and
cadmium from clays by in situ electrokinetic remediation. J. Soil
Contam. 6, 391–407.
Saichek, R.E., Reddy, K.R., 2004. Evaluation of surfactants/
cosolvents for desorption/solubilization of phenanthrene in clayey
soils. Int. J. Environ. Stud. 61 (5), 587–604.
Shosky, D.J., 1996. Emerging technologies for recycling MGP sites.
Fuel Energy Abstr. 37, 56.
United States Environmental Protection Agency (USEPA), 1986. Test
Methods for Evaluating Solid Waste Third Edition. Laboratory
Manual, Physical/Chemical Methods, SW-846, vol. 1A. Office of
Solid Waste and Emergency Response, Washington, D.C.
United States Environmental Protection Agency (USEPA), 2000.
A Resource for MGP Site Characterization and Remediation.
Office of Solid Waste and Emergency Response, Washington D.C.
Williamson, D.G., Raymond, C.L., Kimura, Y.C., 1998. Release of
chemicals from contaminated soils. J. Soil Contam. 7, 543–558.