Desalination 336 (2014) 138–145

Contents lists available at ScienceDirect

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

Controlling swelling behavior of poly (vinyl) alcohol via networked

cellulose and its application as a reverse osmosis membrane
Shaheen Fatima Anis, Boor Singh Lalia, Raed Hashaikeh ⁎
Institute Center for Water Advanced Technology and Environmental Research (iWATER), Masdar Institute of Science and Technology, P.O. Box 54224, Abu Dhabi, United Arab Emirates

H I G H L I G H T S

• Swelling of poly (vinyl) alcohol (PVA) controlled by networked cellulose (NC)

• Swelling capacity of PVA reduced from 340% to 150%.
• The tensile strength of wet PVA improved by 1520%.
• Membranes successfully tested for RO applications
• Optimum results obtained from PVA-15 wt.% NC with 98.9% salt rejection rate

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

Article history: The hydrophilic nature of polyvinyl alcohol (PVA) makes it a promising membrane material for water treatment
Received 14 November 2013 applications but its use is restricted by its high swelling capacity in water. We found that the swelling behavior of
Received in revised form 5 January 2014 PVA can be controlled via aqueous blending of PVA with a modified form of cellulose-networked cellulose (NC).
Accepted 6 January 2014
Dense PVA–NC membranes were obtained by varying NC concentration at 3, 10, 15, 20 and 30 wt.%. Fabricated
Available online 25 January 2014
membranes were characterized for their mechanical, thermal and structural properties. It was found that
Keywords:
15 wt.% of NC was optimum. The swelling capacity of PVA was reduced from 340% to 150%. The tensile strength
Polyvinyl alcohol of wet PVA improved 1520% and the elastic modulus improved 1400%. The fabricated membranes were tested for
Networked cellulose desalination by reverse osmosis (RO) using aqueous solution of sodium chloride with 25,000 ppm feed concen-
Swelling tration and an applied pressure of 24 bars. A 272 μm thick membrane allowed a water flux value of 1.2 l/h·m2
Reverse osmosis with a salt rejection of 98.8%.
Membranes © 2014 Elsevier B.V. All rights reserved.

1. Introduction rejection rate [10]. Many different methods of modifying PVA have been
explored previously which include physically crosslinking the PVA [9],
Polyvinyl alcohol (PVA) has been widely studied for various water chemically crosslinking [7,11,12] and heat treatment [10,13,14].
treatment applications including reverse osmosis, nano-filtration and Among these, chemical crosslinking has been extensively used to
ultra-filtration [1–4]. PVA is a water soluble polymer with inherent impart mechanical stability to the PVA membranes [1]. The problems
hydrophilicity making it an attractive choice for membrane applications associated with these various crosslinking methods have been
[5]. It possesses good film forming ability [6], good anti-fouling proper- explained elsewhere [5]. Therefore, a tight restraint or network
ties [7] and physical and chemical stability. In addition it is also nontoxic that holds onto PVA is required. Such a network should prevent
[8]. Despite such desirable properties, the hydrophilic nature of PVA the polymer from swelling in water without adversely affecting its
increases its swelling capacity and renders it inefficient to be used in mechanical and thermal stability.
its neat condition. Modification of PVA is necessary to prevent it Cellulose is the most abundant biodegradable polymer on earth
from compacting or collapsing under pressure so it is able to retain and can be extracted from plants, fungi, algae and bacteria. It may be
contaminants over the required period of time [5]. modified into several forms with different attractive properties not
Reverse osmosis (RO) requires high quality of water to be produced usually found in its ground state. Modification of cellulose leads to
[1]. For such applications, swelling of PVA poses a serious challenge that changes in structure, morphology and crystallinity [15]. Previous
needs to be overcome [9]. Swollen membranes render the polymer studies show that various forms of cellulosic materials have been incor-
highly permeable to both water and salt. This results in a low salt porated with PVA for membrane applications. These include cellulose
whiskers reinforced with PVA [16], cellulose acetate mixed with PVA
⁎ Corresponding author. Tel.: +971 28109152. [1] and cellulose nano-crystals blended with PVA [17]. Our group has
E-mail address: rhashaikeh@masdar.ac.ae (R. Hashaikeh). reported the fabrication of a modified form of cellulose-networked

0011-9164/$ – see front matter © 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.desal.2014.01.005
 S.F. Anis et al. / Desalination 336 (2014) 138–145 139

Poly (vinyl)
Casting of the Wetting the Reverse
Alcohol- Wetting the as- Freeze drying
solution blend membranes Osmosis Test
Networked cast dried the wetted
in Aluminium before final using NaCl salt
Cellulose membranes membranes
molds testing solution
polymer blend

Fig. 1. Flow scheme for membrane preparation of PVA-networked cellulose.

cellulose (NC), which possess an open structure. This modified, and stabilized to 5 °C. Sulphuric acid (99.99%) was diluted to 70%
regenerated cellulose is more amorphous than the starting micro- (w/w) concentration. The bath temperature was maintained at 5 °C
crystalline cellulose (MCC) from which it is initially made up of. and the diluted acid was added subsequently. 10 g of MCC was
The cellulosic chains in NC arrange themselves in a form of a network added to 100 ml of sulphuric acid and the resulting solution was
after the acid dissolution and regeneration in an anti-solvent. This then stirred for 30 min at 5 °C at 250 rpm until a viscous and
open network structure gives cellulose high mechanical and thermal transparent liquid of cellulose and sulphuric acid was formed. This
stability [18]. cellulose solution was mixed with ethanol at 5 °C to regenerate the
In the present work, we have studied the effect of adding NC on PVA cellulose. The mixture was further mixed for 10 min to allow for
membranes with respect to swelling, crystallinity, hydrophilicity as well complete regeneration of cellulose. Centrifugation of the resulting
as mechanical and structural stability. Membranes with different material at 2400 rpm and 4 °C allowed decanting of the top acidic
concentrations of NC were fabricated and then tested for RO layer. This centrifugation process was repeated three times to remove
desalination. the acid and to separate the precipitated material. The precipitate was
collected and dialyzed in a Spectra/Por (MWCO: 12–14,000) dialysis
2. Experimental membrane under running tap water for three days until the pH of the
suspension reached 6–7. The resulting NC suspension was sonicated
2.1. Materials using Hielscher Ultrasonic Processor UP400S for 1 h at 60 °C. The
mixture was further homogenized by a mechanical homogenizer
PVA (Mw = 145,000 and 99.0–99.8 mol% hydrolysed), ethanol and (IKA-T25 ULTRA-TURRAX).
sulphuric acid (99.99%) were purchased from Aldrich (St. Louis, MO).
MCC was provided by FMC Biopolymer (Philadelphia, PA). 2.3. Fabrication of PVA–NC membranes

2.2. Preparation of networked cellulose suspension PVA solution of 5 wt.% was prepared by mixing PVA powder in
deionized (DI) water at 150 °C until a homogenized solution was
Cellulose was modified via acid Hydrolysis using a Varian® obtained. Solutions of PVA with 3 wt.% NC, 5 wt.% NC, 10 wt.% NC,
dissolution system following a procedure reported earlier [18]. The 15 wt.% NC, 20 wt.% NC and 30 wt.% NC were allowed to cast in Alumin-
temperature of the bath in the dissolution system was adjusted ium molds and left to dry for 3 days at ambient temperature and a
humidity of 50%. The self-supporting as cast membranes were wetted
and then freeze dried using wizard 2.0 by VIRTIS freeze drier system.
Fig. 1 shows the flow chart for membrane preparation.

2.4. Membrane characterization

2.4.1. Mechanical characterization

Tensile testing was performed on vacuum dried and wet samples of
PVA–NC membrane using Instron (model No. 5982). Tensile strength
and Young's modulus were obtained with a load cell of 5 kN at a strain
rate of 1 mm/min. Thickness of the membranes was measured using
a micrometer (Mitutoyo, Japan; detection limit = 0.001 mm). The
thickness of the dog bone specimens used for tensile testing was in
the range of ≈170–280 μm.

2.4.2. Structural and thermal characterization

Prepared PVA–NC membranes were tested for the amount of
water they can retain in themselves. The water content in pure PVA,
PVA-3 wt.% NC, PVA-10 wt.% NC, PVA-20 wt.% NC and PVA-30 wt.%
NC was studied. The dry weight of the membranes was measured
through Citizen weighing machine (model No: CY 204) after which
they were soaked in water for 48 h. The wet membranes were taken
out from water and placed on tissue paper without applying any
pressure to remove excess water from the surface. Their weight was
then measured and the following formula was used to calculate the
percentage of water content in the membranes:

wet weightðgÞ−dry weightðgÞ

% water content ¼  100:
dry weightðgÞ

To study the change in crystallinity and structural integrity,

X-ray diffraction was performed on vacuum dried films of neat PVA,
Fig. 2. In-lab RO unit. PVA-3 wt.% NC, PVA-15 wt.% NC, PVA-20 wt.% NC and PVA-30 wt.%
140 S.F. Anis et al. / Desalination 336 (2014) 138–145

Fig. 3. NC network becoming more evident with increasing concentration of NC (a) 2 wt.% (b) 4 wt.% (c) 10 wt.% and (d) TEM image of networked cellulose showing the open network
structure of the material.

NC using an X-ray diffractometer (PANalytical, Empyrean). The DSC 4000 by heating the sample from 50 °C to 450 °C at a rate of 5 °C/min
machine was operated at 45 kV and 40 mA with Ni-filtered CuKα under a nitrogen flow of 20 ml/min. DSC samples were prepared by
(λ = 1.5056 Å) radiations in 5–70° half angle range. Peak intensities sealing 10–20 mg of membrane material in an Aluminium pan.
were qualitatively compared while the full width at half maximum Thermo-gravimetric analysis (TGA) was also performed on selected
(FWHM) quantitatively indicated the change in crystallinity for the PVA–NC membranes using PerkinElmer TGA 4000 to study the degrada-
selected membrane films. tion behavior of PVA–NC membranes. Samples (20–30 mg) were placed
The structure of NC was viewed under transmission electron micro- in a sample pan and heated from room temperature to 800 °C at a rate
scope (TEM). The TEM sample was prepared by drying a small droplet of 10 °C/min with a nitrogen flow of 20 ml/min.
(b1 μl) of NC in water on the carbon coated copper grid. TEM images
were taken by FEI, Tecnai TF20 TEM. Bright field images were produced
by using a low intensity beam to minimize artifacts induced by the 2.4.3. Morphology
beam. Atomic force microscopy (AFM) was used to characterize the surface
Differential scanning calorimeter (DSC) was used to study variations roughness of the air dried membranes using Asylum Research, MFP-
in melting peak, glass transition and crystallinity of PVA–NC mem- 3DTM Stand Alone. Neat PVA and PVA-20 wt.% NC were compared for
branes with varying NC content. Testing was carried out on PerkinElmer their morphologies. Samples of 1 × 1 cm2 were placed on the glass

Fig. 4. Air dried PVA and PVA-20 wt.% NC membranes (a and d), PVA and PVA-20 wt.% NC membranes after soaking in water for 2 h (b and e) and, PVA and PVA-20 wt.% NC membrane
after 12 h soaked in water (c and f).
 S.F. Anis et al. / Desalination 336 (2014) 138–145 141

450 10
Neat PVA film
9
PVA+3NC
400 PVA+10NC
8 PVA+15NC
PVA+20NC

Tensile Strength (MPa)

7
Water content (%)

350 PVA+30NC

6
300 5

4
250
3

200 2

1
150
0

0 5 10 15 20 25 30 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

wt % of networked cellulose Strain (mm/mm)
Fig. 5. Change in water content with increasing percentage of NC after 48 h of soaking the
Fig. 6. Stress vs. strain behavior of selected PVA–NC membranes containing different wt%
membranes in water.
of NC in their wet states.

and the obtained value was converted to ppm total dissolved solids
substrate and a silicon tip radius of ≈9 nm was used with a resonance (TDS). Salt rejection rate was measured by the following formula:
frequency of ≈300 kHz in a non-contact tapping mode.
% salt rejection
2.4.4. Water contact angle measurement
feed concentrationðppmÞ−permeate concentrationðppmÞ
Surface hydrophilicity was determined using water contact angle ¼  100:
feed concentrationðppmÞ
measurements. The change in hydrophilicity with an increase in NC
concentration was measured by EASY DROP Contact Angle measure-
ment instrument from Kruss (Hamburg, Germany). A micro-syringe
was used to generate the liquid droplets of DI water (volume 2 μl) at 3. Results and discussion
room temperature on the membrane's surface. Measurements of the
contact angle (θ) were made through the tangent line and circle fitting 3.1. Physical observation and swelling behavior
methods and the average of the two values was recorded.
Air drying of PVA–NC solution showed that a more homogeneous
2.5. Reverse osmosis test solution was gradually formed with an increase in the NC concentration.
Moreover, a network formation of cellulose was physically observed as
PVA–NC membranes were tested for RO using aqueous solution of the solution was left to dry. The network formation enhances with
sodium chloride. Feed concentration of 25,000 ppm was used in the increasing percentage of NC in PVA (Fig. 3(a–c)).
in-lab reverse osmosis unit shown in Fig. 2. The osmotic pressure was The neat PVA and PVA-20 wt.% NC membranes were compared for
calculated through the Van't Hoff equation [19,20]. their swelling behavior by soaking them in water for 2 h and 12 h.
The PVA without NC was shown to swell more in water compared to
the PVA-20wt.%NC membrane. The PVA-20 wt.% NC membrane
П ¼ c  iRT remains intact and retains its shape in water even after a period of
12 h. The comparison is shown in Fig. 4. The neat PVA membrane was
where: П = osmotic pressure in Pa, c = molar concentration also seen to absorb more water than its counterpart as shown in Fig. 5.
(mol/m3), i = number of ions formed in the solution, R = universal Percentage of water content with increasing percentage of NC was
gas constant (J/mol-K) and T = temperature (Kelvin). Freeze dried studied for the course of 48 h. The percentage of water retention
membranes of diameter 41.8 cm were wetted before they were run. decreases with increasing percentage of NC implying less swelling but
Testing was performed for 30 min under a pressure of 24 bars. at the same time confirming sufficient water absorption even at a high
No supporting layer was used except the metal flange on which the percentage of NC. This is evident as the water content remains above
membrane was secured in the RO test unit. Permeate conductivity was 100% even with PVA-30 wt.% NC. Fig. 3(d) shows the TEM image of
measured by accumet® XL 50 dual channel pH/Ion/Conductivity meter NC attesting to its open networked structure. The open structure of NC

Table 1
Mechanical behavior of PVA–NC membranes compared with Neat PVA.

Dry state Wet state

Networked cellulose (wt%) Tensile strength (MPa) Tensile modulus (MPa) Tensile strength (MPa) Tensile modulus (MPa)

Neat PVA 25 1004 0.5 0.1

3 40 1595 3.7 0.2
5 17 1231 2.5 0.8
10 16 1352 6.5 1.1
15 20 1356 8.1 1.5
20 20 1670 4.2 2.9
30 25 1266 7.3 2.9
142 S.F. Anis et al. / Desalination 336 (2014) 138–145

Table 2 Table 3
Measure of hydrophilicity of PVA membranes. Thermal properties of selected membranes recorded from the DSC graphs.

Sample Contact angle (θ) Membrane type Tg (°C) Tm (°C)

Neat PVA 28.7 ± 0.23 Neat PVA 89 229

10 wt.%-NC 36.0 ± 2.15 PVA + NC3 101 227
20 wt.%-NC 37.6 ± 1.31 PVA + NC20 97 225
30 wt.%-NC 42.1 ± 2.66

entraps the PVA polymer inside it and does not let the PVA to expand or is spread for very small concentrations of NC in the solution (also
swell. This in turn imparts it stability against swelling. evident from the physical observation) which imparts it mechanical
The stability resulting from adding NC may also be due to the integrity by entrapping the PVA inside it. As the percentage is increased,
hydrogen bonding between the cellulose chains. This provides a NC starts to act as a reinforcing agent and a composite behavior is
compact structure and thus prevents the membrane from swelling. observed. The decrease in the tensile strength with increasing NC after
3 wt.% may be attributed by absorbed water vapor in the specimen
through the atmosphere. Due to the hydrophilic nature of both PVA
3.2. Mechanical stability and NC, this cannot be avoided. Even though strength decreases, it
was observed that PVA-15 wt.% NC, PVA-20 wt.% and PVA-30 wt.% NC
Mechanical properties of PVA–NC membranes are summarized in had strengths comparable to that of neat PVA.
Table 1. Testing was carried out for the membranes in both dry and Mechanical properties of PVA–NC membranes in their wet state
wet states. For dry testing, a very small percentage of NC is able to differ from that of dried membranes for reasons such as moisture
increase the strength of the material as the highest tensile strength is absorption which causes weakening of polymeric chains. Swelling
obtained for PVA-3 wt.% NC. This indicates control of NC content is is also responsible for the lower tensile strength of wet PVA–NC
critical for the formation of PVA–NC membranes. Initially, the network membranes. It is evident that the tensile strength increases with

a) b)
2.9
2.8
FWHM (degrees)

Neat PVA film

PVA+3NC
2.7
PVA+15NC
PVA+20NC 2.6
Intensity

PVA+30NC

2.5
2.4
2.3
2.2
2.1
0 10 20 30 40 50 60 70 80 0 5 10 15 20 25 30
2 -theta Percentage of Networked Cellulose

Fig. 7. (a) X-ray diffractograms of vacuum dried as cast membranes showing a decrease in peak intensity with increasing percentage of networked cellulose and (b) Increase in FWHM
values attesting to the decrease in crystallinity of the membranes with increasing percentage of NC.
Endo

Heat Flow
Heat Flow

70 80 90 100 110 120 130 140 150 160 170

o
Temperature ( C)
200 mW/g
Neat PVA film
PVA+3NC
Exo

PVA+20NC

100 200 300 400 500

Temperature (oC)

Fig. 8. DSC curves showing the melting and decomposition temperatures for selected PVA–NC membranes.
 S.F. Anis et al. / Desalination 336 (2014) 138–145 143

increases slightly with NC content, but the membranes still retain

their hydrophilic property as the contact angle does not exceed 42.1°
even when 30 wt.% NC is used. This is important as the hydrophilicity
of the membranes should not change considerably [5]. NC does not
render the PVA membrane hydrophobic due to the hydrophilic nature
of cellulose. Therefore, NC imparts physical stability to the polymer
without adversely affecting the wettability of the membrane.

3.4. Structural stability

X-ray diffractograms for neat PVA, PVA-3 wt.% NC, PVA-15 wt.% NC,
PVA-20 wt.% and PVA-30 wt.% NC are shown in Fig. 7(a). A decrease in
peak intensity of the peaks confirmed that the membrane became more
amorphous with increasing percentage of NC. Fig. 7(b) shows the
increase in the FWHM values with increasing concentration of NC.
This quantitatively indicates peak broadening and hence a decrease in
crystallinity. This is vital for the performance of the membrane because
the amorphous region lets the water penetrate through it while the
Fig. 9. TGA curves showing the degradation profile for PVA–NC membranes. crystalline region does not [5]. The increase in the amorphousness of
the PVA–NC membrane is expected to in turn increase the performance
for reverse osmosis application.
increasing percentage of NC because of a decrease in the water content Glass transition temperature (Tg) and melting temperature (Tm) for
and swelling (Fig. 5). Although dry PVA-3 wt.% NC membrane possesses neat PVA, PVA-3 wt.% NC and PVA-20 wt.% NC membranes are given
excellent mechanical properties, the wet membrane has a significantly in Table 3. Fig. 8 compares the corresponding DSC curves. The first
lower strength and modulus due to swelling and thus cannot be used endothermic peak corresponds to the glass transition temperature
for RO application. (Tg) of the membrane which increases when NC is added to PVA.
Variation of tensile modulus with increasing NC content shows that Melting was observed at 229 °C which is in close agreement with
the highest modulus is obtained for PVA-20 wt.% NC closely followed by cited values [21]. A slight shift in the melting temperature (from
PVA-3 wt.% NC. NC makes the material stiffer for the same reasons 229 °C to 225.5 °C) is noticed with an increase in NC percentage. This
discussed above. A consistent increase in Young's modulus with an change in the melting temperature is less than 5 °C. Therefore, thermal
increasing NC content was noticed for the wet membranes. stability of the membrane material is retained; addition of NC does not
Fig. 6 shows stress–strain curves for selected PVA–NC membranes in adversely affect the thermal properties of the membranes. The peaks
their wet state. The curves reflect the stretchy behavior of the material obtained just after the Tg may be due to moisture evaporation from
in its wet state. For neat PVA, a very low tensile strength and modulus the sample as PVA is hydrophilic and tends to absorb moisture from
was observed owing to its high water intake and swelling capacity. An the air. PVA degrades just before 350 °C. Degradation of NC occurs
increase of about 1520% in the tensile strength was observed from at around 225 °C [18] which overlaps with the melting of the PVA
neat PVA membrane to the PVA-15 wt.% NC membrane. As reflected polymer. This is why a distinct peak for NC decomposition is not
by the improvement of tensile modulus and tensile strength, the observed.
mechanical stability of PVA–NC membranes increases with increasing NC chains are spread out randomly and this network structure
NC content in both dry and wet states. entraps the PVA polymer inside it. These bundled chains are responsible
for rendering the material amorphous. Therefore, reduction in crystal-
linity increases with increasing NC bundles.
3.3. Water contact angle measurement TGA curves for PVA–NC membranes shown in Fig. 9 confirm the
onset degradation temperatures for NC (≈ 225 °C). Evaporation of
The contact angle data for selected PVA–NC membranes is tabulated residual water takes place from room temperature to about 200 °C
in Table 2. It was found that the contact angle of the membranes after which the PVA starts to rapidly decompose as the temperature

Fig. 10. Three dimensional surface images of (a) neat PVA and (b) PVA-20 wt.% NC air dried membranes.
144 S.F. Anis et al. / Desalination 336 (2014) 138–145

a) 25 Line A
Line B
Line C
b) Line A
Line B
Line C
70
20 60
15

Roughness (nm)
50

Roughness (nm)
10 40
5 30
0 20
-5 10
-10 0
-15 -10
-20 -20
-25 -30
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Distance (µm) Distance (µm)

Fig. 11. Surface roughness profile for (a) Neat PVA and (b) PVA-20 wt.% NC membranes.

increases above 230 °C. Carbonization of the degraded products takes of the feed solution (25,000 ppm) was calculated as 21.2 bars using
place just above 400 °C. Thermal analysis with DSC and TGA confirms the Van't Hoff equation. Therefore, a higher value of 24 bars was chosen
the high thermal stability of PVA–NC membranes which can be used for hydraulic pressure during RO testing. The membranes under test
up to 200 °C. were placed on a metal flange support as shown in Fig. 12(a and c).
All selected membranes showed similar performance with respect
3.5. Morphology to salt rejection while the flux increased slightly with increasing
percentage of NC. Membranes of PVA-20 wt.% NC and PVA-30 wt.% NC
Fig. 10 shows AFM micrographs of the active surfaces of neat PVA gave salt rejection rates of almost 99% and a flux of 1.435 l/h·m2. This
and PVA-20 wt.% NC air dried membranes. It can be seen that the may be attributed to reduction in crystallinity. As previously mentioned,
PVA–NC membranes possess a dense surface. Fig. 11 reveals their it is the amorphous regions in the polymer that allow water to penetrate
cross sectional roughness profiles over 3 chosen lines on the surface through while the crystalline portion does not. According to World
(not shown in the image). The highest roughness obtained for neat Health Organization, drinking water requires a TDS concentration of
PVA membrane was about 23 nm while that for PVA-20 wt.% NC was less than 500 ppm. RO testing with PVA–NC membranes produced
65 nm. These small values attest to the smoothness of the membrane permeates with less than 300 ppm, which falls well into the acceptable
surface in comparison to the total membrane thickness. range for drinkable water [22]. Fig. 12 shows PVA-15 wt.% NC, PVA-
20 wt.% NC and PVA-30wt.%NC membranes after their testing. The
3.6. Reverse osmosis test results membranes remained intact and retained their structural stability
even after being tested for RO. Membranes were soaked in tap water
Results obtained from RO testing including performance parameters for 30 min between runs. They were able to reproduce results for each
such as flux (J) and salt rejection rate (Rs) for PVA-15 wt.% NC, PVA- run tested, indicating their reusability.
20 wt.% NC and PVA-30 wt.% NC membranes have been tabulated in The flux produced was small in comparison to previously reported
Table 4. These membranes were selected according to their swelling values for PVA membranes [5]. This was due to the thicker membranes
behavior (Fig. 5) and mechanical properties (Table 1) because higher used in this study as flux decreases with increasing thickness. However,
percentages of NC in PVA tend to swell less and also possess superior the self-supporting PVA–NC membranes were able to sustain high
strength than ones with a lower wt% of NC in PVA. Osmotic pressure pressures without any additional layers.

Table 4
Reverse osmosis parameters and results.

Membrane Thickness (μm) Pressure (bars) Feed concentration (ppm TDS) Permeate concentration (ppm TDS) Rs (%) J (l/h·m2)

PVA-15 NC 272 24 25,000 295 98.82 1.208

PVA-20NC 185 24 25,000 261 98.96 1.435
PVA-30 NC 255 24 25,000 275 98.90 1.435

Fig. 12. (a) PVA-15 wt.% NC, (b) PVA-20 wt.% NC membrane after the test shows that the membranes retain their shape and stability even after usage (c) PVA-30wt.% NC membrane being
taken off from its reverse osmosis top flange after the test.
 S.F. Anis et al. / Desalination 336 (2014) 138–145 145

4. Conclusion [5] B. Bolto, T. Tran, M. Hoang, Z. Xie, Crosslinked poly (vinyl alcohol) membranes, Prog.
Polym. Sci. 34 (2009) 969–981.
[6] J.M. Yang, W.Y. Su, T.L. Leu, M.C. Yang, Evaluation of chitosan/PVA blended hydrogel
RO membranes were fabricated via slip casting of aqueous blends of membranes, J. Membr. Sci. 236 (2004) 39–51.
PVA–NC. NC when mixed with PVA in controlled amounts was seen to [7] X. Ma, Y. Su, Q. Sun, Y. Wang, Z. Jiang, Enhancing the antifouling property of polye-
thersulfone ultrafiltration membranes through surface adsorption-crosslinking of
impart physical and mechanical integrity to the membrane. Networked poly(vinyl alcohol), J. Membr. Sci. 300 (2007) 71–78.
cellulose successfully controlled the swelling behavior of PVA which [8] G.H. Koops, J.A.M. Nolten, M.H.V. Mulder, C.A. Smolders, Selectivity as a function of
otherwise would require heat treatment or crosslinking. This modified membrane thickness: gas separation and pervaporation, J. Appl. Polym. Sci. 53
(1994) 1639–1651.
form of cellulose entraps the PVA polymer inside its network which
[9] A.S. Hickey, N.A. Peppas, Mesh size and diffusive characteristics of semicrystalline
helps in restricting the swelling. Only a small percentage of NC was poly(vinyl alcohol) membranes prepared by freezing/thawing techniques, J. Membr.
sufficient to significantly increase the mechanical integrity of PVA while Sci. 107 (1995) 229–237.
[10] M.G. Katz, T. Wydeven, Selective permeability of PVA membranes. II. Heat-treated
higher percentages increased its structural and dimensional stability.
membranes, J. Appl. Polym. Sci. 27 (1982) 79–87.
The membranes were tested for RO desalination with aqueous solution [11] A. Hasimi, A. Stavropoulou, K.G. Papadokostaki, M. Sanopoulou, Transport of water
of sodium chloride and showed promising results with high salt rejection in polyvinyl alcohol films: effect of thermal treatment and chemical crosslinking,
rates of 98.8%. The fabrication technique allowed developing self- Eur. Polym. J. 44 (2008) 4098–4107.
[12] V.S. Praptowidodo, Influence of swelling on water transport through PVA-based
supported, but thick membranes. The high membrane thickness is re- membrane, J. Mol. Struct. 739 (2005) 207–212.
sponsible for the low flux data of 1.2–1.4 l/h·m2 reported. Future works [13] C.E. Reid, E.J. Breton, Water and ion flow across cellulosic membranes, J. Appl. Polym.
will focus on fabricating membranes with the same material, but with Sci. 1 (1959) 133–143.
[14] Y. Xianda, W. Anlai, C. Suqin, Water-vapor permeability of polyvinyl alcohol films,
lower thicknesses. Based on the proportionality between flux and mem- Desalination 62 (1987) 293–297.
brane thickness, success in making thin membranes promises to produce [15] A. Asghar, Y. Abdul Samad, B. Singh Lalia, R. Hashaikeh, PEG based quasi-solid
high flux data. The possible challenges associated with these thin mem- polymer electrolyte: mechanically supported by networked cellulose, J. Membr.
Sci. 421–422 (2012) 85–90.
branes will be their mechanical stability and difficulty to handle them. [16] A. Jalal Uddin, J. Araki, Y. Gotoh, Toward “strong” green nanocomposites: polyvinyl
alcohol reinforced with extremely oriented cellulose whiskers, Biomacromolecules
12 (2011) 617–624.
References [17] X. Xu, Y.-Q. Yang, Y.-Y. Xing, J.-F. Yang, S.-F. Wang, Properties of novel polyvinyl
alcohol/cellulose nanocrystals/silver nanoparticles blend membranes, Carbohydr.
[1] M. Amera, Polyvinyl alcohol–cellulose acetate composite reverses osmosis Polym. 98 (2013) 1573–1577.
membranes: I. synthesis and characterization, Hydrol. Curr. Res. 3 (2) (2012) [18] R. Hashaikeh, H. Abushammala, Acid mediated networked cellulose: preparation
1000131, http://dx.doi.org/10.4172/2157-7587.1000131. and characterization, Carbohydr. Polym. 83 (2011) 1088–1094.
[2] D. Bezuidenhout, M.J. Hurndall, R.D. Sanderson, A.J. van Reenen, Reverse osmosis [19] D. Kondepud, Introduction to Modern Thermodynamics, 2008.
membranes prepared from potassium peroxydisulphate-modified poly(vinyl [20] A. Sagle, B. Freeman, Fundamentals of membranes for water treatment, the future of
alcohol), Desalination 116 (1998) 35–43. desalination in Texas, 22004. 137–154.
[3] K. Lang, S. Sourirajan, T. Matsuura, G. Chowdhury, A study on the preparation of [21] Y. Bin, M. Mine, A. Koganemaru, X. Jiang, M. Matsuo, Morphology and mechanical
polyvinyl alcohol thin-film composite membranes and reverse osmosis testing, and electrical properties of oriented PVA–VGCF and PVA–MWNT composites,
Desalination 104 (1996) 185–196. Polymer 47 (2006) 1308–1317.
[4] J. Jegal, K.-H. Lee, Nanofiltration membranes based on poly(vinyl alcohol) and ionic [22] W.H.O., Guidelines for drinking-water quality, Health criteria and other supporting
polymers, J. Appl. Polym. Sci. 72 (1999) 1755–1762. information, 21996.