w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Effects of water chemistry on arsenic removal from drinking

water by electrocoagulation

Wei Wan a, Troy J. Pepping a, Tuhin Banerji b, Sanjeev Chaudhari b, Daniel E. Giammar a,*
a
Department of Energy, Environmental and Chemical Engineering and Center for Materials Innovation, Washington University in St. Louis,
One Brookings Drive, St. Louis, Missouri 63130, United States
b
Centre for Environmental Science and Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

article info abstract

Article history: Exposure to arsenic through drinking water poses a threat to human health. Electro-
Received 17 May 2010 coagulation is a water treatment technology that involves electrolytic oxidation of anode
Received in revised form materials and in-situ generation of coagulant. The electrochemical generation of coagulant
28 July 2010 is an alternative to using chemical coagulants, and the process can also oxidize As(III) to As
Accepted 10 August 2010 (V). Batch electrocoagulation experiments were performed in the laboratory using iron
Available online 17 August 2010 electrodes. The experiments quantified the effects of pH, initial arsenic concentration and
oxidation state, and concentrations of dissolved phosphate, silica and sulfate on the rate
Keywords: and extent of arsenic removal. The iron generated during electrocoagulation precipitated
Arsenic as lepidocrocite (g-FeOOH), except when dissolved silica was present, and arsenic was
Water treatment removed by adsorption to the lepidocrocite. Arsenic removal was slower at higher pH.
Electrocoagulation When solutions initially contained As(III), a portion of the As(III) was oxidized to As(V)
Phosphate during electrocoagulation. As(V) removal was faster than As(III) removal. The presence of 1
Adsorption and 4 mg/L phosphate inhibited arsenic removal, while the presence of 5 and 20 mg/L silica
Lepidocrocite or 10 and 50 mg/L sulfate had no significant effect on arsenic removal. For most conditions
examined in this study, over 99.9% arsenic removal efficiency was achieved. Electro-
coagulation was also highly effective at removing arsenic from drinking water in field trials
conducted in a village in Eastern India. By using operation times long enough to produce
sufficient iron oxide for removal of both phosphate and arsenate, the performance of the
systems in field trials was not inhibited by high phosphate concentrations.
ª 2010 Elsevier Ltd. All rights reserved.

1. Introduction (U.S. EPA, 2001, World Health Organization, 1993). Arsenic

Exposure to arsenic through drinking water is a great threat to
human health. Arsenic is a carcinogen and its consumption
can negatively affect the gastrointestinal tract and cardiac,
vascular and central nervous systems. Considering the high
toxicity of arsenic, the World Health Organization (WHO) and
U.S. Environmental Protection Agency set the maximum
acceptable level of arsenic in drinking water at 10 mg/L
occurs in groundwater primarily as the result of natural
weathering of arsenic-containing rocks, although in certain
areas, high arsenic concentrations are caused by industrial
waste discharges and application of arsenical herbicides and
pesticides (Smedley and Kinniburgh, 2002).
Iron oxides have been widely used as sorbents for arsenic
removal. They have strong adsorption affinities for arsenic
and can have large specific surface areas (Dixit and Hering,

* Corresponding author. Tel.: þ1 314 935 6849; fax: þ1 314 935 7211.
E-mail address: giammar@wustl.edu (D.E. Giammar).
0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2010.08.016
 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2
2003). Arsenic is present in water mainly in the forms of
arsenate (As(V)) and arsenite (As(III)). In the environmentally
relevant pH range 4e10, the dominant As(V) species are
negatively charged (H2AsO 2
4 and HAsO4 ), while the dominant
As(III) species is neutrally charged (H3AsO3). The negatively
charged As(V) species are more likely to be adsorbed and are
generally more easily removed than As(III) in treatment
systems (Balasubramanian and Madhavan, 2001; Kumar et al.,
2004; Parga et al., 2005).
Electrocoagulation is an alternative to using chemical
coagulants for arsenic removal, and thus may be beneficial for
communities with better access to electricity than to chem-
icals. Electrocoagulation involves electrolytic oxidation of an
appropriate anode material and in-situ generation of coagu-
lant (Kumar et al., 2004; Lakshmanan et al., 2009). When
a direct current is applied between two electrodes, metal ions
such as Fe2þ and Al3þ that can contribute to coagulant
formation are released by anode oxidation. With iron elec-
trodes, the Fe2þ released can subsequently be oxidized in
solution to produce an Fe(III) hydroxide or oxyhydroxide
(Lakshmanan et al., 2009). Several previous studies have
reported arsenic removal from water and wastewater by
electrocoagulation (Balasubramanian and Madhavan, 2001;
Kumar et al., 2004; Parga et al., 2005; Thella et al., 2008).
During such processes, arsenic removal by electrocoagulation
involved metal oxide formation followed by arsenic removal
(Balasubramanian and Madhavan, 2001). Electrocoagulation
may also control oxidationereduction reactions; species such
as As(III) may be oxidized on the anode and other species may
be reduced on the cathode.
Several water chemistry factors may affect arsenic
removal by electrocoagulation. The pH can affect arsenic
species distribution and the surface charge of the metal oxides
formed. Arsenic oxidation state can affect arsenic removal,
with As(V) being more easily removed than As(III) (Kumar
et al., 2004). Phosphate may compete with arsenic for
adsorption sites (Meng et al., 2002). Silica can be present at
high concentration in groundwater, and in studies with iron
oxide-based sorbents it has inhibited arsenic removal (Davis
et al., 2001; Zeng et al., 2008a).
The objectives of this study were to assess the impact of
important water chemistry factors on arsenic removal by
electrocoagulation and to examine the performance of elec-
trocoagulation for arsenic removal in laboratory and field
settings. Factors studied were the pH, arsenic oxidation state,
initial arsenic concentration, and the concentrations of dis-
solved phosphate, silica, and sulfate.
2. Materials and methods
2.1. Laboratory-scale electrocoagulation experiments
The electrocoagulation reactor consisted of a 1 L glass beaker
with two iron rods immersed in the aqueous solution. The
rods had diameters of 1.75 cm, lengths of 20 cm, and were
placed 2 cm apart in the arsenic-containing solution. The total
submerged surface area of each electrode was 57 cm2. Before
each experiment, the electrodes were abraded with sand
paper to remove scales and then cleaned with 1 M HNO3 and
385
ultrapure water. A voltage of 12 V was applied to the terminal
electrodes from a direct current power supply. The electric
current was monitored over the course of each two-hour
experiment. To provide enough oxygen for the formation of Fe
(III) precipitates, the solution was sparged with air at a flow
rate of 60 mL/min. The arsenic-containing solution was
magnetically-stirred (200 rpm). Duplicate runs were carried
out for each set of experimental conditions.
The water compositions evaluated are listed in Table 1. All
solutions were prepared with ultrapure water and desired
volumes of stock solutions. An As(V) stock solution was made
from Na2HAsO4$7H2O. The As(III) stock solution was made from
NaAsO2. Tests indicated that the stock As(III) solution contained
about 15% As(V). The phosphate, silica, and sulfate stock solu-
tions were made from Na2HPO4$7H2O, Na2SiO3$9H2O, and
Na2SO4$10H2O, respectively. In order to provide pH buffering,
NaHCO3 was added to achieve a concentration of 1 mM. The pH
of the As-containing solution in each beaker was periodically
readjusted to the target value by adding aliquots of 1 M HNO3 or
1 M NaOH.
For the experiments using As(V), every 5, 15 or 30 min,
15 mL of solution was collected from the beaker. Of this
amount, 7.5 mL was filtered using a 0.45 mm filter membrane
(polyethersulfone), and the filtrate was acidified to 1% HNO3.
Another 7.5 mL of unfiltered suspension was acidified to 1%
HNO3, which completely dissolved the suspended solids. For
the experiments using As(III), an additional sample was
collected for arsenic redox speciation (i.e. separation of As(III)
and As(V)) using an anion-exchange method (Wilkie, 1997).
Before separation, the pH of a 10 mL aliquot of filtered solution
was adjusted to around 3.5 and then passed through a column
containing anion-exchange resin. During As separation, the
first 5 mL of solution were wasted and the remaining 5 mL
were collected. In this method, As(III) eluted through the
column and As(V) was retained on the resin in the column. At
the end of each experiment, 15 mL of unfiltered suspension
was collected for zeta potential measurement. The remaining
settleable solids were collected and freeze-dried in prepara-
tion for solid-phase characterization.
The pH of the solution in the beaker was periodically
measured over the course of the experiment with a pH elec-
trode and adjusted to the target value if necessary. The current
was briefly turned off during the times when the pH was
measured, because the current interfered with the pH
measurement. It took approximately 1 min to measure and
adjust the pH to the desired value. Considering the time used to
adjust the pH, each experiment lasted about 130 min, although
current was only applied during 120 min of the experiment.
Table 1 e Experimental variables evaluated in the
electrocoagulation experiments.
Parameters Range of values
Initial arsenic concentration 100 and 1000 mg/L
Arsenic oxidation state As(III) and As(V)
pH 5, 6, 7, 8 and 9
Phosphate 0, 1 and 4 mg/L as P
Silica 0, 5 and 20 mg/L as SiO2
Sulfate 0, 10 and 50 mg/L as SO2
4
386 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2
The filtered and acid-treated samples from the laboratory
experiments were analyzed for dissolved and total concen-
trations of constituents, respectively. The concentrations of
As, Fe, P and Si were determined by inductively coupled plasma
mass spectrometry (ICP-MS) (7500ce, Agilent Technologies,
Santa Clara, CA). The instrument detection limits for As, Fe, P
and Si were 0.1 mg/L, 0.05 mg/L, 0.01 mg/L, and 0.03 mg/L,
respectively. The specific surface areas (SSA) of the solids were
measured by the BET N2-adsorption method (Autosorb-1-C,
Quantachrome, U.S.A.). X-ray powder diffraction (XRD)
patterns were collected using Cu Ka radiation (D-MAX/A,
Rigaku, Japan). Zeta potential was measured by a nanoparticle
characterization instrument with zeta potential capability
(Nanoseries ZS, Malvern Instruments, U.K.). Dissolved oxygen
for selected samples was measured using a Hach Surface
Water Test Kit (Fondriest Environmental, Inc.).
2.2. Materials and methods of the field study
The electrocoagulation system used for the field studies
consisted of a 50 L plastic bucket, a 12 V direct current source
with a rating of 2 A, two iron plates (10 cm by 15 cm with
a submerged depth of 10 cm), and an aquarium pump with
a diffuser. The iron plates were separated by a distance of
0.5 cm and were connected to a support that fixed them in
place with a non-conductive PVC screw. The electro-
coagulation system was followed by a common ceramic
candle filter assembly (average pore size of 1 mm) for the
removal of suspended particles (Fig. 1).
Field trials for the electrocoagulation system were carried
out in a village in the Nadia district of West Bengal. The village
chosen for the field trials was Ghetugachi, which falls under
Chakdaha block. This village has more than 650 families. The
average arsenic level was found to be 400 mg/L, and the
phosphate levels in the groundwater were also high. Dis-
solved iron in the groundwater ranges from 0.5 to 4.0 mg/L.
Ghetugachi was selected for the field trials because
a predominance of As(III) and the high phosphate concen-
trations presented a challenging scenario for arsenic removal.
A total of 17 electrocoagulation systems were distributed in
Ghetugachi.
After they had been distributed and used by the households
for one week, the field systems were evaluated by treating 50 L
volumes of water collected in the bucket from the hand pumps
located close to each house where an electrocoagulation
system had been installed. The aquarium pump was switched
- 12 V DC +
air pump source iron electrodes
water level
As(III) iron corrosion and
oxidation oxide precipitation
As(V)
adsorption
As
settling
diffuser
Fig. 1 e Domestic electrocoagulation system used for the field study. Water from the electrocoagulation reactor (left) is
further treated using a candle filter unit (right) to remove suspended particles.
on for 15 min to allow the dissolved oxygen in the water to
increase to over 3 mg/L. The electrode assembly was then
immersed in the water and the voltage was applied for 3 h to
provide electrocoagulation treatment. The water was then
allowed to settle for 4 h. The supernatant remaining after
settling was collected and filtered through a ceramic candle
filter that was procured locally. For each system, three water
samples were collected: (1) raw water prior to aeration,
(2) supernatant of the treated water after 4 h of settling, and
(3) water after candle filtration. The water samples were
preserved by acidification with 0.1 N HCl.
The field samples were analyzed by spectroscopic
methods. The arsenic concentrations were determined with
a rapid spectrophotometric method (Dhar et al., 2004). The
iron concentrations were determined using the phenanthro-
line method (Clesceri et al., 1999). Additional samples were
selected randomly for analysis by atomic absorption spec-
trometry (AA 400-FIAS, Perkin Elmer, USA) to verify the iron
and arsenic concentrations determined by the colorimetric
methods. The error of the spectrophotometric methods (for
arsenic and iron analysis) was only 2.0%. The phosphate
concentrations in the field samples were determined using
the ascorbic acid method (Clesceri et al., 1999).
3. Results and discussion
3.1. Production of iron oxide coagulant
During the electrocoagulation process, the solutions changed
from clear and colorless to turbid and reddish brown. The
concentrations of the total iron generated during electro-
coagulation increased linearly with reaction time. In this study,
the reported values are the average plus the standard devia-
tion. For all the electrocoagulation experiments, about 50 mg/L
(average value was 50.5 mg/L) total iron was produced over the
2 h experiment duration. The reactor was operated with
a current of 22 mA, and the total iron produced was consistent
with a value of 52.2 mg/L predicted by Faraday’s law for the
oxidation of the iron electrode to dissolved Fe(II). The current
was nearly identical for the different pH values studied and did
not vary over the course of the experiment, which indicated
that the conductivity of the solutions was not affected by the
electrocoagulation process. The total electricity used to treat
1 L of water was 0.0005 kW h or 0.5 kW h/m3. Fe was released to
solution as Fe(II) (Reaction (1)) and was then oxidized to Fe(III)
unfiltered water
ceramic candle filter
1 µm pore size
water level
filtered water
 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2 387

by the dissolved oxygen. The Fe(III) then precipitated in the 100 ppb As(III), pH = 7, 10 ppm SO4
2-

formation of iron oxides (Reaction (2)).

100 ppb As(III), pH = 7, 5 ppm SiO2

FeðsÞ /Fe2þ þ 2e (1)

100 ppb As(III), pH = 7, 1 ppm P

Normalized Intensity
Fe2þ þ 1:5H2 O þ 0:25O2 /g-FeOOHðsÞ þ 2Hþ (2) 100 ppb As(III), pH = 7

For the experiment using 100 mg/L As(V) at pH 7, the dis- 100 ppb As(V), pH = 5
solved oxygen concentration was measured to be about
100 ppb As(V), pH = 9
10 mg/L at 30, 60 and 90 min, which verified that the solution
was saturated with dissolved oxygen. In each experiment the 1000 ppb As(V), pH = 7
dissolved iron concentrations were very low compared with
100 ppb As(V), pH = 7
the total iron concentrations, which indicated that nearly all
of the iron was present in the solid phases. PDF#00-044-1415, Lepidocrocite: FeO(OH)
The Fe(III) solid was identified as lepidocrocite (g-FeOOH) by
its XRD pattern (Fig. 2). The identity of the iron oxyhydroxides 10 20 30 40 50 60 70 80
o
formed was not affected by As(III) and As(V) or by the presence 2θ( )
of sulfate or phosphate. The lepidocrocite particles were
Fig. 2 e X-ray diffraction patterns of solids generated
100e200 nm long and about 5e20 nm wide. The XRD pattern at
during electrocoagulation. The reference pattern for
pH 5 had no clear XRD peaks, probably resulting from insuffi-
lepidocrocite is included for comparison.
cient collection of solids for XRD characterization. The specific
surface area of the solids was 200 m2/g and was independent of
the solution compositions. The formation of lepidocrocite is
consistent with published synthesis methods involving experimental conditions and that no arsenic-Fe(III) precipi-
oxidation of Fe(II) solutions using dissolved oxygen at ambient tates (e.g. FeAsO4(s)) were expected to form.
temperature (Schwertmann and Cornell, 2000). The presence
of silica significantly affected the iron oxides formed during
electrocoagulation (Fig. 2). Less crystalline iron oxides were 3.2. Effect of pH on As removal
formed in the presence of silica. A previous study found that
silica influenced the type of iron oxides formed in aerated Fe The removal rate of arsenic was affected by the pH. For
(II)- and As(III)-containing water by decreasing the degree of experiments using 100 mg/L As(V), it took less than 30 min for
corner-sharing linkages of Fe(III)-octahedra during Fe(III) the dissolved As(V) concentrations to drop below 1 mg/L at pH
polymerization (Voegelin et al., 2010). Equilibrium calculations 5, 6 and 7 (Table 2). At pH 8, the dissolved As(V) concentration
indicate that Fe(III) oxyhydroxides should precipitate at the only reached this value after 75 min, and at pH 9 the dissolved

Table 2 e Water compositions in electrocoagulation experiments and the time to achieve dissolved arsenic concentrations
below 10 mg/L and 1.0 mg/L.
Arsenic Arsenic pH Phosphate Silica as Sulfate as Time to dissolved Time to dissolved
oxidation state (mg/L) as P (mg/L) SiO2 (mg/L) So2
4 (mg/L
)
As <10 mg/L (min) As <1.0 mg/L (min)

As(III) 100 5 0 0 0 30 90
As(III) 100 6 0 0 0 30 90
As(III) 100 7 0 0 0 30 90
As(III) 100 8 0 0 0 30 90
As(III) 100 9 0 0 0 45 90
As(III) 1000 7 0 0 0 90 120
As(V) 1000 7 0 0 0 45 >120
As(V)a 100 5 0 0 0 15 15
As(V)a 100 6 0 0 0 15 30
As(V) 100 7 0 0 0 5 15
As(V)a 100 8 0 0 0 75 75
As(V)a 100 9 0 0 0 75 >120
As(V) 100 7 1 0 0 60 >120
As(V) 100 7 4 0 0 90 >120
As(V) 100 7 0 5 0 5 60
As(V) 100 7 0 20 0 5 60
As(V) 100 7 0 0 10 5 30
As(V) 100 7 0 0 50 5 45
As(III) 100 7 1 0 0 45 90
As(III) 100 7 0 5 0 30 90
As(III) 100 7 0 0 10 30 90

a The first sample was collected after 15 min in these experiments but after 5 min in all other experiments.
388 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2

As(V) concentrations remained at 4 mg/L after 120 min (Fig. 3).

For experiments using 100 mg/L As(III), it took 30 min for the 100 pH 5-As
dissolved As concentrations to drop below 10 mg/L at pH 5e8, pH 5-As(V)
and at pH 9 it took 45 min (Fig. 4 and Table 2). Although As pH 7-As

Dissolved As(V) ( g/L)

80
removal was slower at higher pH, the final removal efficiency pH 7-As(V)
was independent of pH from 5 to 8. Once sufficient lep-
pH 9-As
60 pH 9-As(V)
idocrocite was produced to provide adsorption sites for As,
low dissolved As concentrations could be obtained. Kumar
et al. (2004) also reported that the final As removal efficiency 40
was independent of pH with increasing pH from 6 to 8 during
electrocoagulation. 20
The influence of pH on adsorption can explain the slower
rate of As(V) removal with increasing pH. Previous studies
0
observed less adsorption of As(V) to hydrous ferric oxide and 0 15 30 45 60 75 90 105 120
goethite with increasing pH (Dixit and Hering, 2003; Meng
Time (min)
et al., 2000). The lepidocrocite produced in the electro-
coagulation reactor had an isoelectric pH of about 7.0 (Fig. 5), Fig. 4 e Dissolved concentrations of total As and As(V)
which is comparable to that reported by Peacock and Sherman during electrocoagulation of solutions initially containing
(2004) from a potentiometric titration method. Below this pH 100 mg/L As(III) at pH 5, 7, and 9.
the surfaces of the particles are positively charged and elec-
trostatic contributions as well as chemical contributions
contribute to As(V) adsorption. Above the isoelectric point,
both the As(V) species and the lepidocrocite surface are was modeled as a first-order reaction with respect to arsenic
(Balasubramanian et al., 2009). Although such a model did not
negatively charged and adsorption is less favorable.
The rate of arsenic removal depends on the rate of lep- account for the change in the Fe(III) oxyhydroxide coagulant
idocrocite production and the rate of arsenic adsorption to the with time and predicts that the arsenic concentration
lepidocrocite. A recent model combined these two processes approaches an unreasonable value of zero with increasing
to interpret As(V) removal from solution during electro- reaction time, this first-order model could still provide a good
coagulation (Wan, 2010). The rate of lepidocrocite generation fit to the experimental data for the treatment times
is essentially constant because of the nearly constant release investigated.
of Fe(II) from the electrode. At the conditions studied Fe(II)
oxidation to produce lepidocrocite is fast relative to Fe(II) 3.3. Effect of oxidation state on As removal
release from the electrode, so the release of Fe(II) at the elec-
trode is the rate-limiting step in lepidocrocite production. As(V) removal was usually faster than As(III) removal (Figs. 3
Arsenic adsorption to the lepidocrocite was successfully and 4). This was in agreement with the results reported by
modeled as a process that was first-order with respect to Kumar et al. (2004) for electrocoagulation and by Meng et al.
lepidocrocite, first-order with respect to dissolved arsenic, and (2002) for arsenic adsorption to iron hydroxides. In all the
that proceeded until equilibrium adsorption was reached. In experiments using As(III), the dissolved As(V) concentration
a separate study, arsenic removal during electrocoagulation increased first and then decreased with increasing reaction

100
pH 5
40
pH 6 No As
pH 7 30 1000 g/L As(V)
D i ss o l v e d A s ( V ) ( g/L)

80 pH 8 100 g/L As(III)

pH 9 20
Zeta Potential (mV)

60
10

0
40
-10
20
-20

0 -30
0 15 30 45 60 75 90 105 120
-40
Time (min) 3 4 5 6 7 8 9 10 11 12
pH
Fig. 3 e Dissolved As(V) concentrations during
electrocoagulation of solutions initially containing 100 mg/L Fig. 5 e Effect of pH on zeta potential of the lepidocrocite
As(V). The data points for pH 5 are partially obscured by suspension generated by electrocoagulation in the
those for pH 6. presence and absence of dissolved arsenic.
 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2 389

time. The increase in As(V) when treating As(III) solutions

a 1 mg/L P
indicated that at least 25% of the As(III) was oxidized to As(V)

Dissolved P(mg/L)
4 4 mg/L P
during electrocoagulation (Fig. 4). The removal mechanism for
As(III) by electrocoagulation was proposed to be the oxidation
of As(III) to As(V) followed by adsorption of As(V) to the iron 2
oxides (Kumar et al., 2004). As(III) oxidation to As(V) has
previously been proposed to occur with dissolved oxygen and
soluble intermediates in Fe(II) oxidation acting as rate- 0
0 15 30 45 60 75 90 105 120
enhancing species (Ciardelli et al., 2008; Sahai et al., 2007). As
(III) oxidation can also occur when Fe(II) is present with Fe(III) b
100 No P

Dissolved As(V)(μg/L)
oxyhydroxides, and the mechanism has been proposed to
1 mg/L P
involve the formation of reactive Fe intermediate species 80
4 mg/L P
(Amstaetter et al., 2010; Bisceglia et al., 2005). 60
40
3.4. Effect of initial As concentrations on As removal
20
0
As(V) and As(III) removal to below 1 mg/L took more time when
0 15 30 45 60 75 90 105 120
the solutions had higher initial As concentrations (Fig. 6 and
Table 2). When the initial arsenic concentrations were higher, c 60
, No P
more iron oxides were needed to decrease the dissolved
, 1 mg/L P
arsenic concentrations. Arsenic removal is consequently 40 , 4 mg/L P
Fe (mg/L)
limited by the production rate of lepidocrocite. However, the
final As removal efficiencies were independent of the initial As
20
(III) concentration and were over 99.9%.

3.5. Effect of phosphate on As removal 0

0 15 30 45 60 75 90 105 120
Time (min)
The presence of 1 mg/L and 4 mg/L phosphate as P inhibited
the removal of As (Figs. 7 and 8). The inhibitory effect was Fig. 7 e Dissolved (a) phosphate (in mg P/L), (b) dissolved As
more significant at higher phosphate concentrations. (V), and (c) total (closed symbols) and dissolved (open
Considerable phosphate was also removed during electro- symbols) iron concentrations during electrocoagulation of
coagulation (Figs. 7a and 8a), which indicates that phosphate solutions initially containing 100 mg/L As(V) at pH 7.
can compete with As species for the surface sites of

a 1.0 100 g/L-As lepidocrocite. A competitive adsorption effect of phosphate on

100 g/L-As(V) As removal agrees with the results of previous studies (Meng
0.8 1000 g/L-As et al., 2002; Zeng et al., 2008b). The inhibitory effect of phos-
1000 g/L-As(V) phate on As removal during electrocoagulation may also be
[As]/[As]0

0.6
caused in part by the slower oxidation of Fe(II) to Fe(III) in the
0.4 presence of phosphate, which can decrease the rate at which
the sorbent is formed (Fig. 7c). Recent work observed forma-
0.2
tion of Fe(III)-phosphate solids during the oxidation of Fe(II) in
0.0 phosphate-rich solutions (Voegelin et al., 2010); however, in
0 15 30 45 60 75 90 105 120
the present study phosphate did not affect the identity of the
b iron oxide formed during electrocoagulation (Fig. 2). Lep-
1.0 100 g/L
1000 g/L odocrocite was the only phase indicated in XRD patterns, and
0.8 there were no peaks for Fe(II)- or Fe(III)-phosphate solids.
[As]/[As]0

0.6
3.6. Effect of dissolved silica on As removal
0.4

0.2 The presence of 5 and 20 mg/L dissolved silica had no signif-

icant effect on As removal, even though considerable silica
0.0
0 15 30 45 60 75 90 105 120 was removed during the electrocoagulation process (Fig. 9)
and silica prevented the formation of lepidocrocite (Fig. 2).
Time (min)
Meng et al. (2002) also observed no significant effects of silica
Fig. 6 e Relative change in dissolved arsenic during on As(V) adsorption to iron hydroxides when silica was
electrocoagulation of solutions initially containing 100 mg/L present at concentrations as high as 36 mg/L. In a separate
and 1000 mg/L (a) As(III) and (b) As(V) at pH 7. Panel (a) also study, Davis et al. (2001) observed silica inhibition of As(V)
shows the concentration of dissolved As(V). adsorption to ferric hydroxide, but only when silica and ferric
390 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2

a 1.0 a 20 5 mg/L SiO

Dissolved SiO (mg/L)

20 mg/L SiO
Dissolved P (mg/L)

0.8 15

0.6 10

0.4 5

0.2 0
15 30 45 60 75 90 105 120
0.0
0 15 30 45 60 75 90 105 120
b 100 No SiO2

Dissolved As(V)(μg/L)
b 100 No P-As 80 5 mg/L SiO
No P-As(V) 20 mg/L SiO
1 mg/L P-As 60
Dissolved As ( g/L)

80
1 mg/L P-As(V)
40
60
20
40 0
15 30 45 60 75 90 105 120
20

0
c 60
0 15 30 45 60 75 90 105 120 , No SiO
, 5 mg/L SiO

Time (min)
Fig. 8 e Dissolved (a) phosphate (in mg P/L) and (b) As
concentrations during electrocoagulation treatment of
solutions initially containing 100 mg/L As(III) at pH 7.
hydroxide had been pre-equilibrated for 50 days; at shorter
contact times, there was much less inhibition. An inhibitory
effect of silica was also observed in a study of arsenic removal
using a porous iron oxide-based sorbed in packed columns
(Zeng et al., 2008a); the inhibition was likely caused by silica
polymerization to physically block access to adsorption sites
within internal pores of the solid and not through any
competitive adsorption phenomenon.
3.7. Effect of sulfate on As removal
The presence of 10 and 50 mg/L SO24 (data not shown) did not
affect the removal of As. Because sulfate did not affect lep-
idocrocite formation and does not adsorb as strongly as As(V)
and phosphate, it is not surprising that sulfate did not affect
the performance of the electrocoagulation process. Similar
observations were made in As adsorption experiments with
sulfate concentrations higher than 200 mg/L (Meng et al.,
2000).
3.8. Arsenic removal in field trials
Almost all of the field units were able to lower arsenic
concentrations to below 10 mg/L after electrocoagulation,
settling, and candle filtration (Table 3 and Fig. 10). For the
small number of systems that were unable to achieve 10 mg/L
(systems 15e17), visits to the households determined that the
units were not being operated as per the instructions
provided. The users of these units were either not aerating the
water (i.e. the aquarium pump was not switched on) or were
carrying out the electrocoagulation for less than the 3 h period
specified in the instructions.
Fe (mg/L)
40
, 20 mg/L SiO
20
0
15 30 45 60 75 90 105 120
Time (min)
Fig. 9 e Concentrations of (a) dissolved SiO2, (b) dissolved
As(V) and (c) total (closed symbols) and dissolved (open
symbols) iron concentrations during electrocoagulation of
solutions initially containing 100 mg/L As(V) at pH 7.
The untreated waters had very high arsenic concentra-
tions, in the range 400e700 mg/L (Table 3 and Fig. 10), and most
also had high phosphate concentrations. Despite the high
initial concentrations, most final filtered water samples had
arsenic concentrations less than 10 mg/L, and the phosphate
concentrations in treated waters were below the detection
limit of 0.5 mg/L. While significant removal of arsenic occurred
with electrocoagulation followed by settling, the filtration step
led to further decrease in the arsenic concentration. In several
cases the filtration step was required to lower the arsenic
concentrations below the 10 mg/L WHO standard and not just
the 50 mg/L Indian standard. Additional arsenic removal
probably occurred on the 1 mm pore size candle filters due to
the removal of microflocs of iron oxide that adsorbed arsenic
but were not separated during the sedimentation phase.
No inhibition of treatment performance from phosphate
was observed in the field trials. The systems with the highest
phosphate concentrations (9 and 10) had settled and filtered
water concentrations that were comparable to those from the
other properly operated systems. The electrocoagulation
treatment time for the field trials was selected to produce
sufficient iron oxide coagulant to remove both As and phos-
phate, so it is not surprising that no effect of phosphate was
seen in these trials. Preliminary laboratory studies with the
reactors used in the field indicated that when water contained
 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2 391

Table 3 e Untreated and treated water compositions for field investigation of electrocoagulation systems.
System pH Initial Conc. (mg/L) Supernatant Conc. (mg/L) Filtrate Conc. (mg/L)

Astot As(V) PO3

4 eP Astot PO3
4 eP Astot PO3
4 eP

1 7.1 526 125 203 13.3 n.d. 0.4 n.d.

2 7.1 526 125 203 9.4 n.d. 0.2 n.d.
3 484 198 12.7 n.d. n.d. n.d.
4 6.8 449 92 182 13.6 n.d. n.d. n.d.
5 6.8 449 92 182 8.2 n.d. n.d. n.d.
6 483 192 12.1 n.d. n.d. n.d.
7 7.0 677 103 253 16.9 n.d. n.d. n.d.
8 526 198 24.1 n.d. n.d. n.d.
9 675 753 11.8 n.d. 3.9 n.d.
10 7.0 677 103 753 11.9 n.d. 0.5 n.d.
11 488 144 7.9 n.d. 4.5 n.d.
12 449 156 12.6 3 n.d. n.d.
13 527 193 14.7 3 n.d. n.d.
14 483 178 30.7 5 5.8 n.d.
15 527 198 48.5 21 10.3 2
16 482 219 50.4 33 26.4 5
17 7.0 677 103 753 50.9 31 25.8 4

*n.d. e not detectable. The detection limits for arsenic and phosphate were 0.2 mg/L and 0.5 mg/L, respectively.

800 80

700 70

As Conc. in Treated Water (ug/L)

Raw water As Conc. (ug/ L)

600 60

500 50

400 40

300 30

200 20

100 10

0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
System ID
Raw Water Settled Water Filtered Water

Fig. 10 e Results of electrocoagulation treatment with domestic systems deployed for the field study. The dashed lines
indicate the WHO guideline value for arsenic in drinking water (10 mg/L) and the Indian standards for arsenic in drinking
water (50 mg/L). Note that the raw water concentration is shown on a separate axis from the settled and filtered water
concentrations.

500 mg/L As(III) and 500 mg P/L phosphate at a pH of approxi-
mately 7, the As concentrations could be decreased below
10 mg/L when 20 mg/L of Fe was released to form the coagu-
lant, which occurred in about 1.5 h. To provide a factor of
safety, the 3 h treatment time was selected for the field tests,
and this reaction time produced 40 mg/L of Fe and overcame
the competitive adsorption with phosphate. The iron
produced during the treatment time generates about 80 mg of
settled solids for each liter of water treated.
The electricity required for treatment of 50 L was
0.036e0.039 kW h or 0.72e0.78 kW h/m3. The time required to
produce sufficient coagulant for arsenic removal depends on
the current, which in turn is affected by the conductivity of
the water. The electrode spacing and treatment time can be
adjusted for the conductivity of a specific groundwater. The
operating costs of treatment are primarily those of the elec-
tricity. Based on typical electricity costs in India, the operating
costs are estimated to be 5 Indian Rupees per m3 of water
treated ($0.11/m3 with a currency conversion of 46 Rupees per
U.S. dollar).
4. Conclusions
The iron generated during electrocoagulation was present in
solid phases as lepidocrocite. Arsenic removal by electro-
coagulation involved lepidocrocite formation followed by
arsenic adsorption. As removal was slower at higher pH and
392 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 3 8 4 e3 9 2
higher initial arsenic concentrations. As(III) was partially
oxidized to As(V) during electrocoagulation. As(V) removal
was faster than As(III) removal. Phosphate inhibited As
removal by acting as a competing adsorbate and possibly by
delaying the oxidation of Fe(II) to produce lepidocrocite.
Although silica prevented the formation of lepidocrocite,
arsenic removal was still very rapid and extensive by
adsorption to the amorphous iron oxides that formed. Sulfate
had no significant effect on As removal or coagulant forma-
tion. Over 99.9% arsenic removal efficiency could be achieved
in both the laboratory experiments and the field trials. The
field trials indicate that electrocoagulation has the potential to
treat arsenic-contaminated water to standards required for
drinking water.
Acknowledgments
The project was supported by the McDonnell Academy Global
Energy and Environment Partnership of Washington
University. The field study was performed with the support of
the Department of Science and Technology of the Govern-
ment of India. The authors are grateful to Kate Nelson for
assistance with laboratory analyses. The authors thank two
anonymous reviewers for their comments that helped
improve this paper.
references
Amstaetter, K., Borch, T., Larese-Casanova, P., Kappler, A., 2010.
Redox transformation of arsenic by Fe(II)-activated goethite
(alpha-FeOOH). Environmental Science & Technology 44 (1),
102e108.
Balasubramanian, N., Madhavan, K., 2001. Arsenic removal from
industrial effluent through electrocoagulation. Chemical
Engineering & Technology 24 (5), 519e521.
Balasubramanian, N., Kojima, T., Basha, C.A., Srinivasakannan, C.,
2009. Removal of arsenic from aqueous solution using
electrocoagulation. Journal of Hazardous Materials 167 (1e3),
966e969.
Bisceglia, K.J., Rader, K.J., Carbonaro, R.F., Farley, K.J., Mahony, J.D.,
Di Toro, D.M., 2005. Iron(II)-catalyzed oxidation of arsenic(III) in
a sediment column. Environmental Science & Technology 39
(23), 9217e9222.
Ciardelli, M.C., Xu, H.F., Sahai, N., 2008. Role of Fe(II), phosphate,
silicate, sulfate, and carbonate in arsenic uptake by
coprecipitation in synthetic and natural groundwater. Water
Research 42 (3), 615e624.
Clesceri, L.S., Greenberg, A.E., Eaton, A.D. (Eds.), 1999. Standard
Methods for the Examination of Water and Wastewater.
American Public Health Association, American Water Works
Association, Water Environment Federation, Washington, DC.
Davis, C.C., Knocke, W.R., Edwards, M., 2001. Implications of
aqueous silica sorption to iron hydroxide: mobilization of iron
colloids and interference with sorption of arsenate and humic
substances. Environmental Science & Technology 35 (15),
3158e3162.
Dhar, R.K., Zheng, Y., Rubenstone, J., van Geen, A., 2004. A rapid
colorimetric method for measuring arsenic concentrations in
groundwater. Analytica Chimica Acta 526 (2), 203e209.
Dixit, S., Hering, J.G., 2003. Comparison of arsenic(V) and arsenic
(III) sorption onto iron oxide minerals: implications for arsenic
mobility. Environmental Science & Technology 37 (18),
4182e4189.
Kumar, P.R., Chaudhari, S., Khilar, K.C., Mahajan, S.P., 2004.
Removal of arsenic from water by electrocoagulation.
Chemosphere 55 (9), 1245e1252.
Lakshmanan, D., Clifford, D.A., Samanta, G., 2009. Ferrous and
ferric ion generation during iron electrocoagulation.
Environmental Science & Technology 43 (10), 3853e3859.
Meng, X.G., Bang, S., Korfiatis, G.P., 2000. Effects of silicate,
sulfate, and carbonate on arsenic removal by ferric chloride.
Water Research 34 (4), 1255e1261.
Meng, X.G., Korfiatis, G.P., Bang, S.B., Bang, K.W., 2002. Combined
effects of anions on arsenic removal by iron hydroxides.
Toxicology Letters 133 (1), 103e111.
Parga, J.R., Cocke, D.L., Valenzuela, J.L., Gomes, J.A., Kesmez, M.,
Irwin, G., Moreno, H., Weir, M., 2005. Arsenic removal via
electrocoagulation from heavy metal contaminated
groundwater in La Comarca Lagunera Mexico. Journal of
Hazardous Materials 124 (1e3), 247e254.
Peacock, C.L., Sherman, D.M., 2004. Copper(II) sorption onto
goethite, hematite and lepidocrocite: a surface complexation
model based on ab initio molecular geometries and EXAFS
spectroscopy. Geochimica Et Cosmochimica Acta 68 (12),
2623e2637.
Sahai, N., Lee, Y.J., Xu, H.F., Ciardelli, M., Gaillard, J.F., 2007. Role
of Fe(II) and phosphate in arsenic uptake by coprecipitation.
Geochimica Et Cosmochimica Acta 71 (13), 3193e3210.
Schwertmann, U., Cornell, R.M., 2000. Iron Oxides in the
Laboratory. Wiley-VCH, New York, NY.
Smedley, P.L., Kinniburgh, D.G., 2002. A review of the source,
behaviour and distribution of arsenic in natural waters.
Applied Geochemistry 17, 517e568.
Thella, K., Verma, B., Srivastava, V.C., Srivastava, K.K., 2008.
Electrocoagulation study for the removal of arsenic and
chromium from aqueous solution. Journal of Environmental
Science and Health Part A-Toxic/Hazardous Substances &
Environmental Engineering 43 (5), 554e562.
U.S. EPA, 2001. National primary drinking water regulations; arsenic
and clarifications to compliance and new source contaminants
monitoring; final Rule. Federal Register 66 (14), 6976.
Voegelin, A., Kaegi, R., Frommer, J., Vantelon, D., Hug, S.J., 2010.
Effect of phosphate, silicate, and Ca on Fe(III)-precipitates
formed in aerated Fe(II)- and As(III)-containing water studied
by X-ray absorption spectroscopy. Geochimica Et
Cosmochimica Acta 74 (1), 164e186.
Wan, W., 2010. Arsenic Removal from Drinking Water by
Electrocoagulation. M.S. thesis, Washington University, St.
Louis.
Wilkie, J.A., 1997. Ph.D. dissertation, University of California, Los
Angeles, Los Angeles, CA.
World Health Organization, 1993. Guidelines for Drinking-water
Quality. In: Recommendations, vol. 1. WHO, Geneva.
Zeng, H., Arashiro, M., Giammar, D.E., 2008a. Effects of water
chemistry and flow rate on arsenate removal by adsorption to
an iron oxide-based sorbent. Water Research 42 (18),
4629e4636.
Zeng, H., Fisher, B., Giammar, D.E., 2008b. Individual and
competitive adsorption of arsenate and phosphate to a high-
surface-area iron oxide-based sorbent. Environmental Science
& Technology 42 (1), 147e152.