WO2015084266A1

WO2015084266A1 - A composite nanofiber membrane for membrane distillation and a method of fabrication thereof

Info

Publication number: WO2015084266A1
Application number: PCT/SG2014/000578
Authority: WO
Inventors: Rong Wang; Yuan LIAO; Anthony Gordon Fane
Original assignee: Nanyang Technological University
Priority date: 2013-12-05
Filing date: 2014-12-05
Publication date: 2015-06-11

Abstract

A composite nanofiber membrane was designed and fabricated for membrane distillation, which comprising: a support layer fabricated by a hydrophobic synthetic material; and a selective layer formed on the support layer. The selective layer was fabricated by electrospinning hydrophobic nanoparticles-a hydrophobic synthetic material blended solution. The hydrophobic synthetic material in the blended solution is the same as the hydrophobic synthetic material of the support layer. A method of fabricating a composite nanofiber membrane is introduced. The method comprises: electrospinning a dope solution comprising a hydrophobic synthetic material onto a collector to form a support layer; modifying nanoparticles to be hydrophobic for dispersion in a solution of the hydrophobic synthetic material; dispersing the modified hydrophobic nanoparticles in the solution of the hydrophobic synthetic material to form a modified dope solution; and electrospinning the modified dope solution onto the support layer to form a selective layer.

Description

A COMPOSITE NANOFIBER MEMBRANE FOR MEMBRANE DISTILLATION AND A

METHOD OF FABRICATION THEREOF

FIELD OF THE INVENTION

This invention relates to a composite nanofiber membrane for membrane distillation and a method of fabrication thereof.

BACKGROUND OF THE INVENTION

Clean water is essential to healthy ecosystems and sustainable socio-economic development. The increasing pressure on freshwater resources calls for effective and novel approaches to produce high-quality water. Among various technologies, membrane distillation (MD) is an emerging process that can utilize low-grade or waste heat to generate high-quality water from impaired water with high recovery (100% in theory) ^m. MD is gaining attention as a low Greenhouse gas (GHG) option for water purification. In its simplest configuration, direct contact membrane distillation (DCMD) involves the transportation of water vapor from a hot feed stream through a hydrophobic micro-porous membrane and the condensation of the water vapor into a cool permeate, where the driving force is the vapor pressure difference across the membrane ^[2].

In the DCMD process, the hydrophobic membrane serves as the barrier between the two liquid phases of feed and permeates and its properties determine the system performance. In order to deliver high vapour permeability and ensure high water quality, the membrane should possess appropriate pore sizes and high porosity ^[3!. To avoid membrane pore wetting, at least one layer of the MD membrane should be hydrophobic or preferably superhydrophobic ^[4]. Moreover, the MD membrane should be physically robust, exhibit excellent thermal stability and maintain a stable performance in long-term usage. However, specially designed MD membranes are not commercially available and most membranes used were initially designed for micro-filtration. Such membranes are suboptimal for MD applications and suffer from progressive membrane wetting in the MD process. Therefore, the development of improved membranes that can fulfill the unique requirements of MD is imperative to facilitate practical applications of the MD process.

SUMMARY OF INVENTION

A novel nanocomposite membrane designed for membrane distillation (MD) application has been developed. A scaffold-like polyvinylidene fluoride (PVDF) nanofiber membrane made by electrospinning, with high porosity and low tortuosity and an adjustable thickness, is used as a mechanical support. A silica-PVDF composite selective layer, which has an extremely high water contact angle (>150°) and water repelling properties, is formed on the top of the support via the same electrospinning technique. The silica-PVDF composite layer has a similar hierarchical structure to the lotus leaf, and is able to form air pockets or water vapour pockets on the membrane surface during the MD process, leading to the lowest contact area between water and the membrane. The unstable contact line points can force the water droplets to roll off the membrane surface spontaneously. According to a first aspect, there is provided a composite nanofiber membrane for membrane distillation, the composite nanofiber membrane comprising: a support layer comprising a hydrophobic synthetic material; and a selective layer formed on the support layer, the selective layer comprising hydrophobic nanoparticles dispersed in a same hydrophobic synthetic material as the hydrophobic synthetic material of the support layer.

The selective layer may have a water contact angle greater than 150°. The selective layer may have a sliding angle lower than 20°. The support layer may comprise a nanofiber membrane. The support layer may have a scaffold-like structure.

The hydrophobic nanoparticles on a surface of the composite nanofiber membrane may be covered and protected by the hydrophobic synthetic material.

The hydrophobic synthetic material may be polyvinylidene fluoride.

The hydrophobic nanoparticles may be modified hydrophobic silica nanoparticles dispersed in a solution of the hydrophobic synthetic material. The silica nanoparticles may be fluorinated.

According to a second aspect, there is provided a method of fabricating a composite nanofiber membrane, the method comprising: electrospinning a dope solution comprising a hydrophobic synthetic material onto a collector to form a support layer; modifying nandparticles to be hydrophobic for dispersion in a solution of the hydrophobic synthetic material; dispersing the modified hydrophobic nanoparticles in the solution of the hydrophobic synthetic material to form a modified dope solution; and electrospinning the modified dope solution onto the support layer to form a selective layer.

The modifying may comprise fluorinating the silica nanoparticles.

The fluorinating reaction may comprise stirring silica nanoparticles in an N-hexane solution comprising a, ω-triethoxysilane terminated perfluoropolyether and tetraethoxysilane.

The method may further comprise centrifuging the fluorinated silica nanoparticles from the solution and annealing the fluorinated silica nanoparticles.

BRIEF DESCRIPTION OF FIGURES

In order that the invention may be fully understood and readily put into practical effect there shall now be described by way of non-limitative example only exemplary embodiments of the present invention, the description being with reference to the accompanying illustrative drawings.

Fig. 1 is a schematic illustration of a nanocomposite membrane for MD according to the present invention.

Fig. 2 is schematic diagram of an electro-spinning setup.

Fig. 3A is a schematic diagram of the mechanism of modification reaction on Si0₂ nanoparticle surface.

Fig. 3B is a chemical structure of Fluorolink FS10.

Fig. 4A is an XPS wide-scan of original Si0₂ nanoparticles.

Fig. 4B is an 0 1s core-level spectra of original Si0₂ nanoparticles.

Fig. 4C is an XPS wide-scan of modified Si0₂ surfaces.

Fig. 4D is an O 1s core-level spectra of modified Si0₂ surface.

Fig. 5A is an XPS C 1s core-level spectra of a PVDF nanofiber membrane.

Fig. 5B is an XPS C 1s core-level spectra of modified silica.

Fig. 5C is an XPS C 1s core-level spectra of a modifed Si0₂-PVDF composite membrane. Fig. 6A is photo image and Field Emission Scanning Electron Microscopy (FESEM) image of the surface morphology of a PVDF nanofiber membrane.

Fig. 6B is photo image and FESEM image of the surface morphology of a superhydrophobic silica-PVDF composite membrane. Fig. 6C is a schematic illustration of a PVDF nanofiber membrane exposed to a liquid.

Fig. 6D is a schematic illustration of a superhydrophobic composite membrane exposed to a liquid.

Fig. 7A is a graph of water contact angles and sliding angles of the composite membranes with various electrospinning times fabricated by PVDF/ small FS10-SiO₂ mixture.

Fig. 7B is a graph of water contact angles and sliding angles of the composite membranes with various electrospinning times fabricated by PVDF/ large FS10-SiO₂ mixture.

Fig. 8A is a schematic illustration of a sliding angle test configuration showing sliding angle calculation.

Fig. 8B is a photograph of a sliding test configuration.

Fig. 9A is a graph of behavior of the water droplets on a superhydrophobic surface of S-

PVDF membrane after ultrasonic-treatment for different times.

Fig. 9B is a graph of behavior of the water droplets on a superhydrophobic surface of L-

PVDF membrane after ultrasonic-treatment for different times.

Fig. 10A is a graph of continuous DCMD test of an electrospun PVDF membrane (3.5 wt%

NaCI solution as feed, T_f=333 K, T_P=293K).

Fig. 10B is a graph of continuous DCMD test of the S-PVDF composite membrane (3.5 wt%

NaCI solution as feed, T_f=333 K, T_P=293K).

Fig. 10C is a graph of continuous DCMD test of the L-PVDF composite membrane (3.5 wt%

NaCI solution as feed, T_f=333 K, T_P=293K).

Fig. 10D is a graph of continuous DCMD test of a commercial PVDF membrane (3.5 wt%

NaCI solution as feed, T_f=333 K, T_P=293K).

Fig. 11A is a schematic illustration of a PVDF nanofiber membrane surface used in a DCMD configuration.

Fig. 11 B is a schematic illustration of the composite superhydrophobic membrane surface used in a DCMD configuration.

Fig. 12 is a flowchart of an exemplary method of fabrication of the composite nanofiber membrane.

DETAILED DESCRIPTION

Exemplary embodiments of the composite nanofiber membrane 10 and its method of fabrication 30 will be described below with reference to Figs. 1 to 12.

The composite flat sheet membrane 10 consists of two layers, a selective layer 100 and a support layer 200. The support layer 200 is a scaffold-like polyvinylidene fluoride (PVDF) nanofiber membrane 200 that serves as the substrate for mechanical support, while the top or selective layer 100 is a silica-PVDF composite layer 100.

The nanocomposite membrane 10 has a similar hierarchical structure to a lotus leaf, having an extremely high water contact angle and water repelling properties suitable for MD application. An exemplary embodiment of the MD nanocomposite membrane 10 with a lotus leaf-liked superhydrophobic selective layer 100 and a scaffold-like nanofiber support layer 200 is shown in Fig. 1. As can be seen in Fig. 1 , during membrane distillation, a feed solution 61 (for example sea water, brine water, waste water) is flowed over the selective layer 100 of the membrane 10 in a direction parallel to the membrane 10, as indicated by arrow 62. Water vapour passes through the membrane 10 as indicated by arrow 63, and is collected as a permeate 64 of pure water that is flowed in a direction parallel to the membrane 10, as indicated by arrow 65, which is opposite to the direction of flow 62 of the feed solution 61. As can be seen in the insert, water in the form of a droplet 99 containing impurities rolls over the surface 100 of the membrane 10 and does not enter air / vapour pockets 66 in the membrane 10 due to the presence of hydrophobic nanoparticles 103 in the selective layer 100 that is supported by the nanofiber layer 200 of the membrane 10.

The scaffold-like PVDF nanofiber membrane or support layer 2.00 has high porosity and low tortuosity and an adjustable thickness. As such, the mass transfer resistance of the bottom layer 200 for water is low, leading to a high water flux.

The silica-PVDF composite selective layer 100 has a similar hierarchical structure to the lotus leaf, and is able to form air pockets or water vapour pockets on the membrane surface during the MD process, leading to the lowest contact area between water and the membrane 0. The selective layer 100 has a high water contact angle above 150° and excellent water repellence with a low sliding angle less than 20°. As such, the membrane 10 has excellent water repellence, which can effectively prevent the membrane 10 from wetting.

The top selective layer 100 is made of a silica-containing PVDF material in which the PVDF material is the same as the material of the substrate or support layer 200. Thus, the two layers 100, 200 have a good adhesion to form a dual-layer composite nanofiber membrane 10 without delamination.

The composite nanofiber membrane 10 showed superior water flux in DCMD compared with other commercial and in-house made hydrophobic membranes. It 10 also showed a stable performance over 50 hours of testing time and produced high quality water with conductivity below 5.0 pS/cm. Electrospinning of PVDF nanofiber membrane 200 as a substrate (31 )

The PVDF nanofiber membrane or support layer 200 was prepared using an electrospinning setup 20 as shown in Fig. 2. In the setup, 20, a polymer dope solution 21 with PVDF concentration between 5 and 10 wt.% in a thin PTFE tube was pushed slowly into high voltage charged sprayers 22 by a syringe pump 23 in a chamber 27. The sprayers 22 were connected to a high voltage supplier 24 which can generate positive voltages of up to 30 kV. A positive voltage of 25 or 28 kV was applied across a distance of 12 cm or 15 cm between the tip of the sprayers 22 and a grounded collector 25. The spinning sprayers can be moved slowly and evenly by a motor during electro-spinning, as indicated by arrow 26. Moisture in the chamber 27 can be altered by pumping dry air into the chamber 27.

During the electro-spinning process (31), nanofibers were produced and collected on non- woven textiles posted on the rotating collector 25 when the applied electric field overcomes the surface tension of the polymer dope solution 21. The collector 25 is preferably cylindrical and rotated about its longitudinal axis L, as shown by arrow 28. Solvents in the nascent nanofibers were evaporated and the nanofibers started to bend concurrently.

Silica nanoparticle modification (32) and dope preparation (33) for the top selective layer 100 In order to prepare a robust superhydrophobic surface for the selective layer 100, silica nanoparticles were first modified by a, ω-triethoxysilane terminated perfluoropolyether ((EtO)₃Si-PFPE-Si(OEt)₃) with an average molecular weight between 1750 to 1950 under the trade name Fluorolink® S10 (FS10), which made them hydrophobic (32) to easily disperse in the second PVDF dope solution and able to stick to the PVDF 'islands' on the substrate, allowing the top surface to be more durable ^[6] .

A desired amount of silica nanoparticles with aggregate diameters of 0.2-0.3 pm or 5-10 nm was stirred rapidly overnight in a N-hexane solution in which FS10 and tetraethoxysilane (TEOS) with a mass ratio of 3:2 was added. The total concentration of FS10 and TEOS was between 2 and 8 wt.%. The modified Si0₂ nanoparticles were then separated by centrifugation and annealed in a vacuum oven at 100 °C for 1 hour. The resultant white powder was stored in a vacuum oven at 50 °C for further experiments and characterization.

Dope preparation (33) and electrospinning of Silica-PVDF top selective layer (34)

The PVDF precursor solutions for the superhydrophobic top layer 100 were prepared by initially dissolving 3 to 8 wt.% PVDF material in N, N-Dimetylformamide (DMF). Then, modified Si0₂ with different diameters as obtained by the modification described above were dispersed in the prepared PVDF solution (23) by stirring rapidly at room temperature. The mass ratio of modified Si0₂ particles/PVDF was fixed at 2:1 to guarantee the superhydrophobic effects of the selective layer 100.

After spinning the porous substrate 200 (31 ), the superhydrophobic surface layers 100 were electrospun (34) using modified silica-blended PVDF dopes on top of the porous substrate 200. As the purpose was to investigate the effects of the superhydrophobic layer 100 precisely on the membrane properties, all the composite membranes 10 had the same substrate layer 200.

Example: Fabrication details of the robust superhydrophobic composite nanofiber membranes

(a) Fabrication of PVDF nanofiber substrate and composite selective layer

Table 1 below summarizes the electrospinning conditions of the PVDF nanofiber membrane 200 and silica-PVDF composite membranes 100. Two types of silica nanoparticles were used. As shown in Table 1 , the PVDF nanofiber support 200 was fabricated by a 8 wt.% PVDF dope. After spinning the porous substrate 200 (31), the superhydrophobic surface layers 100 were electrospun (34) using two modified silica-blended PVDF dopes.

PVDF S-PVDF L-PVDF

Membrane ID

(support layer) (selective layer)

PVDF PVDF

PVDFHSV900/SiO₂

Dope composition HSV900/DMF: HSV900/SiO₂

(large diameter) (wt. %) 8/92 (small diameter)

/DMF: 5/10/95

(0.004% wt LiCI) /DMF: 5/10/95

Dope flow rate

0.03 0.02

(mL min^" )

Travel Speed (mm sec^"1) 0.1 0.1

Travel Distance (cm) 8 8

Distance (cm) 15 12

Voltage (kV) 28 28

Table 1

(b) Silica nanoparticle modification to enhance its hydrophobicity

A 10 wt.% of modified silica nanoparticies was added into the PVDF/DMF solution to prepare the dope for forming the selective layer 100. The modification reaction on silica nanoparticies (32) comprises a hydrolysis reaction of tetraethoxysilane (TEOS), as shown in Fig. 3A(1) and addition of the silinol groups onto the silica surface, as shown in Fig. 3A(2). Grafting occurred by an alcohol condensation reaction between the hydroxyl groups and Si-O-alkyl groups, as shown in Fig. 3A(3). Other Si-O-alkyl groups on the other end of the FS10 molecular chains would either react on the same silica particle or graft onto another silica surface which may increase the particle diameter. The chemical structure of Fluorolink FS10 (B) is shown in Fig. 3B.

(c) Characteristics of the modified silica nanoparticies

To examine the surface enrichment of fluorocarbon chains on the modified silica, an X-ray photoelectron spectroscopy (XPS) analysis, using a Theta Probe XPS provided by Thermo Fisher Scientific (Singapore), was carried out on unmodified and modified Si0₂ surfaces.

As shown in Figs. 4A and 4C, compared with the original silica particles which only have silica and oxygen elements on the surface, the FS10 grafted silica particles (FS10-SiO₂) possess additional carbon, nitrogen and fluoride elements. According to the chemical structure of FS10 shown in Fig. 3B, these additional elements are from the hydrophobic chemical FS10 reacted on the silica surface. The surface content of silica atoms was significantly lower than that of corresponding bulk concentration, indicating that a strong surface enrichment of organic phase such as carbon and fluoride elements was produced after the modification reaction. Furthermore, the high-resolution of XPS 01s core-level spectra of control silica, as shown in Fig. 4B and fluorinated silica, as shown in Fig. 4D, were curve-resolved into peaks to derive more information on the fluorinated segments segregation at the surface. As shown in Fig. 4B,the binding energy oxygen peak at 531.8 eV corresponds to the chemical bonds in Si0₂ ^[7].

Concerning the 01s core-level spectra of modified silica, a figure with double peaks is clearly observed in Fig. 4C. While the oxygen components attributed to the PFPE segments (CF₂-0-CF₂) were centred at a higher binding energy around 536.3 eV, oxygen atoms linked to silica (Si-O-Si) which assisted the reaction between silica surface and FS10 are associated with the less pronounced peak at around 533.1 eV ^[8]. The minor extent of oxygen double bonded to carbon present in FS10 also gives a small contribution to the second peak. By evaluating the chemical composition on silica surfaces before and after modification reaction, it is confirmed that the FS10 successfully covered the silica surface, which shifted the hydrophilic nanoparticles with OH groups on the surface to be fluorinated hydrophobic nanoparticles.

To further demonstrate that the superhydrophobic membrane 100 surface contains modified silica particles and the modified silica was covered and proctected by thin PVDF layers, independent evaluations of the surface characteristics of the original PVDF nanofiber 200 and silica-PVDF composite membranes 10 were conducted by XPS.

As shown in Fig. 5A, the C 1 s core-level spectrum of the PVDF nanofiber membrane 200 can be simply curve-fitted with two peak components, one at a binding energy (BE) of 285.0 eV for the carbon bonded to hydrogen (CHx) and the other at BE of 290 eV for a single carbon bonded to fluorine (C-F₂) ^[9], which is typical of the C1 s core level spectrum for PVDF.

As shown in Fig. 5B, small amounts of hydrocarbon and fluoride carbon were also present on FS10-SiO₂ due to the chemical structure of FS10, and the prominent CF envelope can be fitted with two peaks that correspond to 0-CF₂-0 and O-C2F4-O at 294.9 eV and 293.5 eV, respectively ^[10] Beside these peaks, there are two other peaks that are attributed to the O- CH₂-C (286.6 eV) and -CO-N- (287.9 eV) chemical bonds of the FS10 structure According to the parameters determined from the curve fitting of the superhydrophobic silica- PVDF composite membrane 10, as can be seen in Fig. 5C, the C1s envelope identifies the primary presence of -CH₂- and -CF₂-, which means the surface layer 100 (within 10 nm) was mainly bound and covered by PVDF material.

The surface morphologies of the resultant membranes 10, 200 were observed by a field emission scanning electron microscope (FESEM) which was provided by JSM-7600F, JEOL Asia Pte Ltd, Japan. The water contact angles of the superhydrophobic composite membranes 10 were measured by a goniometer (Contact Angle System OCA, from Data Physics Intruments GmbH in Singapore). One 5 μΙ_ water droplet was dropped onto a levelled membrane surface 100 and the images of the water drop on the membrane surface 100 were captured by an optical system to calculate the contact angle.

Optical photographs of the PVDF nanofiber membrane 200 which was fabricated by 8 wt% PVDF dissolved in N, N-Dimetylformamide (DMF) and the superhydrophobic silica-PVDF composite membrane 10 are shown in Figs. 6A and 6B, respectively. As shown in the FESEM image inserted in Fig. 6A, the PVDF nanofiber membrane 200 surface presents a continuous arrangement of nanofibers with an average diameter of 170 nm. The nanofiber PVDF membrane 200 possesses a contact angle of 142.8°, which is much higher than other PVDF membranes prepared by non-solvent induced phase separation ^[12]. However, the PVDF nanofiber membrane 200 also shows a high contact angle hysteresis. For example, it is observed that a water droplet 99 cannot roll off from its surface even when the membrane 200 is turned upside down, as shown in small inserted pictures in Fig. 6C. This is the so- called petal effect or Cassie impregnating wetting state ^[13]. Water droplets 99 tend to penetrate into the larger-scaled grooves 202 between the nanofibers 201 and remain spherical above the nanofiber membrane 200. It is evident that the water sealed in the micro-sized grooves 202 would assist water droplets 99 in adhering to the membrane 200 surface due to the surface tension force.

On the other hand, after the modified Si0₂-PVDF blended dope (10 wt% Si0₂ and 5 wt% PVDF were mixed in DMF solvent homogeneously (33)) was sprayed via electrospinning onto the nanofiber membrane 200 surface (34), the nanofibers 101 and nano-beads 103 structure could be formed on the composite membrane 10 surface 100, as shown in Fig. 6B. The nanofiber 101 diameter was around 90 nm. The thinner diameter of the modified PVDF nanofibers 101 was probably due to the lower polymer concentration used compared with the pure PVDF dope used for preparing the substrate 200 ^[14]. The water contact angle of the composite membrane 10 is 156.3°. Additionally, the composite membrane 10 possesses excellent water repellence property as shown in Fig. 6D.

Similar to the lotus leaf, the composite silica-PVDF membrane 10 has a microstructure 103 consisting of modified silica-PVDF mixed islands 104 and a nanostructure comprising PVDF- bound silica nanoparticles 105 on the islands' 104 surface. This hierarchical structure 103 of the selective layer 100 prevents water 99 from intruding into the microstructural spaces. The triple contact lines between the solid and liquid or the solid and air on the randomly rough surface are expected to be contorted and unstable, which makes the water droplets 99 easily moved ^{t ]}. The deposition of modified silica-PVDF mixture on the PVDF nanofiber support 200 by electrospinning (34) is the critical step to shift the membrane 200 from hydrophobic to superhydrophobic 100. Figs. 7A and 7B show the effects of the modified silica particle sizes (0.2-0.3 μιτι and 10-15 nm, respectively) and number of spinning layers on water contact angle (black squares) and sliding contact angle (white squares). The technique for measuring sliding contact angle is illustrated in Fig. 8.

It can be seen from Figs. 7A and 7B that the diameter of the modified silica particles has no obvious impact on water contact and sliding angles of as-prepared membranes 10. After spinning 3 layers of the composite dope (34), the membranes 10 have contact angles over 150° and exhibit water repellence properties due to the formation of both microstructures 104 and secondary nanostructures 105. With an increase in spinning layer number, the contact angle of the composite membranes 10 increased progressively to a plateau and the sliding contact angle decreased slightly. This is probably due to the enhancement of hierarchical structures 103. Furthermore, the water sliding angle difference between the PVDF nanofiber membrane 200 and the composite membrane 10 is pronounced. The PVDF nanofiber membrane 200 shows a strong adhesive force for water droplets so that they cannot move, while the composite membrane 10 has a low sliding angel less than 20°. It is believed that the micro- and nano-roughness increased the presence of air pockets on the composite membrane 10 surface significantly, which could effectively trap air between the liquid and solid surface, preventing the liquid from penetrating into the surface cavities. Thus the build-up of a discrete contact between the liquid and the solid leads to the drastic reduction of sliding contact angle ^[15]. Since the composite membranes 10 with nine superhydrophobic layers have the highest contact angle and the lowest sliding angle, the S- PVDF and L-PVDF, which refer to the membranes 10 with small and large modified silica nanoparticles, respectively, were chosen for further study.

The surface maximum and mean pore sizes of as-prepared membranes 200, 10 were determined by a capillary flow porometer (model CFP 1500A, from Porous Material. Inc. (PMI) in Singapore). The membrane porosity was determined by the gravimetric method ^[141. The mechanical property of the membranes 200, 10 was measured using a Zwick/Roell BT1-FR0.5TN.D14 testing machine at a constant elongation velocity of 50 mm min-1 under room temperature (26 °C). Liquid entry pressure (LEPw) points of the membranes 200, 10 were measured using a dead-end cell with Milli-Q water. The pressure on the feed side was increased gradually while allowing it to stabilize for ten minutes after each 14 kPa (2 psi) increment. The pressure at which the first drop of permeate was obtained was the LEPw.

Table 2 lists the characteristic properties of PVDF nanofiber membrane 200, S-PVDF, L- PVDF composite membranes 10 and commercial PVDF membranes. As shown in Table 2, the PVDF nanofiber membrane 200 has 142.8° water contact angle, 0.68 pm surface mean pore size and 1.27 pm max pore size. The composite membranes S-PVDF and L-PVDF 10 have higher contact angles of 156.3° and 153.9°, respectively, with similar surface pore sizes that are much larger than the commercial Millipore hydrophobic membranes. In addition, the three electro-spun membranes 200, 10 exhibit higher porosity, around 83%, than the commercial membrane. For a fair comparison, all the membranes 200, 10 used in the DCMD tests have a similar thickness between 100 to 130 pm. Due to the surface superhydrophobicity, the S-PVDF and L-PVDF membranes 10 present a higher LEPw (liquid enter pressure) than the PVDF nanofiber membrane 200. The commercial PVDF has the highest LEP_W due to its smaller pore size. Other membrane properties are also included in table 2 below. PVDF S-PVDF L-PVDF

Membrane ID

(support layer) (selective layer)

PVDF PVDF

PVDFHSV900/SiO₂

Dope composition (wt. HSV900/DMF: HSV900/SiO₂

(large diameter) %) 8/92 (small diameter)

/DMF: 5/10/95

(0.004% wt LiCI) /DMF: 5/10/95

Dope flow rate

0.03 0.02

(mL min^"1)

Travel Speed (mm sec^"1) 0.1 0.1

Travel Distance (cm) 8 8

Distance (cm) 15 12

Voltage (kV) 28 28

Table 2: Characteristic properties of PVDF nanofiber membrane, S-PVDF, L-PVDF composite membranes and commercial PVDF membranes. Performance of resultant silica-PVDF composite nanofiber membranes

Prior to using the superhydrophobic membranes 10 for MD tests, their durability was examined by ultrasonic treatment. As shown in Fig. 9, compared with the L-PVDF membrane, the water contact angle of S-PVDF membrane decreased slightly after ultrasonic treatment. The possible reason is that a small amount of the modified silica particles on the S-PVDF membrane might be removed during the ultrasonic treatment, while the big particles on the L-PVDF membrane have better inter-tangled force with the polymer chains, making them more stable on the membrane surface. Nevertheless, both the S-PVDF and L-PVDF membranes still possess high contact angles above 150° and excellent water repellent properties after ultrasonic treatment for 30 minutes. Under FESEM, the surface morphologies of the S-PVDF and L-PVDF membranes as shown inset in Figs. 9A and 9B respectively also show no obvious difference before and after 30 min ultrasonic-treatment.

The MD performances of the PVDF nanofiber membrane 200, the composite membranes S- PVDF, L-PVDF 10 and a commercial PVDF membrane (Durapore® Membrane filter, Millipore, Singapore) were tested in a DCMD configuration using a 3.5 wt% NaCI solution as the feed solution under a temperature of 60°C and the permeate side was set at 20°C. The permeate flux of the composite membranes S-PVDF and L-PVDF 0 were 18.1 kg/m²hr and 18.9 kg/m²hr, respectively, while the PVDF nanofiber membrane 200 had a flux around 12.3 kg/m²hr and the commercial PVDF flux was about 10 kg/m²hr.

The composite membranes 10 showed a stable performance over 50 hours of testing time and produced high quality water with conductivity below 5.0 pS/cm. However, the conductivity of the product water from the PVDF nanofiber membrane 200 generally increased during 45 hours usage. The more stable performances of the composite membranes 10 are attributed to their greater hydrophobicity and the better water repellence of the membrane surface.

There are several possible explanations for the higher water flux of the composite membranes 0 compared with the PVDF nanofiber membrane 200. Firstly, when the hot salt solution on the feed side was flowing across the fresh PVDF nanofiber membrane 200 surface during the test, water droplets might have gradually entered into the nanofiber 200 sheet and accumulated between the nanofiber layers due to their loose overlap, which would reduce the temperature difference between the feed and permeation sides and thus decrease water flux significantly ^{[1 ]}. The FESEM images in Fig. 10A illustrate that a large gap appeared between the nanofiber layers due to the feed water accumulation.

Secondly, the superhydrophobic composite PVDF membranes 10 have higher surface porosity than the PVDF nanofiber membrane 200. As depicted in Fig. 11 , the meniscus at the membrane surface represents the effective liquid evaporation area. Compared with the PVDF nanofiber membrane 200 in which the feed solution is entrapped between nanofibers due to the Petal effect, the superhydrophobic composite membranes 10 have larger effective liquid evaporation areas because of the Lotus effect. As described previously, the hierarchical structure with increased roughness on a superhydrophobic surface 100 would provide numerous orifices to trap water vapour inside during the MD process, and this tends to reduce the contact area between the liquid and solid but increase the contact areas between the liquid and vapour. This type of structure not only provides the surface 100 with superhydrophobic and self-cleaning properties, it also provides more effective liquid areas to evaporate the water vapour and enhance the permeation flux in the MD process. Additionally, the superhydrophobic composite membranes 10 have thin regions of increased vapour and air entrapment, which could lower the effective thermal conductivity of the membranes 10 and thus reduce conductive heat losses and provide more driving force for evaporative transfer in the MD process. In contrast, as shown in Fig. 10D, the flux of commercial PVDF was around 10.0 kg/m²hr. The newly developed superhydrophobic PVDF membranes 10 are thus very competitive compared with the commercial flat-sheet PVDF membrane.

The stability of composite PVDF membranes in DCMD process was also investigated. According to the FESEM images shown in Figs. 10B and 10C, after being scoured by salt water for over 50 hours, the superhydrophobic S-PVDF and L-PVDF membranes 10 still have the similar surface morphology as previously. The modified silica-blended PVDF beads 120 could be observed on the composite membrane 10 surface. Additionally, the S-PVDF and L-PVDF membranes 10 still possess high water contact angle above 150° and excellent water repellence, while the surface of the PVDF nanofiber membrane 200 maintains a strong adhesive force with water droplets, which demonstrates that the electrospun membranes 200, 10 are durable in the DCMD application.

The present silica-PVDF composite nanofiber membrane 10 can be used in DCMD applications. It can also be used in other MD configurations which have similar requirements as DCMD to the membrane. In addition, it has the potential to be utilized in membrane contactors, including gas-liquid (G-L) contactor (gas absorption) and liquid-liquid (L-L) contactor processes (liquid-liquid extraction), which use hydrophobic microporous membranes.

Membrane pore size is usually limited by the concern of pore wetting when fabricating membranes for membrane distillation and contactor applications. As a breakthrough, this invention provides a nanofiber membrane substrate 200 with high porosity and large pore sizes, followed by the formation of a silica-PVDF composite selective layer 100, which has an extremely high water contact angle (>150°) and water repelling properties, on the top of the support 200 via the same electrospinning technique. The novel top layer 100 with a similar hierarchical structure to the lotus leaf makes the membrane 10 surface superhydrophobic to prevent the pores from wetting caused by the large pore sizes. As a result, the robust superhydrophobic composite nanofiber membrane 10 was able to outperform conventional pure polymeric hydrophobic membrane in terms of superior water flux and outstanding long term stability, suggesting that the formation of such composite nanofiber membranes 10 is an effective way to enhance the feasibility of membrane distillation processes for practical applications.

Whilst there has been described in the foregoing description exemplary embodiments of the present invention, it will be understood by those skilled in the technology concerned that many variations in details of design, construction and/or operation may be made without departing from the present invention. For example, while PVDF has been described above as the hydrophobic synthetic material from which the composite nanofiber membrane 10 is made, other suitable hydrophobic synthetic materials may be used in alternative embodiments of the membrane 10. While silica nanoparticles have been described as the nanoparticles dispersed in the hydrophobic synthetic material of the selective layer, other suitable hydrophobic nanoparticles may be used. REFERENCES

[I] P. Pal, A. Manna, Water Res. 2010, 44, 5750.

[2] X. Yang, R. Wang, A. G. Fane, J. Membr. Sci. 201 1 , 384, 52.

[3] K. Mohamed, Advances in Colloid and Interface Science 201 1 , 164, 56.

[4] X. Yang, R. Wang, L. Shi, A. G. Fane, M. Debowski, J. Membr. Sci. 2011 , 369, 437.

[5] B. Bhushan, Y. C. Jung, Prog. Mater Sci. 2011 , 56, 1.

[6] S. Wang, Y. Li, X. Fei, M. Sun, C. Zhang, Y. Li, Q. Yang, X. Hong, J. Colloid Interface Sci. 2011 , 359, 380.

[7] T. A. Egerton, G. D. Parfitt, Y. Kang, J. P. Wightman, Colloids and Surfaces 1983, 7, 311.

[8] P. Fabbri, M. Messori, M. Montecchi, S. Nannarone, L. Pasquali, F. Pilati, C. Tonelli,

M. Toselli, Polymer 2006, 47, 1055.

[9] N. Moussaif, C. Pagnoulle, J. Riga, R. Jerome, Polymer 2000, 41 , 3391.

[10] M. Toselli, J. A. Gardella, M. Messori, A. M. Hawkridge, F. Pilati, C. Tonelli, Polym.

Int. 2003, 52, 1262.

[I I] M. Sastry, J. Electron. Spectrosc. Relat. Phenom. 1997, 85, 167.

[12] G. Chen, X. Yang, R. Wang, A. G. Fane, Desalination 2013, 308, 47.

[13] L Feng, Y. Zhang, J. Xi, Y. Zhu, N. Wang, F. Xia, L. Jiang, Langmuir 2008, 24, 4114.

[14] Y. Liao, R. Wang, M. Tian, C. Qiu, A. G. Fane, J. Membr. Sci. 2013, 425^126, 30.

[15] W. Chen, A. Y. Fadeev, M. C. Hsieh, D. Oner, J. Youngblood, T. J. McCarthy, Langmuir 1999, 15, 3395.

Claims

1. A composite nanofiber membrane for membrane distillation, the composite nanofiber membrane comprising:

a support layer comprising a hydrophobic synthetic material; and

a selective layer formed on the support layer, the selective layer comprising hydrophobic nanoparticles dispersed in a same hydrophobic synthetic material as the hydrophobic synthetic material of the support layer.

2. The composite nanofiber membrane of claim 1 , wherein the selective layer has a water contact angle greater than 150°.

3. The composite nanofiber membrane of claim 1 or claim 2, wherein the selective layer has a sliding angle lower than 20°.

4. The composite nanofiber membrane of any preceding claim, wherein the support layer comprises a nanofiber membrane.

5. The composite nanofiber membrane of any preceding claim, wherein the support layer has a scaffold-like structure.

6. The composite nanofiber membrane of any preceding claim, wherein the hydrophobic nanoparticles on a surface of the composite nanofiber membrane are covered and protected by the hydrophobic synthetic material.

7. The composite nanofiber membrane of any preceding claim, wherein the hydrophobic synthetic material is polyvinylidene fluoride.

8. The composite nanofiber membrane of any preceding claim, wherein the hydrophobic nanoparticles are silica nanoparticles modified to be hydrophobic for dispersion in a solution of the hydrophobic synthetic material.

9. The composite nanofiber membrane of claim 8 when dependent on claim 7, wherein the silica nanoparticles are fluorinated.

10. A method of fabricating a composite nanofiber membrane, the method comprising:

a) e!ectrospinning a dope solution comprising a hydrophobic synthetic material onto a collector to form a support layer;

b) modifying nanoparticles to be hydrophobic for dispersion in a solution of the hydrophobic synthetic material;

c) dispersing the modified hydrophobic nanoparticles in the solution of the hydrophobic synthetic material to form a modified dope solution; and

d) electrospinning the modified dope solution onto the support layer to form a selective layer on the support layer.

11. The method of claim 10, wherein the hydrophobic synthetic material is polyvinylidene fluoride.

12. The method of claim 0 or claim 11 , wherein the nanoparticles are silica nanoparticles.

13. The method of claim 12 when dependent on claim 11 , wherein the modifying comprises fluorinating the silica nanoparticles.

14. The method of claim 13, wherein the fluorinating comprises stirring silica nanoparticles in an N-hexane solution comprising a, ω-triethoxysilane terminated perfluoropolyether and tetraethoxysilane.

15. The method of claim 14, further comprising centrifuging the fluorinated silica nanoparticles from the solution and annealing the fluorinated silica nanoparticles.