Desalination 355 (2015) 110

Contents lists available at ScienceDirect

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

Effect of silica on the properties of cellulose acetate/polyethylene

glycol membranes for reverse osmosis
Adnan Ahmad a,, Sidra Waheed a, Shahzad Maqsood Khan a, Sabad e-Gul a, Muhammad Shaq a,
Muhammad Farooq b, Khairuddin Sanaullah c, Tahir Jamil a
a
b
c

Department of Polymer Engineering and Technology, University of the Punjab, Quaid-e-Azam Campus, Lahore 54590, Pakistan
Institute of Chemical Engineering and Technology, University of the Punjab, Quaid-e-Azam Campus, Lahore 54590, Pakistan
Department of Chemical Engineering and Energy Sustainability, University Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia

H I G H L I G H T S

2-stage phase inversion protocol; involving TIPS and controlled evaporation.

Silica is used as an additive to augment permeation performance.
Addition of silica enhanced the fouling resistance remarkably.
Optimum loading of silica particle is required for maximum performance.

a r t i c l e

i n f o

Article history:
Received 23 July 2014
Received in revised form 2 October 2014
Accepted 6 October 2014
Available online xxxx
Keywords:
Reverse osmosis
Polyethylene glycol
Cellulose acetate
Silica
Phase inversion

a b s t r a c t
In this work, a series of cellulose acetate/polyethylene glycol-600 membranes, with varying ratios were prepared
by 2-stage phase inversion protocol. The permeation properties were studied by subjecting membranes in indigenously fabricated reverse osmosis plant. After optimization of different CA/PEG ratios, the membrane with highest
salt rejection capacity was selected and modied with varying amount of silica. The Modied membranes were
characterized for their permeation properties, hydrophilicity, compositional analysis, thermal stability, mechanical
strength and morphological studies. Silica signicantly inuenced the permeation performance of composite
membrane. The ux enhanced from 0.35 to 2.46 L/h m2 along with an 11.41% relative increase in salt rejection.
The hydrophilicity was signicantly enhanced by the addition of silica. In FTIR spectra, the broadening of the
peak around 3500 cm1 and emergence of peak at 950 cm1 specied the incorporation of silica particles. The
thermal analysis indicated the relative increase in degradation temperature (Tmax) and glass transition temperature (Tg) for CPS-5 membrane. The mechanical stability of the modied membranes, increased initially, but
declined with further addition of silica. The results indicated that the incorporation of SiO2 content in the casting
solution improved the fouling resistance of the membranes.
2014 Elsevier B.V. All rights reserved.

1. Introduction
Membranes and membrane processes subsist as old as life [14].
Recently, synthetic membrane processes have appeared among the
most innovative technologies to attain a variety of goals ranging from
molecular separation to concentration of products [5]. The intrinsic
characteristics of membranes like high efciency, simplicity, selectivity,
low energy consumption and good stability make membranes as an
integral part of various chemical industries [6,7].
Membrane based desalination processes involve reverse osmosis
(RO), electro dialysis, forward osmosis, ultraltration and nanoltration
[8,9]. The RO process was commercialized in 1970 and today it is a
Corresponding author. Tel.: +92 3216789792 (Cell).
E-mail address: adnanahmedbaloch@gmail.com (A. Ahmad).

http://dx.doi.org/10.1016/j.desal.2014.10.004
0011-9164/ 2014 Elsevier B.V. All rights reserved.

leading membrane technology representing 80% of total desalination

plants [1012].
An ideal RO membrane should possess the resistance to chemical
attack, excellent separation performance, good mechanical and thermal
stability [13]. The performance of a membrane is notably inuenced by
its constituents, which affect many properties [14,15]. Various polymers
have been used to synthesize membranes, but cellulose acetate (CA)
membranes are well accepted due to their good transport characteristics,
low protein adsorption, excellent water afnity, appropriate mechanical
strength, excellent lm-forming properties and lower cost [16,17]. However, CA has poor thermal stability, lower chemical resistance, inferior
mechanical strength and vulnerable to fouling resistance. Therefore, attempts have been made to improve the performance of CA by developing
hybrid organicinorganic membranes. Few researchers have investigated the effect of alumina (Al2O3), zirconia (ZrO2), titania (TiO2), silver

A. Ahmad et al. / Desalination 355 (2015) 110

Fig. 1. A schematic of reverse osmosis rig.

and silica (SiO2) particles in various polymeric materials [1820]. The resultant hybrid membranes have been reported to increase the ux and
mechanical strength, upon increasing the concentration of inorganic particles [21,22]. SiO2 particles have attracted particular attention owing to
their special properties like small size, thermal resistance, ne suspension formation in aqueous solution, strong surface energy and relatively
inert material.
Polymeric RO membranes are synthesized through phase inversion
protocol. It is the most versatile technique to obtain a variety of morphologies ranging from microltration membranes with very porous
structures, to more dense reverse osmosis membranes [23]. The concept

of phase inversion involves an array of different methods such as

diffusion-induced phase separation, vapor-phase precipitation, phase
inversion by controlled evaporation and thermal-induced phase separation (TIPS) [24]. All the methods involve complex multi-component

Table 1
Effect of concentration of CA/PEG on water content, ux, salt rejection and membrane
permeability.
Membrane Water
type
content (%)

Flux
(L/h m2)

Salt rejection
(%)

Membrane
permeability
(104 L/h m2 kPa)

CA/PEG-1
CA/PEG-2
CA/PEG-3
CA/PEG-4
(Control)

0.87
0.70
0.46
0.35

52.00
60.00
69.20
81.50

2.70
2.50
2.00
1.68

61.50
58.70
53.95
51.11

Fig. 2. Relationship between pressure and ux.

A. Ahmad et al. / Desalination 355 (2015) 110

mass transfer except thermal-induced phase separation which primarily depends on heat transfer [25].
In this paper, a detailed study towards the synthesis of thermomechanically endurable membranes is reported. These membranes
were prepared by using cellulose acetate, polyethylene glycol and silica.
A 2-stage phase-inversion protocol was devised involving thermallyinduced phase inversion followed by a controlled evaporation procedure. The chemical alteration, thermal stability, mechanical strength
and morphological structure of membranes were characterized by
attenuated total reectance-Fourier transform infrared spectroscopy
(ATR-FTIR), thermogravimetric analysis (TGA), differential scanning
calorimetry (DSC), ultimate tensile strength and scanning electron
microscopy (SEM) respectively. The effects of addition of SiO2 particles
on the membrane properties were investigated by examining the
permeation performance, such as pure water ux and solute rejection.
We reported here that the incorporation of silica particles simultaneously
improves the selectivity and permeability of membranes.

2. Materials and methods

Cellulose acetate (CA, Mw 30,000 and acetyl content 39%), polyethylene glycol-600 (PEG) and acetone were supplied by BDH laboratories
supplies, Poole, England. Silica (SiO2) was purchased from Sigma
Aldrich (England). All chemicals and solvents of analytical grade were
used as received.

2.1. Preparation of doped solutions

10 g of CA was dissolved in 80 mL of acetone with constant stirring at
80 C for 2 h. To this homogenous solution, 10 g of PEG was added with
regular stirring at 80 C for 6 h. The viscous and clear solution was
obtained, which was termed as a blended doped Solution, CA/PEG-1
[26]. The different ratios of CA/PEG (12/8, 14/6 and 16/4 g/g) were
used to prepare three additional doped solutions, labeled as CA/PEG-2,
CA/PEG-3 and CA/PEG-4 respectively. The casting solutions were
allowed to cool down to room temperature (25 C) and were kept for
24 h in a sealed ask to remove micro bubbles formed in the solution.

2.2. Casting of membranes

The doped solution was spread slowly on glass plate by maintaining
uniform thickness with the help of micrometer adjustable lm applicator
(Ref: 1117/300 Sheen instruments). The temperature of casted membranes was lowered immediately to 0 C for 15 min to induce (TIPS).
The TIPS causes the formation of dense and asymmetric structure. It
was followed by precipitation under controlled evaporation by increasing temperature up to 60 C [27]. The skinned membranes, thus formed
[28] were carefully removed from the glass plates by using a sharp
knife. The thickness of the resulting membranes was measured and
was found in the range of 50200 microns. These membranes were tested on the basis of their permeation performance and the best performing
CA/PEG-4 (control) membrane was selected for further modication
with silica particles.

Table 2
Effect of different concentration of silica on water content, ux, salt rejection and
membrane permeability.
Membrane
type

Water
content (%)

Flux
(L/h m2)

Salt rejection (%)

Membrane permeability
(104 L/h m2 kPa)

CPS-1
CPS-2
CPS-3
CPS-4
CPS-5

58.11
66.71
75.33
80.33
82.27

1.28
1.49
1.79
2.09
2.46

83.00
85.50
89.50
92.00
90.00

3.18
3.50
4.05
4.93
5.32

3. Characterization
3.1. Membrane performance
3.1.1. Experimental set up of RO rig
The permeation experiments were carried out in the RO experimental rig using a plate and frame membrane module. A schematic of process ow diagram for a plate and frame membrane module is shown
in Fig. 1. It consisted of the feed tank of 10 L capacity. The temperature
of the feed solution was indicated and controlled by a thermocouple
and controller setup. The feed was circulated using a circulation pump
with a max power rating of 1 kW at a max feed rate of 2500 L/min.
The effective membrane area in contact with the feed was 0.018 m2.
The permeate was collected from the sample points provided after the
membrane module. A saline solution (NaCl solution) having a conductance of 20 milli-siemens (mS) and a concentration of 10 g/L was subjected to the feed tank. Each membrane was compacted with distilled
water at 700 kPa for 5 h till it attains steady state. Feed temperature
was kept at 30 C during the operation.
3.1.2. Water content
The ability of membranes to absorb water was determined after
soaking membranes in water for 24 h and weighing afterwards. The
wet membranes were then placed in a vacuum dryer at 75 C for 48 h
and the weights of the dry membranes were recorded [29]. The water
content, absorbed by the membranes, was determined by the following
relation,

Water content %



W w WD
X100
WD

where Ww is wet sample weight and WD is dry sample weight.

2.3. Modication of doped solutions

Silica (15%, w/v) was dispersed in alkaline solution (10 mL) before
adding to the CA/PEG-4 blended dope solution with constant stirring
for 2 h at 80 C. The membranes (CPS-1CPS-5) were casted and dried
following the same procedure as mentioned in Section 2.2. Control
and modied membranes were transparent, exible and clear with
smooth surface.

Fig. 3. Flux of control and modied membranes over time.

A. Ahmad et al. / Desalination 355 (2015) 110

Fig. 4. Droplet images on control and modied membranes.

3.1.3. Permeate ux (J)

The permeate ux (J) represents the amount of pure water collected
per unit time and per unit area at variable pressures. It was calculated by
[30]:

3.1.4. Salt rejection

In reverse osmosis, the solute or salt rejection is dened as the
ratio of the concentration difference across the membrane and
expressed as [31],

J Q=t  A



R % 1Cp =C f x 100

where J is the permeate ux (L/h m2), Q is the amount of permeate (L), t

is the time (h) and A is the membrane area (m2).

where R is the salt rejection (%), C p and Cf are the conductance of

permeate and feed respectively.

Fig. 5. Contact angle of control and modied membranes.

Fig. 6. FTIR spectra of control and modied membranes.

A. Ahmad et al. / Desalination 355 (2015) 110

Fig. 8. Derivative weight loss of control and modied membranes.

Fig. 7. TGA curves of control and modied membranes.

3.5. Thermal analysis

3.2. Membrane permeability
Membrane permeability indicates the tolerance of membrane
towards hydraulic pressure. To determine Pm, the ux was measured
at different trans membrane pressures (P). The pressure during the
operation was increased from 450 to 800 kPa. The permeability of the
membrane was evaluated from the slope obtained by plotting permeate
ux versus membrane pressure difference (P), using Eq. (4)

Pm

Differential scanning calorimetry/thermal degradation measurements

of the membrane samples were performed using TA instruments SDTQ600 thermo gravimetric analyzer at a heating rate of 10 C/min and
temperature programs for DSC were run from 30 C to 800 C. The thermograms for TGA were obtained using a cycle of heating from an ambient
temperature to 800 C. Nitrogen ow was maintained at 15 mL/min in
order to remove all corrosive gas involved in the degradation.

3.6. Mechanical stability

J
P

where Pm is the membrane hydraulic permeability [L h m2/kPa] and

P is a pressure difference (kPa).

Tensile testing of membranes was performed at room temperature

using an Instron-type tensile testing machine (Testometric/M350-10CT
ROCHDALE, England) with a crosshead speed of 3 mm/min. The sample
width was 25 mm and the length between the jaws was 45 mm. The
measurements were performed at 25 C and 25 2% relative humidity.

3.3. Contact angle measurement

3.7. Scanning electron microscopy (SEM)
The contact angles were measured by a Drop Shape Analyzer (DSA
30, KRUSS Germany) through sessile drop method. A dangling droplet
of 5 L of deionized water was dropped carefully from I-shaped microsyringe to membrane surface. The contact angles were calculated by
the software. Each contact angle was measured at ten different points
of each membrane and an average value was calculated [32].

Scanning electron microscopy was performed on Jeol JSM-6480. The

membranes were cut into small pieces and placed on stub and kept in a
specic chamber in a vacuum. The electron beams were sputtered on
sample and images of membranes on varying resolutions were observed.

3.4. Fourier Transform infrared spectroscopy (FTIR)

FTIR of membranes was scanned by using a Shimadzu IR Prestige-21
equipped with Horizontal Attenuated Total Reectance (HATR) accessory. The experiments were run with air as the background. For each
spectrum 100 scans were accumulated with a resolution of 4 cm1.

Table 3
Thermal Stability of control and modied membranes.
Membrane
Type

Tmin
(C)

Tmax
(C)

Weight loss
at Tmax (%)

Residue at
end (%)

Control
CPS-1
CPS-2
CPS-3
CPS-4
CPS-5

270
300
330
340
350
360

400
460
480
500
510
520

87.85
89.19
89.14
89.14
89.14
89.14

1
5.68
5.74
5.80
5.92
6.72

Fig. 9. DSC curves of the control and modied membranes.

A. Ahmad et al. / Desalination 355 (2015) 110

4. Results and discussions

4.1. Membrane performance
The ux of all doped and compacted membranes was calculated at a
pressure of 650 kPa for 7 h. The water content, ux and salt rejection
were measured by using Eqs. (1), (2) and (3) respectively. The water
content, water ux, salt rejection and membrane permeability are
represented in Table 1. CA/PEG-1 membrane showed a maximum ux
of 0.87 L/h m2 while it exhibited minimum salt rejection capacity of
52%. CA/PEG-4 membrane had 81.5% salt rejection and its ux was
0.35 L/h m2.
The CA/PEG-1 membrane, with the highest water content and ux,
showed reduction in salt rejection capacity while CA/PEG-4 membrane
with the lowest water content and ux, had an elevated salt rejection
capacity as indicated in Table. 1. It was observed that with the increased
content of cellulose acetate, the water content, ux and membrane
permeability was decreased while the salt rejection capacity was
increased remarkably.
The transport mechanism within membrane can best be explained
on the basis of the solution diffusion model. According to this model,
transport, process within the membrane involves three steps: sorption
at the surface of the membrane, diffusion into dense membrane under
pressure and then desorption. The hydrophilic nature of PEG acted as
driving force for sorption of water into membrane [3336]. The
decrease in water content was attributed to the decrease in hydrophilic
PEG content.
It was deduced pragmatically that salt rejection was inversely related
to salt passage. Cellulose acetate being desalting in nature was responsible for the removal of salt [37]. This was a reason that CA/PEG-4 had
maximum salt rejection ability. When salt rejection efciency of membrane was at its best, at the same time ux was reduced, depending
upon the varying amount of CA and PEG.
PEG is hydrophilic in nature and acts as pore former [38]. CA/PEG-1
membrane, with maximum quantity of PEG, showed increase in water
ux, but at the same time salt rejection was compromised. This might
be attributed to the formation of macro voids [39] on membrane
which allowed the passage of salt along with water. Moreover, the
diffusion rate of water was accelerated by the presence of PEG due to
its hydrophilic nature. PEG increased the tendency of pore formation
and as a consequence ux was increased.
It was observed that within the pressure range of 450 to 650 kPa, the
ux decreased linearly with increase in trans membrane pressure as
indicated in Fig. 2. These results are in agreement with the reported
data [40].
The decline of ux with the passage of time, even at high pressure
was attributed to the compaction phenomena occurring on the membrane surface. The compaction of membrane under pressure involved
rearrangement of polymeric chains which led to the change in membrane structure with low value porosity which consequently decreased
the ux. Similar results were obtained by Arthanareeswaran et al. [38].
On the basis of salt rejection and ux, CA/PEG-4 membrane was
selected for modication with silica. After modication these membranes
were characterized and compared with CA/PEG-4 (control). The water
content, ux, salt rejection and membrane permeability of modied
membranes are given in Table 2.
The water content of the membrane was related to the hydrophilicity of the membrane. Table 2 indicates that the water content and membrane permeability of modied membrane increased with increase in
silica amount. The relative increase in water content is 37.88%. The
increased water content may be due to the increase in membrane hydrophilicity since the high hydrophilicity facilitates the diffusion of water
molecule into the membrane. Similarly, the ux of all modied membranes also increased with the increase in silica content. CPS-5 membrane
has highest ux of 2.46 L/h m2. Similar results have been obtained by Yin
et al. for silica incorporation in polyamide membranes, which revealed

Fig. 10. Tensile stress and elongation-at-break of control and modied membrane.

that addition of silica nanoparticles increases the ux [41]. The embedding of silica particles in CA/PEG membrane can also disrupt the polymer
chain packing by creating microporous defects between inorganic particle
and polymer interface. The salt rejection increased from 81.5% to 92% for
CPS-4 membrane as a result of incorporation of silica particles [42]. However, a slight decrease was observed at highest silica loading (5% w/v).
This result can be attributed to the agglomeration of silica particles in
certain locations which affect the solute rejection.
4.1.1. Analysis of membrane fouling during permeation
Inorganic compounds can be adsorbed and/or deposited on the
membrane surface, causing membrane fouling which in turn reduce
the ux drastically. Fouling increases energy consumption and results
in high operating cost [42]. Fig. 3 shows the permeation ux of control
and modied membranes. It is attention-grabbing that ux for the control declined with the passage of time and attain a steady state, whereas
for the modied membranes, the ux increased remarkably with the
increase of SiO2 content. The increase in permeation ux was due to
the increase in surface hydrophilicity which was rendered by SiO2
particles. Similar observations were made for CA/SiO2 blend membranes
for ultraltration application [29].
4.2. Membrane hydraulic permeability (Pm)
Membrane hydraulic permeability is an intrinsic resistance of the
membrane. It was determined by subjecting the membranes to various
pressures (450, 500, 550, 600, 650, 700, 750 and 800 kPa) and the Pm
value was calculated from the slope of the linear relationship between
pressures versus ux.
The permeability of unmodied membranes is given in Table 1. The
presence of PEG had considerable effect on the hydraulic permeability.
Thus, when the PEG concentration was decreased from 10 to 4 gm
in the dope solution, there was a corresponding decrease in hydraulic permeability from 2.70 to 1.68 (104 L/h m2 kPa). It might be attributed to
the disappearance of pores which were formed by PEG. More particularly,

Table 4
Mechanical properties and contact angle of control and modied membranes.
Membrane type

TS (Mpa)

Eb (%)

Contact angle (deg)

Control
CPS-1
CPS-2
CPS-3
CPS-4
CPS-5

263.94
278.83
355.55
414.64
437.28
382.46

5.8
6.96
8.73
11.36
12.23
10.33

54.21
48.00
40.11
34.19
28.69
24.10

A. Ahmad et al. / Desalination 355 (2015) 110

CA/PEG-1 membrane with the highest PEG contents offered greatest

permeability. This result was in agreement with that of ux obtained
through permeation experiment.
The Pm values for the modied membranes are given in Table 2,
showing an increase with the increasing amount of silica from 1 to 5%
w/v. For the modied membranes, the permeability increased from
1.28 to 2.46 (104 L/h m2 kPa). The increase in permeability resulted
in an increase in permeate ux as indicated in Table 2. It was mainly due
to increase in membrane hydrophilicity with the incorporation of silica
particles which expedites diffusion and solubilization of membranes.

4.3. Contact angle

The contact angle is an important parameter for measuring surface
hydrophilicity. In general, a smaller contact angle corresponds to a
more hydrophilic material. Fig. 4 represents digital image of a liquid
droplet on a dry surface of all membranes in which the contact angle
() is measured according to the sessile drop method. It is evident
from Fig. 5 that there is decrease in contact angle by increasing silica
particle concentration. These results demonstrated that silica particle
can improve hydrophilicity of membrane. The less hydrophilic surface
shows larger contact angle with the surface and vice versa.

4.4. FTIR analysis

Spectroscopic methods play a crucial role in polymer characterization. The use of methods such as infrared spectroscopy is essential in
order to investigate molecular structure of membranes. FTIR spectrums
of the control and modied membranes are given in Fig. 6.
In case of control, peak at 3469 cm 1 represented stretching of
O\H, 2877 cm1 showed stretching of C\H bond, 1739 cm1 depicted
the strong peak for C_O of carbonyls, 1433 cm1 indicated bending of
C\H followed by peaks at 1369 cm1 and 1224 cm1 which described
rocking and wagging mode of C\H bond. The strongest peak at
1035 cm1 specied C\O\C while peak at 1112 cm1 and 904 cm1
illustrated the presence of saccharide. Similar results had been previously reported in the literature [43,44]. The emergence of new peak
around 950 cm1 in the spectra of all modied membranes could be
attributed to stretching vibration of Si\O\. The broadening of peak
around 32003650 cm1 was due to H-bonded Si\OH [45]. The broadening and emergence of new peak in the spectra of modied membranes
conform the impregnation of silica particles within the membrane
matrix.

4.5. Thermal analysis

The thermal decomposition is expressed in terms of weight loss as a
function of the temperature for the control and modied membranes as
shown in Fig. 7. Each TGA curve shows three main degradation steps.
The rst step took place between 30 C and 250 C and displayed
about 3.5% of weight loss. This degradation was associated with the volatilization of the volatile matter and/or the evaporation of residual
absorbed water. The second stage of degradation started from 260 C
and continued up to 400 C. There was 87.85% weight loss occurring
in second stage due to degradation of cellulose acetate chain. The
temperature at which maximum degradation occurs (Tmax) is 400 C.
Similar results were observed by Arthanareeswaran et al. [46]. The
third step, started at 400 C symbolized the carbonization of degraded
products to ash.
It was observed that the temperature at which degradation initiated
(Tmin) is relatively higher for all the modied membranes as indicated in
Table 3. Similarly, Tmax for all the membranes shifted towards higher
temperature (400 C to 520 C), with increasing amount of silica. The
substantial increase in thermal stability was due to the incorporation
of silica within the matrix of cellulose acetate. For all the modied membranes, when the temperature exceeded higher than 550 C, the mass
remained almost constant, which showed that the residue was mostly
metal oxide [47] whereas in case of control sample 1% residue indicates
the presence of carbon ash.
The rst derivative of the mass-change with respect to time (dm/dt,
called derivative thermogravimetry (DTG)) was calculated and plotted
as a function of temperature (Fig. 8). Using DTG analysis, a series of
peaks were obtained instead of stepwise TG curves. The areas under
the peaks were proportional to the total mass-change of the membrane
sample. The DTG is more precise in showing the onset of a reaction that
is accompanied by mass loss and also permits the reaction temperatures
to be dened more precisely. The DTG peaks are associated with the
onset of the second stage of decomposition. These peaks shifted towards
higher temperature, which indicates an increase in stability of modied
membranes.
4.6. DSC studies
In a DSC analysis, the glass transition temperature (Tg) is an indicative of the structure and packing of the polymer chain in the membrane.
A lower glass transition temperature (Tg) indicated that the membrane
possessed more free volume and therefore a loose structure [43]. The
DSC thermogram of control and modied membranes are given in
Fig. 9. It was observed that the control exhibited a Tg at 60 C. It was

Fig. 11. SEM images of control (a) top surface and cross-section (b).

A. Ahmad et al. / Desalination 355 (2015) 110

interesting to note that Tg of modied membranes, increased from 65 to

80 C. CPS-4 membranes revealed Tg of 80 C. The increase in Tg was
due to the inclusion of silica within CA/PEG membrane. Similar results
were suggested by Nunes et al. [48]. The increase in Tg resulted in higher
packing density and rigidity. These results were in agreement with the
mechanical properties of membranes.
4.7. Mechanical stability
In order to investigate the effect of silica on mechanical properties of
membranes, the tensile stress (TS) and elongation-at-break (Eb) of control and modied membranes were measured and shown in Fig. 10. It
was noteworthy that the elongation-at-break, tensile stress and Young's
modulus initially increased with the addition of SiO2 particles and
reached at peak when the SiO2 particle concentration was 4% (w/v)
and then declined as the SiO2 particle concentration was further
increased. Yan et al. made similar observations for Al2O3/PVDF blend for

UF membrane. The initial increase may be due to increase in interaction

between the silica particles and polymers. However, an excessive concentration may cause the aggregation of particles and prevent it to disperse
uniformly in a polymeric matrix. The large phase separation results in
the decline of mechanical stability of the membrane [49]. These behaviors indicate that an appropriate amount of SiO2 particles in CA/PEG-4
solution can improve the mechanical properties of the membrane.
The salt rejection and mechanical behavior of modied membranes
showed identical pattern i.e. with increasing silica particles there was
increased in salt rejection in addition to mechanical strength. But at
maximum loading of silica particles, there was decline in salt rejection
as well as mechanical behavior as shown in Tables 2 and 4.
4.8. SEM analysis
To attain high performance membranes for specic applications, it is
essential to manipulate the morphological structures of the membranes.

Fig. 12. SEM images of modied membranes (a) top surface (b-f) cross sectional.

A. Ahmad et al. / Desalination 355 (2015) 110

Therefore, the morphological studies of various sections of control and

surface modied membranes were made by using scanning electron
microscope. The scanning electron micrographs of CA/PEG blend membrane are shown in Fig. 11. It reects the dense asymmetric composite
structure, with top, skinned-layer and sub-layer.
Fig. 12 (af) represents the images of modied membranes with
different silica loadings. According to SEM images, the surface morphologies of the membranes were changed signicantly with the addition of
silica particles. The top surface morphology is shown on Fig. 12(a). It
was observed that micro-size silica particle (0.51 m) were distributed
in the membrane. The interaction of silica particles with polymer
disrupts the mobility of polymeric chains resulting in the formation of
macroscopic defects [29]. But as concentration increased the defects
were lled with silica aggregates as observed in Fig. 12(f).
5. Conclusions
In this study, thermo-mechanically stable membranes were synthesized through a 2-stage phase inversion method. A series of CA/PEG
blended membranes showed a change in ux and permeability with
change in ratio of PEG and CA. The membrane with optimum ux and
salt rejection was selected for modication with silica. The modied
membranes resulted into an improved trend in the salt rejection and
permeation properties. CPS-4 had 92% salt rejection which provided
almost 11% increase in salt rejection capacity compared with the control
membrane (81.5%). The mechanical stability of the modied membranes, increased from 1 to 4% (w/v) of silica loading. Further increase
in SiO2 particles in the casting solution resulted in the decrease of tensile
strength and elongation at break. Similarly, Tmax and Tg of all modied
membranes remarkably increased as evident from TGA and DSC analysis.
The SEM images revealed incorporation of silica particles within the
asymmetric composite membrane. The improved hydrophilicity of the
composite membrane also enhanced fouling resistance. This study
showed a critical need for optimizing the silica particle loading, as
overloading of silica may not be advantageous for RO performance and
has negative impact on the RO membrane properties.

Acknowledgments
The authors express their cordial gratitude to the team of Department
of Polymer Engineering and Technology, University of the Punjab, for
their co-operation during the execution of this research project. The authors are also thankful to the University of Malaysia Sarawak (UNIMAS),
which supported this work through grant DPI [Grant ref. no. 02 (DPI23)/
999/2013(06)].

References
[1] W. Lang, J. Shen, Y. Zhang, Y. Yu, Y. Guo, C. Liu, Preparation and characterizations
of charged poly(vinylbutyral) hollow ber ultraltration membranes with
peruorosulfonic acid as additive, J. Membr. Sci. 430 (2013) 110.
[2] H. Strathmann, L. Glomo, E. Drioli, An Introduction to Membrane Science and
Technology, Consiglio Nazionale Delle Richerche, Italy, 2006.
[3] I.G. Wenten, Membr. Sci. Technol. 24 (2002) 10091024.
[4] S. Sejal, R. Desai, Polymer membrane technology, Int. J. Eng. Sci. Innov. Technol. 2
(2013) 400403.
[5] H. Strathmann, Membrane separation processes: current relevance and future
opportunities, AIChE J. 47 (2001) 10771087.
[6] H. Strathmann, Membrane separation processes, J. Membr. Sci. 9 (1981) 121189.
[7] H. Strathmann, Membranes and membrane separation processes, 1, Principles, in
Ullmann's Encyclopedia of Industrial Chemistry, Wiley, Germany, 2012.
[8] H.K. Lonsdale, The growth of membrane technology, J. Membr. Sci. 10 (1982) 81181.
[9] B.V. Bruggen, Desalination by distillation and by reverse osmosistrends towards
the future, Membr. Technol. (2003) 69.
[10] B.V. Bruggen, C. Vandecasteele, T.V. Gestel, W. Doyenb, R. Leysenb, A review of
pressure-driven membrane processes in wastewater treatment and drinking
water production, Environ. Prog. 22 (2003) 4656.
[11] C. Fritzmann, J. Lowenberg, T. Wintgens, T. Melin, State-of-the-art of reverse osmosis
desalination, Desalination 216 (2007) 176.

[12] L.F. Greenleea, D. Lawler, B.D. Freeman, B. Marrot, P. Moulin, Reverse osmosis
desalination: water sources, technology and today's challenges, Water Res. 43
(2009) 23172348.
[13] D. Li, H. Wang, Recent developments in reverse osmosis desalination membranes,
J. Mater. Chem. 20 (2010) 45514566.
[14] K. Lee, T.C. Arnot, D. Mattia, A review of reverse osmosis membrane materials for
desalinationdevelopment to date and future potential, J. Membr. Sci. 370
(2011) 122.
[15] M.R. Garud, S.V. Kore, V.S. Kore, S.G. Kulkarni, A short review on process and
applications of reverse osmosis, Univ. J. Environ Res.Technol. 1 (2011) 233238.
[16] P. Rustemeyer, History of CA and evolution of the markets, Macromol. Symp. 208
(2004) 16.
[17] S. Fischer, K. Thummler, B. Volkert, K. Hettrich, I. Schmidt, K. Fischer, Properties and
applications of cellulose acetate, Macromol. Symp. 262 (2008) 8996.
[18] J.S. Taurozzia, H. Arul, V.Z. Bosak, A.F. Burban, T.C. Voice, M.L. Bruening, V.V.
Tarabara, Effect of ller incorporation route on the properties of polysulfone-silver
nanocomposite membranes of different porosities, J. Membr. Sci. 325 (2008) 5868.
[19] M. Zafar, M. Ali, S.M. Khan, T. Jamil, M.T.Z. Butt, Inuence of glycol additives on the structure and performance of cellulose acetate/zinc oxide blend membranes, Desalination
270 (2011) 98104.
[20] D. Emadzadeh, W.J. Lau, T. Matsuura, M. Rahbari-Sisakht, A.F. Ismail, A novel thin
lm composite forward osmosis membrane prepared from PSfTiO2 nanocomposite
substrate for water desalination, Chem. Eng. J. 237 (2014) 7080.
[21] G. Arthanareeswarana, P. Thanikaivelan, Fabrication of cellulose acetatezirconia
hybrid membranes for ultraltration applications: performance, structure and fouling
analysis, Sep. Purif. Technol. 74 (2010) 230235.
[22] N.M. Wara, L.F. Francis, B.V. Velamakanni, Addition of alumina to cellulose acetate
membranes, J. Membr. Sci. 104 (1995) 4349.
[23] G.T. Caneba, D.S. Soong, Polymer membrane formation through the thermal-inversion
process. 1. Experimental study of membrane structure formation, Macromolecules 18
(1985) 25382545.
[24] D. Li, W.B. Krantz, A.R. Greenberg, R.L. Sani, Membrane formation via thermally
induced phase separation (TIPS): model development and validation, J. Membr.
Sci. 279 (2006) 5060.
[25] R.K. Arya, Drying induced phase separation in multicomponent polymeric coatings
simulation study, Int. J. Sci. Technol. Res. 1 (2012) 4853.
[26] H.K. Lonsdale, U. Merten, R.L. Riley, J. Jay, Transport properties of cellulose acetate
osmotic membranes, J. Appl. Polym. Sci. 9 (1965) 13411362.
[27] 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, J. Membr. Sci. 34 (1987) 6786.
[28] J. Li, K. Nagai, T. Nakacawa, S. Wang, Preparation of polyethyleneglycol (PEG) and
cellulose acetate (CA) blend membranes and their gas permeabiIities, J. Appl.
Polym. Sci. 58 (1996) 14551466.
[29] G. Arthanareeswaran, T.K.S. Devi, M. Raajenthiren, Effect of silica particles on cellulose acetate blend ultraltration membranes: Part I, Sep. Purif. Technol. 68 (2008)
3847.
[30] E. Saljoughi, M. Sadrzadeh, T. Mohammadi, Effect of preparation variables on
morphology and pure water permeation ux through asymmetric cellulose acetate
membranes, J. Membr. Sci. 326 (2009) 627634.
[31] S. Kawk, S.H. Kim, Hybrid organic/inorganic reverse osmosis (RO) membrane for
bactericidal anti-fouling. 1. Preparation and characterization of TiO2 nanoparticle
self-assembled aromatic polyamide thin-lm-composite (TFC) membrane, Environ.
Sci. Technol. 35 (2001) 23882394.
[32] I. Guezguez, B. Mrabet, E. Ferjani, XPS and contact angle characterization of surface
modied cellulose acetate membranes by mixtures of PMHS/PDMS, Desalination
313 (2013) 208211.
[33] M.A. Mazid, Mechanisms of transport through reverse osmosis membranes, Sep. Sci.
Technol. (1984) 357373.
[34] M. Zafar, M. Ali, S.M. Khan, T. Jamil, M.T.Z. Butt, Effect of additives on the properties
and performance of cellulose acetate derivative membranes in the separation of
isopropanol/water mixtures, Desalination 285 (2012) 359365.
[35] H.G. Burghoff, K.L. Lee, W. Pusch, Characterization of transport across cellulose
acetate membranes in the presence of strong solutemembrane interactions,
J. Appl. Polym. Sci. 25 (1980) 323347.
[36] R. Malaisamy, R. Mahendran, D. Mohan, M. Rajendran, V. Mohan, Cellulose acetate
and sulfonated polysulfone blend ultraltration membranes. I. Preparation and
characterization, J. Appl. Polym. Sci. 86 (2002) 17491761.
[37] B. Chakrabarty, A.K. Ghoshan, M.K. Purkait, Effect of molecular weight of PEG on
membrane morphology and transport properties, J. Membr. Sci. 309 (2007)
209221.
[38] G. Arthanareeswarana, D. Mohan, M. Raajenthiren, Preparation, characterization and
performance studies of ultraltration membranes with polymeric additive, J.
Membr. Sci. 350 (2010) 130138.
[39] H. Bai, Y. Zhou, X. Wang, L. Zhang, The permeability and mechanical properties of
cellulose acetate membranes blended with polyethylene glycol 600 for treatment
of municipal sewage, Procedia Environ. Sci. 16 (2012) 346351.
[40] M.T.M. Pendergast, E.M.V. Hoek, A review of water treatment membrane nanotechnologies, Energy Environ. Sci. 4 (2011) 19461971.
[41] J. Yin, E. Kim, J. Yang, B. Deng, Fabrication of a novel thin-lm nanocomposite
(TFN) membrane containing MCM-41 silica nanoparticles (NPs) for water purication, J. Membr. Sci. 426 (2012) 238246.
[42] L. Zoua, I. Vidalis, D. Steele, A. Michelmore, S.P. Low, J.Q.J.C. Verberk, Surface hydrophilic modication of RO membranes by plasma polymerization for low organic
fouling, J. Membr. Sci. 369 (2011) 420428.

10

A. Ahmad et al. / Desalination 355 (2015) 110

[43] C. Liu, R. Bai, Preparation of chitosan/cellulose acetate blend hollow bers for
adsorptive performance, J. Membr. Sci. 267 (2005) 6877.
[44] S. Waheed, A. Ahmad, S.M. Khan, S. Gul, T. Jamil, A. Islam, T. Jamil, Hussain, Synthesis, characterization, permeation and antibacterial properties of cellulose acetate/
polyethylene glycol membrane modied with chitosan, Desalination 351 (2014)
5969.
[45] P. Innocenzi, Infrared spectroscopy of solgel derived silica-based lms: a
spectra-microstructure overview, J. Non-Cryst. Solids 316 (2003) 309319.
[46] G. Arthanareeswaran, P. Thanikaivelan, K. Srinivasn, D. Mohan, M. Rajendran, Synthesis,
characterization and thermal studies on cellulose acetate membranes with additive,
Eur. Polym. J. 40 (2004) 21532159.

[47] Y. Wang, L. Yang, G. Luo, Y. Dai, Preparation of cellulose acetate membrane lled with
metal oxide particles for the pervaporation separation of methanol/methyltertbutyl
ether mixtures, Chem. Eng. J. 146 (2009) 610.
[48] S.P. Nunesa, K.V. Peinemann, K. Ohlroggea, A. Alpers, M. Keller, A.T.N. Pires,
Membranes of poly(ether imide) and nanodispersed silica, J. Membr. Sci. 157
(1999) 219226.
[49] L. Yan, Y.S. Li, C.B. Xianga, S. Xianda, Effect of nano-sized Al2O3-particle addition on
PVDF ultraltration membrane performance, J. Membr. Sci. 276 (2006) 162167.

All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.