CHEMICAL ENGINEERING TRANSACTIONS

VOL. 61, 2017

A publication of

The Italian Association

of Chemical Engineering
Online at www.aidic.it/cet

Guest Editors: Petar S Varbanov, Rongxin Su, Hon Loong Lam, Xia Liu, Jiří J Klemeš
Copyright © 2017, AIDIC Servizi S.r.l.
ISBN 978-88-95608-51-8; ISSN 2283-9216

Experimental Investigation of Nanofiltration Process for the
Separation of Complex Aqueous Electrolyte Mixtures

Muhammad Ali Samee*, Michael Harasek, Anton Friedl
Institute of Chemical, Environmental and Biological Engineering, Technische Universität Wien, Gertreidemarkt 9/166, A -
1060 Vienna, Austria
muhammad.ali.samee@tuwien.ac.at

The era of membrane technology began in 1960s as a viable alternative to the more conventional separation
techniques (distillation, extraction and evaporation processes) and can carry out separations that are difficult
by other means. For these reasons membranes are becoming increasingly important in food and dairy
industries for a variety of desalting, purification, concentration and other separations. Most of the previous
work done on NF process is at room temperature and low pressures. The novelty of this work lies in the high
pressure and high temperature applications in connection with our previous work mentioned in the reference
section. This work focuses on the experimental study of nanofiltration (NF) process for the separation of
binary, ternary and quaternary aqueous mixtures of different salts at a starting concentration of 0.25 w% of
each, whether single or in a mixture form. The experiments proved that NF process is capable of separating
different ions from solution. The evaluation of the data revealed that the rejection of divalent cations and
anions was greater than 96 % in most of the cases.

1. Introduction
The chemical and process industry is governed by environmental legislation to reduce the number of toxic
solvents, reagents and effluents used in processes to minimize the industrial waste which is potentially
harmful to the environment. The need to meet more stringent legislation has led to the search for alternative
for additional separation techniques (Timmer, 2001). The advancement in the research related to development
and application of membrane separation processes is one of the most significant achievement in chemical,
environmental and biological process engineering (Lyshevski, 2014). A membrane can be defined as a
selective barrier that permits the passage of certain species (called permeate) in a fluid. Solute particles are
(partially) retained depending on the properties like size, shape and charge. The four pressure driven
membrane processes are microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse osmosis
(RO). NF is an intermediate process between RO and UF with pore radius of about 0.5 to 2 nm, molecular
weight cut off between 200 to 1,000 Dalton and operating pressure between 5 to 40 bar. NF is advantageous
in different aspects such as it can be operated at low pressure, gives high permeate flux, retention of
multivalent salts and organic solutes, low investment and operation and maintenance costs. In comparison
with UF and RO, NF has always been a difficult process to define and to predict (Li et al., 2008). Transport
mechanisms in nanofiltration membranes are generally described in terms of: convection, diffusion and
electromigration for electrolyte systems (Szymczyk et al., 2003). In literature, for mathematical modelling
purposes, the NF membranes are typically characterised by the pore radius, effective membrane charge
density and effective membrane thickness to porosity ratio (Bowen et al., 1997).

2. Materials and method
In this work the separation performance of NF for aqueous solutions of single inorganic electrolytes (NaCl,
MgCl2.6H2O, Na2SO4 and MgSO4) and different combinations of them to form ternary and quaternary ionic
systems has been studied experimentally. The experiments were conducted on a lab-scale cross-flow
membrane unit OS-MC-01 from Osmota with effective membrane area of 0.008 m2. A negatively charged NF

DOI: 10.3303/CET1761186

Please cite this article as: Samee M.A., Harasek M., Friedl A., 2017, Experimental investigation of nanofiltration process for the separation of
complex aqueous electrolyte mixtures, Chemical Engineering Transactions, 61, 1129-1134 DOI:10.3303/CET1761186

1129

membrane (MPF-34) from Koch Membrane was used. It has a typical operating temperature range of 40 to
70 °C, a pressure range of 14 to 35 bar and a pH range of 0 to 14. All the experiments were performed in
recirculation mode. The retentate was made to flow back to the feed tank to see the effect of increasing
concentration on permeate flux and rejection. The process conditions for all the experiments were 32 bar,
60 °C, feed flowrate of 2.5 L/min and starting feed concentration of 0.25 w% of each electrolyte whether single
or in a mixture form. The feed flow rate is high with respect to compact membrane module design resulting in
crossflow velocities of more than 1.3 m/s. In addition, feed solutions are reasonably dilute, viscosities are low,
and – as an advantage of this process – operating pressures are moderate to high. Specific fluxes are
moderate and so the concentration polarization phenomenon is anticipated not to have a very profound effect
on the separation performance of the NF membrane in most of the cases. The permeate weight was
measured gravimetrically with an electrical balance in specified intervals to calculate the permeate flux. Pure
water permeability was tested before and after each experiment at 32 bar, 2.5 L/min and 30 °C. The cleaning
of the membrane was done accordingly as per manufacturer´s instructions. The flow diagram of the process is
shown in Figure 1.

Figure 1: Experimental setup

The concentrations of dissolved solutes in the feed and permeate samples collected during the experiments
were detectable by conductivity ~ concentration calibration curves for single electrolytes, while the individual
concentrations of anions and cations for single and mixtures of electrolytes were analysed on Dionex 5000+
ion chromatography system.

3. Results and discussion
3.1 Separation of binary ion system

3.1.1 NaCl

The negatively charged NF membrane attracts the cations and repels the anions. This effect is known as
Donnan effect. At lower concentration, the rejection of Na+ in NaCl was high (82 %). As the concentration was
increased, the rejection started to decline (78 %) as shown in Figure 2(a) where the effect of concentration
polarization is negligible but the rejection of Na+ and Cl- ions of the same concentration of NaCl solution in the
presence of sucrose (13 to 25 ⁰Bx) has been reported to be decreased to almost zero (Samee et al., 2016)
due to strong effect of concentration polarization on transport phenomenon. At higher concentration, the effect
of steric hindrance was also negligible due to shielding of charged solutes (Wang et al., 1995). To keep the

electroneutrality condition the Cl- ion follows the same rejection behaviour presented in Figure 2(c). It was also

inferred that at lower concentration, the adsorption of Cl- ions at the membrane surface is negligible and at
higher concentration, more chloride ions will get adsorbed (Schaep et al., 1999) for negatively charged

membrane. Cl- ions adsorption is prevalent on the membrane, with respect to sodium adsorption, since anions
show lower hydration radii than cations (Takagi et al., 1990).

3.1.2 Na2SO4

Na+ ions presented a very high rejection (96 to 98 %) in Na2SO4 solution shown by Figure 2(a) with increasing

trend as concentration of the solution increases followed by SO4
2- shown in Figure 2(d). Since SO4

2- ions cannot

1130

permeate due to strong electrostatic exclusion so Na+ were rejected accordingly to maintain zero current
density.

3.1.3 MgCl2.6H2O

In MgCl2.6H2O solution, Mg
2+ showed a slightly increase in rejection from 96 to 98 % shown in Figure 2(b)

upon the increase of concentration. This behaviour has also been observed by other researchers (Bandini et
al., 2003). As the concentration of MgCl2.6H2O increases, more Mg

2+ will be attracted to the surface and
shield the membrane and results in positive charge. Charge reversion at the surface can also be expected like
positive charge density in this case. The interfacial precipitation of the metal hydroxide which occurs at a lower
pH value than the bulk lead to adsorption phenomenon (Labbez et al., 2003). The adsorption is influenced by
the charge and dielectric constant. It was also explained that at higher concentration, pore shrinking can be

expected (Hunter, 1981). Cl- ions followed the same pattern to keep the electroneutrality presented in Figure
2(c).

3.1.4 MgSO4

In MgSO4 solution, since the SO4
2- ions show a strong electrostatic exclusion so the rejection of SO4

2- ions was
more than 98 % as shown in Figure 2(d). Mg2+ ions followed the same rejection trend to keep zero current
density condition as presented in Figure 2(b).

Figure 2: Rejection of cations and anions as a function of permeate flux in binary ion system

3.2 Separation of ternary ion systems

3.2.1 NaCl+Na2SO4

In this ternary ionic system the rejection of SO4
2- ions remained in the range of 98 to 99 % owing to its large

size and strong electrostatic exclusions between divalent SO4
2- ions and negative membrane charge as

depicted in Figure 3(d). The retention of Na+ ions was decreased from 89 to 82 % as shown in Figure 3(a) and

100

130 140 150 160 170

N
a+

Io
n

R
ej

ec
tio

n
(%

)

Permeate Flux (kg/m2h)
(a)

NaCl

Na₂SO₄

100

130 140 150 160 170

M
g2

+
Io

n
R

ej
ec

tio
n

(%
)

Permeate Flux (kg/m2h)
(b)

MgCl₂.6H₂O
MgSO₄

100

130 140 150 160 170

C
l-

Io
n

R
ej

ec
tio

n
(

%
)

Permeate Flux (kg/m2h)
(c)

NaCl

MgCl₂.6H₂O

100

130 140 150 160 170

S
O
₄²⁻Ion

R
ej

ec
tio

n
(%

)

Permeate Flux (kg/m2h)
(d)

Na₂SO₄
MgSO₄

1131

is strongly dependent on the highly rejected SO4
2- ions due to the electroneutrality condition. Here the

concentration polarization has an appreciable effect on rejection mechanism. With increasing feed
concentration the ionic concentration increases in the polarization layer. Na+ ions contributed to shield the

membrane charge to offer Cl- ions a reduced electrostatic exclusion from the membrane. Also SO4
2- gave a

pull to monovalent co-ion (Cl-) to permeate more to keep the electroneutrality condition. So the rejection of

monovalent co-ion Cl- decreases more (80 to 59 %), as shown in Figure 3(c) and could show negative

retention under certain conditions. Negative retention of Cl- ions has been observed by (Bowen et al., 1996) at
lower molar ratios in the same mixture and also by (Szoke et al., 2002) at certain pH range of feed solution.
Another study by (Samee et al., 2016) with the same feed concentration of salts as that of this experiment but

additionally in the presence of sucrose (13 to 25 ⁰Bx) has elucidated the negative rejection of Cl- ions (up to -
46 %) caused by enhanced concentration polarization effect on transport phenomena. Hereby, the driving
force due to the electroneutrality was higher than the concentration gradient driving force.

Figure 3: Rejection of cations and anions as a function of permeate flux in ternary ion system

3.2.2 NaCl+MgCl2.6H2O

In this ternary ionic mixture the monovalent counter ion Na+ in the presence of divalent counter ion Mg2+

showed the same behaviour as Cl- ion in NaCl+Na2SO4. The rejection of Na
+ ion was minimum amongst all

the ions due to its lower molar ratio and decreased from 61 to 56 % shown in Figure 3(a). Here the effect of
concentration polarization is not very profound as the rejections were not decreased appreciably but Na+ ion
has been reported to show negative retention (up to 21 %) in a mixture of NaCl+MgCl2.6H2O with the same
feed concentration as used in this experiment in the presence of sucrose (13 to 24 ⁰Bx) due to strong effect of
concentration polarization as mentioned by (Samee et al., 2016). Although Mg2+ ion has higher stokes radius

than SO4
2- ion, rejections obtained were lower than SO4

2- ion rejections presented in the previous section. This

can be attributed to the positive charge of Mg2+ ions and strong electrostatic exclusion between SO4
2- ions and

100

120 130 140 150 160

N
a+

Io
n

R
ej

ec
tio

n
(%

)

Permeate Flux (kg/m2h)
(a)

NaCl+MgCl₂.6H₂O
NaCl+Na₂SO₄
Na₂SO₄+MgSO₄

100

120 130 140 150 160

M
g2

+
Io

n
R

ej
ec

tio
n

(%
)

Permeate Flux (kg/m2h)
(b)

NaCl+MgCl₂.6H₂O
MgCl₂.6H₂O+MgSO₄
Na₂SO₄+MgSO₄

100

120 130 140 150 160

C
l-

Io
n

R
ej

ec
tio

n
(

%
)

Permeate Flux (kg/m2h)
(c)

NaCl+MgCl₂.6H₂O
NaCl+Na₂SO₄
MgCl₂.6H₂O+MgSO₄

100

120 130 140 150 160

S
O
₄²⁻Ion

R
ej

ec
tio

n
(%

)

Permeate Flux (kg/m2h)
(d)

NaCl+Na₂SO₄
MgCl₂.6H₂O+MgSO₄
Na₂SO₄+MgSO₄

1132

membrane. The rejection of Mg2+ and Cl- were 90 to 93 % and 74 to 75 % as shown in Figure 3(b) and 3(c)
respectively.

3.2.3 Na2SO4+MgSO4

In this mixture, high rejections of all the ions were observed. In contrast to Mg2+ ion possessing larger stokes
radius, Na+ ion with relatively smaller stokes radius will permeate through the membrane. Nevertheless due to

electroneutrality condition SO4
2- ions should pass through the membrane as well, which was mostly hindered

by electrostatic exclusion and due to relatively large size of this ion. Consequently, the Na+ permeation was
hindered as well to represent a high rejection greater than 95 % as shown in Figure 3(a). The rejections of

Mg2+ ions and SO4
2- ions were greater than 98 % and 97 % as shown in Figure 3(b) and 3(d) respectively.

3.2.4 MgCl2.6H2O+MgSO4

In this ternary ionic mixture the high rejections of the ions were observed which were greater than 95 % for Cl
-

ions as shown in Figure 3(c) and 98 % for and Mg2+ and SO4
2- ions as shown in Figure 3(b) and 3(d)

respectively.

3.3 Separation of quaternary ion systems

3.3.1 NaCl+MgSO4

In this quaternary ion system Na+ ions permeated (52 to 71 %) with Cl- ions (70 to 72 %) as shown in Figure

4(a) and 4(c), whilst Mg2+ ions were rejected with SO4
2- ions at higher values greater than 96 % as shown in

Figure 4(b) and 4(d) respectively. In addition, it is observed that divalent Mg2+ ion worsens the permeation of

Cl- ion in comparison to the ternary NaCl+Na2SO4 system where the rejection decreased from 80 to 59 %. It is
difficult to attain a high performance separation in such systems. Nevertheless, it is possible to partly separate
NaCl from MgSO4 at relatively high concentrations.

Figure 4: Rejection of cations and anions as a function of permeate flux in quaternary ion system

100

115 125 135 145 155 165

N
a+

Io
n

R
ej

ec
tio

n
(%

)

Permeate Flux (kg/m2h)
(a)

MgCl₂.6H₂O+Na₂SO₄
NaCl+MgSO₄

100

115 125 135 145 155 165

M
g2

+
Io

n
R

ej
ec

tio
n

(%
)

Permeate Flux (kg/m2h)
(b)

MgCl₂.6H₂O+Na₂SO₄
NaCl+MgSO₄

100

115 125 135 145 155 165

C
l-

on
R

ej
ec

tio
n

(%
)

Permeate Flux (kg/m2h)
(c)

MgCl₂.6H₂O+Na₂SO₄
NaCl+MgSO₄

100

115 125 135 145 155 165

S
O
₄²⁻Ion

R
ej

ec
tio

n
(%

)

Permeate Flux (kg/m2h)
(d)

MgCl₂.6H₂O+Na₂SO₄
NaCl+MgSO₄

1133

3.3.2 MgCl2.6H2O+Na2SO4

It was observed that in this quaternary ionic system the divalent ions were better rejected than monovalent

ions. The rejections of Mg2+ ion and SO4
2- ion were greater than 98 % as shown in Figure 4(b) and 4(d)

respectively. Cl- ion permeated with Na+ ion to keep the electroneutrality condition as elucidated in Figure 4(a)
and 4(c) with rejections in the range of 76 to 82 % without any considerable effect of concentration
polarization.

4. Conclusions

SO4
2- as a divalent co-ion, showed strong electrostatic exclusion from membrane and resulting in a very high

rejection. In addition, due to the zero current density at steady state, counter-ions Mg2+ was rejected as well.

Monovalent ions Na+ and Cl- showed less rejection than divalent ions and could show negative rejection under
certain conditions. Monovalent ions were permeated together to keep the electroneutrality condition.
Concentration polarization may have significant effects on rejection phenomenon especially in case of
concentrated solutions. In summary, the experimental study of NF separation performance for complex
electrolyte mixtures is beneficial in lieu of process prediction, effective development and optimisation.

Acknowledgments

The authors cordially acknowledge the financial support provided by Higher Education Commission of
Pakistan and technical support provided by Technische Universität Wien, Austria, for the completion of this
project.

References

Bandini S., Vezzani V., 2003, Nanofiltration Modeling: The Role of Dielectric Exclusion in Membrane
Characterization, Chemical Engineering Science, 58, 3303-3326.

Bowen W.R., Mukhtar H., 1996, Characterization and Prediction of Separation Performance of Nanofiltration
Membranes, Journal of Membrane Science, 112, 263-274.

Bowen W.R., Mohammad A.W., Hilal N., 1997, Characterisation of Nanofiltration Membranes for Predictive
Purposes - Use of Salts, Uncharged Solutes and Atomic Force Microscopy, Journal of Membrane Science,
12, 91-105.

Hunter R.J., 1981, Zeta Potential in Colloid Science: Principles and Application, Academic Press, London, UK.
Lyshevski S.E., 2014, Dekker Encyclopedia of Nanoscience and Nanotechnology, 3rd ed., CRC Press, Boca

Raton, USA.
Labbez C., Fievet P., Thomas F., Szymczyk A., Vidonne A., Foissy A., 2003, Evaluation of the “DSPM” Model

in a Titania Membrane: Measurements of Charged and Uncharged Solute Retention, Electro-kinetic
Charge, Pore Size and Water Permeability, Journal of Colloid and Interface Science, 262, 200-211.

Li N.N., Fane A.G., Winston W.S., Matsuura T., 2008, Advanced Membrane Technology and Applications,
Wiley-Blackwell, Oxford, UK.

Samee M.A., Elgohary A., Harasek M., Friedl A., 2016, Experimental Investigation of Nanofiltration Process
for the Separation of Complex Sugar Mixtures Containing Mono- and Multivalent Salts, Chemical
Engineering Transactions, 52, 799-804.

Schaep J., Vandecasteele C., Mohammad A.W., Bowen W.R., 1999, Analysis of the Salt Retention of
Nanofiltration Membranes Using the Donnan Steric Partitioning Pore Model, Separation Science and
Technology, 34, 3009-3030.

Szymczyk A., Labbez C., Fievet P., Vidonne A., Foissy A., Pagetti., 2003, Contribution of Convection,
Diffusion and Migration to Electrolyte Transport Through Nanofiltration Membranes, Advances in Colloid
and Interface Science, 103, 77-94.

Szoke S., Patzay G., Weiser L., 2002, Characteristics of Thin-film Nanofiltration Membranes at Various pH-
Values, Desalination, 150, 123-129

Takagi R., Nakagaki M., 1990, Theoretical Study of the Effect of Ion Adsorption on Membrane Potential and
Its Application to Collodion Membranes, Journal of Membrane Science, 53, 19-35.

Timmer J. M., 2001, Properties of Nanofiltration Membranes: Model Development and Industrial Application,
PhD thesis, Technische Universiteit Eindhoven, Netherlands.

Wang X. L., Tsuru T., Togoh M., Nakao S., Kimura S., 1995, Evaluation of Pore Structure and Electrical
Properties of Nanofiltration Membranes, Journal of Chemical Engineering of Japan, 28, 186-192.

1134