Journal of Hazardous Materials 235–236 (2012) 257–264

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials

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

Removal of natural organic matter and arsenic from water by

electrocoagulation/flotation continuous flow reactor
Emilijan Mohora ∗ , Srdjan Rončević, Božo Dalmacija, Jasmina Agbaba, Malcolm Watson,
Elvira Karlović, Milena Dalmacija
University of Novi Sad Faculty of Sciences, Department of Chemistry, Biochemistry and Environmental Protection, Trg D. Obradovića 3, 21000 Novi Sad, Serbia

h i g h l i g h t s

 A continuous electrocoagulation/flotation reactor was designed built and operated.

 Highest NOM removal according to UV254 was 77% relative to raw groundwater.
 Highest NOM removal accordance to DOC was 71%, relative to raw groundwater.
 Highest As removal archived was 85% (6.2 ␮g/l), relative to raw groundwater.
 Specific reactor energy and electrode consumption was 1.7 kWh/m3 and 66 g Al/m3 .

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

Article history: The performance of the laboratory scale electrocoagulation/flotation (ECF) reactor in removing high con-
Received 20 February 2012 centrations of natural organic matter (NOM) and arsenic from groundwater was analyzed in this study.
Received in revised form 25 July 2012 An ECF reactor with bipolar plate aluminum electrodes was operated in the horizontal continuous flow
Accepted 28 July 2012
mode. Electrochemical and flow variables were optimized to examine ECF reactor contaminants removal
Available online 7 August 2012
efficiency. The optimum conditions for the process were identified as groundwater initial pH 5, flow
rate = 4.3 l/h, inter electrode distance = 2.8 cm, current density = 5.78 mA/cm2 , A/V ratio = 0.248 cm−1 . The
Keywords:
NOM removal according to UV254 absorbance and dissolved organic matter (DOC) reached highest values
Electro-coagulation–flotation
Groundwater
of 77% and 71% respectively, relative to the raw groundwater. Arsenic removal was 85% (6.2 ␮g As/l) rel-
Arsenic removal ative to raw groundwater, satisfying the drinking water standards. The specific reactor electrical energy
Natural organic matter consumption was 17.5 kWh/kg Al. The specific aluminum electrode consumption was 66 g Al/m3 . Accord-
ing to the obtained results, ECF in horizontal continuous flow mode is an energy efficient process to
remove NOM and arsenic from groundwater.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction adsorbents and/or the formation of a complex with the target sub-
stances [2]. Recently, that presence of NOM in surface water was
Natural organic matter (NOM) in water is a major concern reported to negatively influence removal of As(V) by electrocoag-
and should be removed from drinking water from a number ulation/flotation (EFC) treatment [3].
of reasons. NOM affects organoleptic properties of water (color, Coagulation using metal based coagulants is the essential pro-
taste and odor); reacts with most disinfectants used in water cess in the conventional treatment of drinking water [4]. In this
treatment, thus reducing their disinfection power, influencing process coagulating agents and other additives are dosed to pro-
disinfectant demand and disinfection process design; produces dis- duce larger aggregates which can be separated physically. The
infection by-products (DBPs) of various kinds, affects the stability process depends on variety of factors, including but not limited
and removal of inorganic particles; heavily influences coagulant to, pH, coagulant type and dosage and the type and concentration
demand; affects corrosion processes; and affects biostability and of NOM. This is a multi stage process that requires considerable
biological regrowth in distributions systems [1]. NOM often reduces land area and continual supply of chemicals. A more cost-effective
the removal of target substances through its adsorption onto method to clean a wide range of polluted water on site and with
minimum additives is required for sustainable water management
[1]. ECF has been suggested as an alternative. It is the process
∗ Corresponding author. Tel.: +381 63 7599130; fax: +381 21 485 2728. whereby a sacrificial metal anode doses treated water electrochem-
E-mail address: [email protected] (E. Mohora). ically, while the cathode reaction provides afford hydrogen which

0304-3894/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jhazmat.2012.07.056
258 E. Mohora et al. / Journal of Hazardous Materials 235–236 (2012) 257–264

is involved in the flotation process. ECF is characteristic for its ease Table 1
Characteristics of the raw groundwater.
of operation, reduced production of sludge, the floc formed tends to
be much larger, less maintenance and no need for handling chem- Parameters No. of measurements (n) Value ± 
icals [5–7]. ECF has attracted great attention as an eco-friendly pH 10 7.4 ± 0.3
and efficient process, although it has some demerits, such as the Conductivity (␮S/cm) 10 1310 ± 36
need to replace electrodes at regular intervals due to their disso- DOC (mg C/l) 30 9.31 ± 0.51
lution, high cost of electricity, anode passivation and hydroxides UV254 (cm−1 ) 30 0.444 ± 0.026
SUVA (l mg−1 m−1 ) 30 4.64 ± 0.11
which may tend to solubilize in some cases [8]. Recent technical
Arsenic (␮g/l) 15 5.4 ± 16.4
improvements, combined with the growing need for small scale
 – standard deviation based on n measurements.
decentralized water and waste water treatment facilities, have led
to a re-evolution of the ECF [9]. A literature survey indicates that
most of the studies on NOM removal by ECF process have been car- 1730 SB, 0–30 V, 0–3 A), ampere-meter, voltmeter, and sampling
ried out in batch treatment mode using model water [10–13]. Only setup. The electrode dimensions were 14 cm × 13 cm × 0.3 cm and
modest attention has been paid to the ECF treatment of ground- aluminum purity was 99.5%. The distances between the plate elec-
water containing a high NOM concentration and in the continuous trodes varied from 1.2 cm to 2.8 cm, and were numbered in the
flow mode. reactor from 3 to 5.
Arsenic is classified as a Group I carcinogenic substances to In general, each experimental run lasted for 90 min. During the
humans based on epidemiological evidence [14]. The World Health experiments, raw groundwater from the water tank was pumped
organization (WHO) has established 10 ␮g/l as maximum arsenic into the EFC reactor by centrifugal pump. The reactor operated at
concentration for drinking water [15]. Arsenic in groundwater flow rates 4.3 l/h, 6.4 l/h and 10.3 l/h. In order to minimize elec-
occurs in two oxidation states As(III) (arsenite) and As(V) (arse- trode oxidation and passivation, the direction of the power supply
nate). As(III) is more mobile in groundwater and 25–60 times was changed at 90 min time intervals. Due to the geometry of the
more toxic than As(V). Various treatment technologies have been laboratory ECF reactor used, within the single unit coagulation,
applied to remove arsenic including coagulation with precipi- flocculation, flotation and settling occurred in parallel. The treated
tation with iron and aluminum salts, adsorption onto activated groundwater was collected after a desired period of time from the
alumina, activated carbon and activated bauxite, ion exchange and reactor effluent flow and filtered through Whatman No: 42 filter
reverse osmosis. Those technologies showed a medium to low paper before analyses. Effluent groundwater samples were ana-
As(III) removal efficiency and also required pH regulation as it influ- lyzed for NOM, Astotal , pH and temperature. Before and after each
ences arsenic speciation and surface change of adsorbents. Authors experimental run electrodes weight was measured.
thus recommended to oxidize As(III) to As(V) for effective arsenic
removal. However, in the presence of NOM the oxidation step
applied prior to As(III) removal could lead to the formation of toxic 2.3. Analytical methods and calculations
oxidation by products [16]. ECF is offering advantage to oxidize
As(III) in As(V) and improve its removal without required oxidation 2.3.1. NOM analyses
step [3]. Literature results on arsenic removal by ECF showed that UV254 absorbance measurements were performed before and
5 out of 7 studies were based on batch processes [2]. Most of them after ECF treatment in accordance with standard methods [20] by
have used model water and there are only 2 studies concerning UNICAM SP600 UV spectrophotometer at a wavelength of 254 nm
groundwater polluted by arsenic [17,18]. with a 1 cm quartz cell.
Therefore, the objective of the present study was to evaluate After filtration through a 0.45 ␮m membrane filter, the ground-
treatability of groundwater with a high concentration of NOM and water samples were analyzed before and after ECF treatment for
arsenic using an ECF reactor in horizontal continuous flow mode DOC content using an Elementar LiquiTOCII, with oxidation by com-
with bipolar plate aluminum electrodes. The effects of different bustion at 850 ◦ C.
ECF reactor operational parameters on NOM and As removal effi-
ciency like inter-electrode distance, A/V ratio, the initial pH of the 2.3.2. Arsenic analyses
raw groundwater, current density, electrode passivation and the Arsenic analyses were carried by Atomic Absorption Spec-
flow rate were studied, with the goal of optimizing reactor removal troscopy (PerkinElmer AAnalyst 700) according to the standard
efficiency rate for specific energy and electrode consumption. method EPA 7010 (Graphite Furnace Atomic Absorption Spec-
trophotometry, Rev. 0, Feb. 2007).
2. Materials and methods

2.1. Groundwater samples DC

All experiments were performed with groundwater samples V

collected in the Central Banat region (Republic of Serbia), where
A
groundwater is used as a principle water supply source. Its principle
characteristic is listed in Table 1. This groundwater contains unde- E
sirable components like a high concentration of NOM and arsenic P
ECF
as a consequence of its origin [19]. S

S
2.2. Experimental setup and procedure
T

All experiments were carried out in a bipolar laboratory ECF

continuous flow reactor (Fig. 1). It consists of 30 l water tank, cen-
Fig. 1. Diagram of the experiment: T – water tank, P – centrifugal pump, ECF – bipo-
trifugal pump, 4.5 l Plexiglas laboratory scale ECF reactor with a lar continuous flow electro-coagulation–flotation reactor, E – aluminum electrodes,
horizontal continuous flow, DC power supply source (model DF DC power supply, A – ampere-meter, V – voltmeter, S – sampling points.
 E. Mohora et al. / Journal of Hazardous Materials 235–236 (2012) 257–264 259

2.3.3. pH and conductivity

pH measurements were carried out on a WTW InoLab portable
instrument. Conductivity measurements were carried out on a
Hanna model HI 933000.

2.3.4. Calculations
NOM removal efficiency was calculated as:
C0 − Ct
 (%) = × 100 (1)
C0
where C0 – initial concentration, Ct – concentration after ECF treat-
ment time t. The ECF reactor electrical energy consumption (E) was
calculated using the commonly used equation:

E = UItECF (2)

where E is in kWh, U is the applied potential (in V), I is the operat-

ing current in ampere (A) and tECF is the ECF treatment time. The
amount of Al(III) released into the treated groundwater during ECF
treatment was estimated by applying Faraday’s law:
ItECF M
mt = (3)
nF
where mt the dissolved aluminum (g), n is the number of elec-
Fig. 2. Effect of A/V on NOM removal efficiency based on UV254 absorbance. The
trons in the redox reaction, F is Faraday’s constant and M is the ECF treatment was carried out at U = 20 V, Q = 4.3 l/h, d = 1.2 cm, pH 7.7, conductiv-
molecular weight of Al. Current efficiency was calculated as: ity = 1270 ␮S/cm, DOC0 = 9.31 ± 0.51 mg C/l.

me
CE = × 100 (4)
mt 3.2. Effect of initial groundwater pH
The calculation is based on the comparison of experimental
It has been reported that the initial pH of the treated water is
weight loss of the aluminum electrodes me during ECF treat-
important operating factor influencing NOM removal efficiency by
ment and the theoretical amount of aluminum dissolution mt
ECF [23]. The effect may be twofold, first on the distribution of
according to Faraday’s law. The specific electrical energy consump-
aluminum hydrolysis product and secondly, the effect of a passi-
tion (Seec ) is calculated as a function of aluminum electrode weight
vation layer formed especially in the high initial pH of the anode
consumption during ECF in kWh/kg Al [21,22]
surface [24,25]. To examine this effect, raw groundwater samples
nFU were adjusted to the desired pH by adding sodium hydroxide or
Seec = . (5)
MCE × 3.6 × 103 hydrochloric acid. The ECF treatment was carried out at constant
applied potential of 20 V, flow rate 3.6 l/h, interelectrode distance
3. Results and discussion 1.2 cm, and groundwater conductivity of 1284 ␮S/cm.
The results obtained (Fig. 3) indicate that at initial groundwa-
3.1. A/V ratio effect ter pH 5.0, the ECF reactor took approximately 10 min to achieve
a steady state based on UV254 absorbance and approximately
The ratio between the total active electrode area (A) and the 50 min based on DOC. The initial groundwater specific ultravio-
volume of groundwater in the reactor (V) is (crude) measure of the let absorbance (SUVA) at 254 nm value was 4.64 ± 0.11 l mg−1 m−1
potential for delivering coagulant and bubbles to the ECF system (Table 1). This indicates that a large part of the NOM present in the
(on a volumetric bases) [27]. It was reported that an optimal reac- raw groundwater consists of hydrophobic organic compounds with
tor operating current density exists for a fixed A/V ratio [21]. The high molecular masses and a high degree of aromaticity, which can
effect of A/V ratio on NOM removal efficiency was studied between be easily removed by coagulation [1,28–31]. The remainder of the
0.12 cm−1 and 0.31 cm−1 by adding electrodes to the ECF reac- DOC is harder to remove, and thus takes longer to reach a steady
tor from 3 to 5, and the groundwater volume was kept constant. state. The average value of NOM removal efficiency achieved based
The ECF treatment was carried out at the constant applied poten- on UV254 absorbance and DOC for the reactor steady state was 74%
tial of 20 V, flow rate 4.3 l/h, inter-electrode distance of 1.2 cm, and 66%, respectively. However, at initial groundwater pH of 7.0,
initial pH 7.7, and groundwater conductivity of 1270 ␮S/cm. The the EFC reactor took approximately 60 min to achieve a relatively
results obtained (Fig. 2) show that increasing the A/V ratio above steady state treatment regime. The NOM removal efficiency for
value of 0.248 cm−1 or addition of more than four electrodes into the reactor steady state regime time based on UV254 absorbance
the ECF reactor does not improve NOM removal efficiency. This and DOC was 63% and 54% respectively. These results confirm
is associated with the considerable decrease of the operating cur- previously reported studies performed with the model waters con-
rent density as a consequence of A/V ratio increase for the constant taining high NOM concentration and in batch treatment mode, that
potential applied during experimental run. For A/V ratio value the ECF process is more efficient when the initial treated water pH
0.248 cm−1 the ECF reactor in its steady state achieved the high- is adjusted to 5 [10–13]. Therefore, for the following experimental
est NOM removal efficiency of 60% (relative to raw groundwater) work the pH value of the raw groundwater was adjusted to 5.
based on UV254 absorbance.
The obtained results demonstrate that additional electrodes in 3.3. Effect of inter electrode distance (d)
the ECF reactor for the constant applied potential do not assure a
steady increase in the NOM removal efficiency. Therefore, the most NOM removal efficiency and energy consumption was analyzed
favorable A/V ratio of 0.248 cm−1 , or four aluminum electrodes in as a function of inter electrode distance (d). The inter electrode
the ECF reactor, was selected for the following experimental work. distance varied at values of d = 1.2 cm, 2.0 cm and 2.8 cm. The ECF
260 E. Mohora et al. / Journal of Hazardous Materials 235–236 (2012) 257–264

a
a

b
b

Fig. 3. Variation of DOC (a) and UV254 absorbance (b) as a function of ini-
tial groundwater pH value and time. The ECF treatment was carried out at
c
U = 20 V, Q = 3.6 l/h, d = 1.2 cm, conductivity (pH = 7.0) = 1284 ␮S/cm, conductivity
(pH = 5.0) = 1750 ␮S/cm, A/V = 0.248 cm−1 , DOC0 = 9.31 ± 0.51 mg C/l.

treatment was carried out at constant applied potential of 17 V,

flow rate 3.6 l/h, initial groundwater pH 5, and conductivity of
1750 ␮S/cm.
In general, a closer inter electrode distance is desirable, because
electrical resistance (IR drop) increases with inter-electrode dis-
tance increase [18,26]. Results obtained show that ECF reactor
took approximately 10 min (Fig. 4a) to achieve steady state based
on UV254 absorbance and 30 min according to DOC (Fig. 4b). As
explained above (Section 3.2), the organic carbon responsible for
UV254 absorbance is more readily removed than the rest of the
DOC. The highest NOM removal efficiency (Fig. 4a and b) and lowest
specific energy consumption (Fig. 4c) unexpectedly was achieved
for the highest inter electrode distance studied (2.8 cm). This is in
contrast with the results Feng et al. obtained [12], however Yildiz
et al. reported highest NOM removal efficiency for the inter elec-
trode distance of 5 cm [10,11]. The NOM removal efficiency based
on UV254 absorbance and DOC was 77% and 69%, respectively. The Fig. 4. NOM removal efficiency based on UV254 absorbance (a), DOC (b) and specific
result can be explained that flotation of hydrogen bubbles pro- energy consumption (c) as a function of ECF treatment time and inter electrode dis-
tance. The ECF treatment was carried out at U = 17 V, Q = 3.6 l/h, initial groundwater
duced at the cathode as NOM removal path was more stabile and
pH 5.0, conductivity = 1750 ␮S/cm, A/V = 0.248 cm−1 ,DOC0 = 9.31 ± 0.51 mg C/l.
flocks were easy to be floated when the inter electrode distance
was 2.8 cm. Interelectrode distances 1.2 cm and 2 cm were closed to
 E. Mohora et al. / Journal of Hazardous Materials 235–236 (2012) 257–264 261

b Fig. 6. Variation of treated groundwater temperature as a function of the operat-

ing current density and ECF treatment time. The ECF treatment was carried out
at Q = 6.4 l/h, d = 2.8 cm, A/V = 0.248 cm−1 , initial groundwater pH 5.0, conductiv-
ity = 1750 ␮S/cm, DOC0 = 9.31 ± 0.51 mg C/l.

by electric charges. Therefore if additional electrical charge is

supplied to the charged particles via ECF reactor’s electrodes,
the surface charge of NOM molecules is neutralized and several
molecules combine into larger and separable agglomerates. The
increase of the operating current density caused a decrease in reac-
tor DOC removal efficiency. It is a known fact that operating current
density in ECF processes directly determines the coagulant dosage
and the rate of bubble generation, which influences both mixing
and mass transfer in the reactor. It is reported that at low operating
currents settling dominates, and as the current increases, the pol-
lutant fraction that is removed by flotation increases, although the
coagulant appears to be used less effectively [6]. Therefore at higher
operating current densities, bubble densities increase, resulting in a
greater upwards momentum flux and thus faster removal of NOM
Fig. 5. Effect of current density on the performance of the ECF reactor based on and coagulant by flotation from the active reactor volume to the
UV254 absorbance (a) and DOC (b) as a function of treatment time. The ECF treatment surface. Finally we could conclude that for higher operating current
was carried out at Q = 6.4 l/h, d = 2.8 cm, A/V = 0.248 cm−1 , initial groundwater pH 5.0, densities more aluminum was available per unit of time in the ECF
conductivity = 1750 ␮S/cm, DOC0 = 9.31 ± 0.51 mg C/l.
reactor volume but its residence time in the active reactor volume
was shorter, which caused the decline of NOM removal efficiency.
obstruct adequate mass transport in the reactor reducing its NOM The influence of the used operating current density on the varia-
removal efficiency rate and creating instability of the removal effi- tion of effluent groundwater pH value and temperature was studied
ciency over the treatment time (Fig. 4a). Thus, d = 2.8 cm as a most too (Figs. 6 and 7). For operating current densities of 5.71 mA/cm2 ,
favorable was selected for the following experiments. 5.78 mA/cm2 and 11.80 mA/cm2 over 90 min of ECF treatment, the
pH value of the effluent groundwater increased relative to the raw
3.4. Current density effect (i) on NOM and arsenic removal groundwater by 6.9%, 11%, 31%, and the effluent groundwater tem-
perature increased by 13.6% 12% and 34%, respectively. Initial water
In any EFC process the current density is the most impor- pH value affects either distribution of aluminum hydrolyses prod-
tant operational parameter setting the final removal and defining ucts or structure of NOM directly [24,25]. When pH value is higher,
the specific energy and electrode consumption and so eventu- NOM takes more negative charge, and process needs more Al3+ to
ally the ultimate operating cost for the process [23]. The ECF neutralize the negative charge reducing removal efficiency [12].
reactor’s NOM and arsenic removal was studied by conducting Therefore, it is evident that ECF process under higher operating
experiments at current density values: 5.71 mA/cm2 , 5.78 mA/cm2 , current density boost water pH increase over the treatment time
11.80 mA/cm2 while flow rate, inter electrode distance, number of reducing NOM removal efficiency.
electrodes and initial pH were kept constant (Q = 6.4 l/h, d = 2.8 cm, The highest reactor NOM removal efficiency obtained accord-
A/V = 0.248 cm−1 , initial groundwater pH 5.0). ing to UV254 absorbance and DOC was 77% and 71% respectively,
Obtained results (Fig. 5) show that ECF reactor took approx- and was achieved for a current density of 5.78 mA/cm2 . The
imately 10 min to achieve steady state according to UV254 and specific reactor energy and electrode consumption for operating
30 min according to DOC. NOM in water are large organic molecules current density of 5.78 mA/cm2 were 2.01 kWh/m3 , 80 g Al/m3 ,
that carry a negative charge giving them colloidal characteristics. respectively. In addition, the specific reactor electrical energy con-
ECF is based on the fact that the stability of the colloids is influenced sumption (Seec ) was 17.5 kWh/kg Al.
262 E. Mohora et al. / Journal of Hazardous Materials 235–236 (2012) 257–264

Fig. 7. Variation of treated groundwater pH as a function of the operat- Fig. 8. Effect of flow rate on ECF reactor NOM removal efficiency based on
ing current density and ECF treatment time. The ECF treatment was carried UV254 absorbance as a function of treatment time. The ECF treatment was carried
out at Q = 6.4 l/h, d = 2.8 cm, A/V = 0.248 cm−1 , pH 5.0, conductivity = 1750 ␮S/cm, out at d = 2.8 cm, A/V = 0.248 cm−1 , pH 5.0, U = 17 V, i = 5.78 mA/cm2 , conductiv-
DOC0 = 9.31 ± 0.51 mg C/l. ity = 1750 ␮S/cm, DOC0 = 9.31 mg C/l.

The arsenic removal efficiency achieved for operating current

rate of 4.3 l/h in the effluent groundwater sampled after 90 min of
density of 5.78 mA/cm2 was 80% relative to the raw groundwater
treatment.
(residual Astotal concentration in the effluent groundwater sam-
pled after 90 min of treatment was 9.2 ␮g/l). Mechanisms of arsenic
removal by EC involve oxidation of As(III) in As(V) and subsequent 3.6. Effect of passivation
surface complexation with aluminum hydroxides [3].
This study evidently conformed that operating current density One of the greatest operational issues with ECF is electrode pas-
is the key ECF reactor operational parameter defining coagulant sivation. This passive layer increases the applied potential and leads
dose, variation of treated water pH and temperature over treatment to a waste of energy in ECF processes [1]. It is reported that 120 mg/l
time, influencing the dominant NOM separation mode and finally is the critical NOM concentration in the treated water at which the
defining the NOM and As removal efficiency rate. Thus, the follow- passive layer starts to form, from a point 5 min into treatment, on
ing experimental work was carried out using a current density of the surface of the anode, preventing delivery of enough coagulant
5.78 mA/cm2 . [11]. The variation of the operating current as a function of applied
potential and treatment time has been analyzed.
The results obtained at 20 V (Fig. 10) show a significant decrease
3.5. Effect of flow rate on NOM and arsenic removal
of 20% in the operating current, relative to the initial value, during
the 90 min of ECF treatment. This is attributed to aluminum anode
The effect of flow rate on ECF reactor NOM and arsenic removal
passivation. For voltages of 17 and 15 V, the current value change
efficiency was analyzed by conducting experiments at flow rate val-
ues: 4.3 l/h, 6.4 l/h, 10.3 l/h, or residence times 62.8 min, 42.1 min,
26.3 min, while the other selected parameters were kept constant
(d = 2.8 cm, A/V = 0.248 cm−1 , initial groundwater pH 5.0, U = 17 V,
i = 5.78 mA/cm2 , conductivity = 1750 ␮S/cm).
The flow rate does not appear to have a large effect on NOM
removal by ECF reactor (Fig. 8) and after 15–90 min of treatment,
the average UV254 absorbance reductions were 72–77%, depending
on the flow rate values (4.3–10.3 l/h). The best NOM removal (77%)
results were obtained with a flow rate of 4.3 l/h.
The effect of the flow rate on effluent groundwater pH and tem-
perature was also analyzed. In general, increase of the reactor flow
rate slows treated groundwater temperature and pH increase over
the treatment time. For the reactor flow rates of 4.3 l/h, 6.4 l/h,
10.3 l/h, treated water temperature has increased by 13%, 8.9% and
6%, relative to the raw groundwater temperature, for 90 min of
the treatment. In parallel, effluent water pH has increased by 17%
10% and 9%, relative to the initial groundwater pH value that was
adjusted to 5.
The impact of flow rate on Astotal removal was studied (Fig. 9).
The increase of the flow rate reduces reactor’s Astotal removal effi-
ciency. For flow rates of 4.3 l/h, 6.4 l/h and 10.3 l/h, the reduction in Fig. 9. Astotal concentration as a function of ECF reactor flow rate. The ECF treat-
arsenic content was 85%, 77% and 75%, respectively. A minimum ment was carried out at d = 2.8 cm, A/V = 0.248 cm−1 , pH 5.0, U = 17 V, i = 5.78 mA/cm2 ,
Astotal residual concentration of 5.3 ␮g/l was obtained at a flow conductivity = 1750 ␮S/cm, DOC0 = 9.31 ± 0.51 mg C/l.
 E. Mohora et al. / Journal of Hazardous Materials 235–236 (2012) 257–264 263

applied potential 17 V, inter electrode distance 2.8 cm, flow rate

4.3 l/h, A/V ratio 0.248 cm−1 , initial groundwater pH value 5 and
60 mg/l Cl− was present in the treated groundwater. The spe-
cific reactor energy and electrode consumption were 1.7 kWh/m3
and 66 g Al/m3 , respectively. The specific reactor electrical energy
consumption (Seec ) was 17.5 kWh/kg Al. ECF reactor took approx-
imately 10 min to achieve steady state according to UV254 and
30 min according to DOC.
This study showed that ECF treatment is a promising purification
technology for the treatment of groundwater containing high con-
centrations of NOM and arsenic. Outputs of this study will be helpful
for further application in designing economically and environmen-
tally sustainable ECF continuous flow reactors for the treatment of
raw groundwater polluted with NOM and As for small settlements.

Acknowledgements

This research was financed by the Ministry of Education and

Science of the Republic of Serbia (Project No. III43005) and the
Provincial Secretariat for Science and Technological Development
Fig. 10. Variation of the operating current as a function of applied poten- of the Government of Autonomous Province of Vojvodina, Republic
tial and treatment time. The ECF treatment was carried out at Q = 6.4 l/h, of Serbia (Project No. APV 114-451-1985)
d = 1.2 cm, A/V = 0.248 cm−1 , initial groundwater pH 5.0, conductivity = 1750 ␮S/cm,
DOC0 = 9.31 ± 0.51 mg C/l.
References
over 90 min of treatment was insignificant, decreasing by only 4% [1] D. Ghernaout, B. Ghernaout, A. Kellil, Natural organic matter and enhanced
relative to the initial value. coagulation as a link between coagulation and electrocoagulation, Desalin.
A recent study indicated that the electrolyte should contain Water Treat. 2 (2009) 2003–2222.
[2] Y. Seida, Y. Nakano, Removal of humic substances by layered double hydroxide
a minimum of 60 mg/l Cl− to breakdown the anodic passivation
containing iron, Water Res. 5 (34) (2000) 1487–1494.
layer during ECF treatment [25]. In order to explore passivation, [3] V. Pallier, G.F. Cathalifaud, B. Serpaud, Influence of organic matter on arsenic
60 mg/l Cl− (as NaCl) was added to the raw groundwater. The NOM removal by continuous flow electrocoagulation treatment of weakly mineral-
ized waters, Chemosphere 8 (2011) 21–28.
removal achieved according to UV254 absorbance was not signif-
[4] M.M. Emamjomeh, M. Sivakumar, Review of pollutants removed by electro-
icantly improved compared to the results for treatment without coagulation and electrocoagulation/flotation processes, J. Environ. Manage. 90
Cl− addition. However, the operating current did not vary over the (2009) 1663–1679.
90 min of treatment, indicating the breakdown of the anodic passi- [5] C. Riocordel, A. Darchen, D. Hadjiev, Electrocoagulation–electroflotation as a
surface water treatment for industrial uses, Sep. Purif. Technol. 74 (2010)
vation layer, which was more important. Arsenic removal efficiency 342–347.
was 85% (residual Astotal concentration obtained was 6.2 ␮g/l). The [6] P.K. Holt, G.W. Barton, M. Wark, C.A. Mitchell, A quantitative comparison
presence of 60 mg/l Cl− in the treated groundwater reduced reac- between chemical dosing and electrocoagulation, Colloids Surf. A: Physic-
ochem. Eng. Asp. 211 (2002) 233–248.
tor specific energy consumption by 15% (1.7 kWh/m3 ) and specific [7] M.Y.A. Mollah, P. Morkovsky, J.A.G. Gomes, M. Kesmez, J. Parga, D.L. Cocke,
aluminum consumption by 17% (66 g Al/m3 ), compared with the Fundamentals, present and future perspectives of electrocoagulation, J. Hazard.
results of groundwater treatment without addition of supporting Mater. 114 (2004) 199–210.
[8] N.S. Kumar, S. Goel, Factors influencing arsenic and nitrate removal from drink-
electrolyte in the same reactor operational conditions. ing water in a continuous flow electrocoagulation process, J. Hazard. Mater. 173
(2010) 528–533.
[9] P.K. Holt, G.W. Barton, C.A. Mitchell, The future for electrocoagulation as a
4. Conclusions
localised water treatment technology, Chemosphere 59 (2005) 355–367.
[10] Y.S. Yildiz, A.S. Koparal, S. Irdemez, B. Keskinler, Electrocoagulation of synthet-
In this study NOM and arsenic removal was analyzed using the ically prepared waters containing high concentration of NOM using iron cast
laboratory scale continuous flow ECF reactor with bipolar plate electrodes, J. Hazard. Mater. B139 (2007) 373–380.
[11] Y.S. Yildiz, A.S. Koparal, B. Keskinel, Effect of initial pH and supporting elec-
aluminum electrodes. The effects of the reactor’s various opera- trolyte on the treatment of water containing high concentration of humic
tional parameters on NOM and arsenic removal efficiency as well substances by electrocoagulation, Chem. Eng. J. 138 (2008) 63–72.
as energy efficiency and electrode consumption were studied. [12] Q.Y. Feng, X. Li, Y. Cheng, Z. Lai, Removal of humic acid from groundwater by
electrocoagulation, J. China Univ. Mining Technol. 17 (2007) 520, 513-515.
It was found that adding more than four electrodes to the ECF [13] M. Vespsalainen, M. Ghiasvand, J. selin, J. Pienimaa, E. Repo, M. Pulliainen, M.
reactor resulted in a decline in NOM removal efficiency. NOM Sillanpaa, Investigation of the effects of temperature and initial sample pH on
removal is more efficient when the pH of the groundwater was natural organic matter (NOM) removal with electrocoagulation using response
surface methods (RSM), Sep. Purif. Technol. 69 (2009) 255–326.
adjusted to 5. Increase of the operating current density caused a [14] IARC (International Agency for Research on Cancer), IARC Monographs on the
slight decrease in NOM removal efficiency. Effect from flow rate Evaluation of Carcinogenic Risk to Humans: Overall Evaluations of Carcino-
on NOM removal efficiency was not observable, however, increase genicity: An Updating of IARC Monographs, vol. 1–42 (7), International Agency
for Research on Cancer, Lyon, France, 1987, pp. 100–206.
the flow rate was found to decrease Astotal removal efficiency. A
[15] World Health Organization, Guidelines for drinking water quality, vol.1, 2006,
concentration of 60 mg/l Cl− in treated groundwater containing pp. 306.
9.31 ± 0.51 mg C/l of NOM was sufficient for anode passivation layer [16] P.R. Kumar, S. Sanjeev, C.K. Khilar, S.P. Mahajan, Removal of arsenic from water
by electrocoagulation, Chemosphere 5 (9) (2004) 1245–1252.
breakdown.
[17] J.R. Paraga, D.L. Cocke, J.K. Valenzuela, J.A. Gomez, M. Kesmez, G. Irwin, H.
The highest achieved NOM removal according to UV254 Moreno, M. Weir, Arsenic removal via electrocoagulation from heavy metal
absorbance and DOC was 77% and 71%, respectively, (relative to raw contaminated groundwater in La Magunera, Mexico, J. Hazard. Mater. B124
groundwater) as well as arsenic removal efficiency of 85% (6.2 ␮g (2005) 247–254.
[18] J.F. Martinez-Villafane, C.M. Ocampo, A.M. Garcia-Lara, Energy and electrode
As/l), meeting the drinking water standards of 10 ␮g/l. The ECF reac- consumption analyses of electrocoagulation for the removal of arsenic from
tor operational parameters were: current density 5.78 mA/cm2 , underground water, J. Hazard. Mater. 172 (2009) 1617–1622.
264 E. Mohora et al. / Journal of Hazardous Materials 235–236 (2012) 257–264

[19] A. Tubic, J. Agbaba, B. Dalmacija, I.I. Tumbas, M. Dalmacija, Removal of arsenic [26] C. Phalakornkule, S. Polgumhang, W. Tongdaung, Performance of an elec-
and natural organic matter from groundwater using ferric and alum salts: a case trocoagulation process in treating direct dye: batch and continuous upflow
study of central Banat region (Serbia), J. Environ. Sci. Health Part A 45 (2010) processes, World Acad. Sci. Eng. Technol. 57 (2009) 277–282.
363–369. [27] P.K. Holt, Electrocoagulation: unravelling and synthesizing the mechanisms
[20] AWWA–APHA–WEF Standard Methods for the Examination of Water and behind a water treatment process, PhD thesis, Faculty of Engineering, The
Wastewater, 20th edition, American Public Health Association/American University of Sydney, 2003.
Water Works Association/Water Environment Federation, Washington, DC, [28] T. Karanfil, M.A. Schlautoman, I. Erdogan, Survey of DOC and UV measurement
1998. practices with implications for SUVA determination, JAWWA 94 (12) (2002)
[21] N. Mameri, A.R. Yeddou, H. Lounici, D. Belhocine, H. Grib, B. Bariou, Defluo- 68.
ridation of septentrional Sahara water of North Africa by electrocoagulation [29] J.J. Molnar, J.R. Agbaba, B.D. Dalmacija, T.M. Klašnja, M.B. Dalmacija, M.M.
process using bipolar aluminium electrodes, Water Res. 32 (1998) 1604–1612. Kragulj, A comparative study of the effects of ozonation and TiO2 -catalyzed
[22] D. Ghosh, C.R. Medhi, H. Solonki, M.K. Purkait, Decolorization of cri stal violet ozonation on the selected chlorine disinfection by-product precursor content
Solution by electrocoagulation, J. Environ. Prot. Sci. 2 (2008) 25–35. and structure, Sci. Total Environ. 425 (2012) 169–175.
[23] G. Chen, X.M. Chen, P.L. Yue, Electrocoagulation and electroflotation of restau- [30] J.K. Edzvald, J.E. Tobiason, Enhanced coagulation: US requirements and a
rant wastewater, J. Environ. Eng. 126 (2000) 858–863. broader view, Water Sci. Technol. 4 (1999) 47–54.
[24] A.S. Koparal, Y.S. Yildiz, B. Keskinel, N. Demircioglu, Effect of initial pH on [31] N. Her, G. Amy, D. Foss, J. Cho, Y. Yoon, P. Kosenka, Optimization of method
the removal of humic substances from wastewater by electrocoagulation, Sep. for detecting and characterizing NOM by HPLC-size exclusion chromatogra-
Purif. Technol. 59 (2008) 175–182. phy with UV and on-line DOC detection, Environ. Sci. Technol. 36 (2002)
[25] G. Mouedhen, M. Feki, M.D.P. Wery, H.F. Ayedi, Behaviour of aluminium elec- 1069–1076.
trodes in electrocoagulation process, J. Hazard. Mater. 150 (2008) 124–135.