Desalination 363 (2015) 44–50

Contents lists available at ScienceDirect

Desalination
journal homepage: www.elsevier.com/locate/desal

A combined ion exchange–nanofiltration process for water desalination:

I. sulphate–chloride ion-exchange in saline solutions
Nidal Hilal a,b,⁎, Victor Kochkodan b, Hasan Al Abdulgader c, Stephen Mandale a, Saad A. Al-Jlil d
a
Centre for Water Advanced Technologies and Environmental Research (CWATER), College of Engineering, Swansea University, Swansea SA2 8PP, United Kingdom
b
Qatar Environment and Energy Research Institute (QEERI), Dopha, Qatar
c
College of Engineering, King Faisal University, Al Ahsa, Saudi Arabia
d
King Adbul Aziz City for Science and Technology, Saudi Arabia

H I G H L I G H T S

• Sulphate/chloride ion exchange in saline water has been studied both on laboratory and pilot scales.
• Sulphate–chloride exchange depends on the feed salt concentration and the nature of the functional group of the resin.
• Exhausted anion exchange resins were successfully regenerated using 0.2 M Na2SO4 solution.
• The osmotic pressure of sea water was significantly reduced after SO4/Cl ion exchange.

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

Article history: Removal of chloride ions from saline water with seven different ion-exchange (IX) resins in sulphate form has
Received 14 September 2014 been studied both on laboratory and pilot scales. It was found that sulphate–chloride exchange is very fast in
Received in revised form 16 November 2014 aqueous solutions and that the feed salt concentration and the nature of the functional group of the resin play
Accepted 18 November 2014
an important role in this process. It was shown that the chloride/sulphate separation factor depends on salt
Available online 21 November 2014
content in feed water and the higher the substitution of hydrogen atoms in amine functional group of anion
Keywords:
exchange resin the higher chloride over sulphate selectivity. Exhausted IX resins were successfully regenerated
Desalination using 0.2 M Na2SO4 solution and multiple regeneration/saturation cycles proved that this did not affect the resin's
Anion exchange resin performance on chloride ion removal. It was shown that the osmotic pressure of sea water was significantly
Sulphate/chloride exchange reduced after SO4/Cl ion exchange. Due to the drop of osmotic pressure lower energy consuming nanofiltration
Osmotic pressure membranes compared with reverse osmosis membranes might be used for salty water desalination after IX
treatment.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction desalting brackish and/or seawater at the lowest possible cost is of

A sharp growth in the worlds' population coupled with urbanization
has resulted in a rapidly increased demand for fresh water [1]. A lot of
efforts are focused on suitable methods to obtain freshwater by sea
water or brackish water desalination and various desalination technol-
ogies, including thermal multi-stage flash distillation and reverse osmo-
sis, which have been increasingly used to enhance a fresh water supply
around the world [2]. However, the cost-effectiveness of these technol-
ogies is still hampered by high energy consumption [3]. Therefore
the development of a technology capable of producing fresh water by
⁎ Corresponding author at: Centre for Water Advanced Technologies and Environmental
Research (CWATER), College of Engineering, Swansea University, Swansea SA2 8PP, United
Kingdom.
E-mail address: n.hilal@swansea.ac.uk (N. Hilal).
crucial importance.
In this work we have studied the sulphate–chloride exchange in
saline solutions. If monovalent chloride ions of salty water convert
into divalent sulphate ions using ion-exchange (IX), that would reduce
the osmotic pressure of the treated feed and in turn would essentially
reduce the energy required for further membrane desalination, because
nanofiltration (NF) membranes, which efficiently reject the sulphate
ions and operate at significantly lower transmembrane pressures
compared with reverse osmosis membranes, can be used for water
desalination [4].
Many researchers have tested IX resins for the removal of chloride
ions, but in exchange of anions other than sulphate [5–8]. For example,
the removal of chloride, nitrate and sulphate ions from aqueous solu-
tions with a macroporous resin Amberlite IRN 9766 in hydroxyl form was
studied by Dron and Dodi [5]. It was shown that Langmuir adsorption

http://dx.doi.org/10.1016/j.desal.2014.11.016
0011-9164/© 2014 Elsevier B.V. All rights reserved.
 N. Hilal et al. / Desalination 363 (2015) 44–50 45

isotherms provide a good estimation of the sorption capacity on the
contrary to Freundlich and Dubinin–Radushkevitch models.
The strongly basic ion exchanger Amberlite IRA-420 resin in OH form
has been used for chloride ion removal from chloride-polyethylenimine
solution at different temperatures to evaluate the possibility to apply the
IX technology to convert the polyethylenimine flocculant in chloride
form into an adhesive product for printing applications [6]. The equilib-
rium isotherms of chloride ions in aqueous solution on Amberlite IRA-
420 have been obtained and the kinetic studies indicated that the chlo-
ride ions are slowly removed when the polyethylenimine is in solution
due to the ionic pair formation.
The layered double hydroxides such ZnAl–NO3 containing nitrate as
the interlayer anion has been studied as an anion exchanger material to
remove chloride ions from aqueous solution [9]. The effect of dosage of
IX material, solution pH and temperature on the removal efficiency have
been investigated. It was found that the influence of solution pH was not
very significant in the range 5.0–8.0, and the rate of anion-exchange in-
creases upon raising the temperature of aqueous solution.
Chloride–sulphate exchange for sulphate removal from the sea
water fed to an evaporation plant by means of anion exchange resins
has been studied by Boari et al. [10,11]. It was shown that the selectivity
for the bivalent ion depends strongly on the basicity of the resin and the
affinity of every resin for the sulphate ion increases with the dilution of
the aqueous phase and with the equilibrium temperature. The obtained
results showed that in heterovalent exchange processes an important
role is played by the phenomena connected with the electrostatic
interactions.
Direct sulphate–chloride ion exchange for removal of chloride ions
from water has been studied by few investigators. Sarkar and SenGupta
[12] illustrated that sulphate/chloride selectivity depends on degree of
the substitution of hydrogen atom in the amine functional group of
the anion exchange resin. For most of their experiments Purolite A850
anion-exchange resin of a gel type with quaternary ammonium func-
tional group was used.
The aim of this work is to use various commercial anion exchange
resins and test their performance in exchange with chloride ions at
different water salinities. The IX studies were carried out in batch and
continuous modes to investigate the sulphate–chloride exchange equi-
librium, effect of resin dosage and operating conditions on the IX pro-
cess and on the shape of breakthrough curves of the resins. For the
first time we have used two mutually interlinked routes to study sul-
phate–chloride exchange in aqueous solutions, which included both
It should be mentioned that Purolite A500TLSO4, Purolite A400TLSO4
and Ambersep 900SO4 resins are available in SO2− 4 form, while Purolite
A850 resin was shipped in Cl− form and it was transferred to SO2− 4 form
by treatment with 2 M H2SO4.
Purolite A109, Purolite A149S and Purolite A111 resins were provid-
ed in free base form and according to a procedure suggested by the
manufacture the resins were treated with 5 wt.%. Na2SO4 and washed
with distilled water to convert the resin from free base to sulphate form.
Purolite A500TLSO4, Purolite A850, Purolite A109, Purolite A149S,
Purolite A111 and Ambersep 900SO4 are macroporous resins, while
Purolite A400TLSO4 is in a gel form.
2.2. Batch studies with IX resins
The batch studies with IX resins were performed using NaCl
solutions of various concentrations, considering that sodium chloride
(i.e. Na+ and Cl− ions) is the primary constituent of brackish water or
sea water. NaCl stock solution (32.0 g/L) was prepared by weighing
and dissolving an appropriate amount of NaCl salt in Millipore-Q
water. The testing solutions were prepared by an appropriate diluting
of the stock NaCl solution. The pH values of the solutions were about
7, as was measured with a JenWay 3540 pH metre.
For batch IX studies, weighted amounts of air-dried sulphate-forms
of the resins were added in NaCl solutions with concentrations of
30–550 meq/L. Then the samples were agitated with a shaker (Innova
44, New Brunswick Scientific) at a constant speed of 150 rpm for
0.1–24.0 h at room temperature. With time, 0.5–1.0 mL aliquots were
periodically collected and analyzed to determine contents of Cl− and
SO2−4 ions in the probes using an ion chromatograph (Dionex IC900)
with AS14A (4 mm) anion-exchange column.
The removal efficiency of Cl− ions with IX resins was calculated as:
Removal efficiency ¼ ½ðc0 −cÞ=c0   100%;
where c0 and c are chloride concentrations (mg/L) in the solution before
and after IX treatment.
2.3. Column studies with IX resins
Column IX tests were performed to elucidate a shape of break-
through curves of IX resins and to evaluate the effect of operating
−
conditions for optimization of SO2−
4 /Cl exchange process. A glass chro-

the experimental studies on a laboratory scale and further scaling up
and optimization of a pilot scale IX system.
2. Materials and methods
2.1. IX resins
Seven IX resins, Purolite A500TLSO4, Purolite A400TLSO4, Purolite
A850, Purolite A109, Purolite A149S, Purolite A111 and Ambersep
900SO4 which contain quaternary, tertiary, secondary and primary
aminogroups in the polymer matrix have been used in this work. The
characteristics of the resins are presented in Table 1.
matographic type column with an internal diameter of 2.0 cm and the
column aspect ratio (height: width) of 20:1 was used for these
experiments. The feed solution was delivered to the column using a
peristaltic pump (Watson-Marlow 101 U/R) at a flow rate of 0.12–
0.45 mL/s. From the outlet of the column, the effluent fractions of
20 mL each were collected and analyzed with IC.
2.4. Pilot scale IX system
To progress the experimental work to pilot scale trials a pilot scale IX
system has been designed and built. The process and instrumentation
diagram (P&ID) of the designed IX system is presented in Fig. 1. The sys-
tem is constructed from polyvinylchloride pipe work and two 100 L

Table 1
Physical and chemical characteristics of IX resins used in the study.

IX resin Polymer matrix Functional group Shipped form Resin particle size

Purolite A500TLSO4 Polyacrylic crosslinked with divinylbenzene Quaternary ammonium SO2−

4 425–850 μm
Ambersep 900SO4 Polystyrene crosslinked with divinylbenzene Trimethyl ammonium SO2−
4 300–1200 μm
Purolite A400TLSO4 Polystyrene crosslinked with divinylbenzene Quaternary ammonium SO2−
4 425–850 μm
Purolite A109 Polystrene crosslinked with divinylbenzene Primary amine Free base 425–1000 μm
Purolite A149S Polystrene crosslinked with divinylbenzene Secondary amine Free base 425–1200 μm
Purolite A111 Polystrene crosslinked with divinylbenzene Tertiary amine Free base 300–1200 μm
Purolite A850 Polyacrylic crosslinked with divinylbenzene Quaternary ammonium Cl− 300–1200 μm
46 N. Hilal et al. / Desalination 363 (2015) 44–50

Fig. 1. P&ID of a pilot IX system.

poly-glass IX columns. Three large capacity tanks (1000 L) are used to
store and supply the feed, regeneration and rinse solutions required
for IX operation. A speed controllable pump P3 and fixed speed
pumps P4 and P5 have been used for delivering the solutions to the IX
columns. All of the equipment is supported on a stainless steel frame
mounted on castors. Performance monitoring is facilitated by inclusion
of a flow metre and a pressure gauge linked back to a central display on
the control panel.
The designed pilot IX rig for water desalination have been used to
study the effect of flow rate through the IX column, the resins loading
and feed salinity on the efficiency of sulphate/chloride exchange
process.
Osmotic pressure of saline water before and after IX treatment
has been measured with Gonotec Osmomat 030 osmometer (Gonotec
GmbH, Germany).
The very fast kinetics of sulphate–chloride IX process suggests that
in aqueous solutions of high salinity IX resistances from diffusion of
the counterions both through a liquid film surrounding the IX resin in
solution (Nernst film with no convection) and inside IX resins are insig-
nificant. This assumption is confirmed by the data presented in Fig. 3. As
can be seen from this figure, the prolonging of IX treatment from 1 h to
42 h did not practically lead to increase the removal of chloride ions
from water.
The dependencies of dosage of Ambersep900SO4, Purolite A500TLSO4
and Purolite A500TLSO4 resins on concentration of Cl− and SO−2 4 ions in
feed water at IX treatment of NaCl solutions over a concentration range
of 1.0–32.0 g/L are presented in Fig. 4. As expected, a content of chloride
ions in the feed solutions decreases with an increase of the resin dosage.
This is attributed to higher IX capacity with an increase of the resin load-
ing. Addition of more resin beads will increase the total number of avail-
able exchange sites and subsequently, the capacity to take in more

3. Results and discussion 7000

3.1. Batch studies on sulphate–chloride exchange with IX resins 6000

Purolite A500TLSO4
Cl- Concentraon, mg/L

A set of experiments were conducted to determine the time required 5000 Purolite A400TLSO4
for IX resin in sulphate form to reach SO24 −/Cl− equilibrium in saline Ambersep 900SO4
4000
water in accordance with the following reaction:
3000
(RN+)2SO42− + 2NaCl⇔ 2(RN+)Cl− + SO42− + 2Na+,
2000
where (RN+)2SO2− 4 is a polymer matrix of IX resin in sulphate form.
It was found that the chloride–sulphate exchange proceeds very fast 1000
in aqueous solutions of high salinity. For example, as seen in Fig. 2, only
0
5–10 min is needed to reach SO2−/ 4 Cl− equilibrium in 10 g/L NaCl 0 100 200 300 400 500
solution. The chloride removal efficiency after equilibrium is in the Time, min
range of 54 –59% for all three resins. Purolite A500TLSO4 managed to
remove 59% of chloride ions in the solution, while Purolite A400TLSO4 Fig. 2. Concentration of chloride ions over time during IX treatment of 10.0 g/L NaCl solu-
and Ambersep900SO4 removed 56% and 54% respectively. tion with Purolite A500TLSO4, Purolite A400TLSO4 and Ambersep 900SO4 resins.
 N. Hilal et al. / Desalination 363 (2015) 44–50 47

30
where YCl and YSO4 denote equivalent fractions of chloride and sulphate
25 ions in resin, while XSO4 and XCl are equivalent fractions of sulphate and
chloride ions in the aqueous phase, respectively.
Removal, %

20 To investigate a possibility to manipulate the chloride/sulphate sep-

42h
15 aration factor by controlling the degree of substitution of the hydrogen
1h atom in the amine functional groups we have used Purolite A109,
10 Purolite A149S, Purolite A111 and Purolite AT500SO4 resins, which con-
tain primary, secondary, tertiary and quaternary groups in the polymer
5
matrix, respectively. All these resins are of macroporous anion exchange
0 type with polysterene matrix cross-linked with divinylbenzene.
0,2 0,5 1 3 5 8 Based on the data on IX equilibrium, the IX isotherms for sulphate–
Resin dosage, g/100 ml chloride exchange have been calculated. The values of separation factor
can be determined from a constant-temperature equilibrium plot of
Fig. 3. Removal of chloride ions from 32.0 g/L NaCl solution after IX treatment with resin-phase concentration of chloride ion versus its aqueous-phase con-
Ambersep 900SO4 resin at various resin dosages. Contact time is 1 h and 42 h. centration as shown in Fig. 5.
As seen in this figure the relative chloride affinity over sulphate for
chloride ions will increase. It is seen that, all resins performed quite sim- different amine functional groups can be present as Quaternary
ilarly with 10–15% variation between them in chloride removal efficien- (Purolite AT500SO4) N Tertiary (Purolite A111) N Secondary (Purolite
cy. Nevertheless, Purolite A500TLSO4 resin managed to outperform the A149) N Primary (Purolite A109). It may be concluded that the higher
other resins in the chloride removal tests. the substitution of the hydrogen atom in amine functional group by
It might be also seen in Fig. 4 that chloride removal from water with an alkyl group is the higher chloride over sulphate selectivity for the
resins increases with a decrease of feed salinity and, for example, for anion exchange resin. The obtained data are in good agreement with
Purolite A500SO4 resin the removal reaches 80, 68, 54 and 32% for previous studies on anion exchange resins demonstrated that the chem-
NaCl concentrations of 1.0; 5.0; 10.0; and 32.0 g/L respectively. ical nature of amine functionality in IX resins correlated to divalent–
monovalent anion selectivity [12,14,15].
−
3.2. Effect of amine functional group in IX resins on SO2−
4 /Cl exchange in It was found that the efficiency of sulphate–chloride exchange
saline conditions essentially depends on the concentration of feed solutions. As seen in
Fig. 6, a strong base anion exchange Purolite AT500 SO4 resin depicts
The relative preference of chloride over sulphate ion in IX process “favourable” isotherms (convex to X-axis) at NaCl concentrations
with the anion-exchange resin may be evaluated using a separation fac- higher than 0.1 eq./L. At this condition chloride ion, with equivalent
tor αCl
SO4, which represented as [13]: fractions on the resin and in the solution plotted on Y and X axes, is pre-
ferred to sulphate ion during the IX process. At NaCl concentration in
the solution of 0.02 eq./L the resin depicts “unfavourable” isotherm
Cl Y Cl X SO4 (convex to Y-axis) that means that sulphate ions are preferred to chlo-
α SO4 ¼ 
X Cl Y SO4 ride ions. This phenomenon of reducing preference for higher-valent

100 80
Removal effeciency, %

Removal effeciency, %

80
60
60 A500TLSO4
A500TLSO4
40
40 A400TLSO4
A400TLSO4
20 20
Ampers ep 900
0 Ampersep 900
0
0 2 4 6
0 1 2 3 4 5 6
Resin dosage, g Resin dosage, g
a) b)
80 40
Removal effeciency, %

Removal effeciency, %

60 30
A500TLSO4
A500TLSO4
40 20
A400TLSO4 A400TLSO4
20 10
Ampersep 900 Ampersep 900
0 0
0 2 4 6 0 2 4 6
Resin dosage, g Resin dosage, g
c) d)

Fig. 4. Removal of Cl− ions in (a) 1.0 g/L; (b) 5.0 g/L; (c) 10.0 g/L; (d) 32.0 g/L NaCl solutions at different dosages of Purolite A500TLSO4, Purolite A400TLSO4 and Ambersep 900SO4 resins.
48 N. Hilal et al. / Desalination 363 (2015) 44–50

10 3200

2400
Separaon factor

1 Chloride (3)

C, meq/l
0 0,1 0,2 0,3 0,4 0,5 1600 Sulphate (3)
Chloride (2)
Sulphate (2)
0,1 AT500 800 Chloride (1)
A111 Sulphate (1)
A149 0
A109 0 5 10 15 20 25 30
0,01 Bed volume
NaCl concentraon, eq/L
Fig. 7. Breakthrough curves of chloride and sulphate ions during filtration of 5 g/L NaCl
Fig. 5. αCl
SO4 separation factor for Purolite resins with different amino groups at various feed solution through a column with (1) Purolite A400TLSO4, (2) Ambersep 900SO4 and
NaCl concentrations. (3) Purolite A500TLSO4 resins.

ions with increasing ionic strength of the solution has been called
“electroselectivity” and it might be explained in terms of the Donnan
potential [16]. As the aqueous phase dilution increases the Donnan
potential attracts the mobile charges with a force that is proportional
to their charge density. At low feed concentrations therefore, the diva-
−
lent SO2−
4 , ion will be preferred to the Cl ion. The absolute value of
Donnan potential decreases as the aqueous concentration increases
and will eventually become zero at certain values of the concentrations.
These values are typical for each particular resin and once they are
exceeded, there will be an inversion of selectivity when the separation
factor αCl
SO4 = 1.
As seen in Fig. 6, for the strong base anion exchange resin Purolite
AT500SO4 the reversal of chloride–sulphate selectivity occurs at NaCl
concentration of approximately 0.1 eq./L. At higher NaCl concentration
the separation factor αClSO4 N 1 and this means that chloride ion will be
preferred to the sulphate ion for the resin. The higher the solution con-
centration, the higher the αClSO4 separation factor (Fig. 5).
−
3.3. Small scale IX column studies on SO2−
4 /Cl exchange
Fig. 7 illustrates the breakthrough curves of Cl− and SO4 2−.ions with
various IX resins during IX treatment of 5 g/L NaCl solution. It is seen
that all the resins were able to remove almost all chloride ions in the
first 6–7 bed volumes (BV) of the feed solution but the breakthrough
points for the Purolite resins started approximately one BV later than
for Ambersep 900SO4 resin.
Fig. 8 compares breakthrough curves of chloride and sulphate ions at
different NaCl feed concentrations using Purolite A500TLSO4 resin. As
seen in this figure the breakthrough points and the sharpness of the
breakthrough curves are strongly affected by the level of NaCl concen-
tration in the feed. At high concentration of NaCl solutions, saturation
of IX resin proceeds faster (a breakthrough point shifts to smaller BV
values), but breakthrough curves for chloride ions are sharper.
−
3.4. Pilot scale IX column studies on SO2−
4 /Cl exchange
An effect of flow rate through the IX column in a range of 1.0 –5.0
L/min on chloride removal from feed NaCl solution is shown in Fig. 9.
As seen in this figure, the slower flow velocities through the column the
sharper the breakthrough curve for Cl ions and the shorter the length of
mass transfer zone. Higher flow rates through the column tend to flat-
ten the breakthrough curve and lengthen the mass transfer zone.
These finding may be explained by the fact that slow flow rate through
the column facilitates the reaching of ion exchange sites in the resin
beds with targeted chloride ions.
An effect of resin loading in the column in a range of 10.0–50.0 L on
chloride removal from NaCl solution is shown in Fig. 10. As seen in this
figure, an increase in resin dosage sharply elongates the operation time
until the breakthrough point of Cl ions takes place in treated water.
3.5. Osmotic pressure of salty water after IX treatment
It is known that osmotic pressure of the solution depends on the
number of osmotic active particles in the solutions according to the
Vant-Hoff equation [17]:
π ¼ nCRT;
where π is osmotic pressure in kPa; n is number of osmotic active parti-
cles per mole; C is molar concentration in mol/L; R is universal gas con-
stant (8.31441 N·m/mol·K) and T is temperature in K.
1

1
Relave concentraon

0,8 Chloride, 5 g/L

Sulphate, 5 g/L
0,8
0,6 Chloride, 10 g/L
Sulphate, 10 g/L
0,6 0,4 Chloride, 15 g/L
Y,Cl

Sulphate, 15 g/L
0,4 0,2 Chloride, 32 g/L
Sulphate, 32 g/L
0.1 eq/L
0,2 0
0.02 eq/L 0 5 10 15 20
0.2 eq/L Bed volume
0
0 0,2 0,4 0,6 0,8
X, Cl Fig. 8. Breakthrough curves of the chloride and sulphate ions during filtration of NaCl
solution through the column with Purolite A500TLSO4 resin at various feed concentra-
Fig. 6. Chloride IX isotherms for Purolite AT500SO4 resin during treatment of NaCl solu- tions, g/L: 5, 10, 15 and 32. Relative ion concentration corresponds to outlet over inlet IX
tions of different concentrations. concentration.
 N. Hilal et al. / Desalination 363 (2015) 44–50 49

3000
16000
2500
12000

C, m-eg/l
2000
C, mg/L

8000
1500 Cl
1.5 L/min
2.5 L/min 4000 SO4
1000
5 L/min
500 0
0 1 2 3 4
0 Bed volume
0 50 100 150 200 250
Time of IX operaon. min
a)

Fig. 9. Chloride concentrations in the eluent versus time of operation of the IX column with
2500
Purolite AT500SO4 resin at various flow rates through the column, L/min: 1.5; 2.5 and
5.0 Vresin is 10 L. Concentration of NaCl is 2.3 g/L.

Sea water

Osmotic pressure, kPa

If salty water with a high content of chloride ions is passed through 2000 IX-treated sea water
an anion exchange resin in sulphate form then, due to exchange of two
chloride ions with one sulphate ion, the feed after IX treatment will have
lower osmotic pressure and as a result a lower operating pressure might
be used to run the further membrane desalination process. 1500
Fig. 11 illustrates the breakthrough curves of Cl− and SO4 2−.ions and
osmotic pressure of the solutions during IX treatment of Swansea Bay
sea water with Purolite AT500SO4 resin. The total salt content in sea
water was 32.0 g/L, the concentrations of sulphate and chloride ions 1000
were 2.43 and 17.46 g/L, respectively. As can be seen in Fig. 11a, almost 0 1 2 3
all chloride ions were removed from the feed sea water due to exchange Bed volume
with sulphate ions until a break through point of the IX column has been b)
−
reached. As a result of efficient SO2−4 /Cl exchange the osmotic pressure
of sea water dropped essentially after IX treatment (Fig. 11b). This Fig. 11. Concentration of chloride and sulphate ions (a) and osmotic pressure of the solu-
reduction of osmotic pressure gives a possibility to reduce the required tions (b) during IX treatment of Swansea Bay sea water with Purolite A500TLSO4 resin.
energy for further NF desalination of the saline water after IX treatment,
since NF membranes, which efficiently reject the sulphate ions and
operate at significantly lower transmembrane pressure compared solutions will be produced as NF reject stream, when salty water after
with RO membranes, can be used for water desalination. IX treatment will be further desalted with NF membranes. Na2SO4 solu-
tion when passing through an exhausted IX resin in chloride form
3.6. Regeneration of exhausted IX resins would transform it back to the sulphate form according to following
reaction:
A need for frequent regeneration using external regeneration
solutions is a weak side of IX when the process is used to treat water 2(RN+)Cl− + Na2SO4 f (RN+)2SO42− + 2NaCl
with a high salt concentration [13]. The IX operational costs may be
sharply reduced if “internal” reagent solution is used for regeneration. where (RN+)Cl− is a polymer matrix of IX resin in chloride form.
In this study we have evaluated a possibility to use Na2SO4 solutions
for regeneration of the exhausted IX resins. Such concentrated Na2SO4 1

2500 0.8
Relevent concentraon

2000 10 L Chloride (10 g/L)

20 L 0.6
Sulphate (10 g/L)
1500 50 L
C, mg/L

Chloride (28 g/L)

0.4 Sulphate (28 g/L)
1000
Chloride (50 g/L)
500 0.2 Sulphate (50 g/L)

0
0 50 100 150 200 250 0
0 5 10 15 20 25 30
Time of IX operaon, min
Bed volume
Fig. 10. Chloride concentrations in the eluent versus time of operation of the IX column
with Purolite AT500SO4 resin at various resin loading, L: 10, 20 and 50. CNaCl is 2.3 g/L; Fig. 12. Regeneration of Purolite A500TLSO4 resin with 10.0, 28.0 and 50.0 g/L Na2SO4
flow rate through the column is 1.5 L/min. solutions. Relative ions concentration corresponds to outlet over inlet IX concentration.
50 N. Hilal et al. / Desalination 363 (2015) 44–50
5000
4000
Chloride (1 run)
3000
C, meq/l
Sulphate (1 run)
Chloride (2 run)
2000 Sulphate (2 run)
Chloride (3 run)
1000 Sulphate (3 run)
0 Na2SO4 solution and multiple regeneration and saturation cycles proved
0 2 4 6 8 10 12
Bed volume
Fig. 13. Breakthrough curves of chloride and sulphate ions with regenerated Purolite
A400TLSO4 resin during treatment of 10 g/L NaCl solution (three exhaustion runs).
0.2 M Na2SO4 solution has been used for IX resin regeneration after each saturation run.
If successful, using this “internal” regent would sharply reduce the
cost of water desalination because no external regeneration solution is
required for regeneration of an exhausted IX column after SO24 −/Cl−
exchange.
Na2SO4 solutions of concentration of 10.0, 28.0 and 50.0 g/L were
tested for the regeneration of Purolite A500TLSO4 resin. As seen in
Fig. 12, it took around 25 BV of the 10 g/L Na2SO4 solution to regenerate
Purolite A500TLSO4 resin. On the other hand, only 16 BV of the 28 g/L
Na2SO4 solution were needed to regenerate the same resin, while
50 g/L Na2SO4 solution performed only slightly better and 15 BV were
needed for regeneration. Hence, Na2SO4 regeneration solution with
concentration of 28 g/L was used for further studies. It may be assumed
that solutions of similar concentration would be produced as a retentate
during NF treatment of brackish water with sodium chloride content of
2.5 g/L.
To evaluate whether the resins performance changes after the
regeneration the IX resins were exhausted and regenerated multiple
times. The multiple breakthrough curves of chloride and sulphate ions
with regenerated Purolite A500TLSO4 resin during treatment of 10 g/L
NaCl solution are presented in Fig. 13. After each exhaustion run the
resin was regenerated with 0.2 M Na2SO4 solution. As seen in this figure
the regenerated resin performs very consistently on each run after
regeneration. This finding of reversibility of the sulphate–chloride
exchange process is very important because no external regeneration
solution is required for regeneration of exhausted IX column after
−
SO2−
4 /Cl exchange.
4. Conclusions
Seven different IX resins, namely Ambersep 900SO4, Purolite A850,
Purolite A500TLSO4, Purolite A400TLSO4, Purolite A109, Purolite A149S
and Purolite A111, which contain primary, secondary, tertiary and qua-
ternary amino groups in the polymer matrix, were used in the work to
study sulphate–chloride exchange process in saline solutions. It was
found that sulphate–chloride exchange is very fast at saline conditions
and less than 10 min is needed to reach the sulphate–chloride equilibri-
um. Very fast kinetics of sulphate–chloride exchange suggests that in
salty solutions IX resistances from the liquid phase film- and intra-
particle diffusion are insignificant. It was shown that optimization of
sulphate–chloride exchange process may be done by adjusting the
flow rate through the column.
The experimental results clearly indicate that the feed water concen-
tration and the nature of the functional group of the resin play an impor-
−
tant role in the SO2−
4 /CI exchange process. It was found that the higher
the substitution of hydrogen atoms in amine functional group of anion
exchange resin is the higher chloride over sulphate selectivity and
that the chloride/sulphate separation factor depends on salt content in
feed water. It was found that Purolite A500TLSO4 resin is the most ben-
−
eficial among the used IX resins for SO2−
4 /CI exchange process in saline
solutions.
Saturated IX resins were successfully regenerated using 0.2 M
14 have no affected on the IX performance.
It was shown that the osmotic pressure of sea water dropped essen-
tially after IX treatment and due to such drop NF membranes would be
used for further water desalination instead of RO membranes, which are
currently used for removal of chloride ions from brackish water and sea
water. Because NF membranes operate at significantly lower transmem-
brane pressure compared with RO membranes, the required energy for
membrane desalination might be essentially reduced.
Acknowledgement
The authors would like to thank King Adbul Aziz City for Science and
Technology for funding this work.
References
[1] The United Nations World, Water Development Report 3: Water in a Changing
World, UNESCO Publishing, Paris, 2009.
[2] M.A. Shannon, P.W. Bohn, M. Elimelech, J.G. Georgiadis, B.J. Marinas, A.M. Mayes,
Science and technology for water purification in the coming decades, Nature 452
(2008) 301–310.
[3] M. Elimelech, W.A. Phillip, The future of seawater desalination: energy, technology,
and the environment, Science 333 (2011) 712–717.
[4] H.A. Abdulgader, V. Kochkodan, N. Hilal, Hybrid ion exchange—pressure driven mem-
brane processes in water treatment: a review, Sep. Purif. Technol. 116 (2013)
253–264.
[5] J. Dron, A. Dodi, Comparison of adsorption equilibrium models for the study of CL−,
NO− 3 and SO4
2−
removal from aqueous solutions by an anion exchange resin, J. Haz-
ard. Mater. 190 (1–3) (2011) 300–307.
[6] M. Carmona, A. Perez, A. de Lucas, L. Rodriques, J.F. Rodriguez, Removal of chloride
ions from an industrial polyethylenimine flocculant shifting it into an adhesive pro-
moter using the anion exchange resin Amberlite IRA-420, React. Funct. Polym. 68
(2008) 1218–1224.
[7] P. Del Hoyo, F. Moure, M. Rendueles, M. Diaz, Demineralization of animal blood
plasma by ion exchange and ultrafiltration, Meat Sci. 76 (3) (2007) 402–410.
[8] S. Evans, R. Gomes, Desalination of Rhine River water by ion exchange part 1. Pilot
plant studies of lime slurry regeneration, Desalination 19 (1) (1976) 433–438.
[9] L. Lv, P. Sun, Z. Gu, H. Du, X. Pang, X. Tao, R. Xu, L. Xu, Removal of chloride ion from
aqueous solution by ZnAl–NO3 layered double hydroxides as anion-exchanger, J.
Hazard. Mater. 161 (2009) 1444–1449.
[10] G. Boari, L. Liberti, C. Merli, R. Passino, Exchange equilibria on anion resins, Desalina-
tion 15 (1974) 145–166.
[11] A. Aveni, G. Boari, L. Liberti, M. Santori, Sulphaste removal and dealkalization on
weak resins of the feed water for evaporation desalting plants, Desalination 16
(1975) 145–149.
[12] S. Sarkar, A.K. SenGupta, A new hybrid ion exchange-nanofiltration (HIX-NF) sepa-
ration process for energy-efficient desalination: process concept and laboratory
evaluation, J. Membr. Sci. 324 (1) (2008) 76–84.
[13] A.A. Zagorodnii, Ion Exchange Materials: Properties and Applications, Elsevier, 2007.
[14] L. Liberti, D. Petruzzelli, F.G. Helfferich, R. Passino, Chloride/sulfate ion exchange
kinetics at high solution concentration, React. Polym. 5 (1987) 37–47.
[15] G. Boari, L. Liberti, C. Merli, R. Passino, Study of the SO2−/
4 Cl− exchange on a weak
anion resin, Ion Exch. Membr. 2 (1974) 59–66.
[16] F. Helfferich, Ion Exchange, Dover Publications, NY, 1995.
[17] A.V. Adamson, A textbook of Physical Chemistry, Academic Pres, 1973.