Journal of Hazardous Materials 172 (2009) 16171622

Contents lists available at ScienceDirect

Journal of Hazardous Materials

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

Energy and electrode consumption analysis of electrocoagulation

for the removal of arsenic from underground water
J.F. Martnez-Villafane , C. Montero-Ocampo, A.M. Garca-Lara
CINVESTAV-IPN Unidad Saltillo, Carretera Saltillo-Monterrey Km 13.5, A.P. 663, 25900 Saltillo, Coahuila, Mexico

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

Article history: A systematic study of the effect of design and operation conditions of an electrochemical reactor on
Received 26 November 2008 the treatment time for arsenic (As) electro-removal from underground water (GW) was carried out to
Received in revised form 9 July 2009 analyse the energy and electrode consumption. The effects of four factorscurrent density, interelectrode
Accepted 10 August 2009
distance, electrode areavolume ratio, and liquid motion driving modewere evaluated. The response
Available online 18 August 2009
variables were the energy and the electrode consumption and the treatment time to reduce the GW
residual As concentration to 10 g L1 , which is the maximum contaminant level (MCL) established by
Keywords:
the World Health Organization (WHO) in drinking water. The results obtained in this study showed that
Electrocoagulation
Underground water
the factor that had the greatest effect on most of the response variables was the liquid motion driving
Arsenic electro-removal mode. The best residence time was 20 s, which favoured low energy consumption (58.78 Wh m3 ) and
low electrode material loss (9.59 g m3 ).
2009 Elsevier B.V. All rights reserved.

1. Introduction electrode consumption there are not enough proposals reported.

Arsenic contamination of underground water occurs mainly by
natural processes and from uncontrolled industrial discharge. Due
to the high toxic effects of arsenic (As) on human health, the World
Health Organization (WHO) has established an maximum contami-
nant level (MCL) of 10 g L1 for this water contaminant [1]. Among
the As removal techniques available, the electrochemical method of
electrocoagulation (EC) presents important advantages over others
due to the low sludge amount produced and the lack of chemical
reagents needed. However, most of the studies on As removal by
the EC process have been carried out using deionised water (model
water) [26]; consequently, their results can differ with respect to
those obtained with real water because a great number of species
are present in real water that interfere with the As removal process
[7,8]. Most published investigations related to As contaminated real
water treat waste waters [9,10], and there are only a few stud-
ies regarding underground water [11]. The lack of a systematic
approach to EC reactor design/operation has limited its implemen-
tation [12]; some studies have reported the disadvantages of this
technology, including the formation of an oxide layer on the elec-
trode surface [1214], the high energy consumption [1315] and
the electrode material loss [13,14]. The formation of the electrode
oxide layer has been solved by periodically changing the polarity of
the electrodes [3,12,13]; nevertheless, for reducing the energy and
Corresponding author.
E-mail address: mvjfer@gmail.com (J.F. Martnez-Villafane).
Several research reports conclude that the improvement in the As
removal in these studies is due to the increased current density
[2,3,6]. However, this causes a corresponding increase in energy
[16] and electrode consumption. Therefore, it is necessary to study
other design aspects of these reactors to reduce the treatment
times and avoid increasing the energy and electrode consump-
tions. The purpose of this work is to determine the effect of design
and operation factors of the EC process to reduce the energy and
electrode consumption by shortening the treatment time for As
electro-removal from underground water. We present a systematic
study of the factors that are generally known as the main variable
design and operation parameters that affect the EC process: cur-
rent density (i), interelectrode distance (d), electrode area treated
volume ratio (A/V) and liquid motion driving mode (LM).
2. Experimental
The underground water was collected from a well located at
the central part of northern Mexico, and its principle character-
istics are listed in Table 1. The water conductivity was measured
with a CORNING conductivity meter 441 using an electrode
(CAT476501). The analysis of the water included atomic absorp-
tion spectroscopy (AAS, Thermoelemental Instrument, Solaar S4)
for sodium, potassium and iron ions and inductively coupled
plasma spectrophotometry (ICP, Iris Intrepid II X SP) for calcium,
magnesium, silicon and sulphur ions. Potentiometry was used to
determine the F and Cl concentration with the ion selective
electrodes Orion 9609 and 9102, respectively. The initial As concen-

0304-3894/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jhazmat.2009.08.044
1618 J.F. Martnez-Villafane et al. / Journal of Hazardous Materials 172 (2009) 16171622

Table 1
Characteristics of the underground water.

Characteristics Value

pH 7.22
Conductivity 1067 (S cm1 )
Hardness 250.35 (mg L1 )
Ca 56.82 (mg L1 )
Mg 43.32 (mg L1 )
Na 106.48 (mg L1 )
K 20.28 (mg L1 )
SiO2 59.38 (mg L1 )
SO4 2 203.04 (mg L1 )
Cl 60.50 (mg L1 )
F 0.934 (mg L1 )

tration (CAs0 = 131 g L1 ) in the underground water (GW) as well

as the residual As concentration at every stage of sampling for the As
removal rate investigation were analysed by ICP. The samples were
ltered after the EC experiments with a lter paper with a pore size
of 2.5 m to separate the solids present in suspension in the water,
and then the samples were dried at 30 C for 48 h (Table 2). The
experimental metal dissolution value was determined from both
the residual Fe concentration in the GW and the amount of ferric
ocs retained in the lter paper when the GW was ltered after
the EC process. The amount of solids retained in the lter paper
was determined by the weight difference employing an electronic
analytical balance (Sartorius). The pH value of the GW was mea-
sured with a LPH430T (Tacussel) pH-meter before and after each
experiment. It was found that the initial GW pH value was 7.22, and
after each of the EC processes its pH value increased only slightly
by 0.07 0.01. According to Chaudhari et al. [2], the effect of pH on
arsenic removal is not signicant in the pH range of 68; therefore,
the experiments were carried out without any adjustment of pH.
The total cell voltage was measured with a uke multimeter during
the tests.
The EC experiments were carried out in a bipolar batch reactor Fig. 1. Schematic diagram of the EC reactor: (1) water level, (2) electrodes, (3) water
outlet, (4) air diffusers, (5) water pumps.
(Fig. 1) since it provides a simple setup: in this cell arrangement,
there are no connections between electrodes, which facilitates
operation during the EC process [13,15,17]. The reactor can han-
dle a volume of 1.2 L, and it is composed of parallel rectangular
plate electrodes of low carbon steel ASTM 1018 (from 3 to 6 in
3. Background of As removal by the EC process
number) with dimensions of 10 cm 14 cm 0.635 cm, vertically
oriented to promote the internal circulation in the reactor [12]. Pre-
The EC process provides in situ generation of coagulants. The
viously, the electrodes were polished with SiC grinding paper 80
ferric oxyhydroxides (FH) produced by the dissolution of the iron
grade to eliminate polluting agents such as oil, dust and oxide. The
anode operate as As adsorbents and complexing agents. The iron is
As electro-removal was carried out under galvanostatic conditions
dissolved at the anode according to reaction (1), and the water is
using a VIMAR electronics PG-2EV potentiostat/galvanostat.
reduced at the cathode (2). The FH species are obtained by reaction
The effect of four factors on the response variables were studied
(3), and the As electro-removal by ferric ocs is produced by an ion
in this work. The rst was d using three values: 3, 5 and 9 mm. The
exchange reaction (4) [6]:
second was LM, and four different driving modes were used to pro-
duce the GW motion: free convection (f), air injection (a), water Fe Fe2+ + 2e (1)
internal recirculation (w) and air injection-water recirculation
(aw). The water internal recirculation in the reactor was handled by 2H2 O + 2e H2 + 2OH (2)
two submersible multiphase pumps (Aqua 40) that provided a cir- Fe2+ + 3OH FeOOH(s) + H2 O + e (3)
culation ow rate of 2.5 L min1 for each one. The air injection was
controlled by an air pump (Azoo 9500) at a ow rate of 1.6 L min1 , 3FeOOH(s) + HAsO4 2
(FeO)3 AsO4(s) + H2 O + 2OH
(4)
and the discharge was connected to two porous membrane dif-
fusers located at the base of the reactor with insufated bubbles The dosage rate of FeOOH ocs in GW is proportional to the liber-
of 2 mm uniformly distributed along the entire interelectrode ation rate of Fe2+ ions xed by the current density.
gap. This provided a uniform species concentration distribution The energy (E) consumption by the EC process is calculated by
throughout the whole reactor. The third factor studied was the A/V the following equation:
ratio using four values0.350, 0.466, 0.583 and 0.700 cm1 where E = IUt (5)
the number of electrodes was varied from 3 to 6. The fourth factor
studied was i, which was varied between three levels: 1.50, 2.25 where U is the total cell voltage in the reactor, I is the applied cur-
and 3.00 mA cm2 . The response parameters were the treatment rent, and t is the treatment time. Additional energy consumption
time t, the energy consumption E and the electrode weight loss w. corresponding to the air injection (3 Wh) and water recirculation
 J.F. Martnez-Villafane et al. / Journal of Hazardous Materials 172 (2009) 16171622 1619

Table 2
Comparison of the calculated theoretical metal dissolution value wt with respect to the measured experimental value we .

Studied factor Run Operation conditions wt (g m3 ) we (g m3 )

d LM A/V i

1 3 aw 0.466 1.50 4.83 4.76

d 2 5 aw 0.466 1.50 7.25 6.94
3 9 aw 0.466 1.50 10.87 10.27

4 3 f 0.466 1.50 43.50 40.99

5 3 w 0.466 1.50 10.87 10.40
LM
6 3 a 0.466 1.50 9.66 9.42
7 3 aw 0.466 1.50 4.83 4.79

8 3 a 0.350 1.50 13.67 12.97

9 3 a 0.466 1.50 9.66 9.44
A/V
10 3 a 0.583 1.50 9.05 8.88
11 3 a 0.700 1.50 7.26 7.20

12 3 a 0.466 1.50 9.66 9.32

i 13 3 a 0.466 2.25 10.86 10.51
14 3 a 0.466 3.00 9.66 9.59

(6 Wh) devices is considered. The theoretical amount of iron liber-
ated from the anode is given by Faradays law:
ItM
w=
nF
where n is the number of electrons transferred (2) in the reaction, F
is Faradays constant (96,486 C mol1 ), M is the iron atomic weight
(55.85), and t is the treatment time in seconds. From Eqs. (5) and
(6) it is clear that the energy and electrode consumption increase
with longer treatment time and higher applied current.
4. Results and discussion
4.1. Effect of interelectrode distance (d)
The interelectrode distance was varied at values of d = 3, 5 and
9 mm while the rest of the factors remained unmodied at the
following values: LM = aw, A/V = 0.466 cm1 and i = 1.50 mA cm2 .
Fig. 2 shows the effect of d on the residual As concentration as
a function of the EC process time. As can be seen in Fig. 2, as
d increases the treatment time to achieve the MCL of arsenic
(10 g L1 ) increases. The treatment time values increased from 20
to 30 and 45 s when d increased from 3 to 5 and 9 mm. The better
t value obtained with the shortest d is attributed to an increase in
the mass transfer during the EC process as a result of the turbulence
generated. During the experiments with smaller d values, the tur-
bulence at the base and at the top of the electrodes was enhanced
for shorter d values. Hence, for larger d values the movement of
the participant species (including those in reactions (1)(4)) in the
As removal process is slower, the thickness of the concentration
(6) boundary layer in the interelectrode gaps is larger, and the reaction
rate decreases. This behaviour indicates the relative importance of
water motion and turbulence on the treatment time and their large
dependence of the interelectrode gap.
Fig. 3 illustrates the energy consumption as a function of the
charge density (Q) and the measured iron electrode weight loss
when d increases from 3 to 5 and 9 mm. As shown in this gure,
the E value increased almost linearly from 61 to 184 Wh m3 as the
electrode material loss increased from 4.76 to 10.27 g m3 when d
increased from 3 to 9 mm. This behaviour was associated with a
t increase corresponding to an increase in d, which, as mentioned
before, was due not only to the slower movement of the partici-
pant species in the As removal reaction but also the fact that the
distances traversed by the participant species are larger when d
is increased. Additionally, the interelectrode distance also has an
important effect on the cell voltage U according to Eq. (7):
U = EA + A EC + C + dielectrolyte (7)
where EA and EC are the reversible anodic and cathodic potentials,
respectively, and A and C are the corresponding overvoltages,
respectively. The last term is related to the electrolyte IR drop,
where  is the electrolyte resistivity. According to this equation the

Fig. 2. Effect of d on the residual As concentration in GW as a function of the EC Fig. 3. Effect of d on the energy consumption as a function of the charge density and
treatment time. The EC treatment was carried out at i = 1.5 mA cm2 , LM = aw and the electrode weight loss for As removal from GW by the EC treatment carried out
A/V = 0.466 cm1 . at i = 1.5 mA cm2 , LM = aw and A/V = 0.466 cm1 .
1620 J.F. Martnez-Villafane et al. / Journal of Hazardous Materials 172 (2009) 16171622
Fig. 4. Effect of LM produced by free convection (f), water internal recirculation
(w), air injection (a) and air injection-water recirculation (aw) on the residual As
concentration as a function of the EC treatment time. The EC process was carried
out at i = 1.5 mA cm2 , d = 3 mm and A/V = 0.466 cm1 .
increment in the cell voltage under these conditions is attributed
solely to the increase in d, since the electrode material, current
density and type of water are unmodied. The measured U val-
ues increased from 2.4 to 4.2 and 6.8 V when d was increased from
3 to 5 and 9 mm, respectively. In addition, according to Eq. (5), the
obtained U values had a signicant inuence on the energy con-
sumption. Therefore, smaller d values not only promote shorter t
but also reduce U. The reduction of both factors favours a lower
energy consumption and lower electrode weight loss, which also
allows to use GW with smaller amounts of solids in the EC process.
For these results, d = 3 mm was selected for the following experi-
ments.
4.2. Effect of liquid motion driving mode (LM)
The dispersion of the FH ocs and the mass transfer greatly
affect the rate of As removal and are linked to the liquid motion
that results from internal water recirculation (controlled by sub-
mersible multiphase pumps) or air injection (through the reactor
base into a diffuser). Fig. 4 shows a comparison of the residual
As concentration as a function of t using the different LM investi-
gated, water recirculation (w), air injection (a), air injection water
recirculation (aw) and free convection (f), while maintaining the
rest of the examined parameters: d = 3 mm, i = 1.50 mA cm2 and
A/V = 0.466 cm1 . As shown in Fig. 4, the slowest As removal rate
is obtained when the liquid motion is produced by free convec-
tion; here the treatment time to achieve the MCL of 10 g L1 is
180 s. This is attributed to the low dispersion of the FH ocs and
to a feasible electrode anodic passivation, which blocks the active
electrode surface area. When LM was produced by internal water
recirculation, the As removal rate increased greatly, and t was sig-
nicantly reduced (45 s) to a quarter of the time needed with free
convection to achieve the MCL. This behaviour occurs because the
liquid motion favours iron dissolution and OH ion transport from
the cathode towards the anode to form the FH ocs according to
reaction (3). The species produced at the anode form complexes
with the arsenic species present in the water, as indicated in reac-
tion (4). The liquid motion not only promotes the transport of the
soluble species of As towards the anode but also the transport of
the FH ocs towards the bulk of the GW, improving the As removal
rate. When LM was generated by air injection the treatment time
(40 s) was even shorter than that required for water recirculation.
This was attributed to a better species distribution along the reactor
with this LM than that produced by the water recirculation mode
due to the turbulence generated by the air bubbles. Nevertheless,
when both LM types are operating (aw), the treatment time (20 s)
Fig. 5. Effect of LM produced by free convection (f), water recirculation (w), air injec-
tion (a) and air injection-water recirculation (aw) on the energy consumption as a
function of the charge density and the electrode weight loss for As removal from GW
by the EC treatment performed at i = 1.5 mA cm2 , d = 3 mm and A/V = 0.466 cm1 .
is reduced by 50% with respect to that obtained when air is injected
into the GW. This behaviour is attributed to a better species distri-
bution along the entire reactor due to a major turbulence caused by
both LM driving modes, and consequently the As electro-removal
process is enhanced.
Fig. 5 illustrates the energy consumption as a function of
the charge density and electrode weight loss for each LM mode.
As shown in this gure, when LM is provided by air injection,
E is the lowest (56 Wh m3 ). This is slightly lower than the E
value (61 Wh m3 ) required when a combined aw motion mode
is supplied. However, a lower electrode consumption (4.79 g m3 )
is required with the aw motion mode compared with that
required with the air injection mode (9.42 g m3 ). When LM is
carried out exclusively by water recirculation, E increases signi-
cantly up to 100 Wh m3 ; nevertheless, the electrode consumption
(10.40 g m3 ) remains near the value obtained with the air LM
mode. When LM is produced by free convection, the energy con-
sumption increases signicantly by almost 100% with respect to
that required with air injection. The behaviour obtained with free
convection is attributed to the low mass transfer, which gives rise
to the lowest As removal process reactions. Moreover, with this
LM mode the electrode surface is susceptible to forming a passive
layer that promotes the increase of not only the treatment time but
also of the electrode material loss and the total cell voltage (2.84 V),
which give rise to signicant energy consumption. This is consis-
tent with the results reported by several authors [1214]. Actually,
the total cell voltage increases with LM in the order w (2.35 V), a
(2.47 V) and aw (2.52 V) as a consequence not only of the water tur-
bulence but also of the air bubble interference. The formation of the
passive layer is greatly prevented when the bulk liquid is in motion
and the FH ocs are suspended in the interelectrode space. This is
consistent with the results reported by Cocke et al. [13]. Neverthe-
less, when only the internal water circulation mode is provided,
the mass transfer is not as effective as when liquid circulation and
additional turbulence are produced by an air bubble supply. On
the other hand, when both air injection and water recirculation
are supplied, the liquid convection is more effective; however, an
improvement of only 10% is achieved for the energy consumption
reduction with respect to that required with air injection exclu-
sively. In addition, using only air injection when scaling this process
will diminish the reactor cost as well as the device investment and
maintenance compared to those required when the aw mode is
used. In this context, it should be noted that to produce LM solely
 J.F. Martnez-Villafane et al. / Journal of Hazardous Materials 172 (2009) 16171622
Fig. 6. Effect of the A/V ratio on the residual As concentration as a function of the EC
treatment time. The EC treatment was performed at i = 1.5 mA cm2 , d = 3 mm and
LM = a.
by air injection is sufcient to achieve efcient As removal; thus
this LM mode will be used for the following experiments.
4.3. Electrode areavolume ratio (A/V) effect
The effect of the ratio between the total electrode area and the
volume of water in the reactor (A/V) was studied in an interval from
0.350 to 0.700 cm1 by adding electrodes to the reactor from N = 3 to
N = 6, maintaining the GW volume (1.2 L) and increasing the applied
current from 0.630 to 1.255 A to maintain a constant current den-
sity of 1.50 mA cm2 . To determine the effect of this parameter on
the EC process response variables, the rest of the factors remained
unmodied: d = 3 mm, i = 1.50 mA cm2 and LM = a. Fig. 6 shows the
residual As concentration as a function of t for each A/V ratio used.
As shown in this gure, t is reduced progressively from 75 s to 40, 30
and 20 s when the A/V ratio is increased from 0.350 to 0.466, 0.583
and 0.700 cm1 , respectively. This is associated with the larger FH
dosage generated due to an increase in the A/V ratio; therefore,
higher As removal is produced. When A/V 0.350 cm1 , the arsenic
removal kinetics changes from the typical exponential decay to
a sigmoidal prole, which results in a substantial increase in the
treatment time.
Fig. 7 shows the energy consumption of the EC process as a
function of the charge density and electrode weight loss for each
A/V ratio investigated. As shown in this gure, the electrode con-
Fig. 7. Effect of the A/V ratio on the energy consumption as a function of the charge
density and the electrode weight loss for the As removal from underground water
by the EC process performed at i = 1.5 mA cm2 , d = 3 mm and LM = a.
1621
Fig. 8. Effect of i on the residual As concentration as a function of the EC treatment
time during As removal from underground water. The EC treatment was performed
at d = 3 mm, LM = a and A/V = 0.466 cm1 .
sumption increases signicantly as a consequence of decreases
in A/V. It increases from 7.20 to 12.97 g m3 when A/V decreases
from 0.700 to 0.350, respectively. The energy consumption remains
approximately constant (63 Wh m3 ) for A/V values of 0.700 and
0.583 cm1 and decreases to 55 Wh m3 when A/V is decreased to
0.466. However, for the lowest A/V value (0.350 cm1 ) the energy
and electrode consumptions increased abruptly to 81 Wh m3 and
12.97 g m3 , respectively, due to an increase in t as a consequence
of the smaller electrode area in the reactor. Moreover, an increase
in N not only promotes a decrease in the treatment time but also
increases the total cell voltage to values of 1.35, 2.41, 4.49 and 6.65 V
when N is increased progressively from 3 to 6. Consequently, E
increased according to Eq. (5). These results show that the addi-
tion of electrodes into the reactor does not assure the optimal
energy performance of the process. Therefore, the most favourable
A/V ratio, 0.466 cm1 (N = 4), was selected for the following exper-
iments.
4.4. Current density (i) effect
To determine the effect of i on t, E and w, experiments were
carried out with i values of 1.50, 2.25 and 3.00 mA cm2 , while
maintaining the rest of the investigated parameters: d = 3 mm,
A/V = 0.466 cm1 and LM = a. Fig. 8 shows the residual arsenic con-
centration in GW as a function of the treatment time as i was
varied. As shown in Fig. 8, when i increases, the arsenic removal
rate increases. This is consistent with the results reported by sev-
eral authors [2,3,6]. The treatment time decreases from 40, to 30
and 20 s when i is increased progressively from 1.50 to 2.25 and
3.00 mA cm2 , respectively. This was associated with the higher
dissolution of Fe2+ ions and the shorter time period for FH gen-
eration when i increased, with the consequent arsenic removal
increasing.
Fig. 9 illustrates the energy consumption as a function of the
charge density and the electrode weight loss for different values of
i. As shown in this gure, an intermediate i value of 2.25 mA cm2
caused the highest energy and electrode consumption; this is in
part due to the increase in U, since when i increases progres-
sively from 1.50 to 2.25 and 3.00 mA cm2 , U increases to values
of 2.41, 3.52 and 4.54 V, respectively, which is consistent with Eq.
(7). Additionally, t was not short enough to compensate for the
increment in the energy and electrode consumption with respect
to that required at i = 1.5 mA cm2 . However, the highest i value
of 3.00 mA cm2 produces a small increase of energy and elec-
trode consumption with respect to those obtained with the lowest i
value of 1.50 mA cm2 . This behaviour is due to a signicant reduc-
tion of 50% in t (20 s) achieved at i = 3.00 mA cm2 with respect to
1622 J.F. Martnez-Villafane et al. / Journal of Hazardous Materials 172 (2009) 16171622
Fig. 9. Effect of i on the energy consumption as a function of the charge density and
the electrode weight loss during the EC process for As removal from GW by the EC
process carried out at d = 3 mm, LM = a and A/V = 0.466 cm1 .
that obtained at i = 1.50 mA cm2 (40 s), which compensates for the
energy and the electrode consumption. These results show that
the best i value for As removal by the EC process in the interval
of current densities investigated in this work was 3.00 mA cm2 .
5. Conclusions
According to the results obtained in this work, the factor that
had the highest effect on most of the response variables was the
liquid motion driving mode, and the factor that had the smallest
effect in the response variables in all cases was the current den-
sity. The factors that generated a greater effect on the treatment
time in decreasing order are LM > A/V > d > i. The factors that sig-
nicantly affected the energy consumption follow the sequence
d > LM > A/V > i, while the factors that generated a greater effect on
the electrode weight loss in decreasing order are LM > d > A/V > i.
The results of this study show that low values of energy consump-
tion (58.78 Wh m3 ) and electrode material loss (9.59 g m3 ) as
well as the shortest residence time of 20 s were achieved during
the EC process for the arsenic removal from underground water
under the following conditions: d = 3 mm, LM = a, A/V = 0.466 cm1
and i = 3.00 mA cm2 .
Acknowledgements
The authors wish to acknowledge the nancial support to carry
out this project: 38467U from CONACyT.
References
[1] World Health Organization (WHO), Guidelines for drinking water quality, vol.
1, 2006, p. 306.
[2] S. Chaudhari, P.R. Kumar, K.C. Khilar, S.P. Mahajan, Removal of arsenic from
water by electrocoagulation, Chemosphere 55 (2004) 12451252.
[3] H.K. Hansen, P. Nunez, R. Grandon, Electrocoagulation as a remediation tool for
wastewaters containing arsenic, Miner. Eng. 19 (2006) 521524.
[4] H.K. Hansen, P. Nunez, D. Raboy, I. Schippacasse, R. Grandon, Electrocoagula-
tion in wastewater containing arsenic: comparing different process designs,
Electrochim. Acta 52 (2007) 34643470.
[5] D.L. Cocke, J.A. Gomes, M. Kesmez, Daida, M. Weir, H. Moreno, J.R. Parga, G.
Irwin, H. McWhinney, T. Grady, E. Peterson, Arsenic removal by electrocoagu-
lation using combined AlFe electrode system and characterization of products,
J. Hazard. Mater. B139 (2007) 220231.
[6] A. Maldonado-Reyes, C. Montero-Ocampo, O. Solorza-Feria, Remediation of
drinking water contaminated with arsenic by electro-removal process using
different metal electrodes, J. Environ. Monit. 9 (2007) 12411247.
[7] C.Y. Hu, S.L. Lo, W.H. Kuan, Effects of co-existing anions on uoride removal
in electrocoagulation (EC) process using aluminum electrodes, Water Res. 37
(2003) 45134523.
[8] X. Meng, S. Bang, G. Koratis, Effects of silicate, sulfate, and carbonate on arsenic
removal by ferric chloride, Water Res. 34 (1999) 12551261.
[9] J.R. Parga, D.L. Cocke, V. Valverde, J.A. Gomes, M. Kesmez, H. Moreno, M. Weir, D.
Mencer, Characterization of electrocoagulation for removal of chromium and
arsenic, Chem. Eng. Technol. 28 (2005) 605612.
[10] N. Balasubramanian, K. Madhavan, Arsenic removal from industrial efuent
through electrocoagulation, Chem. Eng. Technol. 24 (2001) 519521.
[11] J.R. Parga, D.L. Cocke, J.L. Valenzuela, J.A. Gomes, M. Kesmez, G. Irwin, H. Moreno,
M. Weir, Arsenic Removal via electrocoagulation from heavy metal contami-
nated groundwater in La Comarca Lagunera Mxico, J. Hazard. Mater. B124
(2005) 247254.
[12] P.K. Holt, G.W. Barton, C.A. Mitchell, The future for electrocoagulation
as a localised water treatment technology, Chemosphere 59 (2005) 355
367.
[13] D.L. Cocke, M.Y.A. Mollah, P. Morkovsky, J.A. Gomes, M. Kesmez, J.R. Parga, Fun-
damentals, present and future perspectives of electrocoagulation, J. Hazard.
Mater. B114 (2004) 199210.
[14] Y. Avsar, U. Kurt, T. Gonullu, Comparison of classical chemical and electrochem-
ical processes for treating rose processing wastewater, J. Hazard. Mater. B148
(2007) 340345.
[15] G. Chen, Electrochemical technologies in wastewater treatment, Sep. Purif.
Technol. 38 (2004) 1141.
[16] H. Ayedi, G. Mouedhen, M. Feki, M. Petris, Behavior of aluminium electrodes in
electrocoagulation process, J. Hazard. Mater. B150 (2008) 124135.
[17] W.A. Pretorius, W.G. Johannes, G.G. Lempert, Electrolytic iron occulant pro-
duction with a bipolar electrode in series arrangement, Water SA 17 (1990)
133138.