You are on page 1of 9

Desalination 347 (2014) 175183

Contents lists available at ScienceDirect

journal homepage:

Superhydrophobic PVDFPTFE electrospun nanobrous membranes for

desalination by vacuum membrane distillation
Zhe-Qin Dong, Xiao-hua Ma, Zhen-Liang Xu , Wen-Ting You, Fang-bing Li
State Key Laboratory of Chemical Engineering, Membrane Science and Engineering R&D Lab, Chemical Engineering Research Center, East China University of Science and Technology (ECUST),
130 Meilong Road, Shanghai 200237, China


PVDFPTFE electrospun nanobrous membranes (ENMs) were fabricated by electrospining.

The ENMs with different PTFE micro-powder contents were prepared and characterized.
The PVDFPTFE electrospun nanobrous membranes were superhydrophobic.
The ENMs were used for desalination via vacuum membrane distillation.
The ENMs presented a high permeate ux with stable performance.

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

Article history: In this study, a superhydrophobic nanobrous membrane was prepared on the basis of an electrospun
Received 16 November 2013 polyvinylidene uoride (PVDF)polytetrauoroethylene (PTFE) nanobrous scaffold coupled with a micropo-
Received in revised form 7 May 2014 rous PTFE substrate. The PVDFPTFE nanobrous scaffold was fabricated by electrospining of PVDFPTFE blend
Accepted 8 May 2014
solutions, it was observed that by changing the PTFE micro-powder content in the dope solutions from 0 wt.%
Available online 17 June 2014
to 12 wt.%, the water contact angle (WCA) and the liquid entry pressure (LEPw) of the membrane vary from
130.4 and 84 kPa to 152.2 and 137 kPa, respectively. The superhydrophobic PVDFPTFE nanobrous membrane
Electrospining was then tested for desalination by vacuum membrane distillation (VMD), a stable ux of 18.5 kg/m2 h and salt
PVDFPTFE rejection higher than 99.9% was presented throughout the entire testing period of 15 h, indicating the great po-
Superhydrophobic tential of the PVDFPTFE nanobrous membranes in VMD. For further application of the PVDFPTFE nanobrous
Nanobrous membrane membranes in VMD, a mathematical model was presented to predict the vapor ux of the novel membrane
Membrane distillation under various operation conditions. A good agreement between the experimental and theoretical values for
vapor uxes was obtained; the results indicated that the VMD ux increased with the increase of feed tempera-
ture and ow rate and decreased with the increase of permeate pressure.
2014 Elsevier B.V. All rights reserved.

1. Introduction surface morphology and pore structure. An ideal membrane for MD

should exhibit a superhydrophobic surface to avoid pore wetting and a
Membrane distillation (MD) is a non-isothermal separation process well-designed pore structure to enhance water ux [511]. These proper-
which is driven by a vapor pressure difference through a porous hydro- ties mainly depend on the polymer or additives that are used for prepar-
phobic membrane. There are generally four different MD congurations ing the membrane and also on the method of synthesis of membrane.
including direct contact membrane distillation (DCMD), sweeping gas Most of the materials employed in the preparation of membranes
membrane distillation (SGMD), vacuum membrane distillation (VMD) for MD applications are hydrophobic, such as polypropylene (PP),
and air gap membrane distillation (AGMD). Compared with other sepa- polytetrauoroethylene (PTFE), and polyvinylidene uoride (PVDF).
ration process, MD has a great potential in desalination for possessing However, the commercial membranes fabricated out of these polymers
the advantages of lower operating temperature, less energy consump- are commonly made through a melting process or phase inversion pro-
tion as well as 100% theoretical rejection of non-volatile solute [14]. cess, the surface of these membranes is not highly hydrophobic and the
Although the membrane only acts as a physical barrier in membrane porosity is low, which is the major barrier impeding MD industry appli-
distillation, the performance of MD is strongly affected by the membrane cations [12,13].
Electrospinning is an attractive approach for the fabrication of nano-
Corresponding author. Tel.: +86 21 64253061; fax: +86 21 64252989. meter or submicrometer-sized bers which received an increasing atten-
E-mail address: (Z.-L. Xu). tion in recent years [1417]. Compared with commercial membranes
0011-9164/ 2014 Elsevier B.V. All rights reserved.
176 Z.-Q. Dong et al. / Desalination 347 (2014) 175183

made through conventional process, the electropsun nanober mem- 2. Experimental

branes (ENMs) have showed great potential in membrane distillation
due to their high void volume fraction and interconnected open 2.1. Materials
structure [1821]. Feng et al. rstly attempted to use ENM in MD [18],
up to 22% of saline water was treated with the membrane. The salt re- Polyvinylidene uoride (PVDF, Solef 6010) was purchased
jection was about 99%, and the uxes were as high as or higher than from Solvay Advanced Polymers, L.L.C. (Alpharetta GA, USA) and
those of commercial PVDF membrane, which shows the great potential polytetrauoroethylene (PTFE) micro-powders (Dyneon TF9207)
of PVDF ENM in MD. Essalhi and Khayet [19] investigated the effect of were purchased from Shanghai Xianglan chemicals CO. LTD (China). For
eletrospinning time on ENMs; a considerable increase of both the thick- PTFE micro-powders, the primary particles are 120 nm in diameter, and
ness and the liquid entry pressure of water with the electrospinning the average aggregate length is 4 m. All other solvents and reagents
time was observed, while only a slight enhancement was detected for were analytical grade and purchased from Shanghai Sinopharm Chemical
the void volume fraction. Yuan Liao et al. [20] optimized the pore sizes Reagent CO. LTD (China). The PTFE microporous substrate was purchased
of PVDF ENMs by varying polymer dope compositions, electrospinning from Shanghai Nuozhen New Chemical Materials CO. LTD (China), the
process parameters and heat-press treatment. The DCMD test for the thickness of the substrate was 65 m, with a mean pore size of 1 m
post-treated PVDF ENMs was much better than the commercial PVDF and porosity of 75%. Commercial PTFE membrane manufactured
membranes, suggesting the competency of PVDF ENMs for MD applica- by Sumitomo Electric Industries, Ltd was used to compare with the
tions. However, the uniform electropsun nanober membranes fabri- nanobrous membranes in vacuum membrane distillation.
cated by pure polymer dopes without hydrophobic additives are
usually not superhydrophobic [1820]. Preparation and characteriza- 2.2. Preparation of PVDFPTFE nanobrous membranes
tion of superhydrophobic ENMs for MD were reported by Prince et al.
[21]. The superhydrophobic membrane with a water contact angle of The dope solutions were rst prepared by dissolving 15 wt.% PVDF in
154.2 was obtained by electrospining of PVDF blended with clay nano- the solvent mixture of N, N-dimethylacetamide (DMAC) and acetone
composites, but the PVDFclay membrane presented quite a low ux (50/50 wt.%), different amounts of PTFE micro-powders at 0, 3, 6 and
(b 5.5 kg/m2 h) in direct contact membrane distillation, which limits 12 wt.% were added into the dope solutions to enhance the hydropho-
its application. bicity of the membrane. The compositions of the dope solutions for
The goal of this study was to fabricate a superhydrophobic electrospinning corresponding to the membrane samples S1S4 are
nanobrous membrane with high water ux for membrane distilla- presented in Table 1. The PVDFPTFE solutions were allowed to stir at
tion. The PTFE micro-powders were added in the electrospining of 50 C for about 24 h before electrospinning.
PVDF dope solutions to improve the surface hydrophobicity of the The electrospinning setup used in the present work was the same in
nanobrous membrane, as previous studies showed that the PVDF our previous study [2528]. It consisted of a 20 mL plastic syringe to
PTFE composite membranes fabricated by phase inversion method hold the polymer solution connected to a syringe pump (WC50C6,
or coxial-electrospining method exhibited enhanced hydrophobicity Zhejiang) to control the injection rate of the polymer solution, two
[2225]. In this work, PVDFPTFE nanobrous membranes with dif- electrodes (a metallic needle and a grounded metallic collector kept in
ferent properties were prepared by electrospining of PVDFPTFE horizontal position), and a high-voltage power supply (DWP3031ACD8,
blend solutions on a PTFE microporous substrate, which acts as a Tianjin) which could generate positive direct current voltage up to
support layer to improve the mechanical properties. The schematic 30 kV. The polymer drop from the tip of the needle (0.4 mm inner
diagrams for the assembly of the PVDFPTFE nanobrous mem- diameter) connected to the syringe by a Teon tube is drawn into a
branes were shown in Fig. 1. The prepared membranes were charac- ber due to the high voltage. The whole setup was placed in a temper-
terized by scanning electron microscopy (SEM), atomic force microscopy ature controlled room.
(AFM), water contact angle (WCA) and liquid entry pressure of water The PVDFPTFE nanobrous membranes were fabricated by
(LEPw) to investigate their surface and structure. Subsequently, these electrospinning the dope solutions on a PTFE microporous substrate,
nanobrous membranes were tested in vacuum membrane distillation as shown in Fig. 1. A positive voltage of 18 kV was applied across a dis-
to compare with a commercial MD membrane. tance of 15 cm; the polymer was electrospun at a rate of 0.5 mL/h to get

Fig. 1. Fabrication schematics of PVDFPTFE nanobrous composite membrane.

Z.-Q. Dong et al. / Desalination 347 (2014) 175183 177

Table 1 under identical operating conditions. The inuence of operation condi-

Compositions of electrospining dope solutions. tions of VMD on the performance of nanobrous membranes was also
Sample PVDF (wt %) PTFE (wt.%) DMAC (wt.%) Acetone (wt.%) investigated, the feed temperature, the pressure at the permeate side
and the ow rate were varied in the range of 313343 K, 315 kPa
S1 15 0 42.5 42.5
S2 15 3 41.0 41.0 and 1890 L/h, respectively.
S3 15 6 39.5 39.5 The VMD experimental installation of this work is shown in Fig. 2.
S4 15 12 36.5 36.5 The feed water was driven by a circulation pump (Nanjing Lvhuan
Pumps Co. Ltd.) and the feed ow was measured by a owmeter con-
nected between the pump and the module. A thermostatic bath was
a uniform mat of thickness about 30 m, while the humidity and tem- used to provide the feed with heat energy to perform VMD. The tem-
perature were controlled at 50 5% and 25 1 C, respectively. After perature of the feed water at the inlet of the membrane module was
electrospinning, the nanobrous membranes were dried for a day at continuously measured by a Pt100 probe connected to an automatic
room temperature for the solvents to evaporate. The dried membranes temperature regulator.
were stored in a desiccator before use. In order to maintain the feed concentration, water was added in the
feed tank every 1520 min during experiments. The quantity of the
2.3. Membrane characterization replenishing water was determined by the decrement weight of the
feed tank measured by an electronic balance.
The morphologies of the membranes were observed by scanning At the permeate side of the membrane module, a vacuum pump
electron microscope (SEM; JSM-6360LV, Japan). The average ber (Shanghai Vacdo Vacuum Equipment CO. LTD.) was connected and a
diameter and average pore size of the nanobrous membranes were glass condenser with refrigerating machine was used to recover the
measured from the SEM images using J MicroVision image analyzer. water permeating vapor. The permeate pressure was measured by a
Quantitative surface roughness analysis of the nanobrous mercurial pressure gauge. The permeate ux was calculated, in every
membrane was measured using AFM (Veeco, Nanoscope IIIa Multi- case, by measuring the volume of the condensed water.
mode AFM). Air-dried membrane sample was xed on a specimen
holder and 5 m 5 m areas were scanned by tapping mode in air. 3. Theoretical approach
In the AFM analysis, the average roughness (Ra) was used to evaluate
the morphology of the membrane. The system to be investigated consists of a composite membrane
A JC2000D1 (produced by Shanghai Zhongcheng Digital Technology maintained between hot feed solution and vacuum permeate side. In
Apparatus Co. Ltd., China) system was used for the determination of the VMD processes, evaporation takes place at the hot feed side while con-
dynamic contact angles () of the nanobrous membranes at ambient densation takes place at the vacuum side after water vapor transferred
temperature. A water droplet of 0.2 mL was dispersed from a needle through the pores of the nanobrous top-layer and porous sub-layer.
tip onto the membrane surface. The machine was coupled with a Both mass and heat transfer occur simultaneously across the membrane
camera, enabling image capture at 10 frames/s. The contact angles under vapor pressure difference. Consequently, the temperatures and
were determined from these images using the specic calculation soft- concentrations at the membrane surfaces differ from those at the bulk
ware. To ensure that the results were sufciently credible, the measure- liquid phases, leading to a decrease of the driving force and reduction
ment for each set of samples was performed in triplicate, and the of the VMD ux, as shown in Fig. 3.
average data of contact angles was taken.
The membrane porosity is dened as the volume of the pores divided 3.1. Heat transfer
by the total volume of the membrane. It can be determined by using iso-
propyl alcohol (IPA), which penetrates inside the inter-ber space of the The following heat transfer ux is involved in the VMD system as
membranes and distilled water, which does not enter in the inter-ber described previously [12]:
space. Detailed information was reported elsewhere [29]. Through the feed solution boundary layer:
The evaluation of the membrane wetting resistance was conducted
by liquid entry pressure of water (LEPw) measurements. The LEPw
Q f h f T b T m : 2
value is the minimum applied pressure to water before it penetrates
into the membrane pores, which mainly depends on the pore size and
membrane hydrophobicity according to the Laplace equation [30]: Through the nanobrous top-layer:

2B L cos Q t ht T m T s J w Hv : 3
LEPw NP interface P liquid P vapor 1
r max

Through the porous sub-layer:

where B is a geometric factor determined by pore structure, L is the
liquid surface tension and is the liquid/solid contact angle.
LEPw values were measured using a dead-end ltration set-up Q s hs T s T v : 4
which was designed according to the method described extensively by
Smolder and Franken [31]. The dry hydrophobic membrane was placed At steady state, the overall heat transfer ux through the whole VMD
in a dead end cell and the cell was topped up with DI water. Compressed system, Q, is given by
nitrogen was used to apply pressure in the cell. The pressure was noted
when the rst drop of water came from the cell. The experiment was
Q Q f Q t Q s: 5
carried out thrice using three different membranes made from the
same condition. The results were averaged to obtain the nal LEPw.
In the above equations, h is the heat transfer coefcient, Jw is the per-
2.4. VMD performance test meate ux, Hv is the latent heat of vaporization and T is the absolute
temperature. The subscripts specifying the location of the temperature,
The nanobrous membranes and commercial PTFE membrane were b, m and s refer to the bulk solution, top-layer of the membrane and its
tested in a VMD setup with an effective membrane area of 0.00785 m2 sub-layer, respectively.
178 Z.-Q. Dong et al. / Desalination 347 (2014) 175183

Fig. 2. Schematic diagram of vacuum membrane distillation.

The heat transfer coefcients hf can be estimated with the help of the 3.2. Mass transfer
known empirical correlations:
The permeate ux of water vapor, Jw, is related to the vapor pressure
Nu difference as follows:
hf 6
J w K t P m P s K s P s P v 10
where Nu can be obtained by empirical equations described in Ref. [5].
where Pm is the partial vapor pressure at membrane surface as the
0:74 0:33 function of membrane surface temperature Tm as described in Eq. (11)
Nu 0:0142Re Pr 7 and Ps and Pv are the vapor pressure at the sub-layer and vacuum side,
where Re and Pr are the Reynolds and Prandtl numbers, respectively.
The heat transfer coefcient of the top-layer (ht) and sub-layer (hs), P m P sat T m T m xm 11
can be calculated from the thermal conductivities of the hydrophobic
membrane polymer and the gas present inside the pores: where Psat(Tm) is the saturated vapor pressure of pure water at mem-
brane surface temperature Tm, (Tm) is the activity coefcient of water
kg t kt 1t and xm is the molar fraction of water at membrane surface.
ht 8 Kt and Ks are the permeability of the nanobrous top-layer and
porous sub-layer, respectively. According to the kinetic theory of
the dusty-gas model, the Knudsen diffusion and molecular diffusion
are taken place in the process of the vapor transporting through the
kg s ks 1s
hs : 9 membrane pores. As reported previously [5], the Knudsen diffusion of
water molecules is dominant in VMD process, Kt and Ks can be calculated
by Eqs. (12)(13)
In the above equations, k is the thermal conductivity, is the porosity
of the layer and is the thickness of the layer. The subscripts g, t, s refer 8 t rt 1
to gas present inside the pores, nanobrous top-layer and porous sub- Kt 12
3 t t 2RT m
layer, respectively.

8 s rs 1
Ks 13
3 s s 2RT s

where is the pore tortuosity, is the membrane thickness, R is the gas

constant and M is the water molecular mass. The subscripts t and s
refer to nanobrous top-layer and porous sub-layer, respectively.
According to Eqs. (2)(13), the prediction ux can be obtained by
the iteration method.

4. Results and discussions

4.1. Membrane morphology analysis

The SEM images of the surface of the PVDFPTFE nanobrous mem-

branes prepared with various PTFE micro-powders contents are shown
in Fig. 3. The nanobers fabricated by pure PVDF solution (S1) were
smooth and uniform, due to the good solubility and volatility of the
Fig. 3. Heat and mass transfer through nanobrous composite membrane in vacuum mixed solvents. Rough particles emerged in the nanobers when PTFE
membrane distillation. micro-powders were added into the dope solution, it is noted that the
Z.-Q. Dong et al. / Desalination 347 (2014) 175183 179

amount of rough particles increased with the increase of PTFE concentra- Thus, the addition of the PTFE micro-powders increased the static
tions in dope solutions. Previous studies show that superhydrophobic contact angle 0 since PTFE has a lower surface energy (19 dyn/cm)
surfaces can be produced mainly in two ways. One is to create a rough than PVDF (25 dyn/cm) according to Eq. (2). Furthermore, the AFM im-
structure on a hydrophobic surface (CA N 90), and the other is to ages shown in Fig. 4 indicated that the average surface roughness (Ra)
modify a rough surface by compounds with low surface free energy, increased from 276 nm to 472 nm by changing the PTFE content from
such as uorinated or silicon compounds [32]. The addition of PTFE 0 wt.% to 12 wt.%, which further increased the water contact angle w.
micro-powders not only reduces the surface energy of the membranes The dynamic contact angle of these membranes is shown in Fig. 5.
but also improves the surface roughness, which is benecial to surface The result shows that the water contact angle of all the membranes de-
hydrophobicity. creases slightly over the entire time period, with the smallest decrease
The ber diameters and pore sizes of the fabricated membranes coming from the membrane dope solution containing 12 wt.% PTFE
were determined from the SEM image using the J Micro Vision image micro-powders (S4). This indicates that the membranes with the
analyzer and Image software. The ber diameters and pore sizes for highest PTFE content are able to retain their hydrophobicity over a lon-
various membranes are summarized in Table 2. Both the ber diameter ger period of time.
and pore size of the membranes increase slightly as the concentration of
PTFE micro-powder increases in the dope solution, since a high viscosity
favors the formation of thicker bers and bigger pore size. Similar inves- 4.3. Premilinary evaluation of PVDFPTFE nanobrous membranes in
tigations of dope solution viscosity on membrane properties have been vacuum membrane distillation
conducted previously [19].
A 15 hour continuous separation test was carried out for the PVDF
4.2. Membrane surface analysis PTFE nanobrous membrane samples (S1S4) and the commercial PTFE
membrane (S0), the results are summarized in Fig. 6. As shown in
The water contact angles (WCA) for all the electrospun membranes Fig. 6(a), the conductivity of the permeate water maintains at a quite
(S1S4) are shown in Table. 2, the WCA of the PTFE commercial mem- low level for all the membranes (b 5 S/cm), since the feed solution
brane is also characterized and denoted as S0. The result shows that can hardly penetrate to the condensed water in VMD. It can be seen
all the electrospun membranes (S1S4) have impressive water contact from Fig. 6(b) that the initial permeate ux of the S4 was 18.5 kg/m2 h
angles (130.4 to 151.2) compared with the PTFE commercial mem- for a feed solution containing 3.5 wt.% NaCl, while the feed temperature
brane (121.5). It can also be seen that the WCA of the electrospun and the permeate pressure side were 333 K and 9 kPa, respectively. This
membranes increased as the concentration of PTFE micro-powders in- permeation ux is as competitive as under the same MD conditions, the
creased in the dope solution, the superhydrophobic membrane (S4) permeate ux of the commercial PTFE membrane S0 was around 15 kg/
with a WCA of 151.2 was obtained when the concentration of PTFE m2 h. The high ux performance for the nanobrous membrane (S4) is
micro-powders was 12 wt.%, while the membrane (S1) prepared with mainly due to the high void volume fraction and interconnected open
pure PVDF solution only had a WCA of 130.4. The phenomenon that structure of the nanober layer compared with the commercial PTFE
the WCA increased signicantly with the addition of PTFE micro- membrane.
powders can be explained as follows. The uxes of the nanobrous membranes increased slightly from S1
Generally, the water contact angle for a at surface can be obtained to S4, since the pore size of the nanobrous membranes increased
by the well known Young equation as the following: slightly with PTFE addition. Besides, the uxes of all the nanobrous
membranes decreased over a period operation of around 15 h which
SA SL is an indication of pore wetting as reported previously [29]. As time
cos 0 : 14 progressed, more and more pores become wetted, thus resulted in the
decrease of membrane permeability and water ux. The smallest de-
crease was coming from the membrane S4 (b5%), suggesting that the
The characteristic angle 0 is called the static contact angle. SA and superhydrophobic surface of nanobrous membranes can alleviate the
SL are the surface energies of solid against air and liquid, respectively, pore wetting problem in vacuum membrane distillation. This phenome-
and LA is the surface energy of liquid against air [33]. If a water droplet non can be explained by the LEPw values listed in Table 2. The LEPw
is placed on a rough surface with a homogeneous interface, the area of values of the nanobrous membranes increased with the increase of the
interface increases with respect to that for a smooth surface. The PTFE content, which was mainly due to the change of WCA (Eq. (1)). Be-
Wenzel equation [34] could be used to predict the contact angle: sides, the LEPw value of the superhyphobic membrane S4 is competitive
as the LEPw values of the electrospun nanobrous membrane reported
cos w R f cos 0 15 in the previous literature was around 100 kPa [1820], indicating the su-
perior anti-wetting properties of this membrane in vacuum membrane
where w is the contact angle of a water droplet upon a rough solid sur- distillation.
face calculated by Wenzel equation. Rf is the non-dimensional surface Due to the superhydrophobic surface and inter-connected structure,
factor, equal to the ratio of the surface area to its at projected the PVDFPTFE nanobrous membrane S4 exhibited a stable perfor-
area. This Wenzel model also illustrates that a hydrophobic surface mance and satised water quality over the 15 h testing, which shows
(0 N 90) becomes more hydrophobic with an increase of surface the great potential of the PVDFPTFE nanobrous membranes in vacuum
roughness (Rf). membrane distillation.

Table 2
Properties of nanobrous membranes and commercial MD membrane.

Sample Fiber diameter (nm) Membrane thickness (m) Mean pore size (m) Porosity (%) CA () CA reduction (%) LEPw (kPa)

S0 81 0.1 55% 121.5 1.0 6.0 151 2

S1 460 160 94 0.35 0.10 72 1.5 130.4 1.5 5.6 84 2
S2 490 135 97 0.38 0.13 72 2.0 137.0 1.8 5.1 101 3
S3 510 175 92 0.42 0.08 71 2.0 142.3 1.0 4.1 118 3
S4 580 190 98 0.49 0.15 69 1.5 152.2 2.0 3.0 137 4
180 Z.-Q. Dong et al. / Desalination 347 (2014) 175183

Fig. 4. Surface morphology and water contact angle of PVDFPTFE nanobrous membrane prepared with different PTFE contents.

Fig. 5. AFM images of surface for PVDFPTFE nanobrous membranes S1 and S4.
Z.-Q. Dong et al. / Desalination 347 (2014) 175183 181

Fig. 6. Dynamic water contact angle of PVDFPTFE nanobrous membranes and PTFE
commercial membrane.

4.4. Effect of the operating conditions on VMD performance of PVDFPTFE

nanobrous membranes

To further investigate the operating conditions on VMD perfor-

mance of PVDFPTFE nanobrous membrane, the inuence of feed
temperature, permeate pressure and feed ow rate on the water ux
of the superhydrophobic PVDFPTFE nanobrous membrane (S4) was
As a highly important operating parameter, the effect of feed tem-
perature on the permeate ux of the PVDFPTFE nanobrous mem-
brane was investigated in this study. The variation of the permeate
ux with feed temperatures at different ow rates was summarized in
Fig. 7, the temperatures ranged from 313 K to 343 K at various ow
rates (90 L/h, 54 L/h, 18 L/h), while the pressure at the permeate side
was maintained invariable. A good agreement between the experimen-
tal and theoretical values for vapor uxes was obtained; the exponential Fig. 7. Long time performance of the PVDFPTFE nanobrous membranes and PTFE com-
mercial membrane for VMD application: (a) permeate conductivity; (b) ux.
increase of the MD ux with the increase of the feed temperature is due
to the exponential increase of the vapor pressure (Pm) with membrane
surface temperature (Tm), which increases the transmembrane vapor and the membrane surface is formed due to the large amount of heat
pressure as all the other involved MD parameters are maintained invari- requirement for vaporizing water at the membrane surface, which is
ables. [35] However, considering the limit of membrane material heat called the temperature polarization. Higher ow rate promotes more
resistance and energy consumption for heating, the feed temperature turbulence and eddies in the boundary layer, which would increase the
should be controlled in suitable range.
In VMD, The permeate pressure is an indispensable factor which is
directly related to the transmembrane pressure difference, the inuence
of permeate pressure on water ux was studied and the results are
shown in Fig. 8. The permeate pressure was varied from 3 kPa to
15 kPa at different temperatures (50 C, 60 C and 70 C), while feed
ow rate was maintained constant. It is found that the MD ux de-
creases linearly with the increase of the permeate pressure (Pv), similar
investigations have been reported previously [36]. About 1015 kg/m2 h
water ux decline was observed when the permeate pressure increased
from 3 kPa to 15 kPa, which indicates the obvious effect of permeate
pressure on MD ux. In VMD, both the permeate ux and the trans-
membrane hydrostatic pressure increase with the decrease of the vacu-
um pressure applied in the permeate side (downstream), which results
in the high risk of membrane pore wetting, so the choice of pressure
should be cautious.
Fig. 9 presents the effect of the feed ow rate on the permeate ux of
the PVDFPTFE nanobrous membrane. A practically linear increase of
the MD ux with the feed ow velocity was observed when the feed
ow rate increased from 18 L/h to 90 L/h, the permeate ux improved
about 2 times at 333 K feed temperature and 9 kPa permeate pressure. Fig. 8. Effect of feed temperature on the permeate ux of the PVDFPTFE nanobrous
In VMD process, the difference of temperature between the feed solution membrane.
182 Z.-Q. Dong et al. / Desalination 347 (2014) 175183

membranes in VMD. The inuence of operation parameters, feed temper-

ature, permeate pressure and feed ow rate, on the VMD performance of
the PVDFPTFE nanobrous membrane was also investigated. A good
agreement between the experimental and theoretical values for vapor
uxes was obtained; the results indicated that the VMD ux increased
with the increase of feed temperature and ow rate and decreased with
the increase of permeate pressure.

h heat transfer coefcient
dh hydraulic diameter of channel (m)
Hv latent heat of vaporization (J/mol)
Jw permeate ux (Lh1m2)
K membrane permeability
M molecular mass (mol/kg)
Nu Nusselt number
P pressure (kPa)
Psat saturated vapor pressure (kPa)
Fig. 9. Effect of permeate pressure on the permeate ux of the PVDFPTFE nanobrous Pr Prendtl number
membrane. r size of membrane pore (m)
R gas constant (Jmol1K1)
transfer coefcients hf, thus results in the increase of temperature (Tm) Re Reynolds number
and vapor pressure (Pm) at membrane surface according to Eqs. (5) T temperature (K)
and (11). Therefore, higher ow rate increases permeate ux (Fig. 10). v solution velocity (m/s)
x molar fraction

5. Conclusions
Greek letters
Superhydrophobic PVDFPTFE nanobrous membranes with stable activity coefcient
MD performance were fabricated by electrospining of PVDFPTFE membrane thickness (m)
blend solution on a PTFE microporous substrate. The effect of PTFE porosity of the membrane
micro-powder contents in the dope solution on the properties of the viscosity of feed solution (Pas)
PVDFPTFE nanobrous membranes was investigated, the water con- density of solution (kg/m3)
tact angle (WCA) and the liquid entry pressure (LEPw) of the mem- pore tortuosity
branes increased signicantly as the PTFE concentration increased. The
highest water contact angle 151.2 was obtained at PTFE concentration
12% due to the hierarchical micro/nanostructure membrane surface.
Both the ber diameter and pore size of the nanobrous membranes
b bulk
increased slightly as the PTFE micro-powder concentration increased
f feed
in the dope solution, due to the increase of dope solution viscosity.
g gas
The long time performance of the PVDFPTFE nanobrous mem-
m membrane surface
branes for VMD application was studied. The superhydrophobic
s porous sub-layer
PVDFPTFE nanobrous membrane maintained a stable ux of
t nanobrous top-layer
18.5 kg/m2 h and salt rejection higher than 99.9% after a 15 h continuous
v vacuum side
separation test, suggesting the great potential of PVDFPTFE nanobrous


The authors thank for the nancial support by the Key Technology
R&D Program of Shanghai Committee of Science and Technology in
China (11DZ1205201), Special Project of the Development and Industri-
alization of New Materials of National Development and Reform Com-
mission in China (20132548), the Fundamental Research Funds for the
Central Universities (WA1214062) and the First-class General Financial
Grant from the China Postdoctoral Science Foundation (2013M540336).

[1] T.Y. Cath, V.D. Adams, A.E. Childress, Experimental study of desalination using direct
contact membrane distillation: a new approach to ux enhancement, J. Membr. Sci.
228 (2004) 516.
[2] M. Gryta, M. Tomaszewska, J. Grzechulska, A.W. Morawski, Membrane distillation of
NaCl solution containing natural organic matter, J. Membr. Sci. 181 (2001) 279287.
[3] Y. Xu, B.K. Zhu, Y.Y. Xu, Pilot test of vacuum membrane distillation for seawater
desalination on a ship, Desalination 189 (2006) 165169.
[4] E. El-Zanati, K.M. El-Khatib, Integrated membrane-based desalination system, Desa-
Fig. 10. Effect of ow rate on the permeate ux of the PVDFPTFE nanobrous membrane. lination 205 (2007) 1525.
Z.-Q. Dong et al. / Desalination 347 (2014) 175183 183

[5] J.I. Mengual, M. Khayet, M.P. Godino, Heat and mass transfer in vacuum membrane [21] J.A. Prince, G. Singha, D. Rana, T. Matsuura, V. Anbharasi, T.S. Shanmugasundaram,
distillation, Int. J. Heat Mass Transf. 47 (2004) 865. Preparation and characterization of highly hydrophobic poly (vinylidene uoride)
[6] J. Phattaranawik, R. Jiraratananon, A.G. Fane, Effect of pore size distribution and air clay nanocomposite nanober membranes (PVDFclay NNMs) for desalination using
ux on mass transport in direct contact membrane distillation, J. Membr. Sci. 215 direct contact membrane distillation, J. Membr. Sci. 397398 (2012) 8086.
(2003) 7585. [22] M.M. Teoh, T.S. Chung, Membrane distillation with hydrophobic macrovoid free
[7] M. Gryta, Inuence of polypropylene membrane surface porosity on the perfor- PVDFPTFE hollow ber membranes, Sep. Purif. Technol. 66 (2009) 229236.
mance of membrane distillation process, J. Membr. Sci. 287 (2007) 6778. [23] M. Gryta, M. Barancewicz, Inuence of morphology of PVDF capillary membranes on
[8] F.A.A. Al-Rub, F. Banat, K. Beni-Melhim, Parametric sensitivity analysis of direct the performance of direct contact membrane distillation, J. Membr. Sci. 358 (2010)
contact membrane distillation, Sep. Sci. Technol. 37 (2002) 32453271. 158167.
[9] M. Khayet, Membranes and theoretical modeling of membrane distillation: a re- [24] M.M. Teoh, T.S. Chung, Y.S. Yeo, Dual-layer PVDF/PTFE composite hollow bers with
view, Adv. Colloid Interface Sci. 164 (12) (2011) 5688. a thin macrovoid-free selective layer for water production via membrane distilla-
[10] N. Alkhudhiri, N. Darwish, Hilal, Membrane distillation: a comprehensive review, tion, Chem. Eng. Sci. 171 (2011) 684691.
Desalination 287 (2012) 218. [25] P.P. Lu, Z.L. Xu, H. Yang, Y.M. Wei, Processingstructureproperty correlations of
[11] K.Y. Wang, T.S. Chung, M. Gryta, Hydrophobic PVDF hollow ber membranes with polyethersulfone/peruorosulfonic acid nanobers fabricated via electrospinning
narrow pore size distribution and ultra-thin skin for the fresh water production from polymernanoparticle suspensions, ACS Appl. Mater. Interfaces 4 (2012)
through membrane distillation, Chem. Eng. Sci. 63 (2008) 25872594. 17161723.
[12] R.W. Schoeld, A.G. Fane, C.J.D. Fell, Heat and mass transfer in membrane distilla- [26] P.P. Lu, Z.L. Xu, H. Yang, Y.M. Wei, H.T. Xu, Effects of ethanol and isopropanol on the
tion, J. Membr. Sci. 33 (1987) 299. structures and properties of polyethersulfone/peruorosulfonic acid nanobers fab-
[13] M. Khayet, J.I. Mengual, T. Matsuura, Porous hydrophobic/hydrophilic composite ricated via electrospinning, J. Polym. Res. 19 (2012) 98549864.
membranes: application in desalination using direct contact membrane distillation, [27] P.P. Lu, Z.L. Xu, H. Yang, Y.M. Wei, Preparation and characterization of high-
J. Membr. Sci. 252 (2005) 101. performance peruorosulfonic acid/SiO2 nanobers with catalytic property via
[14] C.Y. Feng, K.C. Khulbe, S. Tabe, Volatile organic compound removal by membrane electrospinning, Ind. Eng. Chem. Res. 51 (2012) 1134811354.
gas stripping using electro-spun nanober membrane, Desalination 287 (2012) [28] X.H. Ma, S.W. Gu, X. Wen, Z.L. Xu, J.L. Zhang, Spinnability of SPPESK and its applica-
98102. tion in esterication, J. Polym. Res. 20 (2013) 155.
[15] R. Wang, Y. Liu, B. Li, B.S. Hsiao, B. Chua, Electrospun nanobrous membranes for [29] K.W. Lawson, D.R. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997) 125.
high ux microltration, J. Membr. Sci. 392393 (2012) 167174. [30] M. Khayet, T. Matsuura, Membrane Distillation: Principles and Applications, Elsevier,
[16] R. Gopal, S. Kaur, Z. Ma, C. Chan, S. Ramakrishna, T. Matsuura, Electrospun nanobrous The Netherlands, 2011.
ltration membrane, J. Membr. Sci. 281 (2006) 581586. [31] K. Smolders, A.C.M. Franken, Terminology for membrane distillation, Desalination
[17] R. Gopal, S. Kaur, C.Y. Feng, C. Chan, S. Ramakrishna, S. Tabe, T. Matsuura, 72 (1989) 249262.
Electrospun nanobrous polysulfone membranes as pre-lters: particulate removal, [32] M. Li, J. Zhai, H. Liu, Y.L. Song, L. Jiang, D.B. Zhu, Electrochemical deposition of con-
J. Membr. Sci. 289 (2007) 210219. ductive superhydrophobic zinc oxide thin lms, J. Phys. Chem. B 107 (2003)
[18] C. Feng, K.C. Khulbe, T. Matsuura, R. Gopal, S. Kaur, S. Ramakrishna, M. Khayet, Pro- 99549957.
duction of drinking water from saline water by air-gap membrane distillation using [33] B. Bhushan, Y.C. Jung, Natural and biomimetic articial surfaces for super-hydrophobic-
polyvinylidene uoride nanober membrane, J. Membr. Sci. 311 (2008) 16. ity, self-cleaning, low adhesion, and drag reduction, Prog. Mater. Sci. 56 (2011) 1108.
[19] M. Essalhi, M. Khayet, Self-sustained webs of polyvinylidene uoride electrospun [34] R.N. Wenzel, Resistance of solid surfaces to wetting by water, Ind. Eng. Chem. 28
nanobers at different electrospinning times: 1. Desalination by direct contact (1936) 988994.
membrane distillation, J. Membr. Sci. 433 (2013) 167179. [35] M.S. El-Bourawi, Z. Ding, R. Ma, M. Khayet, A framework for better understanding
[20] Y. Liao, R. Wang, M. Tian, C.Q. Qiu, A.G. Fane, Fabrication of polyvinylidene uoride membrane distillation separation process, J. Membr. Sci. 285 (2006) 429.
(PVDF) nanober membranes by electro-spinning for direct contact membrane dis- [36] G. Sarti, C. Gostoli, S. Bandini, Extraction of organic components from aqueous
tillation, J. Membr. Sci. 425426 (2013) 3039. streams by vacuum membrane distillation, J. Membr. Sci. 80 (1993) 2133