Desalination 357 (2015) 131139

Contents lists available at ScienceDirect

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

Study of a new process for the efcient regeneration of ion

exchange resins
N.P.G.N. Chandrasekara, R.M. Pashley
School of Physical, Environmental and Mathematical Sciences, University of New South Wales, Canberra ACT 2610, PO Box 7916, Australia

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

A novel process of ion exchange regeneration was developed and evaluated.

Thermal regeneration may be linked to
resin congurational changes.
Ammonium bicarbonate solutions form
the basis of reversible regeneration systems.

a r t i c l e

i n f o

Article history:
Received 15 July 2014
Received in revised form 17 October 2014
Accepted 21 November 2014
Available online xxxx
Keywords:
Desalination
Ion exchange
Sirotherm resin
Sorption
Regeneration
Ammonium bicarbonate

a b s t r a c t
Mixed bead resins comprising weakly acidic (carboxylic acid) and weakly basic (tertiary amine) groups were
produced with properties similar to the thermally regenerable Sirotherm resin. This resin typically had a capacity
of about 0.5 mmol NaCl per gram and the sorption isotherm was found to be consistent with the Langmuir equation. The resin could be partially regenerated by heat (80%) and water washing (40%) as the ion sorption capacity
was much reduced on heating to 80 C. However, the resin could be completely regenerated by a combination of
heat and pre-washing with concentrated ammonium bicarbonate solution. On heating to 60 C or more the bicarbonate salt completely decomposes and can be removed from the resin as carbon dioxide and ammonia gases.
These initial results suggest that this type of ion exchange resin could be used in a continuous process where
the regeneration salt (ammonium bicarbonate) is thermally decomposed, collected and re-used to improve
the efciency of regeneration.
2014 Published by Elsevier B.V.

1. Introduction
Most commercial desalination processes are currently achieved by
two main methods: reverse osmosis (RO) membrane ltration and
Corresponding author.
E-mail address: [email protected] (R.M. Pashley).

http://dx.doi.org/10.1016/j.desal.2014.11.024
0011-9164/ 2014 Published by Elsevier B.V.

thermal evaporation/condensation. Both methods have considerable

technical problems. Seawater reverse osmosis (SWRO) processes require high pressure pumping (up to 65 bar) and use expensive
specialised membranes [1,2] which only last typically for about 5
years [3]. In addition the seawater feed has to be pre-treated to prevent
fouling of the membranes. Thermal methods require the application of
large quantities of thermal energy to overcome the large enthalpy of

132

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

vaporisation of water and so are often dependent on waste heat from

power stations and other industrial plants [4].
Mixed bed ion exchange (IEX) resins have been used for many years
to remove scale-forming ions such as Ca2+ and Mg2+ from feed water
and to produce distilled quality water from tap water. They can also
be used for the desalination of fairly concentrated brackish water, without the need for high pumping pressures, extensive pre-treatment or
high thermal energy input. However, the ion exchange process is limited by the depletion of the resin and the need for large volumes of acid
and base solutions for their regeneration. In this work a novel combined
process of thermal and chemical mixed bead ion exchange regeneration
was developed and evaluated.
Desalting using ion exchangers requires the continuous removal of
the adsorbed salt ions using large volumes of mineral acids and bases.
For example, it has been found [5] that for the removal of one equivalent
of salt, 1.23 equivalents of mineral acid and 1.32 equivalents of base
have to be used. Because of this such desalting processes result in high
operating costs and create substantial volumes of by-product saline
waste and require high levels of acid and base consumption. To overcome these issues new methods to regenerate ion exchange resins by
other means, such as heat energy (Sirotherm process), electrical energy
(electrodialysis) or mechanical energy (piezodialysis) have been studied [610]. The Sirotherm process [6] was developed by the Commonwealth Scientic and Industrial Research Organization (CSIRO) in
Melbourne, Australia. This process was originally based on the use of a
physical mixture of weakly basic and weakly acidic ion-exchange resin
beads. It was later employed for resins containing both weak acid and
weak base components within each bead in the so-called plum pudding structure [7,8]. These resins are thermally regenerable, that is,
they are capable of removing salts from an aqueous solution by sorption
at ambient temperatures and at higher temperatures their sorption capacity is signicantly reduced [11].
The Sirotherm process has been used and studied with many different types of mixed bead resins over the last 40 years [7,1215]. A thermally regenerable composite resin of crosslinked polyacrylic acid and
ethoxylated polyethyleneimine was used for water desalination more
recently by Chanda et al. [16,17]. This type of resin, which is similar to
that used in the present study, can be regenerated thermally with no apparent loss of capacity for up to 10 cycles of operation [16]. Sirotherm
resins are composed of weak acid and weak base groups; thus during
the exchange process a proportion of the Na+ and Cl ions remain in
the solution, due to the hydrolysis of the weak resin salts. Therefore,
high purity deionized water cannot be produced using these resins
and they are suitable only for partial removal of salts. Sirotherm resins
are typically usable only for partial removal of salts which have salinities
in the range of 2000 and 3000 mg/l [16].
The ion adsorption equilibrium conditions in weak acid and weak
base mixed bed resins vary greatly with changes in temperature, pH
and the ionic strength of the solution in equilibrium with the resins
[11,18]. In addition, the capacity of such resins to remove salts also depends on the polymeric structure of the resins, the acidity and basicity
of functional groups, the ratio of acid to basic groups and the resin afnities for counter ions. Titration curve data was reported over a temperature range of 20 C to 80 C for carboxylic acid resins and weakly
basic amine resins [11]. These studies showed that when the temperature is raised in the mixed bed system, which was in equilibrium with
a salt solution, there is a transfer of protons from the base groups to
the acid groups in the mixed bead resin to an extent determined by
the titration curves of the resins [18].
That is, when amine resins are heated they become weaker bases;
hence the amine resins have a plateau in the titration curves which
are displaced to lower pH values with increasing temperature. Carboxylic acid resins have titration curves which demonstrate a reduced affect
with temperature.
These composite resin beads also have a slightly greater adsorption
capacity compared to other typical mixed bed systems but this capacity

is lower than for chemically regenerated strong acid/strong base resin

systems. Recent studies [16,17] on thermally regenerable composite
(Sirotherm-type) resins showed that a weak acid/weak base mixed
bead resin had a substantially reduced ion adsorption capacity at higher
temperatures. Heating from 30 C to 80 C reduced the capacity by a factor of 8 [16].
Studies have been carried out employing ion exchange resins comprising both weakly acidic and weakly basic groups in the resin matrix,
with chemical/thermal regeneration. While using weak acid and weak
base groups, it was found that the desalting capacity of such resins was
lower compared to when using two kinds of cation and anion individual
ion exchange resins [19]. The same study found [19] that conversion of
weakly basic ion exchange functional groups to the carbonate form
would make it likely to maintain a higher level of desalting capacity.
Moreover, the presence of a weak base group in the mixed bed resin
is known to enhance the ion adsorption properties of the mixed bed
resin [20]. According to Weiss et al. [6], tertiary amines have a higher
ability for salt sorption than primary and secondary amines in concurrence with carboxylic acid groups. Numerous studies have been carried
out for the empirical determination of resin pair behaviour under different temperature, pH and ionic strengths [6,11,18,21]. In addition, the
prediction of titration curve behaviour is also useful for resin synthesis
design. Experimental studies have been carried out [11] on the behaviour of ion-exchange pairs.
Ethoxylated polyethyleneimine (EPEI) is considered a good source of
tertiary amine groups, since it consists of large numbers of amine groups
compared to polyethylene amine and studies have been carried out on
this by Chanda et al. [22,23]. Further, the studies carried out by Weiss et
al. [21] showed that variation of acid strength can be achieved by varying
the cross linkage in the resin; by using either methacrylic acid or acrylic
acid or their copolymers during the synthesis of carboxylic acid resins.
Bolto and Siudak [7] showed that resin bead sizes in the range 0.3
1.2 mm should be ideal for thermally regenerable ion exchange resins,
which are composed of both acidic and basic groups together within
the porous particles. Further, they showed that the presence of both
acid and base groups within the same particle absorbed ions rapidly
due to the close proximity of the groups, which shortens the diffusion
path for the transfer of protons [24]. The Plum pudding resin structure, which is prepared by embedding microparticles of polyamines
and microparticles of polyacids in a water and salt permeable matrix,
has been discussed by Bolto and Jackson [8]; who also described the disadvantage of the presence of a high proportion of inert polymeric matrix
material, which tends to limit the thermally regenerable capacity of a
particular resin. Studies have been carried out by Hatch [25], Bolto and
Jackson [8] and Chanda et al. [16] to overcome this issue.
Shimizu [19] stated that the particle diameter of the resin beads is a
key point and it is desirable to be as small as possible, to enhance the
rate of exchange capacity and suggested an optimum size of about
0.5 mm. Generally, the component particles are in the range 0.3
1.2 mm in particle diameter. The smaller sizes are used in uidized
beds but care has to be taken to prevent the dispersion and overow
of ne particles. The larger sizes are generally used in xed bed systems,
where the size distribution can determine the pressure loss across the
bed [19]. Dissociation constants of weak acid/weak base groups also
play a vital role in such desalting processes and these properties directly
contribute to the elution capacity of a resin. The method discussed in
Shimizu's work is generally effective at sodium chloride concentrations
below 3000 mg/l.
Because of the need for large volumes of acid and base solution, the
recovery of regeneration chemicals is an important aim, thus many
groups have tried to develop suitable methods using a variety of techniques. Kadlec et al. [5] used weak acidic cation exchangers during the
desalting process to recover ammonium ions using raised temperatures
between 80 and 150 C.
Concentrated solutions of ammonium bicarbonate have been used as
a draw solution in Forward Osmosis (FO), which was developed [2] to

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

overcome the high pumping pressures required in SWRO. In this process

a concentrated draw solution is used to generate a higher osmotic pressure than seawater. Using an appropriate RO membrane, water is drawn
from seawater across the membrane, without the need for expensive
high pressure pumping. When the aqueous draw solution is made from
a concentrated mixture of dissolved ammonia and carbon dioxide,
these solutes produce a draw solution of concentrated ammonium bicarbonate, which can then be completely decomposed, in solution, by
heating above 60 C [1] to produce the pure, desalted water product.
The gases (NH3 and CO2) can then be collected and used to re-form
new draw solution, either by column distillation or membrane distillation
(MD) [2,26]. In this work we have studied the combined use of thermally
regenerable, mixed weak acid/weak base ion exchange resins together
with ammonium bicarbonate as a regeneration solution, in combination
with heat to remove the adsorbed ammonium and bicarbonate ions.
The process studied in this work was based on the Sirotherm process [6,13] with the addition of a regenerating solution which can
completely regenerate the ion exchange resin and can then be itself almost completely regenerated through a relatively low temperature
chemical decomposition reaction. This process does not demand the consumption of large quantities of acid and base. The regeneration solution is
thermally decomposed into gases and since the resin also has a substantially reduced capacity at higher temperatures, the combined effect offers
a potentially more useful and efcient process. Since ion exchange resin
processes can be considered to include both adsorption and absorption
processes, the more general term sorption was used in this report.
2. Materials and methods
2.1. Materials
80% ethoxylated polyethyleneimine (EPEI) in 37 wt.% aqueous
solution, acrylic acid (AA), the crosslinking agent ethylene

133

glycol dimethacrylate (EGDMA) and initiator 1,1-azobis(cyclohexanecarbonitrile) (ABCN) used for making crosslinked polyacrylic acid (XPAA) were all obtained from Sigma-Aldrich, Australia. 25%
glutaraldehyde was used as a crosslinking agent during this synthesis and this chemical was also obtained from Sigma-Aldrich,
Australia.

2.2. Resin synthesis

The composite resin designated as (XPAA-EPEI-XG) was synthesised
following the procedure outlined in Fig. 1. The weak acid/weak base
mixed bead ion-exchange resin was synthesised by copolymerisation
of 76 wt.% of acrylic acid and 24 wt. % ethylene glycol dimethacrylate
with 0.4 wt.% of 1,1-azobis(cyclohexanecarbonitrile) initiator in
a number of 10 ml sealed glass tubes, which were lled to half
capacity with a nitrogen atmosphere. The reaction mixture was deoxygenated by bubbling nitrogen gas for 10 min, ensuring the
removal of oxygen from the reaction system prior to sealing for the
reaction to obtain a better polydispersity index in the polymer, by
removing its effect on radical polymerization. The mixtures were
then heated in a water bath to 85 C for 1 h and then the glass
tubes were broken and immersed in distilled water, to facilitate the
solid product detaching from the glass. The polyacrylic acid (PAA)
obtained was wet ground to less than 1.2 mm in size and partially
dried. The PAA (100 g) was then reuxed with 10 wt.% ethoxylated
polyethylenimine in 500 ml methanol, with constant stirring for
6 h. The EPEI coated polymer was again crosslinked with 10 wt.%
glutaraldehyde for 3 h and the nal product, which was brown in
colour, was kept in boiling water for 1 h to remove any free EPEI
followed by repeated washing with distilled water. Four batches of
this mixed bead ion-exchange resin were synthesised to ensure
reproducible composition.

Fig. 1. Schematic diagram of the method used for the preparation of the resin XPAA-EPEI-XG by copolymerization with acrylic acid (AA), ethylene glycol dimethacrylate (EGDMA), followed by reaction with ethoxylated polyethyleneimine (EPEI) and crosslinking with glutaraldehyde.

134

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

2.3. Analysis
The moisture content of the resin samples was measured after partial drying with porous lter paper, followed by slow natural drying
in a laminar ow cabinet at room temperature for 10 min. Then the
samples were placed in an oven at 104 C for 18 h and the moisture contents measured.
The resultant wet resin solid product (XPAA-EPEI-XG) had a particle
diameter in the range of 0.31.2 mm and a moisture content of about
40%. The synthesised resin was stored as a wet resin product for
analysis.
The concentration of sodium chloride in solution was determined by
measuring the chloride and sodium ion activity using ion selective electrodes (ISE) (HQ440d-Hach). In these studies, the known concentrations of sodium/chloride ions were measured in millivolts (mV) using
sodium and chloride ISE's and the calibration plots were obtained.
Then the unknown concentration of sodium/chloride ions in the samples was measured as mV values and the concentration of the samples
was found based on the calibration plots. Each measurement was
taken after 10 min of immersion of the ISE electrode in the sample solution. All the measurements were taken at 20.8 0.2 C.
The presence of amine and carboxylic groups in the resin samples
was examined using infra-red (IR) spectroscopy with the KBr disk
method. Resin morphology was studied using a scanning electron microscope (SEM)-TM3000, Hitachi using the bulk sample analysis
technique. The FTIR spectra shown in Fig. 2, reveal the presence
of carboxylic and amine groups in XPAA-EPEI-XG and some
of the basic properties of the synthesised mixed bead resins are
recorded in Table 1. The carboxylic acid content and the amine content
given in Table 1 were obtained from acid/base titrations and pH
measurements.

The morphology of the synthesised resin demonstrates well

established porous structure, as shown in the SEM image in Fig. 3.
2.3.1. Ion exchange equilibrium studies with saline water
The synthesised resin was used for equilibrium studies by immersion in a series of NaCl solutions (in the range 0.1 to 0.5 M, at pH 5.6)
within tightly stoppered tubes at 20 C. These were mechanically shaken for 15 h. In each sorption study, 0.8 g of the wet resin was immersed
in 10 ml of the NaCl solution. Initial and nal concentrations of the chloride were measured for each sample and then the extent of sorption calculated. These equilibrium studies of the mixed bead resin with saline
were performed twice at each concentration to ensure the accuracy of
the results.
2.3.2. Resin regeneration
In the regeneration studies, the mixed bead resin in state 1, as illustrated in Fig. 4, was exposed to 0.5 M NaCl solution and equilibrated for
15 h at 20 C. The resin was then converted into State 2 (see Fig. 4). The
exhausted resin was then washed in distilled water several times until
the rinse water had an electrical conductivity close to that of distilled
water. The washed resin was partially dried in lter paper and then
stored as a wet solid (with determined moisture content, MC, of about
40 w/w %). It was then equilibrated with 0.5 M NaCl for 15 h at 20 C
and the extent of sorption and the percentage of regeneration were obtained from the residual chloride concentration in the equilibrated
solution.
The washed, exhausted resin (0.8 g), in State 2, was heated to
80 1 C for 1 h in 2 L of distilled water, with continuous stirring.
After the treatment the resin sample was washed several times, collected
in a lter paper, pat dry and stored as a wet solid (MC about 40 w/w%).
It was then equilibrated with a solution of 0.5 M NaCl for 15 h at 20 C.

Fig. 2. IR spectrum of the composite sorbent (XPAA-EPEI-XG), composed of copolymerization with acrylic acid (AA), ethylene glycol dimethacrylate (EGDMA), ethoxylated
polyethyleneimine (EPEI) and crosslinked with glutaraldehyde.

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

Table 1
Basic physical and chemical properties of the mixed bead resins used in these
studies.
Property

Value

Moisture content
Particle diameter (mm):
Wet
Dry
Carboxylic content (meq/g) dry
Amine content (meq/g) dry

40%
0.301.2
0.251.1
1.31.5
3.74.0

The extent of sorption and the percentage of regeneration were

calculated by measuring the residual chloride concentration of the
sorbate in the equilibrated solution.
The washed, exhausted resin in State 2 was immersed in 2 M ammonium bicarbonate solution for 15 h at 20 C and then washed several
times in distilled water until the rinse water had conductivity close to
that of distilled water. The sample was then heated for 1 h with continuous stirring at 80 1 C. The resin sample was then washed several
times in distilled water, collected in a lter paper, pat dry and stored
as a wet solid. Samples were then equilibrated with a range of NaCl solutions for 15 h at 20 C and the extent of sorption and the percentage of
regeneration were calculated by measuring the residual chloride concentration of the sorbate in the equilibrated solution. All the regeneration procedures used above were performed twice to ensure accuracy
of the results.

3. Results and discussion

3.1. Sorption isotherms
A typical equilibrium sorption isotherm obtained for NaCl using a
sample of the mixed bead resin (XPAA-EPEI-XG) is illustrated in Fig. 5,
for the adsorption of both Na+ and Cl ions. The maximum NaCl sorption density was found to be typically about 0.5 mmol/g (dry wt) for
the mixed bead resin used in this study.
The sorption isotherms show the classical behaviour of Sirotherm
resin having both weak acid weak base groups in the same bead and the
sorption values obtained were tted well with both Langmuir and
Freundlich isotherms, with the former model tting the data more
closely.

135

The Langmuir isotherm can be written as below, dening the parameters q max and K as the maximum sorption capacity (meq/g dry sorbent) and sorption binding constant (L/meq), respectively.
q


max KC
1 KC 

Where q is the equilibrium sorption (meq/g dry sorbent) and C is

the equilibrium sorbate concentration (meq/L) in solution.
The Freundlich isotherm can be written as follows, where p and n
are Freundlich parameters, and q and C are the same as dened earlier.
qp C


n

The values were determined by least squares t of the adsorption

data and a summary is presented in Table 2.
3.2. A simple thermodynamic model of mixed bead resin ion exchange
processes
The solution ionization properties of the weak acid (carboxylic) and
weak base (amine) groups in a typical WA/WB mixed bead resin, where
the groups are in close proximity, can be modelled using a suitable
amphiprotic molecule, such as glycine. The pKa values as a function of
temperature for the two ionisable groups on the amino acid glycine
have been measured in aqueous solution [27,28]. In this amino acid
the carboxylic (COOH) group is a stronger acid and the amine (NH2) is
a weaker base than the corresponding individual components in solution. When the temperature was increased from 25 C to 75 C; the
COOH groups became slightly more ionized but the NH2 groups become
signicantly less ionized (see Table 3). However, the degree of ionization depends critically on solution pH. At pH values in the range 58
both groups will stay almost fully ionized, even though the ratio of
NH2 groups ionized reduces by 6 times with increase in temperature,
as shown in Fig. 6. By comparison, there is no signicant change in ionization of the COOH groups with temperature. Thus, it appears that the
change in ionization of the amine groups works as a driving force for the
reduction in mixed bead resin capacity with increasing temperature,
presumably by proton exchange.
At higher pH values, greater than 8, the available number of protonated amine groups is signicantly reduced and so the effect of an increase in temperature has a greater effect on total ionization levels.
For example, at around pH 8, protonated amine to non-protonated ratios change from 60:1 to 10:1, as the temperature is increased from 25
C to 75 C. It appears that deprotonation may also tend to reduce the
CO
2 /COOH ratio. Primary, secondary and tertiary amines alone in aqueous solution also behave in a similar manner with temperature, as summarized in Table 4. These ionization constants for primary, secondary
and tertiary amines at 80 C were calculated using van't Hoff equation
based on the experimental values [29] obtained at 25 C.
Measurements of pH changes taken during the equilibrium resin
sorption studies, over a range of NaCl concentrations, were found to
be reduced by less than 0.1 from the initial pH values of around 5.6.
This small increase in hydrogen ion concentration is negligible which
shows that this type of resin has more or less similar sorption capacities
for both Na+ and Cl ions.
3.3. Resin regeneration

Fig. 3. SEM image of a typical weak acid/weak base XPAA-EPEI-XG mixed bead resin, produced by the method described in the text. This image shows the typical pore structure of
the resin, with pores of around 1 m.

The variation in electrical conductivity with the volume of rinse

water (distilled) was measured during the consecutive washing of the
exhausted resins. Typical results are given in Fig. 7. The subsequent
sorption equilibrium studies showed that only a fraction of the
exhausted sites could be regenerated by distilled water washing at
room temperature. The density of available or accessible anionic
(carboxylic) and cationic (amine) binding sites was found to be about

136

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

Fig. 4. Schematic diagram of the functional groups on the ion exchange resin used in this study in various process states.

0.5 meq/g (dry wt) for each of the mixed bead resins produced. By comparison, the total number of sites available was estimated from the reaction scheme to be about 9 and 2 meq/g (dry wt) for cationic and anionic
exchange groups, respectively. These results indicate that there were far
more inaccessible amine groups within the polymer matrix.

The accessible sites presumably have both functionalities (WA/WB

groups) present in close proximity in the mixed bead resin and they
can be readily regenerated either with distilled water, heat or chemical/thermal regeneration techniques. The accessible sites could be divided into easily exchangeable and less readily exchangeable sites.
Thus, only sites which are easily exchangeable, contribute to the reversible regeneration when exposed to excess distilled water. The percentage of regeneration through distilled water washing at room
temperature was typically found to be about 40%, as shown in T1 in
Fig. 8.
Sodium and chloride ions will be physisorbed, electrostatically, onto
oppositely charged groups on the ion exchange resin and so washing
Table 2
Langmuir and Freundlich isotherm parameters for sorption of NaCl on mixed bead resins.
Langmuir isotherm

Fig. 5. Equilibrium sorption isotherm of NaCl on XPAA-EPEI-XG resins at 20 C at a sorbent

loading of 0.8 g (wet resin)/10 ml of equilibration solution. Each solution was equilibrated
with the resin over 15 h, with continuous stirring (typical error for the sorption measurements of about 0.01 mmol Cl or Na/g(dry) sorbent).

Freundlich isotherm

qmax
K
Correlation coefcient
p
n
Correlation coefcient

meq/g dry sorbent

L/meq

0.79
3.17 103
0.998
9.22 103
0.6461
0.976

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

137

Table 3
Ionization constants of glycine in aqueous solutions with temperature.
Glycine

Ka(COOH)

pKa(COOH)

Ka(NH+
3 )

pKa(NH+
3 )

25 C
75 C

4.47E03
4.79E03

2.35
2.32

1.66E10
1.00E09

9.78
9.00

(COO /COOH)
-

(COO /COOH)

Fig. 7. Conductivity decrease on sequential washing of the resin (0.8 g wet weight) with
distilled water. After about 1 l the conductivity approached that of the distilled water
(about 1.0 S/cm).

(NH3 /NH2)
+

(NH3 /NH2)

Fig. 6. Ionization behaviour of COOH groups and NH2 groups on the glycine molecule in
solution, with pH and temperature.

with distilled water should remove these adsorbed ions, replacing them
with hydronium and hydroxide ions. The change in pKa values of the
WA and WB groups with increase in temperature, within the mixed
bead resins where the WA and WB sites are in close proximity, appears
to facilitate the reduction in ion adsorption capacity. This process also
facilitates the regeneration of the resin. In the current study it was
found that the percentage of resin site regeneration, on heating to 80
C was about 80%, as shown in T2 in Fig. 8.
Using the analogy with glycine protonation in aqueous solution, it
might be expected that although there is a 6 drop in the ratio of protonated amine groups to un-protonated on raising the temperature, in
the pH range 68 almost all of the carboxylic acid sites and amine
sites will be ionized. Hence, the reduction in ion exchange capacity
with temperature of these WA/WB resins is hard to explain. It is possible
that the polymer matrix of the resin plays a role in this process, given
the fact that the ion exchange capacity of these resins at room temperature is already much below the number of functional groups expected
from the polymer synthesis composition (see Table 1). The apparent reduction in access to the majority of sites, at room temperature, may increase further as the temperature rises. These effects may be related to
the close proximity of the WA and WB sites within the same beads
and by congurational changes within the polymer matrix on heating.

Fig. 8. Equilibrium sorption isotherm of NaCl on new and regenerated XPAA-EPEI-XG

resins at 20 C at a sorbent loading of 0.8 g (wet resin)/10 ml of equilibration solution.
Each solution was equilibrated with the resin over 15 h, with continuous stirring. Equilibrium ion adsorption for three different regeneration methods was also studied. T1 was
measured after washing with distilled water at room temperature (at 22 C). T2 was measured after distilled water washing and then heating to 80 1 C. T3 was measured after
equilibration with 2 M ammonium bicarbonate solution and then heating to 80 1 C.
(Typical error for the sorption measurements of about 0.01 mmol Cl or Na/g(dry)
sorbent.)

as potassium nitrate, potassium chloride and ammonium chloride (see

Fig. 9). The estimated degree of ionization of ammonium bicarbonate
is about 80% compared to potassium nitrate. Thus, it can be expected
that washing the depleted resins with concentrated ammonium bicarbonate solution should effectively displace adsorbed NaCl, at low temperatures, around 2025 C.
On heating to 80 C, the combination of reduced ion exchange
capacity of the resin and decomposition of the adsorbed ammonium

100000

3.4. Regeneration using ammonium bicarbonate

Conductivity, S/cm

The solution ionization behaviour of ammonium bicarbonate was

studied by measuring the electrical conductivity versus the concentration of the salt. The results obtained show that this salt has a high degree
of ionization in solution, at a similar level to other typical salts [30], such

Ammonium
bicarbonate
10000

Potassium
nitrate
Potassium
chloride
Ammonium
chloride

1000

Table 4
Ionization constants of secondary and tertiary amines in aqueous solution with
temperature.
Base

pKa at 25 C

pKa at 80 C

Ethylamine
Diethylamine
Triethylamine

10.63
10.94
10.72

9.08
9.48
9.51

100
0.001

0.01
0 .1
Concentration of salts, mol/ l

Fig. 9. Electrical conductivities with concentration for several common salts, at 25 C.

138

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

Fig. 10. Schematic diagram of the mechanism proposed for the reduction of ionized sites in the mixed bead resin followed by regeneration by ammonium bicarbonate with temperature.

bicarbonate produces an efcient regeneration. This process was found

to produce about 100% regeneration of the weak acid/base resin used in
this study, as can be seen from the sorption data shown in Fig. 8. These
results suggest that this type of resin combined with ammonium bicarbonate regeneration followed by heating and water washing offers an
efcient and complete resin regeneration.
As discussed earlier, it is possible that congurational changes within the polymer matrix, on heating, bring the two weak acid and base
groups closer together, which both reduces exchange capacity and facilitates ammonium bicarbonate decomposition. As shown in Fig. 10, proton transfer can assist, at higher temperatures (80 C), in ammonium
and bicarbonate ion desorption, along with the decomposition of ammonium bicarbonate, enhancing the mixed bead resin regeneration.
This process enables the sorption capacity of the resin to be returned
to close to its original value, without the need for acid and base washing.
In addition, the emitted gases, carbon dioxide and ammonia, can be
readily collected and re-used to generate the ammonium bicarbonate
solution, which offers the potential for use in both batch and continuous
processes.
4. Conclusions
A mixed bead resin (XPAA-EPEI-XG) was synthesised in order to
study alternative methods for regeneration without the use of conventional acid and base washing. The resin contained weak acid and weak
base groups within the same resin beads. Once the resin was exhausted
by exposure to NaCl solutions, it could be partially regenerated by washing in excess distilled water. These results suggest that about 40% of the
ions bound to the resin were weakly held, probably by electrostatic
forces. These results suggest that there are two types of active binding
sites employed during the ion exchange process within the resin;
some which are accessible and easily exchangeable, and some which
are accessible but less readily exchangeable. The accessible, less readily
exchangeable binding sites could be exchanged using concentrated salts
such as ammonium bicarbonate. Then, at elevated temperatures, it appears that a signicant congurational change occurs within the polymer matrix which both reduces exchange capacity and causes the
complete decomposition of the adsorbed ammonium bicarbonate.
This process can be used for saline water to completely regenerate
this type of resin without the need for acid and base washing. In addition, the decomposition gases can be collected and used to regenerate
the ammonium bicarbonate solution. These results suggest that use of
this salt could be efcient because it readily displaces NaCl from
this type of resin and completely decomposes at relatively modest
temperatures.
Acknowledgments
The authors acknowledge the support of the Australian Research
Council (ARC) (DP120102385) for funding this project.

Reference
[1] H.Y. Ng, W. Tang, W.S. Wong, Performance of forward (direct) osmosis process:
membrane structure and transport phenomenon, Environ. Sci. Technol. 40 (7)
(2006) 24082413.
[2] T.Y. Cath, A.E. Childress, M. Elimelech, Forward osmosis: principles, applications, and
recent developments, J. Membr. Sci. 281 (1) (2006) 7087.
[3] S. Avlonitis, K. Kouroumbas, N. Vlachakis, Energy consumption and membrane replacement cost for seawater RO desalination plants, Desalination 157 (1) (2003)
151158.
[4] M. Rzechowicz, R.M. Pashley, The effect of de-gassing on the efciency of reverse osmosis ltration, J. Membr. Sci. 295 (12) (2007) 102107.
[5] V. Kadlec, P. Huebner, Ion-exchange process for desalting water and subsequent regeneration of the ion exchangers. 1977, U.S Pat.4039442.
[6] D. Weiss, et al., An ion-exchange process with thermal regeneration. II. Properties of
weakly basic resins, Aust. J. Chem. 19 (4) (1966) 561587.
[7] B.A. Bolto, R. Siudak, Further rapidly reacting ionexchange resins, J. Polym. Sci.
Polym. Symp. 55 (1) (1976) 8794.
[8] B. Bolto, M. Jackson, Thermally regenerable ion-exchange resins. I the encapsulation
of polyamine microparticles in a polyacid matrix, React. Polym. Ion Exch. Sorbents 1
(2) (1983) 119128.
[9] R. Nagarale, G. Gohil, V.K. Shahi, Recent developments on ion-exchange membranes
and electro-membrane processes, Adv. Colloid Interf. Sci. 119 (2) (2006) 97130.
[10] B. Bolto, M. Hoang, T. Tran, Review of piezodialysissalt removal with charge mosaic membranes, Desalination 254 (1) (2010) 15.
[11] B.A. Bolto, et al., An ion-exchange process with thermal regeneration. 6. Factors
inuencing titration curve shape of weak electrolyte resins, Aust. J. Chem. 21 (11)
(1968) 27032710.
[12] B.A. Bolto, et al., An ion exchange process with thermal regeneration IX. A new type
of rapidly reacting ion-exchange resin, Desalination 13 (3) (1973) 269285.
[13] B.A. Bolto, et al., An ion-exchange process with thermal regeneration. 14. Thermally
regenerable resin systems with high capacities, Desalination 34 (3) (1980)
171188.
[14] T.B. Hsu, R.L. Pigford, Salt removal from water by continuous ion-exchange using
thermal regeneration, Ind. Eng. Chem. Res. 28 (9) (1989) 13451351.
[15] T.B. Hsu, R.L. Pigford, Mass transfer in a thermally regenerable ion-exchange resin by
continuous cycling, Ind. Eng. Chem. Res. 30 (5) (1991) 10671075.
[16] M. Chanda, et al., A thermally regenerable composite sorbent of crosslinked poly(acrylic acid) and ethoxylated polyethyleneimine for water desalination by Sirotherm process, J. Appl. Polym. Sci. 111 (6) (2009) 27412750.
[17] M. Chanda, et al., Interpenetrating networks of cross-linked poly (acrylic acid) and
cross-linked polyethyleneimine (80% ethoxylated) for desalination of brackish
water by thermoreversible sorption, Ind. Eng. Chem. Res. 49 (16) (2010)
71367146.
[18] D.E. Weiss, et al., An ion-exchange process with thermal regeneration. 4. Equilibria in a mixed bed of weak-electrolyte resins, Aust. J. Chem. 19 (5) (1966)
765789.
[19] H. Shimizu, Ion exchange process for desalination. 1980, U.S. Pat. 40202737.
[20] V.K. Koul, A.K. Gupta, Uptake of sodium chloride by mixture of weakly acidic and
weakly basic ion exchange resins: equilibrium and kinetic studies, Chem. Eng. Sci.
59 (7) (2004) 14231435.
[21] D.E. Weiss, et al., An ion-exchange process with thermal regeneration. 3. Properties
of weakly acidic ion-exchange resins, Aust. J. Chem. 19 (4) (1966) 589608.
[22] M. Chanda, A. Sarkar, J.M. Modak, Ethoxylated polyethylenimine gelcoated on textilegrade acrylic ber. A thermally regenerable superfast sorbent for water desalination, J. Appl. Polym. Sci. 93 (2) (2004) 883893.
[23] M. Chanda, J.M. Modak, A thermally regenerable novel ion-exchanger fabric for
water desalination, J. Polym. Mater. 21 (1) (2004) 3138.
[24] B.A. Bolto, R.E. Warner, An ion-exchange process with thermal regeneration VII.
The rates of neutralization of weak electrolyte resins, Desalination 8 (1) (1970)
2134.
[25] M.J. Hatch, Thermally reversible, amphoteric ion exchange resins consisting of
crosslinked microbeads embedded in crosslinked matrix of opposite exchange
group type. 1976, U.S Pat. 3,957,698.
[26] R.E. Kirts, Method and Apparatus for Producing Potable Water from Seawater Using
Forward Osmosis. 2009, U.S. Pat. 2009/0308727 A1.

N.P.G.N. Chandrasekara, R.M. Pashley / Desalination 357 (2015) 131139

[27] R.G. Clarke, et al., Ionization constants of aqueous amino acids at temperatures up
to 250 C using hydrothermal pH indicators and UV-visible spectroscopy: glycine,
-alanine, and proline, Geochim. Cosmochim. Acta 69 (12) (2005) 30293043.
[28] P. Wang, et al., Thermodynamics of protonation of amino acid carboxylate groups
from 50 to 125 C, J. Solut. Chem. 25 (3) (1996) 243266.

139

[29] M. Cox, et al., The thermodynamics of the acid dissociation of some alkylammonium
ions in water, J. Chem. Soc. B Phys. Org. (1968) 13731379.
[30] G.H. Aylward, T.J.T.J.V. Findlay, SI Chemical Data, 6th ed. John Wiley & Sons Ltd,
Australia, 2008.