ARTICLE IN PRESS

G Model
MEMSCI-9714; No. of Pages 9

Journal of Membrane Science xxx (2009) xxx–xxx

Contents lists available at ScienceDirect

Journal of Membrane Science

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

Preparation, characterization and performance studies of ultrafiltration

membranes with polymeric additive
G. Arthanareeswaran a,b,∗ , D. Mohan b , M. Raajenthiren b
a
Membrane Research Laboratory, Department of Chemical Engineering, National Institute of Technology, Tiruchirappalli 620015, India
b
Department of Chemical Engineering, Anna University, Chennai 600025, India

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

Article history: In the present work, the effect of additive, polyethylene glycol 600 (PEG 600), on the ultrafiltration perfor-
Received 4 September 2009 mance of the polysulfone (PSf)/sulfonated poly(ether ether ketone) (SPEEK) blend flat sheet membranes
Received in revised form prepared in a polar solvent of N,N -dimethylformamide (DMF) by phase separation process was investi-
15 December 2009
gated. The effects of additive concentration on the ultrafiltration performance such as compaction, pure
Accepted 15 December 2009
water flux, water content, hydraulic permeability and morphology were investigated. The addition of
Available online xxx
PEG 600 in the dope solution increases the exchange rate of additive and non-solvent during the mem-
brane formation process, resulting in the appearance of the macrovoids formation as visualized under
Keywords:
Polyethylene glycol
microscope. The pure water flux and Equilibrium water content of the blend membranes were increased
Phase separation while hydraulic permeability was decreased.
Membrane separation © 2009 Elsevier B.V. All rights reserved.
Additive
Ultrafiltration

1. Introduction
For the past several decades, synthetic polymeric membranes
are used for a wide variety of liquid separations such as microfil-
tration, ultrafiltration, reverse osmosis, nanofiltration. A common
method for the preparation of polymeric membranes is the phase
separation process [1,2]. At the same time, blending of two dif-
ferent polymers offers attractive opportunity for the development
of novel materials with valuable combination of properties [3]. In
recent years, several polymer blends have been developed for suc-
cessful application in ultrafiltration separation [4–6]. Alongside, the
addition of organic or inorganic additives as the third component
to the blend polymers has been one of the important procedures
used in membrane preparation to control the morphology and per-
formance of membranes. Studies have been conducted by adding
additives such as polyethylene glycol (PEG), hyperbranched polyg-
lycerol (HPG) polyvinylpyrrolidone (PVP) in the casting solution
to improve the membrane performance [7–9]. Idris and Yet [8]
found that the presence of PEG in the dialysis membrane enhanced
the membrane performance exhibited by the high urea clear-
ance and high urea permeability. Sivakumar et al. [9] investigated
∗ Corresponding author at: Department of Chemical Engineering, National Insti-
tute of Technology, Thanjavur Road, Tiruchirappalli 620015, India.
Tel.: +91 431 2503118; fax: +91 431 2500133.
E-mail addresses: arthanaree10@yahoo.com, arthanareeg@nitt.edu
(G. Arthanareeswaran).
the effects of PVP on CA/PSf blend membranes and showed that
increase in the concentration of PVP in casting solution resulted
in the improved performance of membranes. Polysulfone (PSf) is
widely used as an ultrafiltration, reverse osmosis and pervapora-
tion membrane since polysulfone has good thermal and chemical
stabilities [10]. However, hydrophobic surface of polysulfone mem-
branes has a severe fouling during ultrafiltration processes [11].
Therefore, modifying the PSf membrane, including structure and
cross-section morphology, hydrophilic properties, is significant for
practical applications. Sulfonated PEEK was first applied in mem-
brane formation to improve hydrophilic properties of relatively
hydrophobic polymers [12].
In our previous study [13], SPEEK was introduced to modify
CA membrane in order to obtain CA/SPEEK blend UF membrane.
Introduction of SPEEK in the CA membrane resulted in improved
UF performance. Further, we prepared polysulfone and sulfonated
poly(ether ether ketone) blend membranes and characterized their
ultrafiltration performance [14]. It was found that the extent
of blend compatibility of polymer compositions is 85/15 wt% of
PSf/SPEEK. Phase separation formed beyond 15 wt% of SPEEK into
the PSf/SPEEK polymer blend. Further, 85/15 wt% of PSf/SPEEK
blend membrane shows highest pure water flux, water content, and
porosity compared to other membranes. Su et al. [15] selected five
additives such as butanone (BO), ethyl ether (EE), ethylene glycol
monomethylether (EGME), diethylene glycol (DEGC) and polyethy-
lene glycol and studied the rejection performance of PEG with
molecular weight of 2000–20,000 Da through poly(phthalazinone
ether sulfone)-bisphenol (PPES-B) UF membranes. It focused on

0376-7388/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.memsci.2009.12.020

Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9
2 G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx
the effects of type and concentration of additives on PPES-B mem-
brane and the membranes with PEG400 as additive exhibited better
performance than other four as additives.
In sight of this, the main objective of this work is to investigate
the effect of PEG 600 on the performance of PSf/SPEEK blend ultra-
filtration membrane in terms of pore statistics, compaction, pure
water flux (PWF), equilibrium water content (EWC) and hydraulic
permeability (Pm ). In the initial experiments, effects of various addi-
tives were studied, and PEG 600 was chosen as suitable additive for
this investigation. It has been shown that the introduction of PEG
600 beyond 7.5 wt% in the total polymer composition of PSf/SPEEK
resulted in heterogeneous solution leading to phase separation
during membrane formation. Hence, we selected a maximum of
7.5 wt% of PEG 600 in the PSf, PSf/SPEEK blend composition and
introduced PEG 600 by reducing the composition of DMF as solvent.
The MWCO of the prepared membranes have been determined by
using different molecular weight of proteins. The morphology of
the resulted membranes has also been studied by scanning electron
microscope (SEM).
2. Experimental
2.1. Materials and methods
Commercial grade polysulfone Udel P-3500 (PSf) was procured
from Amoco Polymers Inc., USA. Commercial grade poly(ether ether
ketone) (PEEK) (Victrex PEEKTM 450 PF) was supplied by Abhi-
man Polymers, India. Analytical grade N,N -dimethylformamide
(DMF) and sodium lauryl sulfate (SLS) were obtained from Quali-
gens Fine Chemicals, Glaxo India Ltd., India. Polyethylene glycol
(PEG) 600, oxalic acid, maleic acid and acetone were procured from
SRL Ltd., India. Disodium hydrogen orthophosphate anhydrous and
monosodium dihydrogen orthophosphate heptahydrate were pro-
cured from CDH Chemicals Ltd., India. Bovine serum albumin (BSA),
Mw = 69 kDa, pepsin, Mw = 35 kDa, trypsin, Mw = 20 kDa were pur-
chased from SRL Chemicals Ltd., India and used as received. Egg
albumin (EA), Mw = 45 kDa was obtained from CSIR Bio Chemical
Centre, New Delhi, India. Deionized and distilled water was used
for the preparation of protein feed solution and for the gelation bath
preparation.
2.2. Preparation of blend membranes
Sulfonation of PEEK by using concentrated sulfuric acid is
explained in our earlier reported procedure [13]. The casting solu-
tions (17.5 wt%) were prepared by blending of PSf and SPEEK
polymers of different compositions in presence of different con-
centrations of PEG 600 in DMF under constant stirring for 4 h at
60 ◦ C. A series of such polymer solutions were prepared by varying
the composition of PSf, PSf/SPEEK with an increment of 2.5 wt% of
PEG 600, as shown in Table 2. The obtained homogeneous solution
was allowed to stand for at least 3 h to eliminate the air bubbles. In
the membrane preparation process, two types of phase separation
can be distinguished. The dry phase inversion takes place in the
atmosphere by evaporation of the solvent. The wet phase inver-
sion is carried out by immersing the polymer solution film into a
coagulation bath of a non-solvent, where an exchange of solvent
and non-solvent takes place and the membrane is formed. In this
study, we prepared membranes by a combination of both processes.
The polymer solution was poured over a smooth glass plate and
spread with a metal rod with side runners to produce 0.22 mm
thickness of the membrane. The size of the casting area was approx-
imately 29 cm × 25 cm. For those membranes prepared using a
solvent evaporation period, the glass plate was placed inside a
humidifying chamber. During casting, the relative humidity (RH)
Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
of the casting chamber was maintained at 50 ± 2% and the tem-
perature was kept at 20 ± 2 ◦ C. Prior to casting, a gelation bath of 2 l
consisting of 2.5% (v/v) DMF and 0.2 wt% SLS in distilled water (non-
solvent) was prepared and the bath was ice-cooled to 4 ◦ C. Thirty
seconds was the maximum solvent evaporation time that allowed
the preparation of membrane sheets, longer times resulted in holes
within them. After the solvent evaporation period at humidifying
chamber, the glass plate was immersed into ice-cooled gelation
bath at 4 ◦ C for at least 1 h. The hardened membranes were eventu-
ally peeled from the glass plate and after the gelation, membranes
were stored at 5 ◦ C in distilled water until they were tested. Every
membrane sheet was inspected with a light-box to detect defects,
such as thin spots and pinholes; defective membranes were dis-
carded. Prepared membranes (76 mm diameter) were cut from the
flat sheets for the ultrafiltration experiments.
2.3. Membrane characterization
2.3.1. Morphological studies
The morphological characterization of all the blend membranes
was studied using scanning electron microscopy (SEM) (Leica Stere-
oscan440, UK). The membranes were cut into pieces of varied sizes
and mopped with filter paper and were immersed in liquid nitro-
gen and the frozen membranes were preserved in desiccator for
SEM studies. The top surface and cross-sectional view of the mem-
branes, was observed by SEM under very high vacuum at 25 kV and
at different magnifications.
2.3.2. Pore statistics
The average pore size and surface porosity of membranes were
determined by the ultrafiltration of protein solutions of different
molecular weights. The molecular weight of the solute that has a
solute rejection (SR) above 80% was used to evaluate the average
pore size, R̄, of the membranes using the following equation [16,17]:
 ˛ 
R̄ = 100 (1)
%SR
where R̄ is the average pore size (radius) of the membrane (Å), and
˛ is the average solute radius (Å). The %SR is a percentage solute
rejection of proteins (from Eq. (6)). The average solute radii, also
known as the Stoke radii, were obtained from the plot of solute
molecular weight versus solute radius in aqueous solution, which
was developed by Sarbolouki [17].
The surface porosity, ε, of the membrane was calculated by the
orifice model given below assuming that only the skin layer of the
membrane is effective in separation.
3Jw
ε= (2)
P R̄
where  is the viscosity of the permeate water in Pa s, Jw is the pure
water flux of the membrane in m3 /m2 s, R̄ is the average pore radius
in Å and P is the transmembrane pressure in Pa.
2.3.3. Equilibrium water content
Equilibrium water content is considered to be an important
characterization parameter as it indirectly indicates the degree of
hydrophilicity or hydrophobicity of a membrane [18]. Also it is
related to the porosity of a membrane. Membranes were weighed
in an electronic balance in wet state after mopping the surface
water with a clean tissue paper. The wet membranes were dried
by putting in a vacuum oven for 24 h at a temperature of 50–60 ◦ C
and again they were weighed in dry state. The EWC at room tem-
perature is calculated as follows [18]:
wet membrane weight − dry membrane weight
EWC (%) =
wet membrane weight
× 100 (3)
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9
G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx
2.4. Permeation experiment
2.4.1. Ultrafiltration experiments
The ultrafiltration (UF) experiments were carried out in a 400 ml
batch type stirred cell (ultrafiltration cell – S76-400-Model, Spec-
trum, USA) fitted with a Teflon coated magnetic paddle. The
effective membrane area available for ultrafiltration was 38.5 cm2 .
The solution filled in the cell was stirred at 400 rpm using a mag-
netic stirrer. All the experiments were carried out at 30 ± 2 ◦ C
and compaction at 414 kPa transmembrane pressure while other
permeation studies at 345 kPa transmembrane pressure. The per-
meating solution was collected from the bottom of the cell. The cell
was pressurized using a nitrogen cylinder.
2.4.2. Membrane compaction
Before using each fresh membrane, it was compacted. The mem-
branes were initially pressurized with deionized water for 5 h at a
transmembrane pressure of 414 kPa which is higher than the max-
imum operating pressure designed for the present experiments.
The initial water flux was taken about 30 min after the pressuriza-
tion in the test cell, at 414 kPa. The pure water flux was measured
after every hour in order to monitor the compaction behavior. The
membranes were compacted until a steady flux was observed. Pure
water flux (Jw ) was calculated over measured time intervals using
the following equation. The compact membranes were washed
thoroughly before placing them in the membrane batch cell.
Q find the suitable additive for preparing blend UF membranes with
Jw =
A T
where Q is permeating of pure water (l); A, effective membrane
area (m2 ) and T, the sampling time.
2.4.3. Pure water flux and hydraulic permeability
The compacted membranes were subjected to pure water flux
studies and hydraulic permeability was calculated under steady-
state run at a transmembrane pressure of 345 kPa using Eq. (4).
Membrane hydraulic permeability has got significance particu-
larly for membranes used in pressure-driven separation processes.
Membrane permeability was determined by allowing deionized
water to pass through the compacted membrane. Flux values
of pure water at different transmembrane pressures (ranging
0–414 kPa) were measured under steady-state condition using Eq.
(4). From the slope of the linear relationship between the pure
water flux (Jw ) and transmembrane pressure, hydraulic permeabil-
ity was calculated as
Jw
Pm =
P
where Pm is hydraulic permeability (l m−2 h−1 kPa−1 ); P is trans-
membrane pressure (kPa).
2.4.4. Molecular weight cut-off (MWCO)
UF experiments were conducted in the batch kit cell mentioned
previously to study the influence of concentration of pore former
PEG on molecular weight cut-off (MWCO) and solute rejection of
the prepared membranes. Therefore, proteins of different molecu-
lar weights such as trypsin (20 kDa), pepsin (35 kDa), EA (45 kDa)
and BSA (69 kDa) were chosen for the estimation of MWCO and
solute rejection. All the protein solutions were prepared individu-
ally at a concentration of 0.1 wt% in phosphate buffer (0.5 M, pH 7)
using deionized water and used as standard solutions. The UF cell
was filled with protein solution and pressurized at a constant pres-
sure of 345 kPa and stirred at 400 rpm throughout the experiments
to minimize fouling. During ultrafiltration, the permeate solutions
of corresponding membranes were collected over a period of time
in a graduated tube and were analyzed for the concentration using
Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
3
UV-Visible spectrophotometer (Shimadzu, Model UV-160A) at a
wavelength of 280 nm. From the feed and permeate concentrations
(Cp and Cf ) the percentage rejection was calculated using Eq. (6)
[15].
 
Cp
%SR = 1 − × 100 (6)
Cf
The main objective of the ultrafiltration experiment was to
investigate the rejection behaviour of proteins by different mem-
branes prepared in the present work under constant pressure
conditions. Molecular weight cut-off is an attribute of pore size of
the membranes and is related to the rejection of a spherical solute
of given molecular weight. The molecular weight has a linear rela-
tionship with the pore size of the membrane [19]. In general, the
molecular weight cut-off of the membrane is determined by iden-
tifying an inert solute of different molecular weights that have a
solute rejection of 80–100% in steady-state UF experiments [20].
Therefore, proteins of different molecular weights were chosen for
the estimation of MWCO.
3. Results and discussion
The membrane structure and its properties are influenced by the
composition of the polymer solution (concentration, solvent, addi-
tives). The addition of some pore forming additives to the casting
solution is very effective in changing their properties. In order to
(4)
suitable pore size, the effect of different organic additives as pore
former on pure water flux of PSf membrane was attempted. Thus,
the additives such as maleic acid, oxalic acid and polyethylene gly-
col 600 were chosen. The additives concentration in the casting
solution of blend membranes was maintained at 5 wt%. The effect
of various additives on pure water flux of PSf membrane is given
in Table 1. From Table 1, it is evident that pure PSf has the pure
water flux of 16.8 l m−2 h−1 . Further, it is evident that PEG 600 has
significantly enhanced the water flux to 57.8 l m−2 h−1 , compared
to other additives (pore formers). In addition, it is well known that
PEG 600 acts as a better pore former compared to maleic acid and
oxalic acid in PSf/SPEEK blend system. Zhu et al. suggested that
when PEG is entrapped by membrane materials as additive, PEG
should increase hydrophilicity, and results in important change in
the performance of ultrafiltration membrane such as solute reten-
tion and fouling [21]. Further, the 600 molecular weight PEG was
chosen, because solutions prepared with higher molecular weights
are thermodynamically less stable and coagulation values become
(5) much lower [22]. Hence, for further work, PEG 600 was used as a
pore former additive in PSf/SPEEK blend membranes and the cast-
ing compositions of membranes are shown in Table 2.
3.1. Morphological studies
SEM images for the top surface of different membranes are pre-
sented in Fig. 1. The top surface and cross-section structures of pure
PSf and various compositions of PSf/SPEEK blend membranes have
been reported in our earlier study [15]. It was observed that pure
PSf membranes have a dense top surface [15]. When PEG is added
to the casting solution, pores appear at the membrane surface.
With the increase of PEG content from 2.5 to 7.5 wt% in the casting
solutions, the largest diameter of the micropores at the membrane
surfaces increases from about 38.9 to 42.7 Å. At the same time, the
number pores in the surface of membranes also becomes more.
Similar results were observed by the other researchers [23,24].
Chuanga et al. [22] suggested that the addition of PVP additives
in the system could induce pores in the top layer. However, Wang
et al. [24] showed that the size and number of the macrovoids in
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9

4 G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx

Table 1
Effect of different additives on PWF of PSf (17.5 wt%) membranes.

Total polymer (wt%) Additives (wt%) Solvent (wt%) Pure water flux at 345 kPa (l m−2 h−1 )

PSf Oxalic acid Maleic acid PEG 600 DMF

17.5 – – – 82.5 16.8

17.5 5 – – 77.5 33.3
17.5 – 5 – 77.5 44.4
17.5 – – 5 77.5 57.8

flat membrane decreases with increasing LiCl, PVP and PEG con-
tent in the PES casting solution. From Fig. 1, it is observed that
2.5 wt% of PEG 600 in 95/5 wt% PSf/SPEEK blend compositions of
membranes exhibited smaller pores distributed evenly. However,
the pore size increases with 7.5 wt% of PEG 600 concentrations in
same blend composition. The blend membrane of PSf/SPEEK/PEG
at 85/15/7.5 wt% has the highest surface porosity while the blend
membrane of 85/15 wt% at 2.5 wt% PEG has the lowest surface
porosity (Fig. 1). The formation of pores at the top surface is possibly
due to demixing of the casting solution by means of nucleation and
growth of the polymer rich phase, i.e. the solid phase [25,26]. Smol-
ders et al. [27] reported that appropriate amount of non-solvent
additives enhanced the formation of macrovoids.
It is known that the cross-sectional structure of asymmetric
polymer membrane depends on the solvent and additive used [28].
A finger-like structure was formed when DMF/PEG was used as the
solvent/additive pair with various compositions in blend polymers.
Fig. 2 shows the SEM image of the cross-section of different compo-
sition of PEG in PSf and PSf/SPEEK blend membranes. It may be seen
from the figure that membranes so formed are having asymmetric
structure consisting of a dense top layer and a porous sub layer.
From Fig. 2, when PEG 600 content increased from 2.5 to 7.5 wt% in
95/05 wt% of PSf/SPEEK blend solution, the sublayer seems to have
finger-like cavities as well as macrovoid structure. This finger-like
cavities and macrovoid structure was changed with increasing con-
centration of PEG. Large macrovoids in the substructure typically
result in increased permeability, and therefore, may be desirable
in moderate-pressure UF applications. Similar observations were
also found by Strathmann et al. [29] using LiCl as the additive for
the preparation of polyamide membrane. Due to high affinity of
PEG with water, instantaneous demixing results, leading to the
formation of finger-like cavities in the sublayer of the prepared
membranes [2]. Since decay of the casting solution by adding of PEG
additive somewhat limited the growth of polymer-lean phase, i.e.
the growth of macrovoid, dense layer having sponge-like structure
might be formed. Note that the similar phenomena were observed
with the membrane prepared from the various compositions of
PSf/SPEEK/PEG blend casting solution. Since PSf/SPEEK is water
insoluble, they still remain in the membrane matrix after phase
inversion process. However, part of PEG in the casting solution
might be removed from the membrane matrix due to PEG dis-
solving into the non-solvent (gelation solution) phase during the
process. The growth of PSf/SPEEK-rich phase might be hindered
during the phase inversion process and PEG dissolving induced the
enlargement of the micropore. Similar observations were reported
for effect of PEG on PEI gas membranes [30]. Because of this, the
final morphology of membrane prepared from the PSf/PEEK/PEG
solution is completely different from that of membrane prepared
from the solution containing PSf and PSf/SPEEK blend.
3.1.1. Effect of concentration of PEG on average pore size and
porosity
The average pore size and porosity of the membranes deter-
mined from the protein rejection studies are shown in Fig. 3. It
is evident from these results that the membrane prepared in the
absence of additive has relatively smaller pore size and porosity.
PSf in presence of 2.5 wt% PEG additive membrane had the average
pore size value of 38.9 Å and porosity value was 0.74. The average
pore size of PSf in absence of both SPEEK and PEG was determined
and discussed in earlier work [15]. Further, the average pore size
increases from 38.9 to 42.7 Å and the porosity increases from 0.74
to 1.35 with increase in the concentration of pore former, PEG from
2.5 to 7.5 wt% in 100% polysulfone membrane. When increasing PEG
concentration from 2.5 to 7.5 wt%, an upward trend in average pore
size from 40.2 to 48.7 Å and 41.2 to 50.5 Å for PSf:SPEEK blend com-
positions of 95/5 and 85/15 wt%, respectively, had been observed.
Similarly, the porosity increases from 0.89 to 1.46 and 1.42 to 1.79
for PSf:SPEEK blend compositions of 95/5 and 85/25 wt%, respec-
tively, when increasing the PEG concentration from 2.5 to 7.5 wt%,
respectively. The increase in PEG content in blend may favor the for-
mation of larger pores in membranes. Increase in the pore size and
porosity, in principle, would lead to the increase in the permeate
flux of the membrane. This is in agreement with the results obtained
in this study. Although bigger pore size will favor high permeate

Table 2
Effect of compaction time on PWF of PSf/SPEEK blend membranes with additive concentrations.

Polymer blend composition Additive, PEG 600 wt% Pure water flux (l m−2 h−1 ) at 414 kPa
(17.5 wt%)

PSf SPEEK Time (h)

0 1 2 3 4 5

100 0 2.5 95 80 71 62 45 45
95 5 2.5 115 104 85 70 57 57
90 10 2.5 126 116 102 85 69 69
85 15 2.5 150 131 122 110 92 92

100 0 5.0 123 109 98 79 71 71

95 5 5.0 165 148 127 101 89 89
90 10 5.0 194 183 159 127 113 113
85 15 5.0 234 197 168 145 126 126

100 0 10.0 158 153 136 109 96 96

95 5 10.0 244 189 159 121 110 110
90 10 10.0 297 218 189 152 128 128
85 15 10.0 318 234 203 174 141 141

Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9

G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx 5

Fig. 1. SEM top surface image showing the influence of different composition of PEG 600 in PSf/SPEEK blend membranes.

flux, the solute rejection will drastically fall. The membrane best tion from 2.5 to 7.5 wt%, respectively. The SEM pictures showing
suited for any application would be the one that has more number the top surface of the membranes with different PEG compositions
of pores with smaller size. Because smaller size of pores will favor evidence this trend (Fig. 1). This trend indicates the leachability of
better retention and higher number of pores will favor high perme- the water soluble additive leading to the formation of larger pores
ate flux. Similar observation was also reported by Malaisamy et al.
[31] for cellulose acetate and sulfonated polysulfone membranes.
Table 3
Effect of additive concentration on equilibrium water content of PSf/SPEEK blend
3.1.2. Effect of concentration of PEG on equilibrium water content membranes.
Equilibrium water content of the membranes is an indirect indi-
Blend composition Equilibrium water content (%)
cation of the hydrophilicity and flux behavior of membranes. The (17.5 wt%)
effect of PEG concentration on EWC of PSf/SPEEK blend membranes
PSf (%) SPEEK (%) PEG 600 (wt%)
is illustrated in Table 3. The EMC of PSf is 76.5% in absence of PEG,
which was this result has been also reported in our earlier work 2.5 5.0 7.5
[15]. Further, the EMC increases from 77.2 to 80.2% with increase in 100 0 77.2 (0.6) 79.3 (0.5) 80.2 (0.5)
the concentration of pore former, PEG, from 2.5 to 7.5 wt% in 100% 95 5 80.8 (0.4) 81.7 (0.4) 82.8 (0.4)
of PSf membrane. Further, the EWC increases from 82.5 to 87.5% 90 10 82.5 (0.5) 83.7 (0.4) 87.5 (0.3)
85 15 84.3 (0.4) 85.2 (0.6) 91.0 (0.5)
and 84.3 to 91% for the blend compositions of PSf/SPEEK of 90/10
and 85/15 wt%, respectively, when increasing the PEG concentra- Note: Numbers in the parenthesis are standard deviation.

Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9

6 G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx

Fig. 2. SEM cross-section image showing the influence of different composition of PEG 600 in PSf/SPEEK blend membranes on the morphology.

[9]. The EWC actually represents the fraction of water molecules

occupied in the pores of the membrane. An increase in the water
content indicates that the membrane has become more porous.
Similar trends were also observed by Lv et al. [32] for modified
CA membranes with Pluronic F127 in the casting solutions.

3.2. Permeation experiments

3.2.1. Effect of compaction time on flux

The effect of SPEEK concentration on PWF of PSf membrane
was discussed in our earlier work [15]. The effect of additive con-
centrations on PWF of prepared PSf/SPEEK membranes at various
compaction time is shown in Table 2. In general, the PWF decreases
as compaction time increases for all the prepared membranes. In
the polymer casting solution, hydrophilic additive concentrations
of pore former, PEG 600, were varied, from 2.5 wt% in an increment
of 2.5 wt% for all the polymer blend solutions and the maximum
compatible additive concentration was found to be 7.5 wt%. The Fig. 3. Effect of concentration of PEG on average pore size and porosity of PSf/SPEEK
PWF on compaction at every hour for 4–5 h for PSf and PSf/SPEEK blend membranes (— porosity, . . . average pore size; × 100/0%,  95/05%, 䊉 85/15%).

Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9

G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx 7

Table 4
Initial and final flux ratio during compaction of PSf/SPEEK blend membranes with additive concentrations.

Polymer blend composition (17.5 wt%) Additive, PEG 600 (wt%) Solvent, DMF (wt%) Pure water flux at 414 kPa (l m−2 h−1 )

PSf (%) SPEEK (%) Initial Final Final/initial ratio

100 0 2.5 80.0 95 45 0.47

95 5 2.5 80.0 115 57 050
90 10 2.5 80.0 126 69 0.55
85 15 2.5 80.0 150 92 0.62

100 0 5 77.5 123 71 0.57

95 5 5 77.5 165 89 0.54
90 10 5 77.5 194 113 0.58
85 15 5 77.5 234 126 0.54

100 0 7.5 77.0 158 96 0.61

95 5 7.5 77.0 244 110 0.45
90 10 7.5 77.0 297 128 0.43
85 15 7.5 77.0 318 141 0.44

membranes with various PEG 600 concentrations was measured,
and the initial and steady-state values are reported in Table 4. From
Tables 1 and 4, it is observed that the membrane prepared from
100% PSf in presence of 2.5 wt% PEG 600 as additive has a flux of
95 l m−2 h−1 and almost remains constant at a value of 45 l m−2 h−1
after 4–5 h of compaction time. This constant value is indicated
the completion of compaction. In addition, an increase in pore for-
mer concentration to 5 wt%, increased the constant flux from 45
to 71 l m−2 h−1 . During compaction under hydraulic pressure, the
polymeric chains reorganize and this leads to change in the struc-
ture with a lowered volume of porosity. This leads to an increase
in the hydraulic permeability offered by the membrane, and con-
sequently lowers the flux. Similar observations have been reported
by Chakrabarty et al. [18].
When 7.5 wt% of PEG 600 was added in the 80/15 wt% PSf/SPEEK
blend system in place of solvent DMF, the initial water flux was
318 l m−2 h−1 . After 4 h of compaction, the water flux decreased and
attained a steady-state value of 141 l m−2 h−1 . The SEM images of
PSf/SPEEK membranes support these observations that at 7.5 wt%
PEG 600 concentration numerous pores with higher sizes formed
than with the 2.5 wt% of PEG composition (Fig. 1), which is due to
leaching out of additives in the gelation bath from the membrane.
3.2.2. Effect of concentration of PEG on pure water flux
The effect of PEG concentration on PWF through PSf/SPEEK
blend membranes is illustrated in Fig. 4. The PWF is the key require-
ment of any membrane and must be determined essentially. After
compaction, the pure water fluxes of all the membranes were mea-
sured at a constant sampling period under steady-state conditions,
after 30–45 min of stabilization at 345 kPa transmembrane pres-
sure. The PWF increases from 38.4 to 76.7 l m−2 h−1 with increase
in the concentration of pore former, PEG, from 2.5 to 7.5 wt% in 100%
of PSf membrane. Further, when increasing PEG concentration from
2.5 to 7.5 wt%, an upward trend in PWF from 56.6 to 108.9 l m−2 h−1
and 77.4 to 118.9 lm−2 h−1 for PSf/SPEEK blend compositions of
90/10 and 85/15 wt%, respectively, had been observed. Kim and Lee
[22] showed that the PWF increases exponentially with increasing
PEG/NMP ratio up to a ratio of 1:1. This may be attributed to the
formation of macrovoids by leaching out of PEG from the mem-
brane forming system during gelation, which becomes the domain
of water molecules. Similar observation has been reported by other
researchers for blend membranes with PEG as an additive [7,33].
3.2.3. Effect of concentration of PEG on hydraulic permeability
Hydraulic permeability of clean membrane, Pm , is an indication
of tolerance of membrane towards hydraulic pressure. After the
PWF reached steady state, Pm is determined by measuring the PWF
under various transmembrane pressures from 69 to 414 kPa. The
linear proportionality of PWF to applied pressure can be directly
associated to the transport resistance. The presence of an additive
in the casting solution has considerable effect in the hydraulic per-
meability. Thus when the PEG concentration was increased from 2.5
to 7.5 wt% in the casting solution of PSf, there was a corresponding
decrease in hydraulic permeability from 9.1 to 4.5 l m−2 h−1 kPa−1
as shown in Fig. 5. It is due to the instability of pores which are
formed by leaching out the PEG during the coagulation process in
membrane casting technique. Similar trend had also been observed
for blend membranes prepared from 5, 10 and 15 wt% SPEEK
with PSf in presence of additive. Thus, permeability of PSf/SPEEK
(95/5 wt%) and PSf/SPEEK (85/15 wt%) is reduced from 9.3 to 3.7
and 5.4 to 2.9 l m−2 h−1 kPa−1 , respectively, when PEG concentra-
tion in blend polymer casting solution increased from 0 to 7.5 wt%.
The decreased value of hydraulic permeability may be due to the
fact that the addition of pore former in the casting solution results in
the formation of macrovoids on the membrane surface due to ther-
modynamical instability, which enhances precipitation and porous
nature [29]. Similar observations have also been reported else-
where [27,34].

3.2.4. Effect of concentration of PEG on molecular weight cut-off

Fig. 4. Effect of concentration of PEG on pure water flux of PSf/SPEEK blend mem-
branes at 345 kPa pressure.
The molecular weight cut-off (MWCO) of membranes are gen-
erally defined by the membrane manufacturers as the molecular
weight of the solute at which above 80% of uncharged molecules
are retained [19]. The rejections of above 80% are found in practice,
even for substances of molecular size much larger than the poros-

Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9

8 G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx

macrovoids formation. On addition of PEG in PSf/SPEEK dope solu-

tion, a sharp interface between the PEG and non-solvent water was
formed, resulting in increased solvent–non-solvent exchange rate.
The observations can be summarized as follows:

 All the blend membranes were found to have asymmetric struc-

ture as seen from SEM photographs.
 PEG 600 can be a suitable additive for making asymmetric mem-
branes having a higher PWF.
 It was found that PSf/SPEEK blend polymer solutions having
additive PEG concentration more than 7.5 wt% did not form
membrane.
 With increase in concentration of PEG in PSf/SPEEK blend, the
pore size as well porosity was increased. The hydraulic perme-
ability was seen to reduce greatly with increase in concentration
of PEG.
 A significant increase in MWCO of membrane was observed
when concentration of PEG was increased from 2.5 to 7.5 wt%
in the membrane casting solution.

Overall, PWF experiments and pore statistics suggest that the

Fig. 5. Effect of concentration of PEG on hydraulic permeability of PSf/SPEEK blend membrane prepared with the addition of PEG as additive demon-
membranes. strated the best performance; therefore, it should be considered as
additive in practical applications.
Table 5
Solute rejection and MWCO of PSf/SPEEK blend membranes with PEG as additive.
References
Polymer blend PEG 600 (wt%) %SR of proteins MWCO (kDa)
composition [1] R.W. Baker, Membrane Technology and Applications, McGraw-Hill, New York,
(17.5 wt%) 2000.
[2] T. Matsuura, Synthetic Membranes and Membrane Separation Processes, CRC
PSf (%) SPEEK (%) Press Inc., 1994.
[3] M. Zeng, Z. Fang, C. Xu, Effect of compatibility on the structure of
100 0 2.5 83 39
the microporous membrane prepared by selective dissolution of chi-
95 5 2.5 85 45
tosan/synthetic polymer blend membrane, J. Membr. Sci. 230 (2004) 175–
85 15 2.5 87 45 181.
100 0 5 89 45 [4] W. Zhao, Y. Su, C. Li, Q. Shi, X. Ning, Z. Jiang, Fabrication of antifoul-
95 5 5 87 45 ing polyethersulfone ultrafiltration membranes using Pluronic F127 as both
surface modifier and pore-forming agent, J. Membr. Sci. 318 (2008) 405–
85 15 5 88 >69
412.
100 0 7.5 82 45 [5] M. Wang, L.G. Wu, J.X. Mo, C.J. Gao, The preparation and characterization of
95 5 7.5 85 >69 novel charged polyacrylonitrile/PES-C blend membranes used for ultrafiltra-
85 15 7.5 82 >69 tion, J. Membr. Sci. 274 (2006) 200–208.
[6] W.R. Bowen, S.Y. Cheng, T.A. Doneva, D.L. Oatl, Manufacture and characterisa-
tion of polyetherimide/sulfonated poly(ether ether ketone) blend membranes,
J. Membr. Sci. 250 (2005) 1–10.
ity of the membranes. Thus, molecules or solutes having molecular [7] Y.H. Zhao, B.K. Zhu, X.T. Ma, Y.Y. Xu, Porous membranes modified by hyper-
branched polymers. I. Preparation and characterization of PVDF membrane
weight larger than the molecular weight cut-off (MWCO) of a mem-
using hyperbranched polyglycerol as additive, J. Membr. Sci. 290 (2007)
brane will not pass through the membrane. This is undoubtedly due 222–229.
to polydispersity in the dimensions of the membrane pores. The [8] A. Idris, L.K. Yet, The effect of different molecular weight PEG additives on cel-
MWCO of the membranes can generally be estimated using linear lulose acetate asymmetric dialysis membrane performance, J. Membr. Sci. 280
(2006) 920–927.
polymers such as polyethylene glycol, dextrans or using spherical [9] M. Sivakumar, D. Mohan, R. Rangarajan, Studies on cellulose acetate-
molecules such as proteins, etc. It is seen from Table 5 that mem- polysulfone ultrafiltration membranes. II. Effect of additive concentration, J.
brane made from pure PSf showed a MWCO of 39 kDa due to lower Membr. Sci. 268 (2006) 208–219.
[10] S. Wu, Polymer Interface and Adhesion, Chapman and Hall, London, 1987.
pore size. An increase in additive concentration from 2.5 to 7.5 wt% [11] S. Mok, D.J. Worsfold, A. Fouda, T. Matsuura, Surface modification of polyether-
results in the increase of MWCO of pure CA from 39 to 45 kDa. It sulfone hollow fiber membranes by ␥-ray irradiation, J. Appl. Polym. Sci. 51
was believed that the incorporation of additive in casting solution (1994) 193–199.
[12] M. Peer, UK Patent, GB 2216134 A (1989).
would alter the MWCO of all the PSf and PSf/SPEEK membranes. The [13] G. Arthanareeswaran, K. Srinivasan, R. Mahendran, D. Mohan, M. Rajendran, V.
increase in MWCO with increasing PEG may be due to fast exchange Mohan, Studies on cellulose acetate and sulfonated poly(ether ether ketone)
rate of PEG and water during gelation process, which forms a large blend ultrafiltration membranes, Eur. Polym. J. 40 (2004) 751–762.
[14] G. Arthanareeswaran, D. Mohan, M. Raajenthiren, Preparation and performance
pore on membrane surface as also evidenced from SEM results. On
of polysulfone-sulfonated poly(ether ether ketone) blend ultrafiltration mem-
increasing the SPEEK content in blend from 5 to 15 wt% in presence branes. Part I, Appl. Surf. Sci. 253 (2007) 8705–8712.
of PEG increases the MWCO from 45 to >69 kDa. Similar results have [15] Y. Su, X. Jian, S. Zhang, C. Yan, Preparation of novel PPES-B UF membrane with
good thermal stability: the effect of additives on membrane performance and
also been reported by Malaisamy et al. [35].
cross-section morphology, J. Membr. Sci. 271 (2006) 205–214.
[16] J. Abdoul Raguime, G. Arthanareeswaran, P. Thanikaivelan, D. Mohan,
M. Raajenthiren, Performance characterization of cellulose acetate and
4. Conclusions
poly(vinylpyrrolidone) blend membranes, J. Appl. Polym. Sci. 104 (2007)
3042–3049.
The effects of pore forming additive on the morphology, equi- [17] M.N. Sarbolouki, A general diagram for estimating pore size of ultrafiltration
librium water content pore statistics and permeation performance and reverse osmosis membranes, Sep. Sci. Technol. 17 (1982) 381–386.
[18] B. Chakrabarty, A.K. Ghoshal, M.K. Purkait, Effect of molecular weight of PEG
were investigated. The addition of PEG increases the exchange rate on membrane morphology and transport properties, J. Membr. Sci. 309 (2008)
of additive and non-solvent, resulting in the enhancement of the 209–221.

Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
 ARTICLE IN PRESS
G Model
MEMSCI-9714; No. of Pages 9
G. Arthanareeswaran et al. / Journal of Membrane Science xxx (2009) xxx–xxx
[19] R. Mahendran, R. Malaisamy, D. Mohan, Cellulose acetate and polyethersul-
fone blend ultrafiltration membranes. Part I. Preparation and characterizations,
Polym. Adv. Technol. 15 (2004) 149–157.
[20] B. Jung, Preparation of hydrophilic polyacrylonitrile blend membranes for
ultrafiltration, J. Membr. Sci. 229 (2004) 129–136.
[21] L.P. Zhu, L. Xu, B.K. Zhu, Y.X. Feng, Y.Y. Xu, Preparation and characterization of
improved fouling-resistant PPESK ultrafiltration membranes with amphiphilic
PPESK-graft-PEG copolymers as additives, J. Membr. Sci. 294 (2007) 196–206.
[22] J.H. Kim, K.H. Lee, Effect of PEG additive on membrane formation by phase
inversion, J. Membr. Sci. 138 (1998) 153–163.
[23] W.Y. Chuang, T.H. Young, W.Y. Chiu, C.Y. Lin, The effect of polymeric additives on
the structure and permeability of poly(vinyl alcohol) asymmetric membranes,
Polymer 41 (2000) 5633–5641.
[24] G.J. Wang, L.Y. Chu, M.Y. Zhou, W.M. Chen, Effects of preparation conditions
on the microstructure of porous microcapsule membranes with straight open
pores, J. Membr. Sci. 284 (2006) 301–312.
[25] A.J. Reuvers, C.A. Smolders, Formation of membranes by means of immersion
precipitation. Part II. The mechanism of formation of membranes prepared from
the system cellulose acetate–acetone–water, J. Membr. Sci. 34 (1987) 67–86.
[26] K. Kimmerle, H. Strathmann, Analysis of the structure-determining process of
phase inversion membranes, Desalination 79 (1990) 283–302.
[27] C.A. Smolders, A.J. Reuvers, R.M. Boom, I.M. Wienk, Microstructures in phase
inversion membranes. Part I. Formation of macrovoids, J. Membr. Sci. 73 (1992)
259–275.
Please cite this article in press as: G. Arthanareeswaran, et al., Preparation, characterization and performance studies of ultrafiltration membranes
with polymeric additive, J. Membr. Sci. (2010), doi:10.1016/j.memsci.2009.12.020
9
[28] I.C. Kim, H.G. Yun, K.H. Lee, Preparation of asymmetric polyacrylonitrile mem-
brane with small pore size by phase inversion and post-treatment process, J.
Membr. Sci. 199 (2002) 75–84.
[29] H. Strathmann, K. Kock, P. Amar, R.W. Baker, The formation mechanism of
asymmetric membranes, Desalination 16 (1975) 179–203.
[30] J.J. Shieh, T.S. Chung, R. Wong, M.P. Srinivasan, D.R. Paul, Gas separation per-
formance of poly(4-vinylpyridine)/polyetherimide composite hollow fibers, J.
Membr. Sci 182 (2001) 111–123.
[31] R. Malaisamy, R. Mahendran, D. Mohan, M. Rajendran, V. Mohan, Cellu-
lose acetate and sulfonated polysulfone blend ultrafiltration membranes.
I. Preparation and characterization, J. Appl. Polym. Sci. 86 (2002) 1749–
1761.
[32] C. Lv, Y. Su, Y. Wang, X. Ma, Q. Sun, Z. Jiang, Enhanced permeation performance
of cellulose acetate ultrafiltration membrane by incorporation of Pluronic F127,
J. Membr. Sci. 294 (2007) 68–74.
[33] Y. Liu, G.H. Koops, H. Strathmann, Characterization of morphology controlled
polyethersulfone hollow fiber membranes by the addition of polyethylene
glycol to the dope and bore liquid solution, J. Membr. Sci. 223 (2003) 187–
199.
[34] I. Cabasso, Membranes Encyclopedia of Polymer Science and Engineering, vol.
9, John Wiley & Sons, New York, 1989, pp. 509–579.
[35] R. Malaisamy, R. Mahendran, D. Mohan, Cellulose acetate and sulfonated poly-
sulfone blend ultrafiltration membranes. II. Pore statistics, molecular weight
cutoff, and morphological studies, J. Appl. Polym. Sci. 84 (2002) 430–444.