Electrochimica Acta 55 (2010) 8165–8171

Contents lists available at ScienceDirect

Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta

Industrial wastewaters treated by electrocoagulation

Yunny Meas ∗,1 , José A. Ramirez, Mario A. Villalon, Thomas W. Chapman 1
CIDETEQ, Centro de Investigación y Desarrollo tecnológico en Electroquímica, Parque Tecnológico Querétaro, Sanfandilla, 76703 Pedro Escobedo, Edo. Queretaro, Mexico

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

Article history: In this paper, we describe a novel design of electrocoagulation system for wastewater treatment using
Received 29 October 2009 sacrificial aluminum electrodes. The wastewaters considered were contaminated by a fluorescent pen-
Received in revised form 24 April 2010 etrant liquid used for non-destructive testing of parts in the aircraft industry and could not be treated
Accepted 2 May 2010
satisfactorily by conventional physical, chemical or biological processes. The new electrochemical system
Available online 11 May 2010
clarified the contaminated rinse water and allowed it to be re-used 4 times. The system could reduce 95%
of the chemical oxygen demand, 99% of color and 99% of turbidity in the discharged wastewater. Cost
Keywords:
estimates for the process represent an investment-return time for the electrocoagulation system of about
Electrocoagulation
Industrial
17 weeks.
Water © 2010 Elsevier Ltd. All rights reserved.
Treatment
Recycling

1. Introduction investigated, and the relative advantages of particular arrange-

ments have been reported [13–17].
Various electrochemical treatment methods for wastewaters This work reports results obtained with a new design of elec-
have been reported to reduce the use of chemicals and to remove trocoagulation reactor in terms of its efficiency and its successful
more persistent pollutants relative to traditional physical and integration into a particular industrial wastewater treatment sys-
chemical treatments. Nonetheless, the energy consumption of tem. Among the advantages of this electrocoagulation wastewater
these processes was generally excessive and represented a disin- process are its compact size, low sludge generation, high water
centive for their introduction in the industrial sector. The trade-off recovery and easy operation. Performance parameters are reported
of increased treatment efficiency versus reduction in energy con- for a laboratory-scale reactor, a pilot-plant prototype and an
sumption has therefore become a key issue in the development of integrated industrial installation. Economic and technical results
the field. obtained with the industrial system are reported for a wastewater
The process of electrocoagulation can be applied to a wide range containing a fluorescent penetrant liquid that could not be treated
of wastewaters and is especially effective in removing colloidal satisfactorily by other biological, physical and chemical processes,
inorganic contaminants and pathogens. Successful removal of flu- and the performance is compared with reports from the available
oride and organic compounds has also been reported [1–7]. The literature [18–20].
technique consists of in situ generation of coagulants by the elec-
trodissolution of a sacrificial anode, usually aluminum or iron [8,9],
instead of direct addition of chemical coagulants [10]. Proper design 1.1. Principles of electrocoagulation
of the electrochemical reactor promotes the destabilization of the
suspended pollutants, the formation of larger solid particles by The electrocoagulation process comprises the in situ genera-
aggregation and their efficient separation by sedimentation and/or tion of coagulant by the electrodissolution of a sacrificial anode,
flotation [11], thus exploiting the same mechanisms involved with usually aluminum or iron. This introduces metal ions that are
chemical coagulants [12]. Various electrocoagulation reactors and effective in the destabilization of suspended pollutants. Also, in
integrated systems have been designed for a variety of wastewater the proper pH range, generation of hydroxide species polymer-
applications. Different designs and operating conditions have been izes aluminum cations into Aln (OH)3n that can entrap or adsorb
the pollutants. Finally, the separation of these solids occurs by
sedimentation or by flotation. It is possible to take advantage of
the hydrogen gas generated on the cathode for the latter pur-
∗ Corresponding author. Tel.: +52 442 211 6070; fax: +52 442 211 6042. pose as well as for agitation. Electrocoagulation principles can be
E-mail address: [email protected] (Y. Meas). found in electrochemical technologies applications to wastewaters
1
ISE member. treatment [21].

0013-4686/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2010.05.018
8166 Y. Meas et al. / Electrochimica Acta 55 (2010) 8165–8171

The reactions occurring in an electrocoagulation cell using alu- plaque with electroplating tape (3 M 484 – 3 M). A 25-A, 12-V Novak
minum anodes can be summarized as follows: DC-power supply (112 VAC, 60 Hz) was used to apply the current.
COD was measured with a Hach digester (115 V, 2 A, 100–115 ◦ C,
On the aluminum anode : Al → Al3+ + 3e− P/N 45600-00) using digesting solutions of 0–1500 mg l−1 and of
0–150 mg l−1 strength from Hach Company. Color and turbidity
On the cathode : 2H2 O + 2e− → H2 ↑ + 2OH− were measured with a Hach spectrophotometer (Model DR/4000,
115 V) and are reported in Pt–Co units. Conductivity and pH were
In solution between the electrodes : Al3+ + 3OH− → Al(OH)3 ↓ measured with an Oakton meter (Model CW-35610). Experimen-
tal operating conditions were set at a current density of 333 A m−2
The resulting hydroxide precipitates can agglomerate with pol- and at a solution conductivity of 500 ␮S cm−1 , with conductivity
lutants in the water [22], thus making them easier to capture. adjusted by addition of sodium chloride (NaCl, Fermont, 99.5%).
Better understanding of the fundamental mechanisms would pro- Laboratory samples were taken with a 10-ml pipette 1 min after
vide a basis for improved design and operation [12]. Reports of cessation of current and stirring at the end of the test. It was found
optimization through statistical analysis have recently suggested that samples taken 1 h after the end of the test, after sedimentation
quantitative models to guide process design [23]. of foam, gave the same results.
To identify general guidelines for the technological devel- Industrial samples of 20 l were taken from the process outlets
opment and implementation of electrocoagulation in industry, at the points of interest. The samples were preserved in ice and
various factors were investigated. The first parameter to select sent to the laboratory in less than 60 min. In the laboratory the
when exploring process feasibility with a given wastewater is the samples were treated in the same manner as the laboratory samples
appropriate material to be used for electrodes. This selection will described above.
have an impact on the reactor design, energy consumption, pol-
lutant removal efficiency, and the amount of sludge generated. 2.1.1. Pilot-plant prototype description
Another principal factor to be considered is the stirring intensity This work sought to accomplish these goals of reducing COD,
and the hydrodynamics in the reaction cell, which are related to color and turbidity of the specific wastewater containing the fluo-
two key phenomena: coagulation and flocculation. Coagulation is rescent penetrant oil for its re-use or safe discharge by applying the
enhanced by strong stirring that produces enough turbulence to electrocoagulation unit shown in Fig. 1. Typical operational param-
bring the hydroxide and metal ions into intimate contact with the eters have been reported elsewhere for various electrocoagulation
pollutant. On the contrary, flocculation is facilitated by mild stirring reactors used in the removal of different contaminants [25–27].
which allows the action of Van der Waals forces to attract pollutants A transparent scale-model prototype of a continuous treat-
to the growing flocs. An optimal design should ensure that each of ment process was constructed, made of acrylic material with an
these phenomena is allowed to occur. electrochemical-cell reactor volume of 3.75 l coupled directly to a
clarification chamber. The required peripheral pump and power
1.2. Industrial application supply were provided. This pilot system had the capacity to treat
1.2 m3 per day.
The wastewater considered in this work contains a fluorescent
penetrant oil and may be considered to be an emulsion. This liquid 2.1.2. Industrial-scale process description
penetrant is a commercial formulation produced by MagnaFlux Co. The industrial system comprised the electrochemical cell, a clar-
that is used in aeronautical and other metal-machining industries ification chamber, a sludge-conditioning tank, a horizontal plate
for non-destructive detection of flaws and scratches on metallic pressure filter and filter beds of sand and activated carbon. The
parts. industrial electrochemical cell was scaled to a volume of 23 l and
Until now the industrial operation involved was not able to recy- has the capacity to treat about 6 m3 per day. The cell consists
cle the rinsing-bath water for re-use. The entire bath was simply of a modular electrode design composed of parallel aluminum
sent for disposal, representing around 8000 l of wastewater every plates separated by 0.5 cm to limit energy consumption. Poten-
week depending on the quantity of metal processed. Because local tial is applied with a polarity-inversion frequency of 5 min to avoid
regulations obliged the industry to contain these effluents, their incrustation. The direct coupling to the clarification chamber was
disposal represented an operating cost of $0.35 USD per liter, that modified from the pilot-plant design, with internal conduits serving
is, about $142,800 USD per year. as transition zones that turn turbulent to laminar flow. In this way
Based on the characteristics of the wastewater and local regula- flocculation is enhanced, with flocs conducted toward the top of the
tions and the desire to use treated water for irrigation in green areas chamber by gassing and foam formation. The main flow of flocs is
and gardens, chemical oxygen demand (COD) must be decreased then conducted to the sludge bed, and the remaining flocs migrate
from levels varying between 1000 and 2500 mg l−1 to values below to an ultimate separation by sedimentation. The clear effluent is
75 mg l−1 . Also, the biological oxygen demand (BOD) must be low- filtered with sand and activated carbon. The sludge obtained by
ered to a value below 20 mg l−1 . Moreover, to re-use the rinse water both sedimentation and flotation are sent to the conditioning tank
in the workpiece-testing process, one must decrease the color from and then to a horizontal plate pressure filter to produce dry cake.
350 to 20 Pt–Co units and reduce turbidity from 600 to 10 NTU A small portion of the conditioned sludge stream is mixed with the
while maintaining a conductivity value below 1000 ␮S cm−1 (see inlet water to serve as a coagulant primer or initiator in the elec-
Table 1). trocoagulation cell, lowering somewhat the energy consumption.

2. Experimental 2.2. Selection of operating parameters

2.1. Materials and methods In order to avoid color in the treated wastewater, we chose to use
aluminum because it is colorless in solution. The pH range must be
A feasibility study was carried out in batch at the laboratory controlled to ensure minimal solubility of the relevant hydroxide
scale using 250-ml beakers supplied with a magnetic stirrer (IKA- formed. The minimum solubility of aluminum hydroxide is found
C-MAG, Model H57) and two aluminum electrodes of 5 cm × 6 cm at pH between 5.5 and 7.5, with an optimum at around 6.5. Accord-
active surface area (30 cm2 ) prepared by insulating one side of the ingly, for optimal efficiency the final pH should be located in that
 Y. Meas et al. / Electrochimica Acta 55 (2010) 8165–8171 8167

Table 1
Industrial wastewater characteristics before and after treatment with the industrial prototype (operation parameters: conductivity = 500 ␮S cm−1 , flow = 3.2 l min−1 , and
current density = 110 A m−2 ), local regulation limits [24], and industrial requirements.

Rinsing water characteristics Initial value Allowed limits [24] Requirements to recycle Industrial treatment results Units

Chemical oxygen demand (COD) 1000–2500 <75 <75 37 mg l−1

Biological oxygen demand (BOD) 120 <20 <20 14 mg l−1
Conductivity 140 <3000 500 (<1000) <1000 ␮S cm−1
Color 350 – 20 5 U (Pt–Co)
Turbidity 600 – 10 3 NTU
pH 6.5–7.5 8–10 6.5–7.5 7–7.5 pH U

range, which is also compatible with the pH range recommended dence time for a continuous process. Fig. 3 shows the variation of
in the industrial rinsing process, that is, between 6 and 8 [28–33]. (A) the chemical oxygen demand (COD), (B) color and (C) turbid-
The electrical energy consumption can be reduced by a proper ity and their removal percentages as functions of the electrolysis
design that takes into account the solution conductivity, the dis- time (initial COD = 840 mg l−1 , initial color = 650 Pt/Co Units, initial
tance between the electrodes, the current and the electrode surface turbidity = 600 NTU, conductivity ∼500 ␮S cm−1 and current den-
area. With a minimal distance between the electrodes chosen, sity = 333 A m−2 , final pH ∼7). Optimal removals were reached after
greater electrode surface reduces the potential required and low- five to 6 min, with COD decreasing to values below 75 mg l−1 while
ers the energy consumption. Electrode incrustation is reduced by color and turbidity reached their minima at that same time. There-
alternating the polarity. The results obtained are in agreement with fore, the experiments show that 5–6 min of electrolysis can remove
literature reports [12,20,34,35]. more than 95% of COD, color and turbidity with conductivity of
500 ␮S cm−1 and a current density of 333 A m−2 .
3. Results and discussion Aluminum is dissolved at the anode, and water is electrolyzed
at the cathode, thus producing aluminum hydroxide and hydro-
3.1. Laboratory feasibility tests gen gas. These reactions and the hydroxide precipitation result in
a change of solution pH during the process. Aluminum solubility is
Fig. 2 shows the variations of (A) the final chemical oxygen affected by the solution pH, and therefore the latter affects COD and
demand (COD), (B) final color, and (C) final turbidity as func- color removal. In summary, the final pH tends to increase with the
tions of solution conductivity for three electrolysis durations: electrolysis time and can reach values over 8.5, for which the solu-
5, 10 and 15 min. The left-hand scales give final values while bility of Al(OH)3 increases, reducing the performance of the reactor.
the right-hand scale shows the percentage reduction. The initial This is one reason that the electrolysis time should be limited.
solution conditions and operating parameters were as follows: The relationships of the final pH with the final COD, final color,
initial COD = 1000 mg l−1 , initial color = 350 Pt–Co unit, initial tur- and final turbidity and their removal percentages are shown in
bidity = 600 NTU, current density = 333 A m−2 , final pH ∼7. In all Fig. 4. (Initial COD = 940 mg l−1 , initial color = 1300 Pt/Co U, initial
cases, the best results were obtained with electrolysis time of 5 min turbidity = 300 NTU) We observe that the best results are obtained
and conductivities between 500 and 1500 ␮S cm−1 . Although the when the cell operates at a pH between 7 and 8.5, especially for
best results were obtained with a conductivity of 1000 ␮S cm−1 , color, which reaches values below 20 Pt–Co units.
the conductivity was adjusted in subsequent tests to 500 ␮S cm−1 , The effect of current density was also investigated as it relates
in common agreement with the industry recommendation and to strongly to the energy consumption of the electrocoagulation reac-
reduce the amount of salt addition. A relatively low concentration, tor. Fig. 5 shows the variation of final COD as a function of the
corresponding to around 500 ␮S conductivity, is needed to avoid current density for an electrolysis time of 5 min and a final pH
salt deposits on the parts being inspected. At the unadjusted bath close to 7. The results show that current densities throughout the
conductivity (140 ␮S cm−1 ) sufficient COD removal (above 90%) range of 100–1000 A m−2 , allowed more than 93% COD removal
was achieved but with poor color and turbidity removal (below (<75 mg l−1 ). Although current densities above 400 A m−2 may give
80% and 90%, respectively) at elevated energy cost. Final pH was better COD removal, operating in that range would represent
around 6.5–7.5. unnecessary energy consumption while sufficient COD removal is
To investigate further the effect of electrolysis time, additional reached for our purposes at current densities around 100 A m−2 .
batch experiments were carried out to define the minimal resi- Furthermore, the apparent decrease in COD removal at very high

Fig. 1. Scheme of the pilot prototype set up.

8168 Y. Meas et al. / Electrochimica Acta 55 (2010) 8165–8171

controlling the hydrogen-gas flow because on one hand foam needs

to be separated from flocs while on the other hand separation can
be improved by a proper flow of flocs through the sludge bed that is
formed by sedimentation at the bottom of the clarification cham-
ber. These criteria were taken into account in the design of the
industrial system.

3.3. Results with the industrial-scale process

Fig. 6 shows continuous color and COD removal percentages as

functions of the flow rate, with a conductivity of 500 ␮S cm−1 , a cur-
rent density of 110 A m−2 and a final pH between 6.5 and 7. Initial
COD varied from 725 to 1800 mg l−1 , and initial color varied from
260 to 428 Pt–Co units because the feed was taken during the nor-
mal industrial operations with a variable penetrant liquid content
in the rinsing water. The removal efficiency did not vary signifi-
cantly with the initial COD or color values in the ranges reported.
The results show that the system can remove more than 95% of
both color and COD for a flow range from 1.0 to 3.25 l min−1 , which
correspond to 1.44–4.68 m3 per day. The average flow demand in
the industrial process is 2.4 l min−1 .
Fig. 7 shows final COD and its removal percentage as a function
of the final pH for an average flow of 2.03 l min−1 , a conductivity of
500 ␮S cm−1 , a current density of 110 A m−2 , and an average initial
COD of 1387 mg l−1 . Results show an optimal value of more than

Fig. 2. Variations of (A) the final chemical oxygen demand (COD), (B) final
color and (C) final turbidity and their removal percentages as functions of solu-
tion conductivity for batch electrolysis durations of 5, 10 and 15 min (initial
COD = 1000 mg l−1 , initial color = 350 U (Pt–Co), initial turbidity = 600 NTU, and cur-
rent density = 333 A m−2 ; final pH ∼7).

current density is probably associated with elevated pH and alu-

minum solubility, which are detrimental.
The laboratory feasibility test has shown that it is possible to
purify and recycle this kind of wastewater with its conductivity
adjusted at 500 ␮S cm−1 . The best conditions found were final pH
of 7–8.5, current densities of around 100 A m−2 , with electrolysis
time of 5 min.

3.2. Results with the pilot-plant prototype

Tests on the transparent scale-model prototype of a continu-

ous treatment process demonstrated that removal efficiencies are
Fig. 3. Variation of (A) chemical oxygen demand (COD), (B) color and (C) turbid-
increased by strong turbulence in the electrolysis cell followed by ity and their removal percentages as functions of batch electrolysis time (initial
a flow relaxation in the subsequent clarification chamber. Trans- COD = 840 mg l−1 , initial color = 650 U (Pt–Co), initial turbidity = 600 NTU, conduc-
parent wall permitted us to detect qualitatively the importance of tivity ∼500 ␮S cm−1 and a current density = 333 A m−2 ; final pH ∼7).
 Y. Meas et al. / Electrochimica Acta 55 (2010) 8165–8171 8169

Fig. 6. Chemical oxygen demand (COD) and percentage color removal as functions
of the flowrate (feed COD varied from 725 to 1800 mg l−1 , initial color varied from
260 to 428 U (Pt–Co), conductivity = 500 ␮S cm−1 , current density = 110 A m−2 and
final pH ∼7).

Fig. 4. Variation of (A) the chemical oxygen demand (COD), (B) color and (C) turbid-
ity as functions of final solution pH (initial COD = 940 mg l−1 , initial color = 1300 U
Fig. 7. Chemical oxygen demand (COD) and its percentage removal as a function of
(Pt–Co), initial turbidity = 300 NTU, conductivity ∼500 ␮S cm−1 and current den-
the final pH (average flow = 2.03 l min−1 , conductivity = 500 ␮S cm−1 , current den-
sity = 333 A m−2 ).
sity = 110 A m−2 , and average initial COD = 1387 mg l−1 ).

98% removal at a final pH near 7 or 7.5. As seen in the laboratory

feasibility tests the optimal range of final pH corresponds to the pH rinse water process, the pH levels tend to increase lowering the
of minimum solubility for Al(OH)3 . Therefore an optimal removal COD removal percentage, although it remains above 94%. The rinse
of organic contaminants is enhanced by a low Al(OH)3 solubility to water has been recycled successfully 4 times without upsetting the
increase the entrapment and precipitation into sludge. An increase process.
of 0.5–1 pH unit from the inlet value is observed during the treat- Fig. 8 shows final COD and its removal percentage as a function of
ment under these conditions. As the initial industrial effluent pH the current density with an average flow of 2.4 l min−1 , a conduc-
is adjusted to pH 6.5 according to industrial standards, an opti-
mal removal was facilitated. During the recycling of the industrial

Fig. 8. Chemical oxygen demand (COD) and its removal percentage as a function
Fig. 5. Chemical oxygen demand (COD) as a function of the current density (con- of the current density (average flow = 2.4 l min−1 , conductivity = 500 ␮S cm−1 , and
ductivity = 500 ␮S cm−1 , electrolysis time = 5 min, and final pH ∼7). average initial COD = 1006 mg l−1 ).
8170 Y. Meas et al. / Electrochimica Acta 55 (2010) 8165–8171

Fig. 9. Typical results for removal percentages obtained after each step
in the industrial-scale treatment system (Initial COD = 1060 mg l−1 , initial
color = 380 U (Pt–Co), wet sludge weight = 1.7 kg m−3 with 20% humidity, dry sludge
weight = 1.4 kg m−3 ).
Fig. 10. Specific energy consumption per m3 water as a function of the flowrate
(current density = 110 A m−2 ).
tivity of 500 ␮S cm−1 , and an average initial COD of 1006 mg l−1 .
A diminution of the COD removal percentage was observed with
tion processes without the need of excessive energy consumption
increasing current densities. This phenomenon is due to an exces-
and extra coagulant generation associated with high current levels.
sive generation of aluminum hydroxide; the re-circulating sludge
Table 1 summarizes the results obtained during the operation of
design ensures a basal quantity of coagulant in such a way that
the industrial treatment system for a conductivity of 500 ␮S cm−1 ,
more dissolution becomes counterproductive. Both the foam and
a flow of 3.2 l min−1 and a current density of 110 A m−2 . COD values
the sediment are directed into the sludge-conditioning tank from
decreased from 1000 to 50 mg l−1 and BOD from 250 to 14 mg l−1 .
which a small flow of the supernatant is sent back to the feeding
COD, color and turbidity removal efficiencies were all found to be
pump (Fig. 1), ensuring a basal amount of coagulant and adsorbent
above 98%. A pH range between 7 and 7.5 was adequate to maintain
in the electrochemical reactor.
removal efficiencies above 95%. The sludge-to-treated water ratio
Once the system flow is stabilized the re-circulating sludge
was around 0.1 and 0.2% by weight.
ensures a significant energy saving because a lower current may be
In summary, the results obtained with the optimized operat-
used. Under steady conditions, an excess of aluminum hydroxide
ing conditions, specifically the appearance of the treated water,
and hydrogen production leads to elevated pH where the entrap-
allow its re-use in the rinsing operation instead of sending
ment is diminished due to an increase in hydroxide solubility. The
8000 l week−1 for confinement and disposal. The system generates
results demonstrate that an appropriate design allows working
only 11.2 kg week−1 of sludge, which corresponds to less than half
with minimal current densities thus limiting the energy consump-
a ton of sludge generated each year.
tion of the system.
Taking into account the cumulative costs of confinement with-
Fig. 9 displays a summary of the results for water quality
out treatment, at $2800 USD per week, in comparison with the
obtained at each step of the treatment system, presented in the
costs of treating and re-circulating the rinse waters, at $2.04 USD
form of histogram for one typical tank treated at a flow rate of
per week including energy costs, the consumed electrodes and
2.4 l min−1 , a conductivity of 500 ␮S cm−1 and a final pH of 6.9.
filter materials, along with the basic investment cost, which was
The initial COD of 1060 mg l−1 was reduced significantly after the
about $50,000 USD, for 8 m3 week−1 of water used in the industrial
first step, the electrochemical reactor coupled to the clarification
process, we see that the investment-return time for the electroco-
chamber, reaching a value of 106 mg l−1 . After the following step,
agulation system is about 17 weeks.
the sand filter, COD reached a value of 74.2 mg l−1 . In the last step,
activated carbon-bed filtration, COD reached a value of 53 mg l−1 .
These results correspond to overall reductions of 90%, 93%, and 4. Conclusions
95%, respectively. The initial color of 380 Pt–Co U reached values
of 41.8 Pt–Co U after the first step, 19 Pt–Co U after the second step, The successful development of an electrocoagulation rinse-
and 7.6 Pt–Co U after the third step, corresponding to decreases of water treatment process from laboratory tests to industrial-scale
89%, 95%, and 98%, respectively. The quantity of wet sludge gen- application has been described. In the laboratory feasibility tests
erated was 1.7 kg m−3 with 20% moisture, giving a dry weight of parameters that affect the prototype design were investigated,
1.4 kg m−3 , which is considered to be quite low. including conductivity, pH, current density, and residence time. The
Fig. 10 shows the energy consumption per cubic meter of water study of each parameter separately contributed to development of
processed as a function of the flow rate entering the system for a a feasible design concept. All the batch experiments reported were
constant COD percentage removal, a current density of 110 A m−2 , done with the actual industrial effluent of interest.
and conductivity of 500 ␮S cm−1 . A decrease in the specific energy The laboratory results helped to identify the influence of the
consumption is observed when the flow increases but with con- parameters of interest, the useful ranges for conductivity, cur-
stant COD removal, above 95%. At lower flows (longer residence rent density and residence time, and their inter-relationships. For
times) and for constant current density, an increase in specific example, those tests revealed the value of considering the final pH
energy consumption was observed without a significant increase instead of the initial pH. According to the critical design considera-
in the removal efficiency. To maintain the lowest energy consump- tions, the conductivity was set at a relatively low value, in common
tion and a global efficiency above 95%, it is advisable to maintain agreement with the industry practice, that is, at 500 ␮S cm−1 . Con-
the flow rate between 2.4 and 3.5 l min−1 . The observed behavior ductivity is a key parameter because it impacts directly on the
is attributed to the sludge re-circulation procedure, which helps, energy consumption and also on the number of times a batch can be
once operation is stabilized, to prime the coagulation and adsorp- re-used because the conductivity and the pH increase in each cycle.
 Y. Meas et al. / Electrochimica Acta 55 (2010) 8165–8171 8171

The rise in conductivity with number of cycles of re-use for a given [2] P.K. Holt, G.W. Barton, C.A. Mitchell, Chemosphere 59 (2005) 355.
batch of rinse water is presently the subject of further research. [3] H.A. Moreno-Casillas, D.L. Cocke, J.A.G. Gomes, P. Morkovsky, J.R. Parga, E. Peter-
son, Sep. Purif. Technol. 56 (2007) 204.
The pH is related to the concentration of coagulant, and an effi- [4] K. Yetilmezsoy, F. Ilhan, Z. Sapci-Zengin, S. Sakar, M.T. Gonullu, J. Hazard. Mater.
ciency decrease results from elevated pH that is associated with 162 (2009) 120.
excess aluminum hydroxide production and enhanced solubility. [5] M.M. Emamjomeh, M. Sivakumar, J. Environ. Manage. 90 (2009) 1204.
[6] X. Chen, G. Chen, P.L. Yue, Sep. Filtr. Technol. 19 (2000) 65.
Accordingly, the final pH was monitored, and its optimal range for [7] G. Chen, X. Chen, P.L. Yue, J. Environ. Eng. 126 (2000) 858.
high removal efficiency (>95%) was found to be 7–7.5. The flow in [8] I. Arslan-Alaton, I. Kabdashi, B. Vardar, O. Tunay, J. Hazard. Mater. 164 (2009)
the continuous process was adjusted to match a residence time of 1586.
[9] I. Zongo, J.-P. Leclerc, H.A. Maïga, J. Wéthé, F. Lapicque, Sep. Purif. Technol. 66
5 min, which was found to ensure an optimal removal (>95%) in (1) (2009) 159.
the laboratory level. The inter-electrode distance was reduced to a [10] H. Zhao, J. Peng, A. Xue, J. Ni, Compos. Sci. Technol. 69 (10) (2009) 1629.
minimum to maximize the energy efficiency, which also depends [11] A.H. Essadki, B. Gourich, Ch. Vial, H. Delmas, M. Bennajah, J. Hazard. Mater. 168
(2009) 1325.
directly on the conductivity. An increase in the surface area of the
[12] J.E. Gregor, C.J. Nokes, E. Fenton, Water Res. 31 (1997) 2949.
electrodes allowed a decrease in current density and cell potential, [13] M. Khemis, J.-P. Leclerc, G. Tanguy, G. Valentin, F. Lapicque, Chem. Eng. Sci. 61
which resulted in a reduction in the energy consumption. (2006) 3602.
Several technical factors identified with the pilot-scale proto- [14] S. Zodi, O. Potier, F. Lapicque, J.-P. Leclerc, Sep. Purif. Technol. 69 (2009) 29.
[15] Y.O.A. Fouad, A.H. Konsowa, H.A. Farag, G.H. Sedahmed, Chem. Eng. J. 145 (2009)
type helped to develop the final design of the industrial process. 436.
One example is the partial re-circulation of sludge supernatant, [16] M. Eyvaz, M. Kirlaroglu, T.S. Aktas, E. Yuksel, Chem. Eng. J. 153 (2009) 16.
which enhances coagulation and increases the removal efficiencies [17] J. Ge, J. Qu, P. Lei, H. Liu, Sep. Purif. Technol. 36 (2004) 33.
[18] M. Kobya, M. Bayramoglu, M. Eyvaz, J. Hazard. Mater. 148 (2007) 311.
without increasing the energy consumption. This improvement [19] M. Bayramoglu, M. Eyvaz, M. Kobya, Chem. Eng. J. 128 (2007) 155.
also helped in the diminution of the amount of sludge generated. [20] L.O. Kolarik, C.T. Chin, L.F. David, M.B. Roberts, Water (1991) 23.
Once all these settings and improvements were integrated into [21] G. Chen, Sep. Purif. Technol. 38 (2004) 11.
[22] M.Y.A. Mollah, R. Schennach, J.R. Parga, D.L. Cocke, J. Hazard. Mater. 84 (2001)
the industrial design, the results obtained met the desired water 29.
quality for a very low cost. The improvements also resulted in a [23] Z. Zaroual, H. Chaair, A.H. Essadki, K. El Ass, M. Azzi, Chem. Eng. J. 148 (2009)
very short time for return of the investment as they minimized the 488.
[24] Secretaría de medio ambiente y recursos naturales, norma oficial mexicana,
energy consumption of the system, making it very attractive to the NOM-003-SERMARNAT-1996.
industry. [25] N. Daneshvar, A.R. Khataee, A.R. Amani Ghadim, M.H. Rasoulifard, J. Hazard.
Mater. 148 (2007) 566.
[26] M. Khemis, G. Tanguy, J.P. Leclerc, G. Valentin, F. Lapicque, Trans. IChemE 83
Acknowledgements (2005) 50.
[27] N. Daneshvar, A. Oladegaragoze, N. Djafarzadeh, J. Hazard. Mater. 129 (2006)
We thank the National Science and Technology Council of Mex- 116.
[28] G. Mouedhen, M. Feki, M. De Petris Wery, H.F. Ayedi, J. Hazard. Mater. 150
ico (CONACYT: Consejo Nacional de Ciencia y Tecnología) and the
(2008) 124.
Government of the Mexican State of Hidalgo for their economical [29] Ş. İrdemez, N. Demircioğlu, Y.Ş. Yıldız, Z. Bingül, Sep. Purif. Technol. 52 (2006)
support in the development of the pilot prototype. 218.
We thank also Messier Services Americas for the opportunity [30] A.S. Koparal, Y.Ş. Yildiz, B. Keskinler, N. Demircioglu, Sep. Purif. Technol. 59
(2008) 175.
offered to develop the industrial prototype and especially their [31] P. Cañizares, C. Jiménez, F. Martínez, M.A. Rodrigo, C. Sáez, J. Hazard. Mater. 163
patience in allowing us to test the process within their facilities (2009) 158.
in order to optimize the operation of the system. [32] E.A. Vik, D.A. Carlson, A.S. Eikum, E.T. Gjessing, Water Res. 18 (1984) 1355.
[33] I. Arslan-Alaton, I. Kabdaşlı, D. Hanbaba, E. Kuybu, J. Hazard. Mater. 150 (2008)
166.
References [34] S.H. Lin, C.F. Peng, Water Res. 28 (1994) 277.
[35] B. Merzouk, B. Gourich, A. Sekki, K. Madani, M. Chibane, J. Hazard. Mater. 164
[1] M.M. Emamjomeh, M. Sivakumar, J. Environ. Manage. (2009) 1. (2009) 215.