Journal of Membrane Science 423–424 (2012) 11–19

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science

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

Synthesis and fabrication of nanostructured hydrophobic polyazole

membranes for low-energy water recovery
Husnul Maab 1, Lijo Francis 1, Ahmad Al-saadi, Cyril Aubry, Noreddine Ghaffour,
Gary Amy, Suzana P. Nunes n
Water Desalination and Reuse Center, King Abdullah University of Science and Technology (KAUST), 23955-6900 Thuwal, Saudi Arabia

a r t i c l e i n f o abstract

Article history: Aromatic fluorinated polyoxadiazoles (F-POD) and polytriazoles (F-PT) were synthesized and for the
Received 6 April 2012 first time manufactured into porous membranes by phase inversion and by electrospinning. The phase
Received in revised form inversion F-POD membranes had a mean flow pore size (MFP) of 51 nm, while for F-PT it was around
29 June 2012
74 nm. The electrospun membranes had a much larger pore size, the MFP for F-POD membrane was
Accepted 13 July 2012
Available online 20 August 2012
around 1.7 mm and for F-PT it was 2.7 mm. The membranes were tested for desalination of Red Sea
water using direct contact membrane distillation (DCMD). By combining the high polymer hydro-
Keywords: phobicity and high porosity, apparent contact angles up to 1621 were obtained, assuring the operation
Polyoxadiazole with practically no liquid water leakage under pressure up to 0.9 bar. Salt selectivity as high as 99.95%
Polytriazole
and water fluxes as high as 85 L m  2 h  1 were demonstrated, operating at 80 1C feed temperature and
Membranes
22 1C permeate.
Membrane distillation
& 2012 Elsevier B.V. All rights reserved.

1. Introduction simplicity of the membrane is an advantage. Membranes for MD can

Seawater desalination by reverse osmosis and thermal pro-
cesses is widely used in Middle East countries. However, even in
countries with oil abundance it is clear that the energy consumption
still has to be substantially reduced. In remote regions with water
scarcity the exploration and reuse of water from other sources have
to be considered. The combination of more traditional desalination
methods and less explored technologies like membrane distillation
(MD) could be an effective option. MD can operate at low-grade
thermal or solar heat to treat brackish water, non-condensable
volatile organic compounds and/or, together with other separation
and purification process, to produce fresh water. The principle of MD
has been reviewed in detailed in different papers [1–3], but in
summary saline or wastewater is kept in contact with a porous
hydrophobic membrane and heated to a temperature below the
water boiling point. In the direct contact membrane distillation, cold
water circulates on the other side of the membrane. The partial
vapor pressure difference aroused due to the temperature difference
across the membrane is the driving force for water vapor transport.
The operation at relatively low temperatures (30–80 1C) makes the
process adequate to use low temperature heat sources, including
waste or solar heat. When compared to RO or electrodialysis, the
n
Corresponding author. Tel.: þ966 544700052.
E-mail address: Suzana.nunes@kaust.edu.sa (S.P. Nunes).
1
Authors contributed equally to this research project.
be manufactured in one step with pores in the ultrafiltration or
microfiltration range. They do not need the selectivity of an
additional highly selective layer characteristic of thin-film compo-
site membranes. On the other hand MD membranes need to be
highly hydrophobic (large water–membrane contact angle or high
liquid entry pressure (LEP)) to hinder the transport of liquid water as
depicted in Fig. 1. To promote the water vapor transport through the
membrane pores at acceptable rates, the porosity has to be as high
as possible. Low pore tortuosity and low membrane thickness
minimize the resistance to vapor transport. Materials with low
thermal conductivity as well as thicker membranes help to keep
high temperature differences between both sides of the membrane.
High temperature stability is required to avoid any change during
operation at temperatures close to 100 1C.
Most of the published work uses commercial porous hydro-
phobic membranes based on polypropylene (PP) (e.g. Celgards,
Accurels) and polytetrafluoroethylene (PTFE) (e.g. Gores) [2].
Typical pore diameters are 0.1–1 mm. Both PP and PTFE are hardly
soluble in common solvents at lab temperature. Gores and
Celgards membranes are available as flat sheets and their pre-
paration usually involves extrusion, followed by stretching. These
procedures lead to the formation of membranes with rather large
pores and low porosity [4]. Thermally induced phase separation
(TIPS) and extrusion with expandable/foaming gas are other
methods used for PP membranes. Although PP and PTFE
address the need for high hydrophobicity, their membranes
cannot be manufactured by conventional phase inversion at lab

0376-7388/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.memsci.2012.07.009
12 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19
Fig. 1. Membrane distillation.
temperature due to solubility restrictions. Other more soluble
polymers need to be considered for manufacturing membranes by
phase inversion. Taking this in consideration, the membranes for
MD fabricated by other groups worldwide are in great part based
on poly (vinylidene fluoride) (PVDF) [2,5–7] and modified poly-
ethersulfone (PES) or polyetherimide (PEI) [8]. The first reported
MD membranes were based on modified cellulose acetate
(CA) [9]. All of these polymers can be manufactured into mem-
branes from casting solutions both as flat sheet and hollow fiber
with high porosity using conventional machines. However all the
mentioned polymers are much less hydrophobic than PTFE,
leading to partial surface wetting. Membrane modification has
been proposed by different groups, starting with radiation graft-
ing of polystyrene on CA membrane [9] and addition of fluori-
nated ‘‘surface modifying macromolecules’’ [8] to PES casting
solutions, which migrate to the air/polymer interface during
casting. PVDF still has a higher surface energy (30.3 mN/m) than
PTFE (18.5 mN/m) and the introduction of self-synthesized fluori-
nated silica particles to PVDF hollow fibers during spinning has
been reported in an attempt to improve hydrophobicity [6],
leading to permeation flux as high as 83 L m  2 h  1.
Among other novel membranes are ceramic membranes [10]
modified with perfluoroalkylsilanes [11], membranes with carbon
nanotubes [12,13] and super hydrophobic glasses [14]. Beside the
flat sheet and hollow fiber membranes prepared by phase inver-
sion, new methods of manufacture like electrospinning have been
recently proposed [15] as a possibility to increase porosity and
vapor flux. Electrospinning requires polymer solutions with
optimized viscosity for the membrane manufacture. While the
focus on membrane distillation has increased in the last years, not
much attention has been given to the synthesis of new materials,
particularly tailored for application in this technology. We believe
that, for the full optimization of MD systems, innovations in all
different steps are needed, starting from design and synthesis of
polymers, which would be the best suitable for MD, and going
further to choose the most effective membrane manufacture and
testing under different operational conditions.
High hydrophobicity, low heat transfer, high thermal stability,
and high dimension stability at operation temperature and max-
imum porosity are the important targeted properties for membrane
materials and membranes. We chose a class of materials widely
explored in our group for other membrane applications: polyoxadia-
zoles and polytriazoles [16–21]. They are known for their high
temperature stability and their synthesis has been optimized to
supply high molecular weight polymers with the best film-forming
characteristics for membranes. While poly (phenylene oxadiazole)
[22] is hardly soluble, enabling film casting only from strongly acid
solution, the substitution of the phenylene group by diphenyl ether
or diphenyl hexafluoroisopropylidene increases the solubility in
dimethylformamide and dimethylacetamide, facilitating membrane
casting by phase inversion and electrospinning. Particularly when
fluorinated groups, like diphenyl hexafluoroisopropylidene, are cho-
sen for polyoxadiazole or polytriazole synthesis, highly hydrophobic
products are expected.
A new class of fluorinated polytriazole has been synthesized for
the first time in this study. Polytriazole or polyoxadizole membranes
have not yet been considered for MD process. A series of aromatic
polyoxadiazole and polytriazole polymers (fluorinated and non-
fluorinated) have been synthesized and manufactured into porous
membranes by phase inversion and electrospinning. Synthesized
polymers and fabricated membranes were characterized and tested
in the MD process for the desalination of seawater. Their perfor-
mance in membrane distillation operation was evaluated with very
promising perspectives.
2. Experimental
2.1. Materials
4,40 -(Hexafluoroisopropylidene)bis(benzoic acid) (CAS 1171-47-
7) (HFA) (99% Aldrich), hydrazine sulphate (HS) (CAS 10034-93-2)
(98% Aldrich), N-methyl pyrrolidone (NMP) (99% Aldrich), N,
N-dimethylformamide (DMF), polyphosphoric acid (PPA) (ca 84%
as phosphorous pentaoxide, Alfa Aesar), tetrahydrofurane (THF),
gamma butyrolactone, isopropanol and aniline from Sigma Aldrich
were used as received. PVDF was purchased from Alfa Aesar.
2.2. Synthesis of polyoxadiazole and polytriazole
The synthesis of polyoxadiazole and polytriazole was based on
a procedure reported by our group before for related polymers
[16–18]. For the synthesis of polyoxadiazole, first polyphosphoric
acid (PPA) was added to a 250 mL three necked flask and was
heated at temperature to 100 1C for 1 h under mechanical stirring
along with the passage of dry N2 gas to ensure the removal of
water contents. Hydrazine sulphate (HS) was added to the PPA
and in the meantime the temperature was raised to 160 1C and
the reaction mixture was vigorously stirred to get a homogeneous
solution till it dissolved completely. Once the desired temperature
was attained, then HFA was added and the reaction constitutes
were stirred for about 3 h under dry N2 purge. The molar dilution
ratio of (PPA/HS) and molar monomer ratio of (HS/HFA) was kept
as 10 and 1.2, respectively. The final viscous polymer solution was
poured into cold 1 M NaOH solution under constant stirring for
1 h. The precipitated polymer was transferred into de-ionized
water and was stirred overnight at 80 1C. Finally it was dried in
vacuum oven at 100 1C for 24 h.
For the preparation of polytriazole, aniline was added to the
system 1 h after the addition of HFA at high temperature (180 1C).
The final product is a dark blue colored polymer, the hydrophobic
fluorinated polytriazole. Scheme 1 represents the synthesis routes for
fluorinated (a) polyoxadiazole and (b) polytriazole used in this study.
2.3. Preparation of porous membranes by phase inversion
18 wt% (w/w) polymer solutions in NMP were prepared at
80 1C with constant stirring for 24 h. The polymer solution was cast
with doctor blade on a glass plate (100 mm thick solution) or on
a polyester non-woven support and immersed in water for 1 h.
 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19
Colorless flat sheet porous fluorinated polyoxadiazole membranes
and dark blue fluorinated polytriazole membranes were obtained.
All membranes were dried at room temperature and characterized.
2.4. Preparation of membranes by electrospinning
18% (w/w) polymer solutions were prepared in DMF and added
to a 10 mL glass syringe with a needle tip diameter of 0.5 mm. The
syringe was mounted on the syringe pump in the electrospinning
instrument (Esparyer ES 2000). The feeding rate of polymer solution
was 5 mL/min at room temperature, the electrospining voltage
applied to the needle tip was 30 kV and the distance between the
needle tip and collector was adjusted to be 20 cm. A fast jet of
polymeric solution was ejected from the needle tip in the form of
nanofibers and the fibers were collected on an Aluminum foil as
well as on a non-woven polyester support.
2.5. Polymer characterizations
1
The polymers were characterized by H-NMR and FTIR spectro-
scopy. NMR spectra were obtained in deuterated DMF in a Bruker
Advance III-700 MHz liquid NMR instrument with TCI cryprobe
(Z-axis gradient) model at 25 1C. FTIR spectra were obtained on a
Scheme 1. Synthesis of fluorinated polyoxadiazole and polytriazole.
Fig. 2. Bench scale setup for direct contact membrane distillation (DCMD).
13
Perkin Elmer Spectrum 100 FTIR spectrometer equipped with uni-
versal attenuated total reflectance (ATR) sampling accessory. All
spectra were acquired at lab temperature from 4000 to 550 cm  1
in N2 atmosphere (four scans with spectral resolution of 2 cm  1). Gel
permeation chromatography (GPC), was used for the subject char-
acterization by Agilent 1200 series with RI and UV detectors and the
instrument equipped with column PLgel olexis (7.5 mm  300 mm).
The instrument was calibrated with polystyrene standards (Merck).
THF was used as carrier. 1 mg/mL polymer solutions were injected at
flow rate of 1 mL/min. Thermogravimetric analysis (TGA) was
performed from 30 1C to 700 1C, under argon with heating rate of
0.5 K/min on a Netzsch 209 instrument. The stress–strain curves
were measured on a TA Q800 dynamic mechanical analysis (DMA)
instrument.
2.6. Membrane morphology, contact angle and pore size distribution
The membrane surface and cross-section morphologies were
investigated by field emission scanning electron microscopy
(FESEM) on a Nova Nano FEI microscope. The samples were sputter
coated with Gold using a K575X Emitech equipment. For cross
section the membranes were fractured in liquid nitrogen before. The
hydrophobicity of the membranes was determined by measuring
the water–membrane contact angles, by static drop technique with
a Kruss Scientific goniometry. The Liquid Entry Pressure (LEP) [23],
mean flow pore size (MFP) and pore size distribution were mea-
sured by using a capillary flow porometer, PoroluxTM 1000, after
wetting the membrane sample with perfluoroether (Porefil).
2.7. Membrane distillation
Fig. 2 shows the setup used for the direct contact membrane
distillation (DCMD) experiments. It consists of two parts: part one
is connected to the heating system for controlling the feed water
temperature, while the part two is connected to a chiller (cooling
loop) to condense the permeate vapors. Red Sea water was used
as hot feed solution and was circulated from the bottom side
(selective face) of the membrane, while the permeate vapor was
collected from the other side along with the cold water stream.
The effective membrane area used in all the experiments was
0.0025 m2. The flow rate at feed and permeate side was main-
tained at 1 L/min in the steady state. The salt concentrations of
feed and permeate sides was determined by conductivity meter
(Oakton Eutech Instruments).
14 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19

3. Results and discussion (Co-POD) and poly (diphenyl hexafluoroisopropylidene triazole) (F-
PT). Intense stretching peaks (3500–2500 cm  1) assigned to –OH
3.1. Polymer synthesis and characterization groups of the carboxylic acids and the carbonyl-stretching peak

The following polymers were synthesized according to the

routes depicted in Scheme 1:

 poly (diphenyl hexafluoroisopropylidene oxadiazole) (F-POD),

 poly (diphenyl hexafluoroisopropylidene triazole) (F-PT), and
 poly (diphenyl hexafluoroisopropylidene oxadiazole-co-
diphenylether oxadiazole) (Co-POD).

Fig. 3 shows the 1H NMR spectra of 4,40 -(Hexafluoroisopropyli-

dene)bis- (benzoic acid), poly (diphenyl hexafluoroisopropylidene
oxadiazole), poly (diphenyl hexafluoro isopropylidene oxadiazole-
co-diphenylether oxadiazole) and poly (diphenyl hexafluoroiso-
propylidene triazole), confirming that the reactions have been
completed. The resonance peak at 14 ppm comes from the proton
of the carboxylic group of 4,4-(hexafluoroisopropylidene) bisbenzoic
acid, which is totally absent in the final product. The resonance
peaks at around 7.8 and 8.4 ppm are assigned to the protons (H1 and
H2) of polyoxadiazole. The absence of the resonance peak from 10 to
11 ppm confirms that there is no residual polyhydrazide.
An additional confirmation that the reaction is complete is
provided by Fig. 4, which shows the ATR–FTIR spectra of 4,4’-
Fig. 4. ATR–FTIR spectra of 4,4’-(Hexafluoroisopropylidene)bis(benzoic acid)
(Hexafluoroisopropylidene)bis(benzoic acid) (Monomer), poly (diphe-
(Monomer), poly (diphenyl hexafluoroisopropylidene oxadiazole) (F-POD), poly
nyl hexafluoroisopropylidene oxadiazole) (F-POD), poly (diphenyl (diphenyl hexafluoro isopropylidene oxadiazole-co-diphenylether oxadiazole)
hexafluoro isopropylidene oxadiazole-co-diphenylether oxadiazole) (Co-POD) and poly (diphenyl hexafluoroisopropylidene triazole) (F-PT).

Fig. 3. 1H-NMR spectra of (a) 4,40 -(Hexafluoroisopropylidene) bis(benzoic acid), (b) poly (diphenyl hexafluoro isopropylidene oxadiazole), (c) poly(diphenyl hexafluoro
isopropylidene oxadiazole-co-diphenylether oxadiazole) and (d) poly(diphenyl hexafluoro isopropylidene triazole).
 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19
(1700 cm  1) of the monomer were absent after conversion into
oxadiazole and triazole groups.
The molecular weights (Mn) of the fluorinated polyoxadiazole
and polytriazole, both soluble in THF, were determined by gel
permeation chromatography (GPC) as 51,000 and 33,000 g/mol
with polydispersity Mw/Mn ¼2. No traces of monomers were
found in the final products.
The high thermal stability of the synthesized polymers was
confirmed by thermogravimetric analysis (TGA), as shown in
Fig. 5. Below 400 1C the weight loss is smaller than 0.5% and is
Fig. 5. Thermogravimetric analysis of poly (diphenyl hexafluoroisopropylidene
oxadiazole) (F-POD), poly (diphenyl hexafluoroisopropylidene oxadiazole-co-
diphenylether oxadiazole) (Co-POD) and poly (diphenyl hexafluoroisopropylidene
triazole) (F-PT).
Fig. 6. FESEM images of flat-sheet porous membranes prepared by phase inversion, manufactured from (a) fluorinated polyoxadiazole (b) fluorinated polytriazole; and
(c) cross-section of fluorinated polytriazole membrane.
15
probably caused by loss of absorbed water. The polymer decom-
position starts only above 400 1C.
3.2. Porous membranes
The high hydrophobicity of the synthesized fluorinated polya-
zoles and their solubility in NMP were very advantageous for the
manufacture of flat-sheet porous MD membranes by phase inversion
and by electrospinning, as discussed below. The aim was to fabricate
membranes with maximum porosity and with interconnected pores
for high water vapor transfer from feed to the permeate side. High
water–polymer contact angle helps to avoid liquid water (saline)
penetration into pores, assuring high salt retention. When the
contact angle is high (high LEP), we can afford even larger pores,
without having issues of pore wetting and liquid water transport
through the pores.
Fig. 6a and b shows the surface morphology of membranes
prepared by phase inversion from fluorinated polyoxadiazole and
polytriazole, respectively. Finger-like cavities could be typically
observed for membrane cross-sections of both polymers, as
shown in Fig. 6c for fluorinated polytriazole. The membranes
were cast directly on glass surfaces or on non-woven polyester.
Solutions of poly (diphenyl hexafluoroisopropylidene oxadiazole-
co-diphenylether oxadiazole) in NMP form a gel at room
temperature and were not suitable for manufacture of phase
inversion membranes. In order to have more information on the
porosity of the flat-sheet membranes obtained by phase inversion
with and without polyester support, they were characterized in
the capillary flow porometer and also tested in ultrafiltration
experiments by applying a pressure of 2.7 bar in 10 mL Amicon
cells. The applied pressure (P) in ultrafiltration experiments was
higher than the liquid entry pressure (LEP) for each membrane,
forcing liquid water through the pores. The results can be seen in
Fig. 7b and indicate a higher porosity for fluorinated polytriazole. The
water flux of fluorinated polytriazole membranes in ultrafiltration
16 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19
Fig. 7. (a) Pore diameter distribution of the F-POD and F-PT membranes measured with capillary flow porometer and (b) hydrodynamic water flux measured in
ultrafiltration setups for the analogous membranes prepared with or without polyester non-woven supports.
Fig. 8. Stress–strain curves (room temperature) of asymmetric porous membranes
prepared by phase inversion from fluorinated polyoxadiazole (F-POD), fluorinated
polytriazole (F-PT) and polyvinylidene fluoride.
experiments was 300 L m  2 h  1 bar  1 and increased to 450 L
m  2 h  1 bar  1 by preparing the membrane on non-woven polyester.
For non-supported fluorinated polyoxadiazole the water flux was
around 100 L m  2 h  1 bar  1. The pore size distribution measured on
the Porolux equipment is shown in Fig. 7a, and agrees with this
observation. The mean flow pore size of fluorinated polyoxadiazole
membranes was estimated as 51 nm, while for fluorinated polytria-
zole it was around 74 nm. The membranes surface images observed
by FESEM are for both polymers comparable. Even larger pores seem
to be on the surface of polyoxadiazole, but it is important to mention
that the microscopy, in the way it was performed, images only the
surface and is not able to distinguish open pores dead-end pores. The
porometry and the ultrafiltration experiments give a better estima-
tion of the accessible pores for transport.
Polyoxadiazole membranes supported on polyester nonwovens
have 3.5-fold higher water flux than non-supported membranes,
indicating a much higher porosity. Phase inversion membranes are
usually highly asymmetric and only the top selective layer con-
tributes for the resistance to the water transport, while the rest of
the pore structure is much opener. However during the formation of
the membranes, cast on glass surfaces, a second hydrophobic skin
appears to be formed on the membrane side, which was in contact
with glass. This could be seen also by electron microscopy (image is
not shown here).
An additional advantage of membranes prepared with the
synthesized polyazoles is their strong mechanical stability com-
pared to membranes prepared from other polymers like PVDF.
Fig. 8 shows stress–strain curves measured for porous mem-
branes of fluorinated polyoxadiazole, polytriazole and PVDF in the
same conditions. The area under each curve gives an idea of the
membrane toughness, which is higher for the synthesized poly-
azoles, particularly for F-POD.
Membranes could be prepared with all the three synthesized
polymers by electrospinning. For the polyoxadiazole copolymer the
solution had to be heated at 70–80 1C before electrospinning. For the
other two polymers the electrospinning was conducted at room
temperature without additional solution treatment. The morphology
obtained by FESEM is compared in Fig. 9. All membranes were
prepared with the same solvent and same polymer composition.
Image analysis software (Image J, National Institutes of Health)
was used to evaluate the fibers diameter distribution. It can be
seen very clearly that the morphologies of the electrospun
membranes are bead-free nanofibers. The membranes prepared
from polyoxadiazole have slightly larger fiber diameter ranging
from 160 to 300 nm. In the case of polytriazole the fiber diameter
is equal or less than 100 nm. For the polyoxadiazole copolymer
the fiber diameter range is broad and ranges from 70 to 420 nm
with the most frequent diameter being around 140 nm. The
difference in diameter size distribution can be assigned to the
polymer properties (solution viscosity, surface tension) since the
experimental conditions were kept the same for all preparations.
The pore diameter distributions of the electrospun membranes
measured by capillary flow porometer are shown in Fig. 9 (right
side). The pores are much larger than in the case of the phase
inversion membranes. A very narrow distribution of the pore
diameters can be observed for polyoxadiazole copolymer membrane
with a mean flow pore size (MFP) of 1.3 mm. For the fluorinated
polyoxadiazole membrane, the MFP is around 1.7 mm. However a
broad pore size distribution can be seen for polytriazole membrane,
ranging from 2.5 to 3.3 mm with a MFP of 2.7 mm.
In summary the MFP measured with the porometer increases in
the following sequence: Co-POD (1.3 mm)oF-POD (1.7 mm)oF-PT
 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19 17

Fig. 9. FESEM images of fluorinated (a), polyoxadiazole (b), polytriazole and (c) polyoxadiazole copolymer nanofibrous membranes prepared by electrospinning (left
column); fiber diameter distribution (middle column); pore diameter distribution measured with capillary flow porometer (right column).

(2.7 mm); the FESEM images in Fig. 9 indicate that the pores increase
in the sequence; F-PToCo-PODoF-POD; the values of Liquid Entry
Pressure (LEP), measured with water (first bubble point) in the same
porometer increase as 0.62 bar for Co-PODo0.76 bar for F-PTo
0.90 bar for F-POD.
For the LEP value there is a strong contribution of pore size
itself (higher pores, lower LEP), but also of the hydrophobicity.
Membranes made of more hydrophobic materials with the same
pore size would have higher LEP values. LEP values are the most
important as far as the MD membrane selectivity is concerned.
High LEP is aimed to hinder any liquid water leakage through the
membrane during operation. It is here confirmed that even
though the electrospun membranes prepared with fluorinated
polyoxadiazole have larger pores than when prepared by phase
inversion, they still can afford a large operation feed pressure,
without any liquid water penetration into pores. High porosity
and high hydrophobicity makes the membrane particularly pro-
mising for MD performance.
Fig. 10 shows a comparison between water contact angles
measured for dense films of the different synthesized polymers
and commercial poly (vinylidene fluoride) (PVDF), as well as the
contact angles for the respective porous membranes obtained by
phase inversion and electrospinning. Fluorinated polyoxadiazole
is the most hydrophobic of considered polymers, but the apparent
contact angle depends also on porosity. The increase of apparent
contact angle yapp with the increase of porosity follows the
classical Cassie equation [24]:
cosyapp ¼ f 1 cosy1 f 2 ð1Þ
where f1 is the fraction of the surface occupied by the polymeric
material and f2 is the fraction of the open area (pore), y1 is the
contact angle of water on the dense film of the same polymer.
18 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19

Fig. 10. Contact angle of dense polymer films and porous membranes prepared by
phase inversion and electrospinning from PVDF, polyoxadiazole copolymer,
fluorinated polyoxadiazole and fluorinated polytriazole.

During electrospinning, the fibers are deposited on a collector

plate in a random manner and it is possible to achieve a very high
porosity compared to phase inversion membranes. The highly
porous electrospun membranes (high f2 value) have therefore much
higher contact angle, even though the polymer used for their
manufacture is the same as that used for a porous phase inversion
membrane or for a dense film. The combination of high porosity
(high f2 value) and high hydrophobicity (high y1) of electrospun
fluorinated polytriazole membranes, which has larger pores than the
analogous fluorinated polyoxadiazole membrane, leads to particu-
larly high apparent contact angle (high yapp) values and confirms the
expectation of good results in MD operation.

3.3. Membrane distillation

The flat-sheet membranes prepared by phase inversion and by

electrospinning were tested for 1 h in a direct contact membrane
distillation (DCMD) setup for using real Seawater as a feed solution.
Comparison of water fluxes for different membranes as a function of
feed inlet temperature is represented in Fig. 11. The general
observation is that the fluxes increase, as the temperature increases
for the various membranes, as expected, since the driving force for
water transport in MD is mainly the partial vapor pressure differ-
ence due to the temperature difference between both sides of the
membrane. The electrospun membranes exhibited the highest water
flux, independent of the polymeric material.
For electrospun fluorinated polyoxadiazole at 80 1C, the high-
est flux value obtained was 78 L m  2 h  1; for the polyoxadiazole
copolymer the flux was slightly lower (72 L m  2 h  1) under the
same conditions, since the pore size and porosity are also lower.
Higher values were obtained for electrospun fluorinated poly-
triazole membranes (85 L m  2 h  1). The pore size distribution for
these membranes (measured with capillary flow porometer) is
shifted to higher values (2.5–3.3 mm) when compared to fluori-
nated polyoxadiazole (around 1.7 mm). MD water fluxes as high as
83 L m  2 h  1 have been reported in the literature for PVDF
hollow fiber [6] at 80 1C and 195 L m  2 h  1 for PTFE flat-sheet
membranes [25] at 130 1C. We prefer to keep the operation
temperatures lower than 80 1C, since the idea is to prove that
the membranes would be effective enough for solar thermal-
driven membrane distillation, where the operating temperature
(60–80 1C) can be obtained easily from the solar collectors. We
believe that the porosity of the membranes reported here can still
be further optimized, by changing the manufacturing conditions.
Relevant is that, in all cases, the salt rejection was more than
Fig. 11. Membranes performance in MD experiments with permeate temperature
22 1C: (a) Membranes made of fluorinated polyoxadiazole by different methods
and electrospun polyoxadiazole copolymer and (b) membranes made of fluori-
nated polytriazole.
99.95%. This is an advantage of the high hydrophobicity of all
polymers used for membrane manufacture. As reported before,
the LEP values for the electrospun membranes explored here are
between 0.62 and 0.90 bar (with pore size varying from 1.3 to
3.3 mm).
The membranes prepared by phase inversion had much lower
porosity and therefore also lower water flux. For fluorinated
polyoxadiazole, supported phase inversion membrane had a
water flux of 25 L m  2 h  1, while for the non-supported mem-
brane the flux goes down to 6 L m  2 h  1. For fluorinated poly-
triazole the values for the supported phase inversion membrane
was 31 L m  2 h  1, practically the same as for the non-supported
membranes. The same exceptionally high values of salt selectivity
were confirmed. However since the pore diameters are below
100 nm, more than 10-fold smaller than for the electrospun
membranes, it is expected that these membranes could be
successfully used even at much higher temperature driven vapor
pressure difference, without suffering decrease of salt selectivity.
The big change in water flux for polyoxadiazole membranes by
adding a polyester support follows the same pattern observed in
 H. Maab et al. / Journal of Membrane Science 423–424 (2012) 11–19
ultrafiltration experiments (Fig. 7b). The open polyester support
hinders the formation of an additional skin in the membrane and
keeps the resistance to the water vapor transport low. For fluori-
nated polytriazole the effect of the support on the MD performance
is not significant. Taking into account all of the tested flat-sheet
membranes, the fluorinated polytriazole was the best material
reported here for MD, independent of whether the manufacturing
technology was phase inversion or electrospinning.
4. Conclusions
A new class of membranes have been manufactured and tested
for membrane distillation, based on hydrophobic polyoxadiazoles
and polytriazoles. Flat sheet membranes were prepared by phase
inversion and electrospinning. The membranes were fully char-
acterized in terms of porosity. In all cases high hydrophobicity
and high porosity were achieved with excellent thermal stability.
The performances of supported and unsupported flat-sheet asym-
metric phase inversion and electrospun membranes were com-
pared in DCMD operation. Salt selectivity as high as 99.95% was
confirmed in DCMD tests, operating with seawater feed at
different temperatures. Water flux as high as 85 L m  2 h  1 was
demonstrated at 80 1C feed temperature and 22 1C permeate.
Acknowledgment
We thank KAUST Core Labs for supporting the characterization
and KAUST central workshop for the fabrication of test setups and
cells. We thank Prof. Peng Wang for access to the electrospinning
equipment.
References
[1] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensive
review, Desalination 287 (2011) 1–362.
[2] M. Khayet, Membranes and theoretical modeling of membrane distillation: a
review, Adv. Colloid Interface Sci. 164 (1) (2011) 56–88.
[3] G. Mannella, V. La Carrubba, V. Brucato, Some features of polymeric
membranes for water purification via membrane distillation, J. Appl. Polym.
Sci. 122 (2011) 3557–3563.
[4] S.P.N. Peinemann, K.V. Pienemann, Membrane Technology in the Chemical
Industry, 2nd ed., Wiley-VCH, 2006.
[5] M. Gryta, M. Barancewicz, Influence of morphology of PVDF capillary
membranes on the performance of direct contact membrane distillation,
J. Memb. Sci. 358 (1) (2010) 158–167.
19
[6] F. Edwie, M.M. Teoh, T.S. Chung, Effects of additives on dual-layer hydro-
phobic hydrophilic PVDF hollow fiber membranes for membrane distillation
and continuous performance, Chem. Eng. Sci. 68 (2012) 567–578.
[7] P. Wang, M.M. Teoh, T.S. Chung, Morphological architecture of dual-layer
hollow fiber for membrane distillation with higher desalination performance,
Water Res. 45 (17) (2011) 5489–5500.
[8] M. Khayet, T. Matsuura, J. Mengual, M. Qtaishat, Design of novel direct
contact membrane distillation membranes, Desalination 192 (1–3) (2006)
105–111.
[9] D.Y. Cheng, S.J. Wiersma, Composite Membrane for a Membrane Distillation
System, US Patent 4,316,772, 1982.
[10] Z. Hendren, J. Brant, M. Wiesner, Surface modification of nanostructured
ceramic membranes for direct contact membrane distillation, J. Memb. Sci.
331 (1–2) (2009) 1–10.
[11] S. Cerneaux, I. Struzynska, W.M. Kujawski, M. Persin, A. Larbot, Comparison
of various membrane distillation methods for desalination using hydropho-
bic ceramic membranes, J. Memb. Sci. 337 (1–2) (2009) 55–60.
[12] L.F. Dumee, K. Sears, J. Schutz, N. Finn, C. Huynh, S. Hawkins, M. Duke, S. Gray,
Characterization and evaluation of carbon nanotube Bucky-Paper mem-
branes for direct contact membrane distillation, J. Memb. Sci. 351 (1–2)
(2010) 36–43.
[13] K. Gethard, S. Mitra, Carbon nanotube enhanced membrane distillation for
online preconcentration of trace pharmaceuticals in polar solvents, Analyst
136 (12) (2011) 2643–2648.
[14] Z. Ma, Y. Hong, L. Ma, M. Su, Superhydrophobic membranes with ordered
arrays of nanospiked microchannels for water desalination, Langmuir 25 (10)
(2009) 5446–5450.
[15] C. Feng, K. Khulbe, T. Matsuura, R. Gopal, S. Kaur, S. Ramakrishna, M. Khayet,
Production of drinking water from saline water by air-gap membrane
distillation using polyvinylidene fluoride nanofiber membrane, J. Memb.
Sci. 311 (1) (2008) 1–6.
[16] D. Gomes, S.P. Nunes, Fluorinated polyoxadiazole for high-temperature
polymer electrolyte membrane fuel cells, J. Memb. Sci. 321 (1) (2008)
114–122.
[17] M.L. Ponce, J. Roeder, D. Gomes, S.P. Nunes, Stability of sulfonated poly-
triazole and polyoxadiazole membranes, Asia Pac. J. Chem. Eng. 5 (1) (2010)
235–241.
[18] M. Boaventura, M. Ponce, L. Brandao, A. Mendes, S. Nunes, Proton conductive
membranes based on doped sulfonated polytriazole, Int. J. Hydrogen Energy
35 (21) (2010) 12054–12064.
[19] M. Ponce, M. Boaventura, D. Gomes, A. Mendes, L. Madeira, S. Nunes, Proton
conducting membranes based on benzimidazole sulfonic acid doped sulfo-
nated poly (oxadiazoleäıtriazole) copolymer for low humidity operation, Fuel
Cells 8 (3–4) (2008) 209–216.
[20] D. Gomes, J. Roeder, M.L. Ponce, S.P. Nunes, Characterization of partially
sulfonated polyoxadiazoles and oxadiazole–triazole copolymers, J. Memb.
Sci. 295 (1–2) (2007) 121–129.
[21] D. Gomes, R. Marschall, S.P. Nunes, M. Wark, Development of polyoxadiazole
nanocomposites for high temperature polymer electrolyte membrane fuel
cells, J. Memb. Sci. 322 (2) (2008) 406–415.
[22] E. Hensema, J. Boom, M. Mulder, C. Smolders, Two reaction routes for the
preparation of aromatic polyoxadiazoles and polytriazoles: syntheses and
properties, J. Polym. Sci. Part A: Polym. Chem. 32 (3) (1994) 513–525.
[23] K. Smolders, A.C.M. Franken, Terminology for membrane distillation, Desa-
lination 72 (1989) 249–262.
[24] A. Cassie, Contact angles, Discuss. Faraday Soc. 3 (1948) 11–16.
[25] D. Singh, K.K. Sirkar, Desalination of brine and produced water by direct
contact membrane distillation at high temperatures and pressures, J. Memb.
Sci. 389 (2011) 380–388.