1-s2.0-S0043135412002138-main

ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 6

Available online at w

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

An evaluation of a hybrid ion exchange electrodialysisprocess in the recovery of heavy metals from simulateddilute industrial wastewater

Akrama Mahmoud a,*, Andrew F.A. Hoadley b

a Laboratoire de Thermique Energetique et Procedes (EAD 1932), ENSGTI, rue Jules Ferry, BP 7511, 64075 Pau, FrancebDepartment of Chemical Engineering, Building 35, Clayton Campus, Monash University, Victoria 3800, Australia

a r t i c l e i n f o

Article history:

Received 10 August 2011

Received in revised form

25 December 2011

Accepted 20 March 2012

Available online 28 March 2012

Keywords:

Copper ions

Ion exchange

Electrodialysis

Electrodeionization

Electromigration

Wastewater

Treatment of dilute solutions

* Corresponding author. Tel.: þ33 0540175193E-mail addresses: akrama_mahmoud@ho

0043-1354/$ e see front matter ª 2012 Elsevdoi:10.1016/j.watres.2012.03.039

a b s t r a c t

Hybrid ion exchange electrodialysis, also called electrodeionization (IXED), is a technology in

whichaconventional ionexchange (IX) is combinedwithelectrodialysis (ED) to intensifymass

transfer and to increase the limiting current density and therefore to carry out the treatment

process more effectively. It allows the purification of metal-containing waters, as well as the

productionof concentratedmetal salt solutions,whichcouldbe recycled.Theobjectiveof this

paperwas to investigate the ability of the IXED technique for the treatment of acidified copper

sulphate solutions simulating rinsing water of copper plating lines. A single-stage IXED

process at lab-scale with a small bed of ion exchanger resin with a uniform composition was

evaluated, and the treatment performance of the process was thoroughly investigated. The

IXED stack was assembled as a bed layered with the ion exchanger resin (strong acid cation-

exchange Dowex�) and inert materials. The stack configuration was designed to prevent

a non-uniform distribution of the current in the bed and to allow faster establishment of

steady-state in the cell for IXED operation. The influence of operating conditions (e.g. ion

exchanger resin with a cross-linking degree from 2 to 8% DVB, and current density) on IXED

performancewas examined. A response surfacemethodology (RSM)was used to evaluate the

effects of the processing parameters of IXED on (i) the abatement yield of the metal cation,

which is a fundamental purification parameter and an excellent indicator of the extent of

IXED, (ii) the current yield or the efficiency of copper transport induced by the electrical field

and (iii) the energy consumption. The experimental results showed that the performance at

steady-state of the IXED operationwith a layered bed remainedmodest, because of the small

dimension of the bed and notably the current efficiency varied from 25 to 47% depending on

the conditions applied. The feasibility of using the IXED in operations for removal of heavy

metals frommoderately dilute rinsing waters was successfully demonstrated.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction mainly due to more stringent legislations for the protection of

The recovery of heavy metals from industrial aqueous solu-

tions has received great attention in recent years. This is

; fax: þ33 0559407801.tmail.com, amahmoud@ier Ltd. All rights reserve

the environment. Most heavy metals are very toxic and cause

great environmental damage (JUttner et al., 2000; Janssen and

Koene, 2002).

gmx.fr (A. Mahmoud).d.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 6 3365

The conventional techniques for metal ions abatement,

such as hydroxide precipitation, or direct electro-reduction do

not provide sufficient removal efficiency and as a conse-

quence secondary treatment processes are required down-

stream. An important technique for secondary treatment

below ppm levels is ion exchange: the metal ions contained in

the waste solution are exchanged by the less toxic ions con-

tained in the fixed matrix of an ion exchange bed. Once the

bed has reached capacity, it must be regenerated with

a concentrated electrolyte such as a strong acid. This produces

a concentrated waste stream which has to be treated and

ultimately must be disposed of somewhere.

One alternative technique is electrodialysis, which is an

electrically driven process involving the use of ion-selective

membranes. This technique not only concentrates metals

from the rinse streams, but also helps to maintain the quality

of a plating bath. However, electrodialysis cannot be carried

out at ppm concentrations. Therefore, current research tends

to propose potential alternatives to enhance the treatment

ability of conventional processes. Different options have been

investigated to enhance the treatment of metal-containing

rinsing waters, such as electrodeionization, electro-

coagulation (Adhoum et al., 2004; Dermentzis et al., 2011a),

complexation/ultrafilration-electrolysis (Baticle et al., 2000),

and reverse osmosis/nanofiltration-flotation (Sudilovskiy

et al., 2008). Proper choice of treatment technologies for

a given system is essential, but a good choice can lead to

higher abatement yield, and lower energy and environmental

costs.

Hybrid systems combining ion exchange (IX) with electro-

dialysis (ED) have been suggested to combine the advantages

of the two individual techniques and in particular, to remove

heavy metals from dilute solutions in a continuous process.

The IXED process is not a new idea, but the practical

applications have been limited to treatment of radioactive

waste (Glueckauf, 1959), and production of ultrapure water,

with the removal of both anions and cations (Ganzi et al., 1987,

1997; Thate et al., 1999), or more simply for water purification

(Dejean et al., 1997; Souilah et al., 2000). The well-known

Millipore� single IXED cell for production of ultrapure water

is shown schematically in Fig. 1 as an example.

Since the initial commercialization in 1987, IXED systems

have found worldwide application in industries with themost

demanding high purity water requirements such as in the

manufacture of pharmaceuticals, semiconductors, and high

quality optics, in surface finishing of electronic components,

in the generation of electric power, and in food processing

(Ganzi and Parise, 1990; Soria et al., 1993; Gifford and Atnoor,

2000; Wood and Gifford, 2002).

More recently, the combined technique was considered for

treatment of metal-containing rinsing waters, where the IXED

process yields concentrated solutions of metal ions in the

cathode compartment and metal-free acidified water leaving

the cationic resin bed: the example of Ni salts was extensively

treated by Janssen’s group (Spoor et al., 2001, 2002) and by

Dzyazko and Belyakov (2004), Dzyazko (2006), Lu et al. (2007,

2010, 2011) and Priya et al. (2009). Other metals were also

investigated, including, zinc (Grebenyuk et al., 1998), cobalt

(Yeon et al., 2003; Song et al., 2004), copper (Mahmoud et al.,

2003; Guana and Wang, 2007; Feng et al., 2008; Arar et al.,

2011), chromium (Dzyazko et al., 2008a,b; Alvarado et al.,

2009; Xing et al., 2009a,b), cadmium (Dermentzis et al.,

2011b), arsenic (Basha et al., 2008) or mixtures of metals ions

(Souilah et al., 2000; Feng et al., 2007; Smara et al., 2007).

Interestingly, the possible competing adsorption of divalent

ions in the process water, e.g. Ca2þ and Mg2þ, was studied

(Spoor et al., 2001; Vasilyuk et al., 2004; Fu et al., 2009) and its

significance was shown to be reduced by using inorganic Zr-

containing sorbents (Vasilyuk et al., 2004). Most of these

studies are empirical and carried out in either lab-scale

devices or at pilot-scale. The equipment used for most IXED

studies comprised a number of test cells of plate and frame

design, spiral wound design, cylindrical and rectangular

shape and a variety of sizes. A number of these IXED cells,

with more chambers in the cell and with different ion-

exchange resin arrangements (mixed bed, clustered or

layered cation and anion-exchange resins, separate bed), have

been engineered by researchers in the field, and have reached

varying degrees of development (Kunz, 1987; Guiffrida et al.,

1990; Ganzi et al., 1997; Tessier et al., 1997; Gifford and

Atnoor, 2000; Grabowski et al., 2006).

The present paper investigates the effects of the processing

parameters (current density, resin nature) of the ion exchange

electrodialysis (IXED) process for Cu2þ removal from acidified

copper sulphate solutions, simulating the rinse water from

copper plating lines. The abatement yield of the copper cation

is a fundamental treatment parameter and an excellent indi-

cator of the extent of IXED. The current yield or the efficiency

of copper transport induced by the electrical field is also

analysed and is an indicator of the operating cost of the

process. A short bed of ion-exchange resin is used to prevent

a non-uniform distribution of the current in the bed and to

allow faster establishment of steady-state in the cell for IXED

operation. Finally, in order to design efficient processes, the

response surface methodology (RSM) is used to test for

statistical significance and determine the optimum treatment

performance.

2. Theoretical background

In this study, a hybrid system is obtained by inserting a bed of

ion-exchange resin initially in Hþ form into the central

compartment of an electrodialysis cell. This compartment is

located between two ion-selective membranes that essen-

tially divide the cell into three separate compartments. The

copper solution to be treated is injected continuously through

a packed-bed of ion exchange resins. During IXED, metal

cations are sorbed by the resins, and are transported to the

cathode compartment under the action of the applied elec-

trical field. Water electrolysis occurs in the two external

electrode compartments, and Hþ formed at the anode is

transferred through the membrane into the resin bed and

replace the metal ions which will then migrate through the

second membrane to the electrode compartment. This

phenomenon is known as electro-regeneration. In order to

explain more closely the mechanism of IXED, the role of the

ion-exchange resin bed in the presence of an electrical field is

first discussed. Following this, a theory for the ion exchange

equilibrium step and IXED efficiency is presented.

Fig. 1 e Schematic of the Millipore (Ionpure) single IXED configuration showing the mixed resins in the dilute compartment

[adapted from Ganzi et al. (1987); TSS Water Course; Mahmoud, 2004].

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 63366

2.1. Ion-exchange resin bed role

In a hybrid system such as IXED, the ion-exchange resin bed

plays a major role in the reduction of the high electrical

resistance in the dilute compartment, while the ion exchange

membranes allow depletion and concentration of solutions in

their dedicated compartments. Ion-exchange resins are

polymers that have fixed ionic sites prone to reactionwith free

ions of the opposite charge. The ionic groups of the resin

provide a location at which the dissolved ions can be

exchanged. Ion-exchange resins in electrolyte solutions are

electrically conductive, and counter ions can transfer across

the polymer under an electric field, allowing mass transfer

and associated current flow. The electrical conductivity of ion-

exchange resins varies with the mobility and affinity of the

counter ions with which the resins are in contact (Helfferich,

1962). In the case of a very low specific conductivity of the

interstitial solution, the specific conductivity of the bed is

enhanced by the presence of ion exchange resins. Therefore

the main advantage of using ion exchange resins in the dilute

compartment is the substantial reduction in electrical resis-

tance which is achieved when very low concentration solu-

tions are concerned. This result in an increase in the limiting

current density and, consequently, the treatment process

proceeds more effectively compared to electrodialysis alone.

Demkin et al. (1987), from their studies in which an ion-

exchange fibrous filler was used in an ED cell, observed that

the limiting current density was increased considerably from

4 to more than 24 A/m2 for the low and high concentration

cases, respectively.

2.2. Cu2þ/Hþ ion exchange isotherm

Themonodivalent ion exchange process betweenCu2þ andHþ

is represented by Equation (1):

2Hþresin þ Cu2þ

solution4Cu2þresin þ 2Hþ

solution (1)

The selectivity coefficient of the ion-exchange process can

be defined using the equivalent ionic fractions as follows:

KCu;H ¼�XCu2þ

XCu2þ

���XHþ

XHþ

�2

(2)

where Xi and Xi are the equivalent ionic fractions of species i

in the solution and in the resin phase, respectively. The

equivalent fractions of Cu2þ and Hþ are defined as:

XHþ ¼ ½Hþ�2½Cu2þ� þ ½Hþ� XCu2þ ¼ 2½Cu2þ�

2½Cu2þ� þ ½Hþ� (3)

XHþ ¼ ½Hþ�2½Cu2þ� þ ½Hþ�

XCu2þ ¼ 2½Cu2þ�2½Cu2þ� þ ½Hþ�

(4)

where ½Cu2þ� and ½Hþ� are the ion concentrations in the ion-

exchange resin, and ½Cu2þ� and ½Hþ� are the ion concentra-

tions in the solution phase.

2.3. IXED efficiency

The IXED performance can be evaluated in terms of (i)

abatement yield ðhCu2þ Þ in the metal cation and (ii) current

efficiency (or transport) yields (CE) of Cu2þ ions. The

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 6 3367

abatement allowed by IXED process was effectively assessed

by the difference between the copper concentration at the

inlet and the outlet divided by the copper concentration at the

inlet of cell, hCu2þ . On the other hand, the ratio of the quantity

of copper transferred to the cathode compartment and the

charge passed during electrodialysis taking into account

Faraday’s constant, expresses the efficiency of copper trans-

port induced by the electrical field:

CEcathodeCu2þ ¼ zCu � F� ncathode

Cu2þ

Q(5)

where ncathodeCu2þ is the number of moles Cu2þ species in the

cathode compartment and Q the electrical charge passed in

the circuit at time t. For the present case of constant current

operations, the electrical charge is given by: Q ¼ I� dt. A

similar yield can be defined at the anode:

CEanodeCu2þ ¼ zCu � F� nanode

Cu2þ

Q(6)

where nanodeCu2þ is the number of moles Cu2þ species in the

cathode compartment.

Moreover, a fraction of the copper cations transferred are

reduced into metal copper on the Pt-coated cathode. The

current yield of this copper deposited at the cathode can be

given by:

CEmetalCu ¼ zCu � F� nmetal

Cu

Q(7)

where nmetalCu is the mole amounts of copper deposited at the

cathode. Finally, the total current efficiency of the ion transfer

to the cathode side can be defined as follows:

CEcathodeCu ðtotalÞ ¼ CEcathode

Cu2þ þ CEmetalCu ¼ zCu � F� �

ncathodeCu2þ þ nmetal

Cu

�Q

(8)

This efficiency represents the transference number of

copper species through the membrane to the cathode

compartment.

Table 1 e Physical properties of the resins (Mahmoud,2004; Mahmoud et al., 2003, 2007).

Parameter Dowex 50WX (2%)

Dowex 50WX (4%)

DowexHCR-S (8%)

Granulometry (Mesh) 50e100 50e100 20e50

Capacity (eq./1) 1.81 1.17 0.76

Selectivity coefficient 9.5 � 2.8 10.1 � 3.9 13.7 � 3.9

k (Hþ-form) (mS/cm) 265 242 200

k (Cu2þ-form) (mS/cm) 29 20 10

k: The electrical conductivity of the solid phase.

3. Materials and methods

3.1. Experimental procedures

3.1.1. Feed solutionsThe experiments were carried out at room temperature with

mixed CuSO4 and H2SO4 solutions of various concentrations.

The solutionswere preparedwith reagents of analytical purity

(Merck) and conductimetry water. Concentration of copper, in

the form of copper sulphate was fixed at 1.57 mmol/1, corre-

sponding to 100 ppm Cu ions. The solutions were acidified

with sulphuric acid, whose concentration in the resulting

solutionswas 1mmol/l, corresponding to pH equal to 3.2� 0.1.

3.1.2. Ion-exchange resinsThree styrene-base cation exchangers Dowex� were used,

with a DVB cross-linking degree varying of 2, 4 and 8%. Prop-

erties of the resins were determined previously (Mahmoud,

2004; Mahmoud et al., 2003, 2007) are given in Table 1. In

particular, the copper capacity was measured at room

temperature with mixed CuSO4 and H2SO4 solutions of

various concentrations by fixing the total normality at 0.5 eq./l

Cu2þ solution. The composition of these solutions was calcu-

lated using a thermodynamic model of the mixed solution for

various proportions of copper salt over sulphuric acid, with

the total normality at 0.5 eq./l (Mahmoud et al., 2003). This

model derived from previous investigations of the

ZnSO4eH2SO4 system (Zouari and Lapicque, 1992) takes into

account the non ideal behaviour of the various species by

means of the Pitzer model for the various interactions. In

addition, the partial dissociation of copper sulphate was also

accounted for, as recommended in Wasylkiewicz (1990).

Fig. 2 shows the ion exchange isotherm of copper ions on

the resins investigated. The isotherm profile indicates that

copper is strongly preferred by those resins over protons, as

shown in the Fig. 2. For these three resins, the simple laws for

sorption isotherms fit well with the experimental data. For any

ionic composition, the separation factor equals the ratio of the

two rectangular areas of Hþ and Cu2þ touching one another in

the corresponding point on the isotherm. The diagonal dotted

line is the isothermof a fictitious ion exchange resinwhich has

no preference for either counter ion.

3.2. Experimental set-up

For the IXED experiments, a lab-scale three-compartment cell

was used (Mahmoud, 2004; Mahmoud et al., 2003, 2007;

Monzie et al., 2005). The electrode compartments and the

packed bed were 100 mm high and 10 mm wide, and the

membranes were Nafion� 117. The membrane gap, corre-

sponding to the bed thickness, was 15 mm. Both electrodes

were platinum-coated titanium plates which are dimension-

ally stable (not consumed coulometrically). For most of the

experiments, it was preferable to work with a smaller ion

exchange bed for the sake of its more uniform composition.

This bed was 15 mm high and was situated between two inert

resin beds of the same particle size, each of which was

42.5 mm high. The reduced active bed length allowed faster

transient loading of the resin upon percolation of the copper

solution, in addition to more uniform distributions of vari-

ables in the bed. The experimental set-up with the cell

geometry, including the beds inserted in the central

compartment, is illustrated in Fig. 3. A DC power supply

(HewlettePackard E3612A), operating under constant current

(or voltage), was used for the IXED operation. Two multi-

meters (ITT instruments MX20) were used to control the

current and to monitor voltage fluctuations in the IXED cell.

Fig. 2 e Ion-exchange isotherms of copper ion on the resins

investigated. Solid lines correspond to fitting with average

value of the selectivity coefficient.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 63368

Finally, to control the treated solution quality, the pH and the

electrical conductivity were measured by a pH/Ion meter

(pHM692, MeterLab) and a conductivity meter (CDM210,

MeterLab), respectively.

One litre of sulphuric acid solutions at 0.5 M were circu-

lated batchwise in the electrode compartments at 30 l h�1

using a centrifugal pump (Iwaki). The bed was continuously

percolated downwards by the prepared solutions at 0.60 l h�1

using a peristaltic pump (BVK, Ismatec). The resinwas initially

in the Hþ-form, and runs were carried out at fixed current at

ambient temperature (22 �C � 1 �C).

3.3. Measurements and analytical methods

The range of variables for the IXED conditions was based on

the preliminary studies conducted for the estimation of the

minimum current that would be required for complete

Cu2+

30 l/h

1 litre Tank

H2SO4

CuSO4

Solution

CM

-

42.5 mm

42.5 mm

15 mm

15

Treated

CM: Cation Membrane

H+-form ion-exchange bed

Inert bed

Fig. 3 e Schematic view of experimental set-up for IXED experi

compartment.

removal of the copper ion flux fed to the cell assuming a 100%

current efficiency (Mahmoud, 2004). Three constant current

densities 10, 13.3 and 20 mA/cm2 were used to investigate the

influence of the current density on the IXED treatment.

Experiments were carried out for 10 h with the three resins

considered.

The solutions of the electrode chambers and that leaving

the central compartment were collected at regular intervals.

The pH, conductivity, and copper concentration of these

samples were measured. The cell voltage was continuously

recorded. After the run, the bed was percolated with 2 M

H2SO4 solution for desorption of the copper ions, then rinsed

with deionized water. The cell was taken apart and the traces

of copper deposited on the Pt/Ti cathode were dissolved with

dilute nitric acid. The copper contents of the various liquid

fractions were determined by flame atomic absorption. At the

end of each run, a mass balance for Cu2þ was conducted,

taking into account the amount of metal ions sorbed on the

resin, the mole numbers transferred to both compartments

and the mole amount of metal deposited. Mass balances were

shown to hold within 7% in most cases.

4. Results and discussion

4.1. Effect of the resin nature in the dilute compartment

4.1.1. Measurements of the electrical resistance and thepower consumptionThe advantages of the IXED process in terms of voltage fluc-

tuation or electrical resistance and power consumption over

the ED process on its own are evaluated. The impact of the

resin nature was carried out at a constant current density of

20 mA/cm2 and for the 1.57 mmol/1 copper solution. Fig. 4

shows the transient behaviour of the electrical resistance

and power consumption for the three resins. During the

loading of the resin with copper ions, Period (I), there was

(0.6 l/h)

Ti-Pt

10 mm

100

mm

H2SO4

Solution

1 litre Tank

Electrode

30 l/h

CM

+

mm

Solution

ments with the beds inserted in the central (dilute)

0.0E+00

2.0E-04

4.0E-04

6.0E-04

8.0E-04

1.0E-03

1.2E-03

1.4E-03

1.6E-03

1.8E-03

2.0E-03

0 1 2 3 4 5 6 7 8 9 10

Time (h)

Pow

er c

onsu

mpt

ion

(kW

h)

0

200

400

600

800

1000

1200

1400

1600 ED

IXED with Dowex 2%

IXED with Dowex 4%

IXED with HCR-S 8%

Ele

ctri

cal r

esis

tanc

e (

Ohm

)

(I) (II)

58.4

%

67.8

%

70.7

%

56%

66%

69%

Fig. 4 e Electrical resistance and power consumption variation during ED and IXED runs for the three resins considered at

a constant current density of 20 mA/cm2. Period (I): Loading of the resin, Period (II): Steady-state.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 6 3369

a regular increase in cell voltage. In Period (II) steady values in

the range 11.9e14.5, 12e13.2, and 16.9e18.7 V were attained

for 2, 4 and 8% DVB, respectively, after 2 h or so. As shown in

Fig. 4, the corresponding electrical resistance was in range

395.7e482.7, 402e439, and 563e623 Ohm, respectively. High-

est cell voltages were recorded with 8% DVB resins, in agree-

ment with their lowest electrical conductivity (or highest

electrical resistance, see Table 1). Moreover, it was found that

the highest reduction in the electrical resistance (about 71%)

was offered by the 4% resin (with moderate degree of cross-

linking). This can be explained by its high capacity and its soft

structure. It was also found that the IXED technique requires

much less energy than ED technique alone. For example, this

technique ismore energy efficient requiring less than 32% and

44% of the ED energy for the moderate and high DVB cross-

linking degree, respectively. These findings seem to indicate

that the hybrid process seems to be an interesting technique

for improving the electrical efficiency of treating dilute

wastewater. An overview of the specific electric energy

consumptions based on copper removal (kW h/molremoved

copper) is discussed in Section 4.3.

4.1.2. Effect of the resin nature on the pH in the dilutecompartmentThe use of ion-exchange resins may also give rise to changes

in the pH caused by the production of hydrogen and hydroxide

ions, which may reduce problems typically associated with

concentration polarization near the membranes (Spoor et al.,

2002).

However, the experimental results show that the increase

in DVB cross-linking degree does not have a significant effect

on the pH values. As a matter of fact, the pH of the three

solutions sampled, changed very little during the batch

experiments. In particular, the pH of the sulphuric acid

solutions remained in the range 0.6e0.7, with a slight

increase at the cathode side due to production of OH� ions.

a

b

c

Cu

conc

entr

atio

n (m

mol

/l)

0.0

0.4

0.8

1.2

1.6

2.0

0 1 2 3 4 5 6 7 8 9 10

Outlet with Dowex 2%Outlet with Dowex 4%Outlet with HCR-S 8% Inlet

Time (h)

Cu2+

conc

entr

atio

n (m

mol

/l)

20 mA/cm

31%33%

41%

(I) (II)

0.0

0.4

0.8

1.2

1.6

2.0

0 1 2 3 4 5 6 7 8 9 10

Dowex 2% Dowex 4% HCR-S 8%

Time (h)

20 mA/cm

(I) (II)

0.0

0.1

0.2

0.3

0.4

0.5

0 1 2 3 4 5 6 7 8 9 10

Dowex 2% Dowex 4% HCR-S 8%

Time (h)

Cu

conc

entr

atio

n (m

mol

/l)

20 mA/cm

Fig. 5 e Treatment of the copper sulphate solution with the

small resin bed at a constant current density of 20 mA/cm2,

depending on the resin stiffness. (a): Copper ion

concentration at the outlet of the central compartment

versus time. (b): Copper ion concentration in the cathode

compartment versus time. (c): Copper ion concentration in

the anode compartment versus time. Period (I): Loading of

the resin, Period (II): Steady-state.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 63370

The solution leaving the central compartment had a pH

ranging from 2.6 to 2.8, due to the flux of Hþ transferred from

the anode side.

4.1.3. Effect of the resin nature in the dilute compartment onthe abatement yieldThe concentration of copper ions in the cathode liquid is

plotted against time in Fig. 5(b), whereas Fig. 5(a) gives the

outlet concentration of copper species. In the first period, the

increase in Cu2þ concentration in the cathode chamber was

rather low, due to the progressive loading of the resin; and as

a consequence the abatement of copper ions was complete.

After this, the accumulation in the cathode chamber pro-

gressed at a regular, higher pace, whereas the resin bed was

not sufficient for total removal of Cu2þ. After 2 h or so, the

concentration in the cathode chamber, levelled off to attain

a steady level at the end of the run. Both the accumulation rate

in the cathode chamber and the steady-state concentration at

the outlet depended on the resin stiffness: the resin with 4%

DVB cross-linking degree allowed the best water purification

performance and produced the most concentrated copper

sulphate solutions. As shown in Fig. 5(a), the abatement yield

ðhCu2þ Þ attained 41%with this resin, 33 and 31%with resins at 2

and 8% DVB cross-linking degree, respectively. The final cor-

responding value for copper concentration in the cathode

chamber was found at 1.79mmol/l to be compared to 1.72 and

1.35 mmol/l with resins at 2 and 8% DVB cross-linking degree,

respectively.

The best performance offered by the 4% resin can be

explained by its high capacity and its soft structure, which is

expressed indirectly by its electrical conductivity. The softest

grade gave the highest transport properties to the cations, but

its poor capacity limits the sorption flux. Finally, the HCR

grade with 8% DVB is far too rigid to allow efficient ion

transport, in spite of its high capacity. The poor ability of the

rigid resin was previously observed in electromigration

experiments (Mahmoud et al., 2003).

4.1.4. Effect of the resin nature in the dilute compartment onthe current yieldThe amount of copper ions transferred to the cathode

compartment is plotted against time in Fig. 5(b). Asmentioned

already, the amount of copper ions increased regularly with

time, depending on the resin grade.

On the other hand, copper ions were observed to be

transferred to the anode compartment, with the overall

transfer resulting from antagonistic migration and diffusion

fluxes. Concentration of copper ions in the anode chamber

increased regularly with time, with slightly higher rates in the

loading period of each experiment. Despite its high transport

property, the 2% resin gave the lowest side-transfer rates in

the anode chamber. As shown in Fig. 5(c), the 4% resin allowed

the highest side-transfer rates to the anode during the first 5 h

of testing. Thereafter the 8% resin gave the highest side-

transfer rates to the anode chamber. The corresponding final

concentration of copper ions in the anode chamber was

0.425 mmol/1, corresponding to 27 ppm.

Dismantling of the cell structure revealed a significant

copper deposit on the Ti/Pt surface of the cathode with

a mainly dendritic morphology. This was particularly evident

in the case of the 4% resin in spite of the supporting electrolyte

in the catholytic chamber flowing at 30 l h�1. The amount of

metal recovered at the cathode was comparable to the mole

amount of copper ions in the cathode chamber as shown in

Table 2. Moreover the deposition rate was likely to be

enhanced at the end of the runs by the increasing Cu2þ

concentrations in the cathode chamber. Corresponding

Faradic yields and values for the removal efficiency are given

in Table 2. The current efficiency for the copper ion transfer

through the membrane, CEcathodeCu ðtotalÞ, being the sum of the ion

Table 2 e Current efficiency of transport to the external compartments (CE); removal yield ðhCu2D Þ; the mole amount ofcopper ions in the anode and cathode chambers; and the mole amount of copper deposited at the cathode after 10 h ata constant current of 20 mA/cm2.

Resin nanodeCu2þ mmol nmetal

Cu mmol ncathodeCu2þ mmol CEanode

Cu2þ % CEmetalCu % CEcathode

Cu2þ % CEcathodeCu ðtotalÞ % hCu2þ %

50 WX (2%) 0.162 0.170 1.72 2.89 3.04 30.72 33.72 33

50 WX (4%) 0.339 0.114 1.79 6.06 2.03 31.94 33.97 41

HCR-S (8%) 0.426 0.0378 1.35 7.62 0.67 24.06 24.73 31

Copper sulphate injected into the bed is 9.63 mmol within 10 h.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 6 3371

and metal contributions, is also given. This efficiency corre-

sponds to the transference number through the bed and the

cation-sidemembrane and attained 33.97%with the 50WX 4%

resin (Table 2).

4.2. Effect of the current density

Table 3 summarizes the experimental results for the three

resins considered and for three processing current densities

10, 13.3 and 20mA/cm2. The amount of Cu2þ ions recovered in

the cathode compartment increased with the applied current,

but lower current efficiencies were obtained for the highest

current for the three resin grades. The moderately flexible 4%

resin appears to allow the best results. The amount of copper

deposited was generally far lower than the quantity of copper

ions, in particular at low current densities. The yield for

copper deposition increased stronglywith the current applied,

varying from 0.5% at 10 mA/cm2 to a few percents for the

highest current density (Table 3).

The removal yield was averaged over the period of steady

behaviour of the bed. This yield depended on the resin grade

and the current density (Table 3). The best results were also

obtained with the 4% grade. Moreover the efficiency of the

water treatment was enhanced with higher current densities;

however, for the 4% grade, the removal yield was observed to

be barely affected by the current density over 10 mA/cm2.

Fig. 6(a) shows theoutlet concentrationof copper species. In

thefirst period, the loadingof the resinparticles allowednearly

complete removal of copper ions from the feed solution. Then,

the copper concentration at the outlet increased rapidly to

attain a steady level after 2 h or so, as shown in Fig. 6(a).

Table 3 e Removal yield ðhCu2D Þ; current efficiency of transportcopper ions in the anode and cathode chambers; the mole am

Resin Current densitymA/cm2

nanodeCu2þ

mmolnmetalCu

mmolncathodeCu2þ

mmol

Dowex (2%) 10 0.532 0.0157 1.07

13.3 0.373 0.0375 1.45

20 0.162 0.0170 1.72

Dowex (4%) 10 0.469 0.0110 1.29

13.3 0.556 0.0894 1.50

20 0.339 0.0114 1.79

HCR-S (8%) 10 0.459 0.0118 1.161

20 0.427 0.0378 1.35

Copper sulphate injected to the bed is 9.63 mmol within the 10 h.

Concentrations of copper ions in the external compart-

ments followed a predictable trend over the course of the

experiment, as shown in Fig. 6(b), (c). In most cases the linear

variations indicate a constant ion flux, at least in the 10-h

experiments. Because of the moderate size of the resin bed

and the restricted periods of each experiment, Cu2þ concen-

trations in both compartments were below 2 mmol/l. Values

of the average flux to the external compartments were

deduced by linear regression of the data during the period

mentioned. The slight non-linearity in the copper concentra-

tion observed with the 8% resin reduced the accuracy of the

determination. Values for fluxes into the external compart-

ments were shown in Fig. 7.

In the cathode compartment, the accumulation fluxwas an

increasing function of the current density. As for the removal

efficiency, the best results were obtained with the 4% resin, as

shown in Fig. 7(a). The flux of copper ions transferred to the

anode compartment decreased with the current density, as

shown in Fig. 7(b). High current densities favour migration

from the anode towards the cathode through the resin bed,

hindering transport by diffusion from the central compart-

ment. This tendency observed with the three resins did not

allow accurate interpretation of the phenomenon. However,

extrapolation of the data to zero current led to estimates for

the diffusion flux to the anode compartment with values

ranging from 5.76 to 9 10�3 mmol/cm2 h. Finally, the current

efficiency of the electrolytic operation varied from 24 to 47%,

depending on the current density and the resin grade. The

best current efficiencies for the transfer of Cu2þ to the cathode

were obtained with the mid-range of current density, in

particular for the two softest resins. For all resins, the current

efficiency rapidly decreased for the high current densities.

to the external compartments (CE); the mole amount ofount of copper deposited at the cathode after 10 h.

CEanodeCu2þ

%CEmetal

Cu

%CEcathode

Cu2þ

%CEcathode

Cu ðtotalÞ%

hCu2þ

%

18.99 0.57 38.35 38.91 29

9.98 1.00 38.77 39.77 28

2.89 3.04 30.67 33.72 33

16.76 0.39 46.23 46.63 39.89

14.91 2.39 40.08 42.48 38.96

6.06 2.03 31.94 33.97 41

16.42 0.42 41.34 41.76 31

7.62 0.67 24.06 24.73 31

a

b

c

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 1 2 3 4 5 6 7 8 9 10

InletOutlet at 10 Outlet at 13.3 Outlet at 20

Time (h)

Cu

conc

entr

atio

n (m

mol

/l)

40%

mA/cmmA/cm

mA/cm

Dowex 4%

0.00.20.40.60.81.01.21.41.61.82.0

0 1 2 3 4 5 6 7 8 9 10

10 mA/ 13.3 mA/ 20 mA/

(Dowex 4% ; Cathode compartment)

Time (h)

cm

cmcm

Cu

conc

entr

atio

n (m

mol

/l)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 1 2 3 4 5 6 7 8 9 10

10 mA/ 13.3 mA/ 20 mA/

Time (h)

(Dowex 4% ; Anode compartment) cm

cmcm

Cu

conc

entr

atio

n (m

mol

/l)

Fig. 6 e Treatment of the copper sulphate solution, with (Dowex 50 WX 4%). (a): Effect of the current density on the

concentration at the outlet of the bed versus time. (b): Effect of the current density on the copper ion concentration in the

cathode compartment versus time. (c): Effect of the current density on the copper ion concentration in the anode

compartment versus time.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 63372

4.3. Electrical energy consumption

Fig. 8, gives an overviewof the different specific electric energy

consumptions based on copper removal, calculated from the

sum of the total number of mole of copper transferred to the

external compartments taking into account the moles of

copper deposited at the cathode. As shown in Fig. 8, it was

found that the specific electric energy consumptions increased

with the current density. This specific electric energy

consumption depended also on the resin grade. Themoderate

flexible 4% resin gave the best results for the highest current

density. The corresponding electric energy consumption at

20mA/cm2was 1.93 and 2.67 kWh/molremoved copper in the case

of 2%, and 8% resins, respectively. The electrical energy

consumption was only 1.54 kW h/molremoved copper in the case

of the 4% resin.

On the other hand, Fig. 8 also reveals that the IXED

technique required much less energy than the ED technique

alone. For example, this technique is much more energy

efficient requiring less than 32 and 44% of the ED energy for

the low and high current density cases, respectively. These

findings seem to indicate that the application of the IXED is

a promising technique for the purification of metal-

containing waters.

a

b

0.0000.0020.0040.0060.0080.0100.0120.0140.0160.0180.020

8 10 12 14 16 18 20 22

Dowex 2% Dowex 4%HCR-S 8%

Spec

ific

flu

x of

cop

per

ion

(mm

ol/c

m2 h

)

Cathode compartment

Current density (mA/cm2)

0.000

0.001

0.002

0.003

0.004

0.005

0.006

8 10 12 14 16 18 20 22

Dowex 2% Dowex 4%HCR-S 8%

Current density (mA/cm2)

Anode compartment

Spec

ific

flu

x of

cop

per

ion

(mm

ol/c

m2 h

)

Fig. 7 e Specific flux of copper ion recovered in the external

compartments. (a): Specific flux of copper ion transferred to

the cathode compartment. (b): Specific flux of copper ion

transferred to the anode compartment.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 6 3373

4.4. Analysis of the CCD design

4.4.1. Standardized Pareto ChartIn order to design efficient processes, the response surface

methodology (RSM) is used to establish an empirical model

thatmaps the response surface using data from a designed set

of experiments to identify the direction and parameter ranges

for the response optimization.

The objectives of the present study were (i) to evaluate the

effects of the processing parameters in a large parameter

space, (ii) to determine if the separation enhancement, if any,

results only from electrical effects or from coupled IXED

Current density (mA/cm2)

Ene

rgy

cons

umpt

ion

(kW

h/m

ol re

mov

ed c

oppe

r)

0

2

4

6

8

10

12

023.3101

Dowex 2%Dowex 4%HCR-S 8%ED

Fig. 8 e Energy consumption during IXED as a function of

both the current density and cross-linking DVB.

effects, (iii) to maximize the abatement yield ðhCu2þ Þ and the

current yield ðCEcathodeCu ðtotalÞÞ and (iv) to minimize the energy

consumption. As a consequence, RSM is a most suitable

method for this optimization.

A central composite design (CCD) is one of the most useful

approaches in determining optimum conditions of many

processes (Cochran and Cox, 1957; Khuri and Cornell, 1987;

Mahmoud et al., 2008). Therefore, in this study, CCD was used

to point out the relationship existing between the response

function, abatement yield ðhCu2þ Þ, current yield ðCEcathodeCu ðtotalÞÞ

and the energy consumption, process variables, current

density (X1) and DVB cross-linking degree (X2). The range of

the independent variables for the IXED process conditions

were the current density, X1 (8e20 mA/cm2), DVB cross-

linking degree X2 (2e8%).

The model calculation based on the standardized values

allows a comparison of the relative influence of the factors on

the response, by comparing the square roots of the sum of

squares of all the coefficients related to a factor. To determine

which factors have a significant impact on the response

variable, the Pareto Chart was used. The standardized Pareto

Chart contains a bar for each effect, classed from the most

significant to the least significant. The length of each bar is

proportional to the standardized effect. The length of each bar

indicates the effect of these factors and the level of their

effects on responses. A vertical line (reference line) is drawn at

the location of the 0.05 critical value. Any bars that extend to

the right of that line indicate effects that are statistically

significant at the 5% significance level. Fig. 9 shows the stan-

dardized Pareto Chart and depicts the main effect of the

independent variables on the (i) abatement yield, (ii) current

yield and (iii) the energy consumption calculated per the total

amount of copper transferred to the external compartments

and copper deposited. It can be inferred, as shown in Fig. 9(a),

that the factor X1 (current density), X1X1, X1X2 and X2X2

extending behind the reference line, have a significant effect

on the abatement yield.

As shown in Fig. 9(a), the quadratic effect of DVB (X2X2) has

a bigger impact on the response than both the linear and

quadratic effect of the current density (X1) and the interaction

effect of X1X2.

Note that the factor X1 (current density) has the most

significant main effect at the 95% confidence level on the

current yield, followed by the quadratic effect of DVB and the

interaction effect of X1X2, as can be seen from Fig. 9(b).

Moreover, the factor X1 (current density) has the most signif-

icant main effect at the 95% confidence level on the energy

consumption, followed by the interaction effect of X1X2, as

can be seen from Fig. 9(c).

4.4.2. Optimization stepIn order to optimize the IXED variables for the response Y, the

RSM was used. The response surfaces for the abatement yield

ðhCu2þ Þ, the current yield ðCEcathodeCu ðtotalÞÞ and the energy

consumption generated by STATGRAPHICS� Centurion XVI.I

software are displayed in Fig. 10.

Using the analysis options dialog box (energy consumption

STATGRAPHICS� Centurion XVI.I software) of the RSM, the

range of each factor allowing the optimum abatement yield

and current yield to be obtained is determined. After

a

b

c

0 2 4 6 8 10 12

+

-

0 2 4 6 8 10

+

-

0 2 4 6 8 10

+

-

Standardized effect

Standardized effect

Standardized effect

Reference line

Reference line

Reference line

X1

X2

X2X2

X1X2

X1X1

X1

X2

X2X2

X1X2

X1X1

X1

X2

X2X2

X1X2

X1X1

Fig. 9 e Standardized Pareto Chart showing the effects of

the independent variables X1 (current density) and X2 (DVB)

and their combined effects on the abatement yield (a), the

current yield (b) and the energy consumption (c).

Fig. 10 e Estimated response surface for the abatement

yield (a), the current yield (b) and the energy consumption

(c) as a simultaneous function of X1 (current density) and X2

(DVB).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 63374

generating the polynomial equations relating the dependent

and independent variables, the process was optimized for the

response Y. The optimization was performed to obtain the

levels of X1 � X2 which maximize Y (abatement yield and

current yield).

As discussed previously, the energy consumption is very

important as well. Therefore, the best combination of process

variables for the energy consumption response function was

determined in order to limit the energetic cost of the process

and simultaneously obtain a satisfactory abatement yield and

current yield. The optimum processing conditions were

a current density of 9.6 mA/cm2 and DVB of 5%, which gave an

abatement yield, current yield and energy consumption of

36.63%, 45.13% and 0.531 kW h/molremoved copper, respectively.

5. Conclusions

The IXED was successfully used for the removal of heavy

metals from moderately acidified copper sulphate solutions

simulating rinsing water of copper plating lines, under

a variety of processing conditions ranging from 10 to 20 mA/

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 6 3375

cm2 and from 2 to 8% DVB. Importantly, it has been illustrated

that the IXED can be used to remove a significant proportion of

the copper ions from moderately dilute copper solutions, as

well as to produce a concentrated metal salt solution, which

could be recycled. It was also found that the IXED technique

requires less than 32 and 44% of the ED energy for the

moderate and high DVB cross-linking degree, respectively.

These findings seem to indicate that the hybrid process seems

to be a promising technique for improving the electrical effi-

ciency of treating dilute wastewater.

The analysis from the RMS appeared very well suited to

predict the optima set of operating variables in order to ach-

ieve a satisfactory treatment (abatement yield and current

yield) using the minimum amount of electricity.

Acknowledgement

This research was performed at Laboratoire des Sciences du

Genie Chimique, CNRS-ENSIC, Nancy, France. One of the

authors (A. Mahmoud) gratefully acknowledges F. Lapicque

and L. Muhr, my supervisors, for their kind help and

suggestions.

r e f e r e n c e s

Adhoum, N., Monser, L., Bellakhal, N., Belgaied, J.E., 2004.Treatment of electroplating wastewater containing Cu2þ, Zn2þ

and Cr(VI) by electrocoagulation. Journal of HazardousMaterials 112 (3), 207e213.

Alvarado, L., Ramırez, A., Rodrıguez-Torres, I., 2009. Cr(VI)removal by continuous electrodeionization: study of its basictechnologies. Desalination 249 (1), 423e428.

Arar, O., Yuksel, U., Kabay, N., Yuksel, M., 2011. Removal of Cu2þ

ions by a micro-flow electrodeionization (EDI) system.Desalination 277 (1e3), 296e300.

Basha, C.A., Selvi, S.J., Ramasamy, E., Chellammal, S., 2008.Removal of arsenic and sulphate from the copper smeltingindustrial effluent. Chemical Engineering Journal 141 (1e3),89e98.

Baticle, P., Kiefer, C., Lakhchaf, N., Leclerc, O., Persin, M.,Sarrazin, J., 2000. Treatment of nickel containing industrialeffluents with a hybrid process comprising of polymercomplexation ultrafiltration-electrolysis. Separation andPurification Technology 18 (3), 195e207.

Cochran, W.G., Cox, G.M., 1957. Some methods for the study ofresponse surfaces. In: Experimental Designs, second ed. JohnWiley and Sons, New York.

Dejean, E., Laktionaov, E., Sandeaux, J., Sandeaux, R.,Pourcelly, G., Gavach, C., 1997. Electrodeionization with ion-exchange textile for the production of high resistivity water:influence of the nature of the textile. Desalination 114 (2),165e173.

Demkin, V.I., Tubashov, Y.A., Panteleev, V.I., Karlin, Y.V., 1987.Cleaning low mineral water by electrodialysis. Desalination64, 367e374.

Dermentzis, K., Christoforidis, A., Valsamidou, E., 2011a. Removalof nickel, copper, zinc and chromium from synthetic andindustrial wastewater by electrocoagulation. InternationalJournal of Environmental Sciences 1 (5), 697e710.

Dermentzis, K., Christoforidis, A., Papadopoulou, D., AnthimosDavidisa, A., 2011b. Ion and ionic current sinks for

electrodeionization of simulated cadmium plating rinsewaters. Environmental Progress & Sustainable Energy 30 (1),37e43.

Dzyazko, Yu.S., Belyakov, V.N., 2004. Purification of a dilutednickel solution containing nickel combining ion exchange andelectrodialysis. Desalination 162, 179e189.

Dzyazko, Yu.S., 2006. Purification of a diluted solution containingnickel using electrodeionization. Desalination 198 (1e3),47e55.

Dzyazko, Yu.S., Rozhdestvenskaya, L.M., Vasilyuk, S.L.,Belyakov, V.N., Kabay, N., Yuksel, M., Arar, O., Yuksel, U.,2008a. Electro-deionization of Cr (VI)-containing solution. PartI: chromium transport through Inorganic ion-exchanger.Chemical Engineering Communications 196 (1e2), 3e21.

Dzyazko, Yu.S., Vasilyuk, S.L., Rozhdestvenskaya, L.M.,Belyakov, V.N., Stefanyak, N.V., Kabay, N., Yuksel, M., Arar, O.,Yuksel, U., 2008b. Electro-deionization of Cr (VI)-containingsolution. Part II: chromium transport through inorganic ion-exchanger and composite ceramic membrane. ChemicalEngineering Communications 196 (1e2), 22e38.

Feng, X., Wu, Z., Chen, X., 2007. Removal of metal ions fromelectroplating effluent by EDI process and recycle of purifiedwater. Separation and Purification Technology 57 (2), 257e263.

Feng, X., Gao, J.-S., Wu, Z.-C., 2008. Removal of copper ions fromelectroplating rinse water using electrodeionization. Journalof Zhejiang University: Science A 9 (9), 1283e1287.

Fu, L., Wang, J., Su, Y., 2009. Removal of low concentrations ofhardness ions from aqueous solutions usingelectrodeionization process. Separation and PurificationTechnology 68 (3), 390e396.

Ganzi, G.C., Egozy, Y., Guiffrada, A.J., 1987. Deionisation-highpurity water by electrodeionization performance of the ion-pure continuous deionization system. Ultrapure Water 4 (3),43e50.

Ganzi, G.C., Parise, P.L., 1990. The production of pharmaceuticalgrades of water using continuous deionization post-reverseosmosis. Journal of Parenteral Science and Technology 44 (4),231e241.

Ganzi, G.C., Jna, A.D., Dimascio, F., Wood, J.H., 1997.Electrodeionization-theory and practice of continuouselectrodeionization. Ultrapure Water 14 (6), 64e69.

Gifford, J.D., Atnoor, D., 2000. An innovative approach tocontinuous electrodeionization module and system design forpower application. In: International Water Conference,Pittsburgh, PA. (Oct. 23e25, 2000).

Glueckauf, E., 1959. Electro-deionisation through a packed bed.British Chemical Engineering 4 (12), 646e651.

Grabowski, A., Zhang, G., Strathmann, H., Eigenberger, G., 2006.The production of high purity water by continuouselectrodeionization with bipolar membranes: influence of theanion-exchange membrane permselectivity. Journal ofMembrane Science 28 (1e2), 297e306.

Grebenyuk, V.D., Chebotareva, R.D., Linkov, N.A., Linkov, V.M.,1998. Electromembrane extraction of Zn from Na-containingsolutions using hybrid electrodialysis-ion exchange method.Desalination 115 (3), 255e263.

Guana, S., Wang, S., 2007. Experimental studies onelectrodeionization for the removal of copper ions from dilutesolutions. Separation Science and Technology 42 (5), 949e961.

Guiffrida, A.J., Ganzi, G.C., Oren, Y., 1990. EletrodeionizationApparatus and Module, US Patent 4956071.

Helfferich, F., 1962. Ion Exchange. McGraw-Hill, New York, 624 pp.Janssen, L.J.J., Koene, L., 2002. The role of electrochemistry and

electrochemical technology in environmental protection.Chemical Engineering Journal 85 (2e3), 137e146.

Juttner, K., Galla, U., Schmieder, H., 2000. Electrochemicalapproaches to environmental problems in the processindustry. Electrochimica Acta 45 (15e16), 2575e2594.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 3 6 4e3 3 7 63376

Khuri, A.I., Cornell, J.A., 1987. Response Surface, Designs andAnalysis. 703 Marcel Dekker, New York.

Kunz, G.K., 1987. Process and Apparatus for Treatment of Fluids,Particularly Desalinization of Aqueous Solution, US Patent4636296.

Lu, H., Yan, B., Wang, J., Fu, X., 2007. Recovery of nickel ions fromdilute solutions by electrodeionization process. HuagongXuebao/Journal of Chemical Industry and Engineering (China)58 (5), 1259e1261.

Lu, H., Wang, J., Yan, B., Bu, S., 2010. Recovery of nickel ions fromsimulated electroplating rinse water by electrodeionizationprocess. Water Science and Technology 61 (3), 729e735.

Lu, H., Wang, J., Bu, S., Fu, L., 2011. Influence of resin particle sizedistribution on the performance of electrodeionizationprocess for Ni2þ removal from synthetic wastewater.Separation Science and Technology 46 (3), 404e408.

Mahmoud, A., Muhr, L., Vasiluk, S., Aleynikoff, A., Lapicque, F.,2003. Investigation of transport phenomena in a hybrid ionexchange-electrodialysis system for the removal of copperions. Journal of Applied Electrochemistry 33 (10), 875e884.

Mahmoud, A., 2004. Phenomenes de transfert dans un proceded’electrodeionisation de solutions diluees de cuivre: etudeexperimentale et modelisation, PhD dissertation, INPL, Nancy,France (in French).

Mahmoud, A., Muhr, L., Grevillot, G., Lapicque, F., 2007.Experimental tests and modelling of an electrodeionizationcell for the treatment of dilute copper solutions. CanadianJournal of Chemical Engineering 85 (2), 171e179.

Mahmoud, A., Fernandez, A., Arlabosse, P., 2008. Thermally-assisted mechanical dewatering (TAMD) of suspensions of fineparticles: analysis of the influence of the operating conditionsusing the response surface methodology. Chemosphere 72(11), 1765e1773.

Monzie, I., Muhr, L., Lapicque, F., Grevillot, G., 2005. Mass transferinvestigations in electrodeionization processes using themicrocolumn technique. Chemical Engineering Science 60 (5),1389e1399.

Priya, P.G., Basha, C.A., Ramamurthi, V., Begum, S.N., 2009.Recovery and reuse of Ni(II) from rinsewater of electroplatingindustries. Journal of Hazardous Materials 163 (2e3), 899e909.

Smara, A., Delimi, R., Chainet, E., Sandeaux, J., 2007. Removal ofheavy metals from diluted mixtures by a hybrid ion-exchange/electrodialysis process. Separation and PurificationTechnology 57 (1), 103e110.

Song, J.H., Yeon, K.H., Moon, S.H., 2004. Transport characteristicsof Co2þ through an ion exchange textile in a continuouselectrodeionization (CEDI) system under electro-regeneration.Separation Science and Technology 39 (15), 3601e3619.

Soria, P., Saporta, M., Berdun, C., 1993. Preparation of deionizedfruit juice using continuous deionization. Flussiges Obst 60(10), 368e371.

Souilah, O., Akretche, D.E., Amara, M., 2000. Water reuse of anindustrial effluent by means of electrodeionisation.Desalination 167, 49e54.

Spoor, P.B., ter Veen, W.R., Janssen, L.J.J., 2001.Electrodeionization 2: the migration of nickel ions absorbed ina flexible ion-exchange resin. Journal of AppliedElectrochemistry 31 (10), 1071e1077.

Spoor, P.B., Grabovska, L., Koene, L., Janssen, L.J.J., ter Veen, W.R.,2002. Pilot scale deionisation of a galvanic nickel solutionusing a hybrid ion-exchange/electrodialysis system. ChemicalEngineering Journal 89 (1e3), 193e202.

Sudilovskiy, P.S., Kagramanov, G.G., Kolesnikov, V.A., 2008. Use ofRO and NF for treatment of copper containing wastewaters incombination with flotation. Desalination 221 (1e3), 192e201.

Tessier, D.F., Toupin, J.D., Towe, I.G., 1997. ElectrodeionizationApparatus Having Geometric Arrangement of Ion ExchangeMaterial, Patent WO 9725147.

Thate, S., Specogna, N., Eigenberger, G., 1999. Electrodeionization-a comparison of difference EDI concepts used for theproduction of high-purity water. UltrapureWater 16 (8), 42e56.

TSS Water Course electro-deionization: www.techsup.com.Vasilyuk, S., Maltseva, T.S., Belyakov, V.N., 2004. Influence of

water hardness on removal of copper ions by ion-exchange-assisted electrodialysis. Desalination 162, 249e254.

Wasylkiewicz, S., 1990. Ion association in aqueous solutions ofelectrolytes. II. Mathematical model for sulphates of bivalentmetals. Fluid Phase Equilibria 57 (3), 277e296.

Wood, J., Gifford, J., 2002. Improvements in continuouselectrodeionization for plant applications. In: Industrial WaterConference, Orlando, FL. (Dec. 11e12, 2002).

Xing, Y., Chen, X., Yao, P., Wang, D., 2009a. Continuouselectrodeionization for removal and recovery of Cr(VI) fromwastewater. Separation and Purification Technology 67 (2),123e126.

Xing, Y., Chen, X., Wang, D., 2009b. Variable effects on theperformance of continuous electrodeionization for theremoval of Cr(VI) from wastewater. Separation andPurification Technology 68 (3), 357e362.

Yeon, K.H., Seong, J.H., Rengaraj, S., Moon, S.H., 2003.Electrochemical characterizationof ion-exchange resinbeds andremoval of cobalt by electrodeionization for high purity waterproduction. Separation Science and Technology 38 (2), 443e462.

Zouari, I., Lapicque, F., 1992. An electrochemical study of zincdeposition in a sulfate medium. Electrochimica Acta 37 (3),439e446.