Desalination 413 (2017) 127135

Contents lists available at ScienceDirect

Desalination

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

Anti-fouling membranes by manipulating surface wettability and their

anti-fouling mechanism
Xuemei Lu a, Yuelian Peng a,, Haoran Qiu a, Xinrui Liu a, Lei Ge b
a
Beijing Key Laboratory for Green Catalysis and Separation, College of Environmental and Energy Engineering, Beijing University of Technology, Beijing 100124, China
b
School of Chemical Engineering, the University of Queensland, Brisbane 4072, Australia

H I G H L I G H T S

Two routes were used for superhydrophobic and superhydrophilic surface modication.
Three organic foulants examined in membrane distillation.
The anti-wetting and anti-fouling behaviors of three membranes were compared.
Superhydrophobic modication displayed much better self-cleaning property.
Superhydrophilic modication only slowed down wetting and fouling.

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

Article history: This study aims to compare the anti-wetting and anti-fouling surface behavior in membrane distillation (MD) by
Received 26 August 2016 two surface modication routes on porous polyvinylidene uoride (PVDF) membranes. A superhydrophobic
Received in revised form 13 January 2017 membrane (SiO2-PFTS/PVDF) was obtained by dynamically forming 1H,1H,2H,2H-peruorooctyl trichlorosilane
Accepted 26 February 2017
containing SiO2 nanoparticles on the membrane surface as in our previous study; whereas a superhydrophilic
Available online xxxx
membrane (PVA/PVDF) was developed by attaching a thin layer of poly(vinylalcohol) hydrogel onto the mem-
Keywords:
brane's surface. The effects of surface modication on their anti-wetting or anti-fouling properties were exam-
Anti-wetting ined in MD using an aqueous NaCl solution with various organic foulants (e.g., kerosene, humic acid (HA), and
Anti-fouling sodium dodecyl benzene sulfonate (SDBS)). The results showed that the superhydrophobic SiO2-PFTS/PVDF
Superhydrophobic membrane displayed excellent self-cleaning characteristics and wetting resistance against all three studied
Superhydrophilic foulants. In contrast, the hydrophilic surface layer of the PVA/PVDF membrane only slowed down wetting and
Membrane distillation fouling when in contact with kerosene and HA. Nonetheless, when dealing with SDBS, its anti-wettability perfor-
mance was comparable to that of the SiO2-PFTS/PVDF membrane. The superhydrophobic SiO2-PFTS/PVDF mem-
brane exhibited anti-fouling and anti-wetting behaviors even though the extended DerjaguinLandauVerwey
Overbeek theory indicated the attraction force between the membrane surface and all three foulants.
2017 Elsevier B.V. All rights reserved.

1. Introduction
Membrane distillation (MD) is an emerging separation membrane
technology based on the vapor pressure gradient across a porous hydro-
phobic membrane [1]. The advantages of applying MD for desalination
include relatively low operating temperatures, lower operating pres-
sures and fewer requirements on membrane mechanical properties
[2]. With these unique advantages, MD processes have demonstrated
promising results in seawater desalination, wastewater treatment and
many other applications [3].
Corresponding author.
E-mail address: pyl@bjut.edu.cn (Y. Peng).
The issues of membrane fouling and wetting have been frequently
discussed for improving the MD performance. Membrane fouling,
which is dened as the deposition of contaminants (organic and/or in-
organic materials) on the membrane's surface or into the membrane's
pores, decreases the water purication performance of the membrane
and increases operational cost by adding extra processing requirements,
such as periodical membrane cleaning and/or replacement [4]. In partic-
ular, membrane fouling can be accelerated by a hydrophobic interaction
between contaminants and membrane surface [5]. Membrane wetting
includes two steps: surface wetting and partial wetting [6]. An increase
of the permeate conductivity usually suggests the occurrence of partial
wetting [7,8]. Generally, membrane wetting depends on the membrane
materials, the content and type of foulants [9], membrane's pore size
and porosity [10], as well as membrane surface morphology and

http://dx.doi.org/10.1016/j.desal.2017.02.022
0011-9164/ 2017 Elsevier B.V. All rights reserved.
128 X. Lu et al. / Desalination 413 (2017) 127135

structure [11]. Moreover, membrane wetting can be caused by the ad- insights in developing high performance membranes via membrane
hesion of bio-organisms (biofoulants) [12], membrane degradation by modications.
long time operation [13], as well as mechanical and thermal damage
by heating and cooling cycling [8]. 2. Theory
As a popular hydrophobic membrane material, polyvinylidene uo-
ride (PVDF) is widely used in MD [14]. The hydrophobic PVDF mem- 2.1. Concept of surface tension
branes are susceptible to fouling when treating aqueous solutions
containing various organic matters, e.g. oily substances, humic acid, sur- The wettability of a solid surface can be expressed by the contact
factants [7,8,15]. Furthermore, the long-term stability for an enhanced angle, , which is governed by Young's Eq. [7]:
PVDF membrane performance in MD is a signicant issue, so further
membrane modication becomes critical [16]. Therefore, chemical S SL
cos 1
modication of the membrane's surface is necessary to improve anti- L
fouling and anti-wetting properties. There are two major strategies for
hydrophobic surface modication: changing the PVDF's intrinsic hydro- where S, SL and L refer to the surface tensions of solid in contact
phobicity towards either super-hydrophobicity or super-hydrophilicity. with air, solid in contact with liquid, and liquid in contact with air, re-
The superhydrophobic nature of the surface renders the membrane spectively. The surface tension of a solid surface can be written as the
with anti-wetting properties, while retaining high ux. Super hydro- following when the GoodVan Oss model is applied [31]:
phobicity helps to introduce an air gap between the liquid droplet and
the membrane's surface [17], thereby increasing the allowable pore S TOT LW AB
S S 2
size prior to the occurrence of pore wetting, and consequently ensuring
higher mass ux and throughput [18]. Generally, a superhydrophobic q
surface is acquired for the modication with uorosilanized nanoparti- AB
S 2 S S
3
cles [1923]. In a previous work, we have also fabricated amphiphobic
q q q
PVDF composite membranes with excellent anti-wetting and anti-
TOT
L 1 cos 2 LWS L
LW S L
S L 4
fouling properties against various organic foulants such as humic acid
and surfactant [24].
Shifting the PVDF's hydrophobicity towards super-hydrophilicity is a where is the contact angle, and the surface tension parameters
major strategy for hydrophobic surface modication. A pure water layer (LW +
S , S , S ) of a solid can be obtained by determining the contact an-
is easily formed on a highly hydrophilic surface, preventing the adsorp- gles between the solid and three probe liquids with well-known surface
tion and deposition of hydrophobic pollutants onto the membrane's tension properties (LW +
L , L , L );
TOT
is the total surface tension; and
surface [25]. Surface grafting [7] and coating [26] can be effective in im- AB is the AB components of surface tension [32]. In this study, the pris-
proving the hydrophilicity of PVDF membranes. A highly hydrophilic tine and modied membrane surfaces were treated as the solid. The cal-
membrane surface was developed via plasma induced grafting of poly- culation of all components of the surface tension of a polymer (LW S ,
ethylene glycol (PEG) and TiO2 particles deposition onto the PVDF +
S ,S ) was carried out using Eq.(4) for three liquids with known values
membrane surface, and the resulted hydrophilic-modied composite of LW +
L , L , L . The liquid surface tension data of water, diiodomethane
membrane was tolerant to salty feed solution containing mineral oil and glycerol are listed in Table S1.
[7]. Liang et al. [27] combined the plasma treatment with the coating
of silica nanoparticles for the hydrophilic modication of PVDF ultral- 2.2. The XDLVO theory
tration (UF) membranes. The surface modication signicantly im-
proved the wettability and anti-biofouling properties of PVDF The extended Derjaguin-Landau-Verwey-Overbeek (XDLVO) theory
membranes. Fane et al. [28] successfully developed composite mem- describes the total interaction energy, or adhesion energy, between a
branes with a poly (vinylalcohol)/PEG hydrophilic layer on a PVDF sub- colloid and a membrane surface as the sum of Lifshitz-van der Waals
strate for desalination. A similar composite membrane was synthesized (LW), acidbase (AB) and electrostatic double layer (EL) interactions
by attaching a thin layer of agarose hydrogel on the surface of a Teon [33].
membrane, which displayed anti-fouling properties against the studied
surfactant [16]. Du et al. [29] modied a PVDF membrane by coating GXDLVO GLW GEL GAB 5
with a poly (vinylalcohol) (PVA) solution followed by crosslinking.
The treated membrane displayed both a high ux and a slow fouling
where GXDLVO is the total interaction energy between a membrane
rate. Rahimpour et al. [30] immobilized acrylic and amino moieties on
surface and a colloid immersed in water, GLW is the LW interaction
a PVDF membrane surface, the hydrophilicity, antifouling properties,
term, GEL is the EL interaction term, and GAB is the AB interaction
and ux recovery of the composite membrane were all enhanced.
term. In Eq.(8), 0 is the permittivity of free space, r is the relative di-
As above discussion, previous literatures have shown that both hy-
electric constant of water, is the reciprocal of the Debye length, p
drophobic and hydrophilic treatment can contribute to the membrane
and m are Zeta potentials of membrane and foulant respectively, d0 is
anti-fouling, however it is difcult to evaluate various surfaces modi-
the minimum contact distance between foulant and membrane. The
cation methods based on different membrane supports. The aims of
input values used for the XDLVO module are listed in Table S2. And
this study are to compare the anti-fouling and anti-wetting mechanisms
the interfacial free energy for the interaction between solid and liquid
of the PVDF based composite membranes derived from hydrophobic
immersed in an aqueous medium can be further expressed as following
and hydrophilic treatment in same direct-contact membrane distilla-
[26]:
tion (DCMD) process. Humic acid, oily substances and surfactants
were chosen as the organic pollutants in DCMD, representing natural q q q 
organic matter, hydrophobic and amphiprotic pollutants, respectively. GLW 2 S W
LW LW LWL W S L W
LW LW LW LW
6
Since the process of fouling and wetting on membrane surface can be
a slow process, the long-term stability test (N 100 h) is critical to evalu- q q

 p
 q
ate the effectiveness of different modication routes. The XDLVO theory GAB 2 W S L W
W S L W S L S L

was also applied in this study to explore the mechanisms of wetting and
fouling by various organic matters in MD process, providing some 7
 X. Lu et al. / Desalination 413 (2017) 127135
2 3
0 r 2
2 4 2 m p
GEL m p 1 cothd0
cschd0 5 8 put a 5 L droplet on the coated fabric surfaces. The slides were slowly
2 2 2
m p
According to thermodynamic theory, the negative value of GXDLVO
indicates a force of attraction. The lower the value, the stronger the at-
traction between the liquid and the solid surface that one can expect.
In contrast, if the value is positive, there is a repulsive force between
the solid and the liquid. For a situation in which a membrane is in con-
tact with organics-containing water, a high negative value of GXDLVO
would suggest a severe membrane fouling by organics, but a high posi-
tive value would imply less membrane fouling tendency. In order to cal-
culate GXDLVO, the van der Waals, acidbase and electrostatic double
layer interactions should be obtained from Eqs. (6), (7) and (8), after
contact angle measurements.
3. Experimental
3.1. Materials
All the chemicals were of reagent grade and used as received:
1H,1H,2H,2H-peruorooctyl trichlorosilane (CF3(CF2)5(CH2)5SiCl3,
PFTS, K Scientic Ltd., China), SiO2 particles (~30 nm, Maikun Chemical
Co. Ltd., China), Kerosene (3#, 99.9%, Shenzhen Dekang Petrochemical
Products Co. Ltd., China), Humic acid (HA, Sigma-Aldrich, USA), Sodium
chloride (NaCl, Beijing Yili Fine Chemicals Co. Ltd., China), Sodium dode-
cyl benzene sulfonate (SDBS), poly (vinylalcohol) (PVA) and glutaralde-
hyde (Tianjin Fuchen Chemicals Co. Ltd., China), Cyclohexane and
ethanol (Sinopharm Chemical Reagents, China), and PVDF membrane
(IPVH00010, 0.45 m, Millipore Corporation, USA). The deionized
water utilized was produced in our laboratory (RiOs16, Millipore,
USA) with a conductivity of 3.6 S/cm.
3.2. Preparation of superhydrophobic PVDF composite membrane
The surface modication of hydrophobic PVDF membrane was per-
formed according to our previous reported method [22]. The SiO2-
PFTS layer was deposited onto the PVDF surface by ltering at a pressure
of 0.02 MPa. After rinsing with ethanol and pure water, the derived
membrane was dried at 80 C for 5 h and referred as SiO2-PFTS/PVDF.
The detailed procedure can be found in our previous study [24].
3.3. Preparation of superhydrophilic/hydrophobic composite membrane
PVA was dissolved completely in hot deionized water (95 C) at
5.0 wt%. 20 mL of PVA solution (5.0 wt%) and 5 mL of glutaraldehyde so-
lution (1.0 wt%) were mixed with deionized water under stirring for
15 min, followed by the drop-wise addition of 0.5 mL of hydrochloric
acid under stirring at ambient temperature. Then, the mixture was
dip-coated on the PVDF membrane rapidly, so a hydrogel layer was
formed after cooling. Finally, the PVA hydrogel layer-coated PVDF com-
posite membrane was immersed in deionized water before using. The
developed membrane was marked as PVA/PVDF.
3.4. Characterizations
The sessile drop method was used to measure the contact angle (CA)
of three foulants on the membranes, each measurement was taken im-
mediately after placing a droplet from the probe on the membrane. Con-
tact angle measurement was performed using a standard Contact Angle
Goniometer (DSA100, KRUSS, Germany), an instrument that has image
analysis attachments, including video camera, monitor and image-
analysis software. Quick water rolling is another important aspect of
superhydrophobic materials. To evaluate water rolling on the mem-
branes, the sliding angles of water droplets for all the samples were
129
measured. For measuring the sliding angle, we rst placed the fabric
on a glass slide uniformly using a double-side adhesive tape, and then
tilted until the water droplet started to roll down. The Zeta potential
of the membranes was determined using a SurPASS electrokinetic ana-
lyzer (Anton Paar GmbH, Graz, Austria). During the process of measur-
ing the zeta potential, the KCl solution concentration was maintained at
0.83 mmol/L, while the operation pressure was 0.03 MPa. The zeta po-
tentials of kerosene solution, HA solution and SDBS solution were ob-
tained by a laser scattering size analyzer (NICOMP380ZLS, USA). The
membrane surface functionalities were examined by Fourier transform
infrared spectroscopy (FTIR) spectrum, which was collected at room
temperature over a scanning range of 4004000 cm1with a resolution
of 4.0 cm1, using a Bruker Vertex70 spectrometer (USA).
3.5. Direct-contact membrane distillation tests
The DCMD set-up was composed of a feed reservoir, a cold reservoir
and the membrane distillation module [24]. The actual volume of the
feed was 25 L. The effective area of the membranes was 26.4 cm2. Firstly,
The feed uid, with an electrical conductivity of about 50 mS/cm, was
introduced in a polypropylene (PP) vessel with a capacity of 25 L, and
then was circulated by a peristaltic pump (BT600-2J, Longer Pump,
China) at 100 rpm. On the other side of the membrane, the permeate
was cooled down by a cooling circulator and was cycled by another peri-
staltic pump (BT600-2J, Longer Pump, China) at 100 rpm. The electrical
conductivity of the permeate was measured via a conductivity meter
(DDS-11A, Ridao, China). During the experiment, the feed temperature
of the hot side was xed at 76 C, while that of the cold-side was xed at
21 C. The ow rates were 3 L/min for both the hot-side and the cold-
side.
Four kinds of synthetic feed solutions were used in MD. Sodium
chloride solution (3.5 wt%) was chosen to simulate seawater. Kerosene,
SDBS and HA in synthetic seawater (10 mg/L) was used to simulate nat-
ural organic foulants. A detailed procedure can be found in our previous
publication [24]. The permeate was determined using a chemical oxy-
gen demand (COD) rapid tester (5B-1, Lanzhou Lianhua Environmental
Protection Technology Co. Ltd., China), a total organic carbon analyzer
(TOC, V-CSH, Shimadzu, Japan) and a UVvisible spectrophotometer
(UV-2550, Shimadzu, Japan) at a xed period to detect kerosene, SDBS
and HA, respectively.
4. Result and discussion
4.1. Characterization of the modied membrane surface
Fig. 1 compares the FTIR spectra of the pristine PVDF, the SiO2-
PFTS/PVDF and the PVA/PVDF membranes. As for the PVDF mem-
brane (Fig. 1a), the vibration bands at 1170 cm1 and 1204 cm 1
can be assigned to the asymmetrical and symmetrical stretching of
CF2, respectively. The vibration of CH2 groups appeared at the frequency
of 1402 cm1. As shown in our previous study [24], the enhanced asym-
metrical (1170 cm1) and symmetrical (1204 cm1) stretching of CF2
and the vibration of CH2 (1402 cm1) indicates the existence of PFTS
(Fig. 1b). The newly appeared vibration band at 1070.9 cm1 assigns
to the Si-O-Si characteristic peaks, suggesting the successful uorinated
silica coating on the PVDF membrane surface. On the other hand, for the
PVA/PVDF membrane (Fig. 1c), the presence of the bands at
3288.6 cm1 is attributed to OH stretching vibration of the PVA inter-
molecular condensation. The strong peaks at 2918.6 cm1 (the asym-
metrical stretching vibration of CH2) and 1144 cm1 (the stretching
vibration of C\\C groups) relate to the crystalline of PVA. In addition,
the weak CF2 peak (1204 cm1) of the PVA/PVDF membrane also indi-
cates the attachment of hydrophilic PVA coating onto the PVDF mem-
brane surface.
130 X. Lu et al. / Desalination 413 (2017) 127135

4.2. The anti-fouling mechanism of the composite membranes

4.2.1. Surface characteristics of the composite membranes

The wettability of the surface depends on the surface energy and the
roughness of the surface [34]. Table S3 shows the CA values of the pris-
tine and composite membranes. Fig. 2 presents the SEM surface mor-
phologies of three membranes. As can be seen, the pristine PVDF
membrane showed a water contact angle (WCA) of 124.2 and a
diiodomethane contact angle (DCA) of 87.9. In contrast, the modied
SiO2-PFTS/PVDF membrane displayed higher WCA (167.3) and DCA
(140.9) values, which is mainly attributed to the rougher surface struc-
ture [24] and to the lower surface energy of the uoro-silane coating
(Fig. 2c & Fig. S1c). SiO2-PFTS/PVDF membrane also had a lower sliding
angle and the droplets slide easily on the membrane surface displaying
excellent self-cleaning characteristic, what is known as lotus effect [35].
Compared to the poor anti-fouling behavior of pristine PVDF, SiO2-PFTS/
PVDF displays better anti-fouling characteristics for a variety of foulants
like kerosene, HA or SDBS [24]. The surface adsorption of organic
foulants on the pristine PVDF membrane can further accelerate the con-
Fig. 1. ATR-FTIR spectra of PVDF (a), SiO2-PFTS/PVDF (b) and PVA/PVDF (c) membranes tact angle reduction and the formation of a hydrophilic surface, while
surface. the retention of superhydrophobicity by the SiO2-PFTS/PVDF mem-
brane, may be due to its anti-fouling behavior. On the other hand, for

(a) PVDF

(b) PVA/PVDF

(c) SiO2-PFTS/PVDF

Fig. 2. Surface morphologies (left) and the schematic diagrams (right) of water droplet contact with three membrane surfaces.
 X. Lu et al. / Desalination 413 (2017) 127135
the modied PVA/PVDF membrane, the water droplets spread out easily
on the surface due to the superhydrophilicity of PVA hydrogel layer
(Fig. S1b, Table 1 and Fig. 2b). In such case, the wetting of the PVA/
PVDF membrane by foulants with hydrophilic groups is thermodynam-
ically feasible because the foulants are in the PVA hydrogel and closer to
the PVDF support.
Since the surfaces of pristine and superhydrophobic composite
membranes were rough, the contact area ratio of water and membrane
surface (f) was calculated by the Cassie-Baxter equation [36]:
0
cos rfcos f 1
where is the contact angle on a smooth surface, is the contact
angle on a rough surface. In our study, of 80 was used [37], the mea-
sured value of was from Table 1. r is 1.7 for polymer sphere in this
study [36]. Therefore, according to the above equation, the f value of
pristine PVDF and SiO2-PFTS/PVDF membranes were 0.34 and 0.02,
respectively.
4.2.2. The interaction between the composite membranes and foulants
During the DCMD process, the organic foulants may adsorb on the
membrane surface because the feed contacts directly with the mem-
brane. The physicochemical interaction between the foulants and mem-
branes can be described by the interfacial free energy according to the
XDLVO theory.
Three types of liquids (water, diiodomethane and glycerol) have
been used for contact angle measurements on three membrane surfaces
and three foulants. The averaged contact angle values and standard de-
viations are given in Table S3. Compared with the pristine PVDF, the
SiO2-PFTS/PVDF surface shows superhydrophobicity, whereas the
PVA/PVDF surface exhibits very high hydrophilic properties
(Table S3). Table S4 shows the calculated surface tension parameters
of the three membranes. The data shows that the pristine PVDF has a
high electron acceptor component (+) and a zero electron donor com-
ponent (), resulting in a non-polar component of surface tension.
Similar results were also reported in the open literature [38,39]. PVDF
presents a large negative charge of 45.51 mV due to the electronega-
tivity of F atom. In contrast, the SiO2-PFTS/PVDF exhibits lower surface
energy (L, LW +
L , L , L ). As the atomic F content decreases with the ex-
istence of Si and O elements, the surface potential changes to a less neg-
ative value of 36.69 mV. As for the PVA/PVDF, since PVA is a nonionic
polymer and completely covers the membrane surface, the zeta poten-
tial of the membrane surface increases to 67.38 mV. HA and SDBS have a
high electron donor component () and a low electron acceptor com-
ponent (+), indicating that they are polar and likely to be ionized to
donate electron in aqueous media.
With the surface tension data from Table S4, the interfacial free en-
ergy between the membrane surface and the liquid droplet in aqueous
phase can be determined according to the XDLVO theory, and the calcu-
lated results are shown in Table S5 & Fig. 3. Since the interfacial free en-
ergy value (GEL) of the electrostatic double layer is negligible, the
interfacial free energy (GXDLVO) is mainly determined by the interfacial
free energy value of van der Waals (GLW) and polar interactions
(GAB). The PVDF and SiO2-PFTS/PVDF membranes exhibit a negative
GXDLVO with the contact of all three foulants, suggesting the existence
of an attraction force between the membrane surface and each foulant.
Table 1
Static angles and sliding angles of three membranes.
Water Diiodomethane
Membrane Static angle Sliding angle Static angle Sliding angle
() () () ()
PVDF 124.2 1.2 26.0 4.0 87.9 1.1
SiO2-PFTS/PVDF 167.3 0.6 3.0 2.0 140.9 1.2 12.0 3.0
PVA/PVDF 0 64.6 1.6
131
9
Fig. 3. Interfacial free energy between foulants and membrane surfaces.
On the contrary, the PVA/PVDF membrane possesses a positive value of
GXDLVO, implying repulsion forces between the three foulants and PVA
hydrogel. By comparing the GLW and GAB, it was observed that the
acidbase force has a larger signicance in the overall interaction for
PVDF and PVA/PVDF membranes.
4.3. Membrane performance during the DCMD process
4.3.1. Sodium chloride solution
The DCMD test was performed using 3.5 wt% NaCl solution as the
feed. The permeate ux and electrical conductivity of the three mem-
branes were showed in Fig. 4. Because of the mass transfer resistance
from the modication layer, the two modied membranes displayed
slightly lower permeate ux than the pristine PVDF [28,40]. The mean
ux values for PVDF, SiO2-PFTS/PVDF and PVA/PVDF membranes are
30.0, 25.2 and 26.5 kg/(m2 h), respectively. The salt rejection of PVDF,
SiO2-PFTS/PVDF and PVA/PVDF membranes are all over 99.9%, resulting
of a stable permeate electrical conductivity.
4.3.2. Sodium chloride solution containing oil matter
As an oil matter, 10 mg/L kerosene were added in the feed to study
the anti-fouling features of the membranes. The stability of the kerosene
emulsion was examined by oil droplet size analysis every 24 h, showing
no measurable change (~1.56 m) during the whole DCMD test. Fig. 5
shows the time-dependent permeate ux and conductivity of the pris-
tine PVDF and modied membranes. As shown in our previous work
[24], the conductivity increased sharply within 20 h when testing the
pristine PVDF membrane. The test was purposely ended after 20 h due
to the low desalination efciency and the extremely high conductivity
(1200 S/cm) in the permeate solution. It suggested that the pristine
PVDF membrane was wetted by kerosene, resulting in increasing salt
penetration rate through the membrane. The wetting phenomenon
can be explained by a two-step fouling process [7]. Firstly, the oil drop-
lets spontaneously attached on the PVDF membrane surface because of
the large negative interfacial free energy (G = 32.38 mJ/m2), mainly
arising for the acidbase interaction (GAB = 29.05 mJ/m2). Secondly,
the oil droplets fully covered the membrane surface after the DCMD op-
eration. Based on Table S5, both the van der Waals interactions and the
acidbase interactions contribute to the oil fouling.
The PVA/PVDF membrane showed stable permeate ux in the rst
40 h. However, the ux began to decrease slowly after 40 h. The conduc-
tivity of the PVA/PVDF membrane started to increase after 60 h owing to
the salt break through. Similarly modied hydrophilic/hydrophobic
membranes in the references presented a stable ux over 24 h of oper-
ation without fouling and wetting due to combined effects of enhanced
surface hydrophilicity and reduced membrane pore size [12]. It is note-
worthy that a long-term DCMD test is necessary to evaluate the stability
132 X. Lu et al. / Desalination 413 (2017) 127135
Fig. 4. The permeate ux (a) and conductivity (b) of PVDF, PVA/PVDF and SiO2-PFTS/PVDF
membranes by feeding 3.5 wt% NaCl solution.
because partial wetting is actually a slow process. The negative GLW
and positive GAB (Table S5) values indicated that the van der Waals
forces imposed an attraction between the kerosene and the membrane,
but a repulsive force by the acidbase forces. The combined total effect
can be repulsive though, resulting in less kerosene adsorption and stable
permeate conductivity in the rst 40 h while the PVA layer can act as a
barrier. However, the sublayer of the PVA/PVDF membrane was still
gradually wetted after 60 h when increasing the kerosene concentration
on the feed/membrane interface, thereby deteriorating salt rejection
(Fig. 5).
Compared with the pristine PVDF and the PVA/PVDF membranes,
the SiO2-PFTS/PVDF membrane showed a more stable ux in 200 h
(Fig. 5). Salt rejection remained at 99.8% and the conductivity change
was negligible [24]. Although the negative interfacial free energy
(G = 48.63 mJ/m2) between SiO2-PFTS and kerosene suggests that
SiO2-PFTS tends to adsorb kerosene, the oil/water droplets slides away
from the SiO2-PFTS/PVDF surface before contacting due to its slow
water sliding angle of 3.0 2.0 (Table 1). The f value of the SiO2-
PFTS/PVDF surface was only 0.02, far lower than 0.34 of the pristine
PVDF. After performing the DCMD test, the characteristic peaks of kero-
sene at 30002850 cm1 were observed on both the pristine PVDF and
PVA/PVDF membranes, but not on the SiO2-PFTS/PVDF surfaces
Fig. 5. Dynamic permeate ux (a) and conductivity (b) of the pristine PVDF, PVA/PVDF and
SiO2-PFTS/PVDF membranes by feeding 3.5 wt% NaCl solution with 10 mg/L kerosene.
(Fig. S2). Therefore, the contact between SiO2-PFTS/PVDF and kerosene
displayed signicant anti-wetting and anti-fouling properties during
the MD process.
4.3.3. Sodium chloride solution containing humic acid
Humic acid, a natural organic compound, widely exists in most
wastewater causing membrane fouling or wetting [41] owing to its ad-
sorption onto hydrophobic surfaces [42]. With the existence of HA in the
NaCl solution, the permeate ux and conductivity of the pristine PVDF
and modied membranes are presented in Fig. 6.
As shown before [24], the permeate ux of pristine PVDF membrane
started to decrease gradually after 40 h, reaching approximately 73.4%
of the initial ux at 100 h. Compared to kerosene fouling, the partial
wetting by HA was slower (Fig. 6).
Although there is a higher adsorption of HA (G = 43.43 mJ/m2)
than of kerosene (G = 32.38 mJ/m2), but a HA cake (Fig. S4a) slows
down HA breakthrough. As displayed in Fig. 6b, the permeate conduc-
tivity remained below 100 S/cm for the rst 40 h, then dramatically in-
creased to 1143 S/cm with the reduction of salt rejection to 97.7% at
100 h. The existence of HA in the permeate solution indicates the pene-
tration of HA through the membrane (Table S6), which is in line with
the salt breakthrough in the references [8,24].
For the PVA/PVDF membrane, the permeate ux decreased signi-
cantly during the DCMD test, and it only maintained approximately
 X. Lu et al. / Desalination 413 (2017) 127135 133

dendritic-network structures of membrane surface, which greatly re-

duces the contact area between liquid-solid interface like the lotus leaf
(f = 0.02) and to the low surface energy of the SiO2-PFTS/PVDF modi-
ed surface. It is noteworthy that no HA was detected in the permeate
solution (Table S6), indicating no HA penetration. Even though a
brown HA deposited layer formed on the membrane's surface due to
concentration polarization, the HA gel was easily washed out and bro-
ken down into discrete partitions (Fig. S4c).

4.3.4. Sodium chloride solution containing surfactant

Fig. 7 demonstrates the time-dependent permeate ux and conduc-
tivity of the pristine and modied membranes during the DCMD process
in the presence of an amphiprotic surfactant (SDBS). As reported previ-
ously [24], after adding 10 mg/L SDBS into the feed, the conductivity in-
crement and permeate ux reduction of the pristine PVDF indicated
that the membrane was partly wetted by SDBS(Fig. S5c). In contrast to
this, only a slight permeability reduction and conductivity increment
were observed for the two modied membranes during a 100-h test.
A 99.9% of salt rejection indicated that there was no NaCl penetration
across SiO2-PFTS/PVDF and PVA/PVDF membranes (Table S6).
As shown in Table S5 & Fig. 3, the negative interaction energy be-
tween PVDF and SDBS (G = 70.82 mJ/m2) indicates a strong attrac-
tion force. After SiO2-PFTS modication, the attraction force was

Fig. 6. Dynamic permeate ux (a) and conductivity (b) of the pristine PVDF, PVA/PVDF and
SiO2-PFTS/PVDF membrane by feeding 3.5 wt% NaCl solution with 10 mg/L HA.

36.1% of its initial ux after 140 h, with a permeate conductivity in-

creased to 1117 S/cm. HA penetrated through the PVA layer and
PVDF sublayer and was detected in the permeate after 48 h (Table S6).
Compared with the pristine PVDF, a less pronounced conductivity incre-
ment and a lower HA concentration in the permeate indicate that the
PVA layer only slows down the penetration of HA and NaCl. Based on
the negative GLW and positive GAB (Table S5), HA is subjected to a re-
pulsive force of the PVA layer (G = 18.32 mJ/m2), resulting in less HA
adsorption (Fig. S3b). However, HA accumulated on the feed/mem-
brane interface due to severe concentration polarization and gradually
formed a brown cake (Fig. S4b), leading to pore blockage [14] and de-
crease of permeate ux (Fig. 6). Due to the water solubility of HA, HA
can penetrate the PVA layer, and then adsorb on the PVDF support
and wett the membrane pores, leading to the penetration of HA and
NaCl through membrane. An even higher repulsive force from hydro-
philic coating is needed to resist the HA breakthrough.
For the SiO2-PFTS/PVDF membrane, owing to the self-cleaning abili-
ty of the uoro-silane layer on uorinated SiO2, a stable permeate ux
with high salt rejection (99.9%) was observed [24]. Though the attrac-
tion force existing between the SiO2-PFTS layer and HA
(G = 35.21 mJ/m2) is of importance, the anti-wetting and anti- Fig. 7. Dynamic permeate ux (a) and conductivity (b) of PVDF, PVA/PVDF and SiO2-PFTS/
fouling properties of the membrane can be mostly attributed to the PVDF membranes by feeding 3.5 wt% NaCl solution with 10 mg/L SDBS.
134 X. Lu et al. / Desalination 413 (2017) 127135
signicantly reduced (G = 2.07 mJ/m2). Under quase-neutrality, the
self-cleaning features of the SiO2-PFTS coating increased the resistance
of SDBS adsorption and wetting. On the other hand, PVA modication
introduced a strong repulsive force between the PVA layer and SDBS
(G = 54.75 mJ/m2). Thus, the hydrophilic PVA layer could effectively
prevent SDBS adsorption and wetting. A similar effect was also observed
on a hydrophobic porous Teon modied by a thin hydrogel during 24 h
of MD processing [16]. Overall, both the self-cleaning character of the
SiO2-PFTS/PVDF and the strong repulsive force of the PVA/PVDF were
effective in MD for the surfactant foulant.
Among the three membranes (Figs. 57), the SiO2-PFTS/PVDF mem-
brane displays the highest wetting resistance against all three typical
foulants. The hydrophilic PVA hydrogel of PVA/PVDF membrane, on
the other hand, can effectively prevent wetting by SDBS, but only
slows down wetting and fouling by kerosene and HA.
In synthesis, when in contact with a NaCl solution containing hydro-
carbons, organic acids and surfactants, membrane fouling and wetting
processes turned out to be different for the three studied membranes.
If the organic foulant adsorbs onto the membrane's surface, the hydro-
phobicity of membranes will be reduced and wetting cannot be avoided.
In such a case, NaCl and other soluble organic matter in the feed are like-
ly to permeate across the membrane. Schematic representations of the
(a) PVDF
(b) PVA/PVDF
(c) SiO2-PFTS/PVDF
Fig. 8. Schematic diagram of anti-fouling mechanism in PVDF, PVA/PVDF and SiO2-PFTS/PVDF membranes.
transport mechanisms for the different studied organic substances pass-
ing through the membranes with hydrophobic, superhydrophobic and
superhydrophilic surfaces are shown in Fig. 8, respectively.
For the hydrophobic pristine PVDF membrane (Fig. 8a), the organic
matter (HA), the hydrocarbon (kerosene), and the surfactant (SDBS)
rst adsorb onto the membrane's surface and then diffuse into the
membrane bulk, causing membrane wetting. As to the absorption of
surfactant, the hydrophobic moieties of the surfactant are preferentially
adsorbed on the membrane's surface with the hydrophilic moieties fac-
ing towards the water phase. Thus, the hydrophobic membrane surface
gradually becomes a hydrophilic surface, allowing the penetration of
foulants.
For the superhydrophilic PVA/PVDF membrane (Fig. 8b), the PVA/
PVDF membrane exhibited strong repulsion for the three kinds of sub-
stances, making the contact between such and the membrane's surface
more difcult. However, the soluble organic matter (HA) can penetrate
across the PVA layer and come in contact with PVDF support, leading to
membrane wetting and fouling. In contrast, the repulsion force between
the PVA layer and kerosene is very small, but still causes the contact of
kerosene with the PVDF support and the consequential membrane wet-
ting/fouling. Unlike in the hydrophobic PVDF membrane, the hydrophil-
ic moieties of SDBS preferentially adsorb on the PVA layer with the
 X. Lu et al. / Desalination 413 (2017) 127135
hydrophobic moieties in the water phase, preventing the SDBS contact
with the PVDF support, and showing a better anti-fouling resistance as
a result.
For the superhydrophobic SiO2-PFTS/PVDF membrane, as depicted
in Fig. 8c, because of its good self-cleaning properties, water and all
the foulants in the feed nd great difculty in making contact with the
membrane's surface. The exhibited non-selectivity, and excellent anti-
fouling and anti-wetting performance suggest that SiO2-PFTS modica-
tion of PVDF membranes can be a potentially effective method to fabri-
cate high performance membranes for MD.
5. Conclusions
Two surface modication routes have been explored on PVDF mem-
branes to obtain superhydrophobic and superhydrophilic surfaces. By
dynamically forming peruorooctyltrichloro silane (PFTS) coated SiO2
nanoparticles onto the PVDF substrate, the static contact angle (167.3
for water, 140.9 for diiodomethane) and the low sliding contact angle
(3 for water, 12 for diiodomethane) were signicantly increased. On
the other hand, a superhydrophilic membrane's surface was obtained
by attaching a thin layer of PVA hydrogel onto the PVDF surface.
When treating industrial wastewater (mainly containing inorganic elec-
trolytes, oily matter, organic acid and surfactants) by the membrane dis-
tillation process, rapid wetting and fouling of the pristine PVDF
membrane was observed. A PVA coated membrane displayed an effec-
tive anti-fouling ability for SDBS (a surfactant), but only slowed down
the fouling process for kerosene (a typical oil) and humic acid (organic
acid). In contrast, a superhydrophobic surface derived from the SiO2-
PFTS modication exhibited excellent self-cleaning properties and wet-
ting resistance against all three organic foulants. Moreover, the salt re-
jection rate remained unchanged after over 100 h of membrane
distillation. Therefore, the SiO2-PFTS surface modication can be an ef-
fective method to solve fouling problems for a variety of existing MD
membranes.
Acknowledgments
The authors thank the National Natural Science Foundation of China
for the nancial support (21576004).
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.desal.2017.02.022.
References
[1] K.W. Lawson, D.R. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997) 125.
[2] L.M. Camacho, L. Dume, J. Zhang, J.D. Li, M. Duke, J. Gomez, S. Gray, Advances in
membrane distillation for water desalination and purication applications, WaterSA
5 (2013) 94196.
[3] E. Curcio, E. Drioli, Membrane distillation and related operations - a review, Sep.
Purif. Rev. 34 (2005) 3586.
[4] A. Drews, Membrane fouling in membrane bioreactors-characterisation, contradic-
tions, cause and cures, J. Membr. Sci. 363 (2010) 128.
[5] Y.F. Yang, Y. Li, Q.L. Li, L.S. Wan, Z.K. Xu, Surface hydrophilization of microporous
polypropylene membrane by grafting zwitterionic polymer for anti-biofouling, J.
Membr. Sci. 362 (2010) 255264.
[6] M. Gryta, Wettability of polypropylene capillary membranes duringthe membrane
distillation process, Chem. Pap. 66 (2012) 9298.
[7] G. Zuo, R. Wang, Novel membrane surface modication to enhance anti-oil fouling
property for membrane distillation application, J. Membr. Sci. 447 (2013) 2635.
[8] S. Meng, Y. Ye, J. Mansouri, V. Chen, Fouling and crystallization behaviour of
Superhydrophobic nano-composite PVDF membranes in direct contact membrane
distillation, J. Membr. Sci. (2014) 102112.
[9] A.C.M. Franken, J.A.M. Nolten, M.H.V. Mulder, D. Bargeman, C.A. Smolders, Wetting
criteria for the applicability of membrane distillation, J. Membr. Sci. 33 (1987)
315328.
135
[10] R.B. Saffarani, B. Mansoor, R. Thomas, H.A. Arafat, Effect of temperature dependent
microstructure evolution on pore wetting in PTFE membranes under membrane
distillation conditions, J. Membr. Sci. 429 (2013) 282294.
[11] E. Guillen-Burrieza, A. Servi, B.S. Laliaa, H.A. Arafat, Membrane structure and surface
morphology impact on the wetting of MD membranes, J. Membr. Sci. 483 (2015)
94103.
[12] A. Burgoyne, M.M. Vahdati, Direct contact membrane distillation, Sep. Sci. Technol.
35 (2000) 12571284.
[13] M. Gryta, J. Grzechulska-Damszel, A. Markowska, K. Karakulski, The inuence of
polypropylene degradation on the membrane wettability during membrane distilla-
tion, J. Membr. Sci. 326 (2009) 493502.
[14] F. Liu, N.A. Hashim, Y.T. Liu, M.R.M. Abed, K. Li, Progress in the production and mod-
ication of PVDF membranes, J. Membr. Sci. 375 (2011) 127.
[15] Po-Ju Lin, Ming-Chien Yang, Yu-Ling Li, Prevention of surfactant wetting with aga-
rose hydrogel layer for direct contact membrane distillation used in dyeing waste-
water treatment, J. Membr. Sci. 475 (2015) 511520.
[16] M. Gryta, M. Barance wicz, Inuence of morphology of PVDF capillary membranes
on the performance of direct contact membrane distillation, J. Membr. Sci. 358
(2010) 158167.
[17] A. Lafuma, D. Quere, Superhydrophobic states, Nat. Mater. 2 (2003) 457460.
[18] Z. Ma, Y. Hong, L. Ma, M. Su, Superhydrophobic membranes with ordered arrays of
Nano spiked micro channels for water desalination, Langmuir 25 (2009)
54465450.
[19] H. Zhang, R. Lamb, J. Lewis, Engineering nanoscale roughness on hydrophobic
surface-preliminary assessment of fouling behaviour, Sci. Technol. Adv. Mater. 6
(236) (2005).
[20] S. Meng, J. Mansouri, Y. Ye, V. Chen, Effect of templating agents on the properties
and membrane distillation performance of TiO2-coated PVDF membranes, J.
Membr. Sci. 450 (2014) 4859.
[21] A. Razmjou, E. Arin, G. Dong, J. Mansouri, V. Chen, Superhydrophobic modication
of TiO2 nanocomposite PVDF membranes for applications in membrane distillation,
J. Membr. Sci. 415416 (2012) 850863.
[22] Z. Liu, H. Wang, Superhydrophobic poly (vinylideneuoride) membranes with con-
trollable structure and tunable wettability prepared by one-step electrospinning,
Polymer 82 (2016) 105113.
[23] Y. Liao, R. Wang, A.G. Fane, Engineering superhydrophobic surface on
poly(vinylideneuoride) nanober membranes for direct contact membrane distil-
lation, J. Membr. Sci. 440 (2013) 7787.
[24] X. Lu, Y. Peng, L. Ge, R. Lin, Z. Zhu, S. Liu, Amphiphobic PVDF composite membranes
for anti-fouling direct contact membrane distillation, J. Membr. Sci. 505 (2016)
6169.
[25] G.D. Kang, Y.M. Cao, Development of antifouling reverse osmosis membranes for
water treatment: a review, Water Res. 46 (2012) 584600.
[26] B.A. Jonathan, A.E. Childress, Assessing short-range membranecolloid interactions
using surface energetics, J. Membr. Sci. 203 (2002) 257273.
[27] K. Suhan, M. Mark, J.B. Heon, H.M.V. Eric, Crossow membrane ltration of
interacting nanoparticle suspensions, J. Membr. Sci. 284 (2006) 361372.
[28] P. Peng, A.G. Fane, X. Li, Desalination by membrane distillation adopting a hydro-
philic membrane, Desalination 173 (2005) 4554.
[29] J.R. Du, S. Peldszus, P.M. Huck, X. Feng, Modication of poly(vinylideneuoride) ul-
traltration membranes with poly(vinyl alcohol) for fouling control in drinking
water treatment, Water Res. 43 (2009) 45594568.
[30] A. Rahimpour, S.S. Madaeni, S. Zereshki, Y. Mansourpanah, Preparation and charac-
terization of modied nano-porous PVDF membrane with high antifouling property
using UV photo-grafting, Appl. Surf. Sci. 255 (2009) 74557461.
[31] C.J. van Oss, Acidbase interfacial interactions in aqueous media, Colloids Surf. A
Physicochem. Eng. Asp. 78 (1993) 149.
[32] K. Suhan, M. Mark, J.B. Heon, H.M.V. Eric, Crossow membrane ltration of
interacting nanoparticle suspensions, J. Membr. Sci. 284 (2006) 361372.
[33] B.A. Jonathan, A.E. Childress, Assessing short-range membranecolloid interactions
using surface energetics, J. Membr. Sci. 203 (2002) 257273.
[34] G. McHale, N.J. Shirtcliffe, M.I. Newton, Super-hydrophobic and super-
wettingsurfaces: analytical potential, Analyst 129 (2004) 284287.
[35] L. Feng, Y. Zhang, J. Xi, Y. Zhu, N. Wang, F. Xia, L. Jiang, Petal effect: a
superhydrophobic state with high adhesive force, Langmuir 24 (2008) 41144119.
[36] S. Tan, Q. Xie, X. Lu, N. Zhao, X. Zhang, J. Xu, One step preparation of
superhydrophobic polymeric surface with polystyrene under ambient atmosphere,
J. Colloid Interface Sci. 322 (2008) 15.
[37] X.X. Shen, W.F. Yang, Q.F. Zhang, Hydrophobic modication of PVDF, Mater. Sci.
Forum 687 (2011) 658661.
[38] F. Liu, N.A. Hashim, Y. Liu, M.R.M. Abed, et al., Progress in the production and mod-
ication of PVDF membranes, J. Membr. Sci. 375 (2011) 127.
[39] N. Subhi, A.R.D. Verliefde, V. Chen, P. Le-Clech, Assessment of physico chemical in-
teractions in hollow bre ultraltration membrane by contact angle analysis, J.
Membr. Sci. 403404 (2012) 3240.
[40] J. Zhang, Z. Song, Fabrication and characterization of superhydrophobic poly (vinyl-
idene uoride) membrane for direct contact membrane distillation, Desalination
324 (2013) 19.
[41] S. Goh, J. Zhang, Y. Liu, A.G. Fane, Fouling and wetting in membrane distillation and
MD-bioreactor (MDBR) for wastewater reclamation, Desalination 323 (2013)
3947.
[42] M. Khayet, A. Velzquez, J.I. Mengual, Direct contact membrane distillation of humic
acid solutions, J. Membr. Sci. 240 (2004) 123128.