Journal of Membrane Science 428 (2013) 104–115

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science

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

Fabrication and characterization of

polyvinylidenefluoride-co-hexafluoropropylene (PVDF-HFP)
electrospun membranes for direct contact membrane distillation
Boor Singh Lalia a, Elena Guillen-Burrieza b, Hassan A. Arafat b, Raed Hashaikeh a,n
a
Materials Science and Engineering Program, Masdar Institute of Science and Technology, Masdar City, PO Box 54224, Abu Dhabi, United Arab Emirates
b
Water and Environmental Engineering Program, Masdar Institute of Science and Technology, Masdar City, PO Box 54224, Abu Dhabi, United Arab Emirates

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

Article history: Polyvinylidene fluoride-co-hexafluoropropylene (PVDF-HFP) membranes were prepared via the elec-
Received 3 July 2012 trospinning method. Three sets of mats, with different fiber diameters and physical properties, were
Received in revised form obtained by varying polymer concentration from 10 to 15 wt%. Membranes were formed by hot
30 October 2012
pressing the prepared mats to fuse the fibers. As prepared mats were hot pressed in order to control the
Accepted 31 October 2012
membranes dimensional stability and mechanical strength and to alter the pore size and pore size
Available online 16 November 2012
distribution. The effects of hot pressing and variation of fiber diameter on pore size distribution,
Keywords: porosity, contact angle, gas permeation, and liquid water entry pressure (LEP) of the fabricated
PVDF-HFP membranes were all studied. Optimized membranes were obtained with mean pore size, porosity,
Electrospun membrane
contact angle and LEP of 0.26 mm, 587 5%, 12517 2.41 and 19.1 psi, respectively. Fabricated mem-
Desalination
branes were tested for desalination by direct contact membrane distillation and a water flux value of
Membrane distillation
20–22 L h  1 m  2 was obtained with a salt rejection (SR) ratio of 98%.
& 2012 Elsevier B.V. All rights reserved.

1. Introduction angle of 101.21 for the microporous PVDF-HFP hollow fiber

membranes [14]. PVDF-HFP flat sheet and hollow fiber mem-
Membrane distillation (MD) is an emerging technique for branes have been prepared and characterized for its use in
desalination of water due to its high salt rejection, no extensive membrane distillation [12,15].
pretreatment and low energy consumption using waste heat Recently, electrospun membranes have been gaining consider-
source [1,2]. The major drawback of this technology, compared able attention in various applications such as fuel cells, filtration,
to commercial available reverse osmosis, is the low flux rate and catalysis, tissue engineering, solar cells and separators for bat-
wetting of pores during membrane distillation [3,4]. The pre- teries [16]. The unique features of these non-woven membranes
requirements for polymers used in fabrication of MD membranes are controllable aspect ratios (aspect ratio¼L/d; L-length of the
are the high hydrophobicity, chemical inertness and low thermal fiber and d-diameter of the fiber) and morphology of the nano/
conductivity. Different polymers such as polytetrafluoroethylene micro fibers, which is achieved by varying the solution viscosity,
(PTFE), polypropylene (PP), polysulfone and poly(vinylidene fluor- environmental conditions, applied electric potential and the flow
ide) (PVDF) have been used to make MD membranes [5–8]. rate of the solution [17]. Due to these unique properties, non-
The MD membranes are prepared by using the common methods woven electrospun membranes have been demonstrated for use
viz. stretching, phase inversion, and dry jet-wet spinning [9–11]. in water filtration and chemical separation in liquids [18].
PVDF membranes are given particular consideration due to their Recently, electrospun PVDF membranes were also proposed for
asymmetric morphology which assists in improving the water MD [19].
permeation flux [7]. Also, copolymers of PVDF, such as hexafluor- Electrospun fibers are randomly oriented and not intercon-
opropylene (HFP) and tetrafluorethylene (TFE) possess superior nected at nodes. Due to this, electrospun membranes have poor
hydrophobicity and high free volume compared to PVDF [12,13]. mechanical strength and structural integrity [20]. Heat treatment
The contact angle increased from 80 70.781 to 86.670.511 with of most polymeric electrospun membranes for several hours
copolymerization of PVDF with TFE. Shi et al. reported a contact above the glass transition temperature and below melting point
helps to fuse the fibers together. This enhances the structural
integrity and mechanical properties of the membranes [18,21].
n
Corresponding author. Tel.: þ971 28109152. Kaur et al. studied the influence of hot pressing on the bubble
E-mail address: [email protected] (R. Hashaikeh). point, thickness, mechanical properties and pure water flux

0376-7388/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.memsci.2012.10.061
 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115 105

of polyacrylonitrile membranes used in nanofiltration applica- refers to the upper surface of the membrane and bottom surface
tions [20]. refers to the membrane bottom face the aluminum foil.
To the best of our knowledge, there is no report in the
literature concerning the fabrication of electrospun membranes
2.4. Membrane characterization
composed of PVDF-HFP for MD. Electrospun membranes with
different fiber diameters were prepared by controlling the poly-
2.4.1. Pore size distribution, porosity, mean pore size and bubble
mer concentration in the solution. The electrospun membranes
point
were hot pressed using a modified hot pressing procedure in
A Capillary Flow Porometer (PMI, Ithaca, USA) was used to
order to reduce the heat treatment time. The electrospun mem-
characterize the mean pore size, bubble point or maximum pore
branes with and without hot pressing were characterized using
size and pore size distribution of the membranes. This equipment
scanning electron microscopy (SEM), capillary flow porometry
can also perform the Mullen burst, LEP and gas permeability tests,
(CFP), porosity measurements, contact angle measurements,
as will be described later. The wet/dry flow method was used
Mullen burst test and finally tested in direct contact MD (DCMD).
where the sample was initially wetted with a low surface tension
wetting liquid (GalwickTM) and subsequently placed in a sealed
chamber through which gas (dry air) flows. The pressure of the
2. Experimental
gas was increased stepwise and the through flow was recorded.
The flow rate is zero in the beginning because all the pores are
2.1. Materials
blocked by the wetting liquid. When the pressure is sufficient to
remove the liquid from the largest pores, the flow is no longer
PVDF-HFP (Kynar Powerflexs LBG, Mw ¼450,000 g mol  1), acet-
zero and the bubble point is then recorded. As pressure is
one (Aldrich, CHROMASOLVs, for HPLC,Z99.8%), and dimethyla-
increased, smaller pores become unblocked and the gas flow rate
cetamide (Aldrich, CHROMASOLV s plus, for HPLC, Z99.9%) were
increases until the whole sample becomes dry. The cumulative
used as received. PVDF Transfer Membranes (0.2 mm marketed by
pressure needed for gas flow through the wet sample is used to
Thermo Scientific) were used for comparison.
calculate pore diameters. The relation between the pressure
needed to displace liquids of very low surface tension from pores
2.2. Fabrication of PVdF-HFP membranes and its radius is given by the following relation.
2s
Clear and uniform polymer solutions of 10, 12 and 15 wt% DP ¼
r
PVDF-HFP were prepared in a binary mixture of acetone and
dimethylacetamide with a weight ratio of 7:3 using a magnetic where, s is the surface tension of the wetting liquid. After the wet
stirrer at room temperature. PVDF-HFP is referred to as PH in the curve is recorded, the process is repeated for the dry sample. The
following text for simplicity. The PH solution was loaded in a gas flow rates through the wet and dry samples are used to
10 mL glass syringe with a luer lock steel tip connected to a compute pore distribution. The mean flow pore diameter corre-
stainless steel needle ((Suzhou Lanbo Needle Co., Ltd, China) sponds to 50% of the total flow. Further details about this method
Gauge 18, inner diameter¼0.838 mm) by a PTFE pipe. A Nanon- can be found in literature [23,24]. The total porosity of the
01A electrospining setup (MECC, Japan) was used for electrospin- membrane samples was also measured. This was achieved by
ning the polymer solution. The relative humidity of electrospining completely filling the pores of a pre-weighed membrane sample
chamber was 60 72% and the temperature was 2571 1C. using a highly-wetting liquid with density 0.93 kg L  1 (Silwicks,
A potential difference of 20 kV was generated between the needle provided by PMI, Ithaca, NY-USA). The sample with liquid was
and a rotating aluminum drum grounded target placed 15 cm then weighed and the volume of pore-filling liquid was obtained
from the tip of the needle. A solution feed rate of 1 ml h  1 was by means of the following equation.
used. The electrospun membranes of PH were collected on the mwet_s mdry_s
aluminum drum at 25 1C and dried at 50 1C for 24 h in a e¼
rsilwick  V s
conventional oven. The thickness of electrospun membrane was
measured by using a micrometer (Mitutoyo, Japan; detection where Vs is the volume of the sample, rsilwick is the wetting liquid
limit¼ 0.001 mm) and was in the range of 110–150 mm. density and mdry_s, mwet_s are the weights of the dry and wetted
samples, respectively. The volume of the samples was calculated
using sample thickness, which was determined with a precision
2.3. Hot pressing of PVdF-HFP membranes micrometer and was later confirmed by SEM images. Three
samples of each membrane were used to measure the porosity.
The membrane were hot pressed using a household iron.
The surface temperature was measured by placing a thermocou-
ple between the two aluminum foils with hot iron kept on top for 2.4.2. Liquid entry pressure (LEP) measurement
one minute. The surface temperature of the iron was 20075 1C. LEP measurements using DI water were done using the same
Before pressing, the electrospun membranes were sandwiched capillary flow porometer. In this experiment, the membrane
between two A4-size papers to avoid the direct contact of hot sample was placed in a sealed chamber and 2–3 mL of DI water
surface with the electrospun membrane. The iron was moved was poured onto the active surface of the membrane to comple-
from one end to another for 1–2 s. The surface temperature of the tely cover it. The air pressure in the chamber (above the sample)
iron was above the melting point of the polymer ( 150 1C [22]); was then increased gradually by means of a compressed air until
hence the iron was not placed directly on the electrospun fiber to the pressure was high enough to enable the air to start passing
avoid the damage of the membrane. The electrospun membrane through the membrane. At this initial passage, the gas flow was
peeled off from the aluminum foil and folded in a way that the detected using sensors on the backside of the membrane, and the
bottom of one membrane comes in contact with bottom of the corresponding pressure was recorded as LEP. Measurements
other membrane. After hot pressing two layers of the membrane, were taken at room temperature. To reduce the error, multiple
the top and bottom surface remained the same. The electrospun measurements were taken and averaged. Experimental error was
membranes were collected on the aluminum foil, top surface found to be around 71 psi.
106 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115

2.4.3. Contact Angle measurements

To measure the contact angle (y), an EASY DROP Contact Angle
measurement instrument from Kruss (Hamburg, Germany),
equipped with image-processing software was used. A piece of
membrane was placed on a platform and deionized (DI) water
was used as liquid at room temperature. A micro-syringe was
then used to generate the droplets (1–2 mL) on the membrane
surface. The instrument captures a digital image of the
drop on the membrane and estimates the y by geometrical
methods (sessile drop, circle-fitting, etc.). A minimum of three
measurements for each membrane were taken and averaged to
minimize the uncertainties of the measurements (partial evapora-
tion, chemical interactions between the solid and liquid, etc).
The standard deviation for all measured contact angles was lower
than 31.

2.4.4. Scanning electron microscopy (SEM)

The electrospun membranes were examined by scanning
electron microscopy. A Quanta FEG 250 SEM by FEI (Hillsboro,
OR, USA) operating at 10–20 keV was employed in this study.
Samples taken from each membrane were gold and palladium
coated using a GATAN, Model 682, precision etching coating
system (PECS).

2.4.5. Atomic force microscopy (AFM)

The surface roughness of the electrospun membranes (with
and without hot pressing) was characterized using AFM (Asylum
Research, MFP-3DTMStand Alone). The samples were cut into
small square pieces (1  1 cm2) and pasted on the glass substrate.
AFM was performed using a Silicon tip of radius  10 nm, with a
resonance frequency of  70 kHz and a spring constant of
2 N m  1 in the non-contact tapping mode.

2.4.6. Mullen burst tests

A Mullen-burst test was performed in order to simulate the
actual forces a membrane will be subjected to during MD
operation. Mullen burst is a type of test to determine how much
force is required to rupture the membrane as it is distended
[25,26]. A membrane sample (50  50 mm) was securely clamped
over a rubber diaphragm. The rubber diaphragm was used to keep
the water from escaping the system or touching the sample (as
depicted in Schematic 1). During the test, the diaphragm was
steadily pressurized with water controlled by a pressure regula- Schematic 2. (a) DCMD setup used for membrane distillation and (b) schematic of
tor. Once the membrane ruptured, the rubber diaphragm would the DCMD setup .

suddenly expand to the point where it broke through the

membrane. At the same time, the test would automatically be stopped by the controlling software, and the corresponding water
pressure within the diaphragm at the point when the media
ruptured was recorded by the software as the bursting strength.

2.4.7. Gas permeability

The sample is sealed in the sample chamber of the capillary
flow porometer (described earlier), such that gas flows through
the sample only in the direction in which permeability is to be
determined. The pressure of the test gas on one side of the sample
is increased in small steps. The differential pressure across the
sample and the corresponding gas flow rate are accurately
measured as a function of increasing differential pressure by flow
and pressure transducers. Gas permeability is defined by Darcy’s
law. According to this law, flow of fluids through porous media is
proportional to the pressure gradient causing flow.

2.5. MD experiments

Direct contact MD (DCMD) experiments were carried out

Schematic 1. Illustration of Mullen burst measurement setup . under controlled conditions using a custom built MD system.
 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115 107

Table 1
Properties of the fabricated PVDF-HFP electrospun membranes.

Membrane Thickness Fiber size LEP Gas flow rate Pore size range Mean pore size Largest pore size
(lm) (lm) (PSI) (at 40 psi) L min  1 (lm) (lm) (bubble point) (lm)

15 PH 150 0.2–1.3 5.77 – 0.9–2.25 1.49 2.25

15 PH-hot press 85 0.2–1.6 10.57 – 0.38–1.07 0.63 1.07
15 PH-hot press-2layers 170 0.2–1.6 14.50 555 0.3–1.04 0.51 1.04
12 PH 150 0.1–0.8 7.2 – 0.9–1.94 1.16 1.94
12 PH-hot press 112 0.1–0.9 13.37 – 0.48–1.01 0.61 1.02
12 PH-hot press-2layers 1707 5 0.1–0.9 10.7 – 0.25–0.89 0.50 0.89
10 PH 135 0.1–0.5 11.9 – 0.45–1.39 0.73 1.39
10 PH-hot press 65 0.1–0.65 15.1 – 0.35–0.96 0.55 0.96
10 PH-hot press-2layers 110 0.1–0.65 19.1 190 0.21–0.77 0.26 0.77
0.2 PVDF (commercial membrane) 197 N/A 42.27 20 0.15–0.2 0.2 0.48

The system is fully instrumented and computer-controlled (see 3. Results and discussion
Schematic 2). Throughout the experiments, all the operational
parameters such as pressure, temperature and flow rates were 3.1. Morphology of electrospun membranes
monitored using sensors and recorded using software developed
for this purpose. The membrane sample used was a flat sheet The pore size, porosity and mechanical strength of the electro-
piece of 5  11.5 cm. Metallic mesh-type spacers (purchased from spun membranes can be altered by varying several factors, includ-
www.smallparts.com, USA) were used on both sides to promote ing the viscosity of the solution, applied potential, the internal
turbulence and to support the membrane. The feed solution was diameter of the needle and flow rate of the solution [17]. The
prepared with DI water and table salt with a final concentration literature suggests that reduction in fiber diameter reduces the pore
of 10 g L  1 and a conductivity of around 18 mS cm  1. The size of the membrane [27]. Hence, the effect of fiber diameter of the
permeate solution was tap water with a conductivity of around PVDF-HFP membranes on the pore size distribution was investi-
110 mS cm  1. All the MD experiments were carried out using a gated in this work. In this study, the viscosity of the solution was
flow rate of 0.25 L s  1 on both sides, which given the dimensions changed by varying the concentration of PH in the solution to study
of the chamber, corresponds to a linear flow velocity of the effect of polymer concentration on the pore size distribution,
0.32 m s  1. The selected flow rate (0.25 L s  1) was chosen based porosity and its mechanical properties, while keeping all other the
on the LEP of the prepared membranes (discussed later) and the parameters constant. The measured key properties of the mem-
corresponding pressure reached in the chamber (around 18 psi), branes produced in this study are tabulated in Table 1, along with
which is a function of the feed flow rate. Only the membranes that of commercial membrane for comparison. These properties
that showed the best features for the MD process (high LEP, will be referred to during the following discussion.
narrow pore size distribution and small mean pore size) and most Fig. 1(a) shows the SEM image of electrospun membranes
similar characteristics when compared to the PVDF commercial prepared from 15 wt% PH solution. The fibers obtained have
membrane (see Table 1) were chosen for the MD experiments, diameter in the range of 200 nm to 1.3 mm, as measured by the
and these were 10 PH-2 layers and 15 PH-2 layers. During the SEM software. The viscosity of the solution was further diluted by
experiments, the feed inlet temperature was increased gradually using 10 wt% PH and the SEM images are shown in Fig. 1(c). SEM
from 30 to 50 1C while the permeate temperature was monitored images of this latter membrane show that a decrease in viscosity
and recorded but not controlled. However, for all the experi- of the polymer solution reduces the fiber diameter down to a
ments, the initial permeate-side temperature was the same range of 100 to 500 nm. The decrease in polymer concentration
(  24 1C). Total permeate and feed volume variations were con- from 15 to 10 wt% in the solution leads to the formation of beads
tinuously measured from the start of the test by means of two as seen in Fig. 1(c). Doshi and Reneker investigated the relation-
penetrometers installed on the feed and permeate tanks, respec- ship between bead formation and polymer concentration of the
tively. These sensitive penetrometers provided a water-level solution [28]. Deitzel et al. found that fiber thickness increases
reading every 2 s, from which the variation in total water volume with an increase in solution viscosity and relates the solution
was calculated. After correcting for volume expansion due to viscosity and fiber diameter by the power law relationship [29].
temperature change, the water flux per square meter of mem- Hence, in this study, the PH electrospun membranes with differ-
brane was calculated using the penetrometer reading. Thus, two ent fiber sizes were obtained using different concentrations of
flux readings were obtained every 10–15 s, one using the feed polymer solution and the effect of fiber size on the pore size
(hot) side penetrometer and one using the cold (permeate) side distribution, porosity, and mechanical properties of the mem-
penetrometer. The two flux readings were found to match very branes was investigated. The electrospun fibers are randomly
well and their average was reported. The results of the tested distributed and not interconnected. Heat treatment of the elec-
membranes were compared to those of commercial PVDF mem- trospun membranes for several hours has been reported to
branes with nominal pore size of 0.2 mm, also tested using our enhance the dimensional integrity and mechanical properties of
DCMD system. The flux results were plotted against the difference the membranes [18]. In the proposed method, a rapid hot
of temperature between the cold and hot stream as the permeate pressing of the membrane above the melting point of the polymer
temperature was not kept constant during the experiments. As is performed for 1–2 s. It helped to fuse the fibers together and
explained in later sections, the mass transfer in MD is governed reduce the thickness of the membrane. This formed a dimension-
by the difference in vapor pressure across the membrane which is ally stable network of electrospun fibers and reduced the thick-
in turn created by a difference in temperature between the hot ness of the membrane which helped to improve the mechanical
feed stream and the cold permeate, gas or vacuum on the other properties discussed later using Mullen burst test. Table 1 showed
side of the membrane, rather than the absolute feed temperature. that the electrospun membrane prepared from 15 wt% solution
108 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115

5μm 3μm

5μm
2μm

Fig. 1. SEM images of ((a) and (c)) electrospun fibers and ((b) and (d)) electrospun hot pressed fibers prepared from 15 and 10 wt% polymer solution, respectively.

Fig. 2. Cross-sectional SEM images of hot pressed 10 PH 2-layers electrospun membranes.

has thickness of 150 mm, which was reduced to 85 mm after hot improved liquid entry pressure (Table 1) and Mullen burst
pressing. The two hot pressed layers of 15 PH membrane was pressure (Fig. 7).
170 mm thick. The SEM images of the hot pressed membranes The mean pore size and pore size distribution are crucial
obtained from 15 to 10 wt% PH solution are illustrated in Fig. 1(b) parameters to the performance of MD [30]. As stated by Schofield
and (d), respectively. Both images show that hot pressing of the et al. [31] membranes for MD should have a reasonably small
membranes fuses the fibers at the interlay points. Cross-sectional pore size (mm range) and be preferably smaller than 0.5 mm,
SEM images of membrane were used to study the adhesion to prevent wetting. Also for Knudsen and viscous diffusion, which
between the two layers of hot pressed membranes. Fig. 2 presents is preferred in MD processes, larger pores are needed and there is
the cross-section images of the hot pressed 10 PH-2 layers an agreed range of; 0.2 to 0.5 mm [31]. For commercial mem-
membrane. It was found that after hot pressing, the fibers branes currently used in MD processes, mean pore sizes range
of two layers diffused together and could not be distinguished from 1 to 0.2 mm with corresponding LEP values between 7 and
from one another. The two layers hot pressed membranes had 58 psi [32].
 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115 109

electrospun membranes prepared from 10, 12 and 15 wt% solu-

tions. The membrane obtained from 15 wt% polymer solution
(15 PH) shows (Fig. 3(a)) a broad pore size distribution in the 0.9–
2.2 mm range with a mean pore size of 1.49 mm. This is due to the
large fiber diameters of the electrospun fibers as discussed in SEM
results above. However, the thickness of the membranes (15 PH,
12 PH and 10 PH) was kept almost the same as shown in Table 1.
By keeping the thickness of the membranes constant, the effects
of fiber diameter on some of the membrane properties (viz. pore
size distribution, porosity, contact angle, Mullen test etc.) were
investigated. The pore size distribution curve of 15 PH-hot press
membranes shifts to 0.4 to 1.1 mm with mean pore size of
0.63 mm. Further, to study the effect of thickness of the membrane
on pore size distribution, two layers of 15 PH membranes
were hot pressed. The increase in thickness of the hot pressed
membrane further reduced the pore size distribution to 0.3 to
1.07 mm with mean pore size of 0.51 mm. Similar pore size
distribution behavior was observed for membranes obtained from
10 and 12 wt% polymer solutions as shown in Fig. 3(b) and (c).

14
10PH Hot press 2 Layers
12 0.2 PVdF

Distribution (%) 10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Pore size (μm)

Fig. 4. A comparison of pore size distribution of 10 PH hot pressed 2 layers

membrane with commercial 0.2 PVdF membrane (mean pore size 0.2 mm).

120
Electrospun
110 Hot press
Hot press 2 layers
100
Porosity (%)

90

80

70

60

50

Fig. 3. Pore size distribution of 15, 12 and 10 wt% solution in acetone/DMA 40

(7:3 w/w). 9 10 11 12 13 14 15 16
Concentration of PH in solution (%)
The effect of polymer concentration and hot pressing on the Fig. 5. Variation of porosity with concentration of PH in acetone/DMA solution
pore size distribution of the electrospun PVDF-HFP membranes (7:3; w/w) of electrospun membrane, electrospun-hot pressed and electrospun-
was studied. Fig. 3 shows the pore size distribution curves for PH hot pressed-2 layers.
110 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115

However a very narrow pore size distribution with a mean pore on the 10 PH, 12 PH and 15 PH hot pressed 2 layers membranes
size of 0.26 mm was obtained for 10 PH hot pressed 2 layers and their results are shown in Fig. 6. For all membranes, the flow
membrane. These findings are well in agreement with the rate varies linearly with the air pressure, as expected. The 15 PH
reported results [20] that hot pressing is a facile method for hot pressed 2 layers membrane has highest gas permeability. For
improvement of porous properties of electrospun membranes. example, the gas flow rate was found to be 555 cm3 s  1 at 40 psi.
First, it effectively reduces the mean pore size into the preferred This is due to the large pore size distribution of the 15 PH
range for MD and second, it narrows the pore size distribution, membrane as shown in Fig. 3. The 10 PH membrane has a
contributing to reduce the presence of big pores and therefore narrower pore size distribution as compared to the 15 PH mem-
reducing the risk of wetting. branes, and, therefore, has lower gas permeability ( 165 cm3 s  1
The pore size distribution of 10 PH hot press 2 layers mem- at 40 psi). The permeability for commercial PVDF membrane was
brane was compared with the commercial PVdF membrane with found to be 88.6 cm3 s  1. Pore size and flow rate are inversely
mean pore size of 0.2 mm. The CFP plots are illustrated in Fig. 4. related and a compromise has to be made between the pore size
The results revealed that the pore size distribution of the hot and gas permeability for optimizing the MD process. Thickness of
pressed electrospun membrane is very close to the commercial the membrane is another factor that influences the gas flow rate.
membrane. A thin membrane with small pore size is preferred for the MD
Fig. 5 illustrates the variation of porosity with polymer con- process. A thin membrane reduces the transportation path of
centration for the electrospun membrane for both hot pressed water vapor through the membrane and assists in increasing the
electrospun and hot pressed electrospun 2 layers. The calculated water flux. In this study, the hot pressing of two layers of
porosity of electrospun membranes lies in the 90 75% range. Rao membrane contribute to the narrower pore size distribution with
et al. reported a porosity of 86% for poly(acrylonitrile)/poly(- mean pore size 0.26 mm (10 PH-hot press 2 layers) and a reduc-
methyl methacrylate) (PAN/PMMA) electrospun membranes used tion of thickness from 270 mm (135 mm thickness of one layer) to
as separator for lithium ion batteries [33]. Qi et al. obtained 92% 110 mm. The modification achieved by hot pressing improves the
porosity for the electrospun poly(phthalazinone ether sulfone prerequisite parameters for membrane distillation.
ketone) membrane [34]. A variation of 10% was observed in the
porosity of electrospun membrane. This is due to the sponge like 3.2. Mechanical properties of electrospun membranes
behavior of the membrane which leads to inaccuracies in thick-
ness measurements. The hot pressing of electrospun membranes The mechanical properties of the electrospun membranes
reduces the porosity to 60–70%. This is due to thickness reduction, were studied by Mullen burst test. The burst pressure for PH,
which ultimately decreases the pore volume after hot pressing. PH hot pressed and PH hot pressed 2 layers membranes prepared
The porosity of the membrane further goes down to 55–60%, from different concentrations of polymer solution is given in
when 2 layers of membranes are hot-pressed together. However, Fig. 7. The results reveal that the burst pressure increases with an
after hot pressing, the membranes have reasonably high porosity increase in polymer concentration. This is due to the fact that an
required for MD application. In general, the porosity of the increase in polymer concentration results in an increase in fiber
membranes used for MD ranges between 30% and 85% [4]; diameter as discussed earlier. The increase in fiber thickness
however, the higher the porosity the larger the area for evaporation ultimately increases the mechanical strength of the spun mats
and the lower the thermal conductive losses of the membrane. and hence of the membranes. The hot pressing of electrospun
In the DCMD processes, the water vapors permeate through membranes added more strength to the membranes. The hot
the membranes and condenses into the cold liquid on the liquid- pressed electrospun membrane with 2 layers exhibited the high-
membrane interface on the permeate side. Thus, the gas perme- est burst pressure 28.2 psi. This is due to interlocking of the fibers
ability of the membrane is an important parameter for estimating after hot pressing and the effect of thickness of the membrane,
the vapor flux through the membrane. It depends on various two layers of membrane increases the thickness of the membrane
parameters of the membrane like porosity, pore size, thickness hence the Mullen burst pressure is higher than a single layer.
etc. Gas permeability tests using compressed air were performed
35
3000 Electrospun
10PH Electrospun & hot pressed
30 Electrospun & hot pressed 2 Layers
Mullen burst pressure (psi)

2500 15PH
0.2PVdF
Flow rate (cc sec-1)

25
2000

1500 20

1000 15

500 10

0 5
0 20 40 60 80 100 120 140 160 9 10 11 12 13 14 15 16
Air pressure (psi) Conc. of PH in Acetone/DMA solution (%)

Fig. 6. Gas permeability tests of electrospun hot pressed-2 Layer membranes Fig. 7. Variation of burst pressure with concentration of PH in acetone/DMA
prepared from 10, 12 and 15 wt% solution in acetone/DMA (7:3 w/w), and solution (7:3; w/w) of electrospun membrane, electrospun-hot pressed and
0.2 PVdF commercial membrane for comparison. electrospun-hot pressed-2 layers.
 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115 111

160
E lectrospun
E lectrospun hot press
E lectrospun hot press 2 Layers
150

Contact angle (θ)

140

130

120

110
9 10 11 12 13 14 15 16
Conc. of PH in solution (%)
Fig. 8. (a) Variation of contact angle with concentration of PH in acetone/DMA solution (7:3; w/w) of electrospun membrane, electrospun-hot pressed and electrospun-hot
pressed-2 layers, (b) a drop of water drop on the surface of electrospun membrane and (c) electrospun and hot pressed membrane.

3.3. Hydrophobicity of electrospun membranes PVDF (mean pore size 0.2 mm) membrane. The flux results have
been plotted as a function of the difference in temperature
The contact angle (CA) results for the prepared membranes are between the hot and cold streams. Maximum flow rates obtained
shown in Fig. 8. In general, all the prepared membranes showed ranged between 20–21 L h  1 m  2 for the PVDF commercial
high CA values (CAZ1201), well within the range reported for membrane (at a maximum Tfeed of 68 1C); 9–8 L h  1 m  2 for
PVDF and PTFE membranes [35]. However, there is a clear 15 PH (at a maximum Tfeed of 50 1C) and between 20 and
difference between the electrospun membranes with and without 22 L h  1 m  2 for 10 PH (at a maximum Tfeed of 65 1C). In general,
hot pressing. The hot pressed membranes have considerably the flux values were better for the electrospun PVDF-HFP mem-
lower CA (a 10% decrease was observed with hot pressing of the branes, especially at higher delta T ( 415 1C) where 10 PH showed
electrospun membranes). This can be explained by the effect of around a 60% higher permeate flow rate when compared to the
hot pressing process on the roughness of the membrane surface. commercial PVDF. At lower DT (o5 1C), 10 PH and 15 PH showed
The electrospun membranes have a sponge-like structure and an 62% and a 37% higher flow rate, respectively, when compared to
irregular surface. On the other hand, hot pressed membranes the PVDF commercial membrane. The operational parameters
show a smoother surface as the process helps to bring the fibers chosen for the MD experiments, feed flow of 4 GPM and tem-
closer together, resulting in a much regular surface [36]. peratures ranging between 30 and 68 1C) correspond to an
The effect of hot pressing on the surface roughness was char- absolute pressure in the chamber of around 18 psi. The salt
acterized by AFM topography. Fig. 9(a) and (d) illustrate the three rejection for a MD membrane depends mainly on the LEP pressure
dimensional height retrace AFM images of 10 PH electrospun which in turn depends on the hydrophobicity and the pore size of
membrane before and after hot pressing, respectively. The surface the membrane (as the Cantor Laplace equation predicts). The LEP
roughness is directly correlated to the Z-value (i.e. the piezo- value of the PVDF commercial membrane (42 psi) is much higher
electric sensor moves vertically across the surface). The cross- compared to the absolute pressure reached during the experi-
section analysis of the AFM images was performed at three ments (18 psi). The salt rejection factor for this membrane, under
different regions (line A, B and C in Fig. 9(b) and (c)) to study given conditions, is expected to be 100%. This was proven
the variation of height across the cross-sectional profile of the experimentally, as the conductivity of the permeate tank was
membrane. The change in height was  2.5 mm for 10 PH electro- decreasing throughout the experiment. It can be assumed that
spun as shown in Fig. 9(c). The cross-sectional profile of hot almost pure vapor was being collected. For the same reason, the
pressed membrane is given in Fig. 9(f). The results showed that single layer membranes have comparatively lower LEP values (see
the variation of height across the cross-section decreased from Table 1) and have a broader pore size distribution which can lead
2.5 to  1 mm after hot pressing. The variation of height across to a higher variability over the membrane sheet, were not
the cross-section is directly related to the surface roughness. The selected for the MD experiments. In the case of the experiments
AFM results revealed that the hot pressing of the electrospun conducted with the electrospun membranes (15 PH-2 layers and
membrane leads to a smoother surface. In addition, the effect of 10 PH-2 layers), the conductivity of the permeate stream was
polymer concentration on the CA is much less significant, as constant and around Cpermeate ¼524 mS cm  1. The percentage of
revealed by the data in Fig. 8. High hydrophobicity is desirable in salt rejection (%SR) can be calculated using the following formula:
MD since it is one of the factors affecting LEP and therefore pore

C feed  C permeate C tap_water
wetting. In addition, water repellency diminishes the risk of %SR ¼  100 ð1Þ
C feed
fouling [37].
where the conductivity of the tap water used as cold stream
3.4. Performance in MD process (Ctap_water) was 110 mS cm  1 and the conductivity of the feed
water (Cfeed) was 18 mS cm  1. The result is a SR of 98%. The flux
Fig. 10 shows the distillate fluxes of two electrospun mem- results presented here were then compared to the reported
branes (10 PH and 15 PH) in comparison with the commercial ones that could be found in literature for MD using PVDF-HFP
112 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115

2.0
LineA
Line C 1.5 LineB
LineC
1.0

Height / μm
0.5
Line B
0.0

-0.5
Line A
-1.0

-1.5
0 1 2 3 4 5 6 7 8 9 10
Distance / μm

1.0
LineA
0.8 LineB
Line C LineC
0.6
0.4
Height / μm

0.2
Line B 0.0
-0.2
-0.4

Line A -0.6
-0.8
0 1 2 3 4 5 6 7 8 9 10
Distance / μm

Fig. 9. Three dimensional surface images of (a) 10 PH membrane and (d) hot pressed 10 PH membrane. Cross-section profile (e) indicated by line A, B and C in (b) before
hot pressing and (f) indicated by line A–C in (c) after hot pressing.

membranes (prepared via phase-inversion) or electrospun mem- noting that in this work, the authors do not report the pore size, LEP,
branes using other polymers. gas permeability or thickness of their membranes and neither the
First attempt to use electro-spun nanofiber membrane in MD experimental parameters (i.e. absolute pressure or linear velocity),
was done by Feng et al. [19]. These authors prepared PVDF which are needed for a fair comparison.
nanofiber membranes to produce drinking water using aqueous Prince et al. [38] prepared electrospun membranes consisting
solutions of NaCl at concentrations of 1, 3, and 6 wt% as feed and of PVDF and clay nanocomposites and tested them under
air-gap membrane distillation as the process. The maximum reported DCMD conditions reaching a maximum permeate flow rate below
permeate flux, achieved at a temperature difference of 60 1C and a 6 kg m  2 h  1 and a SR of 99% for a DT of 65 1C and a Tfeed
feed flow rate of 0.1 GPM, was around 11–12 kg m  2 h  1and a SR of of 80 1C using a 3.5 wt% NaCl aqueous feed solution. In the case
98.7–99.9%, depending on the feed salt concentration. It is worth of PVDF-HFP membranes for MD, only one reference was found by
 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115 113

30 30
10PH Hot Pressed 2layers
15PH Hot Pressed 2layers
10PH-hot pressed 2 layers
25 PVDF 0.2 25
Vapour flux (L h-1 m-2)

Vapor Flux (L h-1 m-2)

20
20

15
15

10
10

5
5 C = 1.49810-3 L h-1 m-2 Pa-1
R2 = 0.98
0
0 5 10 15 20 25 30 0
Delta T (°°C) 0 5 10 15 20 25
ΔP (KPa)
Fig. 10. DCMD specific distillate production [L h  1 m  2] as a function of delta T
(between hot and cold streams) for the prepared membranes 10 PH and 15 PH hot 30
pressed 2 layers and the commercial PVDF 0.2. Experimental parameters; Vfeed/
1
permeate: 0.32 m s ; Cfeed: 18 mS cm  1; Ctap_water: 110 mS cm  1, Tfeed: 30–50/ 0.2 PVDF
68 1C; Tpermeate_0: 24 1C.
25

Garcı́a-Payo et al. [39], in which hollow fiber membranes were Vapor flux (L h-1 m-2) 20
prepared by the dry/wet spinning technique and tested under
DCMD conditions. Maximum permeate flow rates reported were
15
below 1.44 kg m  2 h  1 for a DT of 25 1C and a Tfeed of 45 1C. The
general MD mass transfer model proposed by Schofield et al.
suggests that the vapor transfer rate is proportional to the vapor 10
pressure difference across the membrane [40]
N ¼ C ðPT1 P T2 Þ ¼ C DP ð2Þ 5 C = 1.00410-3 L h-1 m-2 Pa-1
where N is the vapor flux, P is the vapor pressure at temperature T R2 = 0.99
and C is the mass transfer coefficient. Subscripts 1 and 2 denote
0
the feed and permeate sides of the membrane, respectively. The
0 5 10 15 20 25
value of C depends on the characteristics of the membrane and is
ΔP (KPa)
a key parameter to assess its performance and compare it to other
membranes. However, it is difficult to experimentally measure Fig. 11. DCMD permeate flux [L h  1 m  2] as a function of DPav (as calculated by
the temperature at the interface between the vapor and liquid Eq. (3)) for electrospun 10 PH and 0.2 PVDF membranes Experimental parameters:
phase, which is affected by the temperature-polarization phe- Vfeed/permeate: 0.32 m s  1; Cfeed: 18 mS cm  1; Ctap_water: 110 mS cm  1, Tfeed: 30–50/
68 1C; Tpermeate_0: 24 1C.
nomenon. In this study, as suggested by Zhang et al. [41], the
same hydrodynamic and thermal conditions and hence theoreti-
cally the same boundary layer conditions were used to assess and 15 PH is in the range of the commercial membrane (C ¼0.912 
compare the performances of commercial and electrospun PVDF- 10-3 L h  1 m  2 Pa  1 for PH-15 and C ¼1.004  10  3 L h  1
HFP membranes. A global mass transfer coefficient was calculated m  2 Pa  1 for the commercial PVDF membrane), that of 10 PH
experimentally from the bulk temperatures by plotting N against (C ¼1.489  10  3 L h  1 m  2 Pa  1) is about 48% higher. The com-
DP. To do so, a logarithmic average DP (DPav) for counter current paratively lower thickness (110 against 170 mm) and higher
flow was calculated as expressed below: porosity (58% against 55%) of the latter compared to the 15 PH


explains the slightly higher mass transfer coefficient of the 10 PH
P P pi  P fi Ppo
DP av ¼ fo   ð3Þ membrane.
ðP P Þ
LN PfoP pi It is rare to find experimental data on mass transfer coeffi-
ð fi po Þ
cients for MD membranes in literature. However, when compar-
where Pfi, Pfo, Ppi and Ppo are the vapor pressures at inlet and ing the results of this work with published, for example by Wu
outlet temperatures of the feed and permeate sides, respectively. et al. [42], they present different composite membranes and
The vapor pressures were calculated by means of the well-known report mass transfer coefficients ranging from 0.14 to 0.96 
Antoine’s equation. The results were then plotted as shown in 10  3 L h  1 m  2 Pa  1 for a polystyrene grafted cellulose acetate
Fig. 11 for 10 PH, 15 PH and commercial PVDF membrane. A linear membrane and 0.82  10  3  1.32  10  3 L h  1 m  2 Pa  1 for OFCB
fit was performed with the slope corresponding to the C value. For (octafluorocyclobutane) plasma modified cellulose nitrate mem-
the fit, the intercept was fixed at zero as established in Eq. (2). brane. However, these membranes also showed a rejection factor
The obtained C values and their respective adjustments (R2) from 66.7% to 99.1% in the case of the polystyrene membranes and
were: C¼1.489  10  3 L h  1 m  2 Pa  1 and R2 ¼0.98 for 10 PH, from 92.1% to 99.5% for the plasma modified. Zhang et al. [41]
C¼0.912  10  3 L h  1 m-2 Pa  1 and R2 ¼0.58 for 15 PH, and tested five different commercial membranes under similar opera-
C ¼1.004  10  3 L h  1 m-2 Pa  1 and R2 ¼0.99 for the commercial tional conditions (DCMD: Cfeed of 10 g L  1; Vfeed of 0.36 m s  1 and
0.2 mm PVDF membrane. While the mass transfer coefficient for max. DT of 40 1C) and report mass transfer coefficients of around
114 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115

0.7  10  3 L h  1 m  2 Pa  1 for a PVDF supported membrane with [2] K.Y. Wang, S.W. Foo, T.-S. Chung, Mixed matrix PVDF hollow fiber mem-
mean pore size of 0.3 mm, and 1.5  10  3 L h  1 m  2 Pa  1 for a branes with nanoscale pores for desalination through direct contact mem-
brane distillation, Ind. Eng. Chem. Res. 48 (2009) 4474–4483.
0.2 mm supported PTFE membrane. The flux values achieved using [3] K.W. Lawson, D.R. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997)
PVDF-HFP are either in the range of published membranes or, in 1–25.
many cases, higher. The characteristics of the optimized PVDF-HFP [4] M.S. El-Bourawi, Z. Ding, R. Ma, M. Khayet, A framework for better under-
standing membrane distillation separation process, J. Membr. Sci. 285 (2006)
membranes 10 PH and 15 PH hot pressed 2 layers (i.e. porosity,
4–29.
pore size, pore size distribution and thickness) are within the [5] C.-Y. Kuo, H.-N. Lin, H.-A. Tsai, D.-M. Wang, J.-Y. Lai, Fabrication of a high
optimal ranges for MD. As a result, the membranes showed a fairly hydrophobic PVDF membrane via nonsolvent induced phase separation,
high LEP, which allows MD operation at relatively high flow Desalination 233 (2008) 40–47.
[6] J.-J. Kim, T.-S. Jang, Y.-D. Kwon, U.Y. Kim, S.S. Kim, Structural study
velocities without jeopardizing their salt rejection factor. When of microporous polypropylene hollow fiber membranes made by the
compared to the commercial PVDF membrane, the improved melt-spinning and cold-stretching method, J. Membr. Sci. 93 (1994) 209–215.
results of membranes gas permeability (Fig. 6), lower thickness [7] M. Tomaszewska, Preparation and properties of flat-sheet membranes from
poly(vinylidene fluoride) for membrane distillation, Desalination 104 (1996)
and slightly bigger mean pore size (Table 1) explained their better
1–11.
MD results in terms of vapor flux (Figs. 10 and 11). [8] B. Li, K.K. Sirkar, Novel membrane and device for direct contact membrane
distillation-based desalination process, Ind. Eng. Chem. Res. 43 (2004)
5300–5309.
4. Conclusions [9] S. Loeb, S. Sourirajan, Sea water demineralization by means of an osmotic
membrane, in: Saline Water Conversion? II, American Chemical Society,
1963, pp. 117–132.
PVDF-HFP mats with different fiber diameters were synthe- [10] D. Wang, K. Li, W.K. Teo, Preparation and characterization of polyvinylidene
sized via the electrospining method. The structure and morphol- fluoride (PVDF) hollow fiber membranes, J. Membr. Sci. 163 (1999) 211–220.
[11] S. Simone, A. Figoli, A. Criscuoli, M.C. Carnevale, A. Rosselli, E. Drioli,
ogy of the resultant mats was studied using scanning electron
Preparation of hollow fibre membranes from PVDF/PVP blends and their
microscopy and was then related to the polymer concentration application in VMD, J. Membr. Sci. 364 (2010) 219–232.
used in electrospinning. Scanning electron micrographs indicated [12] M.C. Garcı́a-Payo, M. Essalhi, M. Khayet, Preparation and characterization of
that by varying the polymer concentration in the solvent (i.e. PVDF-HFP copolymer hollow fiber membranes for membrane distillation,
Desalination 245 (2009) 469–473.
viscosity of the solution) the fiber diameter and pore size of the [13] C. Feng, B. Shi, G. Li, Y. Wu, Preparation and properties of microporous
resulting mats can be optimized. It was found that membranes membrane from poly(vinylidene fluoride-co-tetrafluoroethylene) (F2.4) for
fabricated using the 10 wt% concentration (10 PH-hot pressed membrane distillation, J. Membr. Sci. 237 (2004) 15–24.
[14] L. Shi, R. Wang, Y. Cao, C. Feng, D.T. Liang, J.H. Tay, Fabrication
2 layer) had the optimum properties for MD application. The
of poly(vinylidene fluoride-co-hexafluropropylene) (PVDF-HFP) asymmetric
properties and performance of the fabricated membranes were microporous hollow fiber membranes, J. Membr. Sci. 305 (2007) 215–225.
compared with commercial PVDF membrane and results indi- [15] C. Feng, R. Wang, B. Shi, G. Li, Y. Wu, Factors affecting pore structure
cated the suitability of the fabricated membranes for MD applica- and performance of poly(vinylidene fluoride-co-hexafluoro propylene)
asymmetric porous membrane, J. Membr. Sci. 277 (2006) 55–64.
tions. DCMD results revealed a maximum flow rate of 20– [16] Z.-M. Huang, Y.Z. Zhang, M. Kotaki, S. Ramakrishna, A review on polymer
22 L h  1 m  2 for 10 PH-hot pressed 2 layer (at a maximum Tfeed nanofibers by electrospinning and their applications in nanocomposites,
of 65 1C) compared 20–21 L h  1 m  2 for the PVDF commercial Compos. Sci. Technol. 63 (2003) 2223–2253.
[17] N. Bhardwaj, S.C. Kundu, Electrospinning: a fascinating fiber fabrication
membrane (at a maximum Tfeed of 68 1C). The mass transfer technique, Biotechnol. Adv. 28 (2010) 325–347.
coefficient value obtained for the optimized 10 PH hot pressed [18] R. Gopal, S. Kaur, Z. Ma, C. Chan, S. Ramakrishna, T. Matsuura, Electrospun
2 layers (C ¼1.489  10–3 L h  1 m  2 Pa  1) was found to be 48% nanofibrous filtration membrane, J. Membr. Sci. 281 (2006) 581–586.
[19] C. Feng, K.C. Khulbe, T. Matsuura, R. Gopal, S. Kaur, S. Ramakrishna,
higher than that of the PVDF commercial membrane (C ¼1.004 
M. Khayet, Production of drinking water from saline water by air-gap
10  3 L h  1 m  2 Pa  1) while salt rejection factor results under membrane distillation using polyvinylidene fluoride nanofiber membrane,
DCMD conditions (0.32 m s  1 and 18 psi) were around 98%. J. Membr. Sci. 311 (2008) 1–6.
The results demonstrated that hot pressing and double layering [20] S. Kaur, R. Barhate, S. Sundarrajan, T. Matsuura, S. Ramakrishna, Hot pressing
of electrospun membrane composite and its influence on separation perfor-
improve the characteristics of an electrospun membrane in pursu- mance on thin film composite nanofiltration membrane, Desalination 279
ing the features sought for a MD membrane (i.e. small mean pore (2011) 201–209.
size, narrow pore distribution, mechanical strength while keeping [21] S.-S. Choi, S.G. Lee, C.W. Joo, S.S. Im, S.H. Kim, Formation of interfiber bonding
in electrospun poly(etherimide) nanofiber web, J. Mater. Sci. 39 (2004)
high permeability and porosity). As a result, higher LEP and
1511–1513.
remarkable vapor fluxes were achieved by using this methodology. [22] B.S. Lalia, Y.A. Samad, R. Hasheikah, Nanocrystalline cellulose-reinforced
The membranes prepared in this way showed a vapor flux compar- composite mats for lithium-ion batteries: electrochemical and thermomecha-
able to that of the PVDF commercial membrane. In the case of the nical performance, J. Solid State Electrochemistry, http://dx.doi.org/10.1007/
s10008-012-1894-1, in press.
10 PH-hot press 2 layers the vapor flux was even higher than that of [23] M. Khayet, T. Matsuura, Preparation and Characterization of Polyvinylidene
commercial PVDF membrane. This is also reflected in the mass Fluoride Membranes for Membrane Distillation, Ind. Eng. Chem. Res. 40
transfer coefficient values or B values which resulted to be higher in (2001) 5710–5718.
[24] D. Li, M.W. Frey, Y.L. Joo, Characterization of nanofibrous membranes with
comparison with the PVDF commercial’s one. Our preliminary
capillary flow porometry, J. Membr. Sci. 286 (2006) 104–114.
assessment of PVDF-HFP electrospun and hot pressed membranes [25] B.A. Almanza, K.A.Y. Cooksey, R.C. Cooksey, W.F. Jaffe, Effect of reprocced
shows that this method has a high potential to tailor make MD material on selected physical characteristics of corrugated shipping cartons
membranes and use them in desalination processes. used in foodservice operations, Foodserv. Res. Int. 7 (1993) 177–183.
[26] R.H. Tang, W.A.W. Jr., Non-formaldehyde durable press finishing for cellulosic
textiles with phosphinocarboxylic acid, US Patent number 5,496,477, 1996.
[27] K. Singha, S. Maity, M. Singha, P. Paul, D.P. Gon, Effects of fiber diameter
Acknowledgement distribution of nonwoven fabrics on its properties, Int. J. Text. Sci. 1 (2012)
7–14.
[28] J. Doshi, D.H. Reneker, Electrospinning process and applications of electro-
The authors would like to thank Mr. Sahin Onder (Arkema Turkey spun fibers, J. Electrostat. 35 (1995) 151–160.
and Middle East) for providing sample of Kynar Powerflexs LBG. [29] J.M. Deitzel, J. Kleinmeyer, D. Harris, N.C. Beck Tan, The effect of processing
variables on the morphology of electrospun nanofibers and textiles, Polymer
References 42 (2001) 261–272.
[30] A.M. Alklaibi, N. Lior, Membrane-distillation desalination: status and poten-
tial, Desalination 171 (2005) 111–131.
[1] J.H. Hanemaaijer, J. van Medevoort, A.E. Jansen, C. Dotremont, E. van [31] J. Phattaranawik, R. Jiraratananon, A.G. Fane, Effect of pore size distribution
Sonsbeek, T. Yuan, L. De Ryck, Memstill membrane distillation—a future and air flux on mass transport in direct contact membrane distillation, J.
desalination technology, Desalination 199 (2006) 175–176. Membr. Sci. 215 (2003) 75–85.
 B.S. Lalia et al. / Journal of Membrane Science 428 (2013) 104–115 115

[32] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensive [38] J.A. Prince, G. Singh, D. Rana, T. Matsuura, V. Anbharasi, T.S. Shanmugasundaram,
review, Desalination 287 (2012) 2–18. Preparation and characterization of highly hydrophobic poly(vinylidene fluor-
[33] M. Rao, X. Geng, Y. Liao, S. Hu, W. Li, Preparation and performance of gel ide)—clay nanocomposite nanofiber membranes (PVDF–clay NNMs) for desali-
polymer electrolyte based on electrospun polymer membrane and ionic nation using direct contact membrane distillation, J. Membr. Sci. 397–398
liquid for lithium ion battery, J. Membr. Sci. 399–400 (2012) 37–42. (2012) 80–86.
[34] W. Qi, C. Lu, P. Chen, L. Han, Q. Yu, R. Xu, Electrochemical performances [39] M.C. Garcı́a-Payo, M. Essalhi, M. Khayet, Effects of PVDF-HFP concentration
and thermal properties of electrospun poly(phthalazinone ether sulfone on membrane distillation performance and structural morphology of hollow
ketone) membrane for lithium-ion battery, Mater. Lett. 66 (2012) fiber membranes, J. Membr. Sci. 347 (2010) 209–219.
239–241. [40] R.W. Schofield, A.G. Fane, C.J.D. Fell, Gas and vapour transport through
[35] M. Qtaishat, M. Khayet, T. Matsuura, Guidelines for preparation of higher flux microporous membranes. I. Knudsen–Poiseuille transition, J. Membr. Sci. 53
hydrophobic/hydrophilic composite membranes for membrane distillation, J. (1990) 159–171.
Membr. Sci. 329 (2009) 193–200. [41] J. Zhang, N. Dow, M. Duke, E. Ostarcevic, J.-D. Li, S. Gray, Identification of
[36] B. Bhushan, Y. Chae Jung, Wetting study of patterned surfaces for super- material and physical features of membrane distillation membranes for high
hydrophobicity, Ultramicroscopy 107 (2007) 1033–1041. performance desalination, J. Membr. Sci. 349 (2010) 295–303.
[37] A. Nakajima, K. Hashimoto, T. Watanabe, Recent studies on super- [42] Y. Wu, Y. Kong, X. Lin, W. Liu, J. Xu, Surface-modified hydrophilic membranes
hydrophobic films, Monatshefte für Chemie/ Chem. Mon. 132 (2001) 31–41. in membrane distillation, J. Membr. Sci. 72 (1992) 189–196.