Journal Pre-proof

Biomimetic superhydrophobic membrane for membrane distillation with robust wetting

and fouling resistance

Deyin Hou, Kofi S.S. Christie, Kai Wang, Min Tang, Dewu Wang, Jun Wang

PII: S0376-7388(19)31931-3
DOI: https://doi.org/10.1016/j.memsci.2019.117708
Reference: MEMSCI 117708

To appear in: Journal of Membrane Science

Received Date: 27 June 2019

Revised Date: 27 October 2019
Accepted Date: 29 November 2019

Please cite this article as: D. Hou, K.S.S. Christie, K. Wang, M. Tang, D. Wang, J. Wang, Biomimetic
superhydrophobic membrane for membrane distillation with robust wetting and fouling resistance,
Journal of Membrane Science (2019), doi: https://doi.org/10.1016/j.memsci.2019.117708.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier B.V.

Graphical Abstract
 Biomimetic superhydrophobic membrane for membrane
distillation with robust wetting and fouling resistance

Revised Manuscript

MEMSCI_2019_1834_R1

Submitted to Journal of Membrane Science

Oct. 26th 2019

Deyin Houa,b*, Kofi S.S. Christiec, Kai Wangd, Min Tangb, Dewu Wangd, Jun Wanga,b

a
Key Laboratory of Drinking Water Science and Technology, Research Center for
Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China
b
College of Resources and Environment, University of Chinese Academy of Sciences,
Beijing 100190, China
c
Department of Civil and Environmental Engineering, Vanderbilt University, Nashville,
Tennessee 37235-1831, United States
d
School of Chemical Engineering and Technology, Hebei University of Technology,
Tianjin 300130, China

*Corresponding author: Deyin Hou

E-mail: dyhou@rcees.ac.cn
Tel.: +86 10 62917207

1
 1 Abstract: Membrane distillation (MD) is a promising membrane-based thermal process

2 capable of desalinating highly saline water. However, its application is limited by fouling

3 and wetting of commercial hydrophobic membranes. Inspired by the lotus leaf, we

4 developed a biomimetic superhydrophobic polyvinylidene fluoride (PVDF) membrane

5 for robust MD via self-assembly method. The hierarchical micro-nanoscale texture on

6 the membrane surface was constructed by grafting the spherical

7 polyvinylsilsesquioxane (PVSQ) nanoparticles onto micron-sized silica particles

8 (SiPs). The membrane surface energy can be simultaneously lowered due to the

9 hydrophobic groups including vinyls and methoxyls created from the condensation

10 reaction of the vinyltrimethoxysilane (VTMOS). The resulting membrane showed a

11 very high water contact angle (~160°) and a low water sliding angle (<15°),

12 demonstrating the strong hydrophobicity that is expected for a lotus-leaf-like surface.

13 Compared to the commercial PVDF membrane, the fabricated superhydrophobic

14 membrane exhibited superior resistance against wetting by a surfactant (sodium

15 dodecyl sulfate) and fouling by humic acid during MD experiments. Our results

16 suggest that biomimetic superhydrophobic membranes can enhance the operational

17 robustness of MD and facilitate the efficient use of MD for desalinating saline

18 wastewaters in favor of beneficial water reuse and resource recovery.

19 Keywords: Membrane distillation; Superhydrophobic surface; Self-assembly; Membrane

20 wetting; Membrane fouling

2
 1 1. Introduction

2 With increasing stress on water resources and growingly stringent regulations on

3 liquid waste discharge, zero liquid discharge (ZLD) has been on the rise globally [1].

4 Essential to ZLD are effective brine management technologies that are both

5 economical and operationally reliable [2]. Membrane distillation (MD) has been

6 recognized as one of the promising desalination technologies for further recovering

7 water and concentrating the brine after a feed solution has been treated using reverse

8 osmosis in a ZLD treatment system [3]. The major advantages of MD include the

9 capability of treating hypersaline brine using low-grade thermal energy, near-perfect

10 rejection of non-volatile species, and the relatively low capital cost compared to

11 conventional thermal distillation processes especially when it is employed as a

12 small-scale and modular unit process [4-7]. For these reasons, research and

13 development related to MD for hypersaline brine treatment has increased dramatically

14 in recent years.

15 Vapor transfer in MD is driven by either a temperature- or vacuum-induced

16 partial vapor pressure difference across the MD membrane [8]. Regardless of specific

17 system configurations, an MD membrane always serves as a barrier for direct liquid

18 permeation but a medium for vapor transfer [9]. Therefore, an MD membrane must be

19 hydrophobic to prevent the intrusion of feed solution [10]. Materials of commercial

20 membranes that are typically used in MD studies include polyvinylidene fluoride

21 (PVDF), polypropylene (PP), and polytetrafluoroethylene (PTFE) [11]. These

22 hydrophobic commercial membranes perform well in desalinating relative clean

3
 1 saline water such as seawater and brackish water [12,13]. However, their performance

2 can be significantly compromised by feed waters containing hydrophobic or

3 amphiphilic contaminants. Specifically, hydrophobic contaminants are prone to attach

4 onto the surface of these membranes via long-ranged and attractive

5 hydrophobic-hydrophobic interactions [14], resulting in pore blocking and flux

6 decline in long-term MD operation [15]. Amphiphilic contaminants, on the other hand,

7 can actively adsorb onto pore walls within a membrane and change the wetting

8 properties of the membrane pores, which can eventually lead to pore wetting and poor

9 salt rejection [16,17]. These limitations prevent the use of commercial hydrophobic

10 membranes in MD for treating complex wastewaters such as oil and gas produced

11 water and wastewater from a variety of industries [18,19].

12 Nature provides many inspirations for developing novel biomimetic materials.

13 A well-known example is the lotus-leaf which has inspired material scientists to

14 develop superhydrophobic surfaces with similar structures. According to the

15 well-documented “lotus effect,” the microscopic surface architecture minimizes the

16 adhesion of water to the surface, resulting in very high water contact angle and very

17 low sliding angle [20]. Spherical water droplets can roll on a lotus leaf easily and

18 collect any contaminants on the surface, which is described as a self-cleaning effect

19 [21]. In general, the development of a superhydrophobic surface requires both large,

20 preferably hierarchical, surface roughness and low surface energy [22]. Following this

21 principle, a wide variety of the artificial superhydrophobic surfaces have been

22 developed using different methods such as lithographic patterning, plasma etching,

4
 1 anode oxidation, sol-gel method, phase separation, electrospinning, and chemical

2 etching [23-25].

3 The development of superhydrophobic membranes have also attracted significant

4 interest in MD applications. A superhydrophobic glass membrane for MD was

5 prepared through a process that involved glass fiber drawing, dissolving template

6 material and chemical etching [26], however, the fabricated membrane presented a

7 low permeate flux. As a versatile approach that can create nanometer- or

8 submicrometer-sized fibrous structures, electrospinning was also introduced to

9 fabricate superhydrophobic membranes [27,28], but the mechanical strength of the

10 fibrous membranes was unsatisfactory for industrial MD applications. Additionally,

11 surface modification techniques, such as spray coating and plasma etching [29,30],

12 were used to prepare superhydrophobic MD membranes. It has been a common

13 practice for the aforementioned methods to introduce fluoro-silanes to lower

14 membrane surface energy to obtain superhydrophobic surfaces [31,32], which are

15 costly and environmentally unfriendly. Among the many modification techniques,

16 self-assembly is one of the simplest for the creation of hierarchical structures on a

17 membrane surface. The self-assembly technique displays a good potential for

18 scalability, and it is relatively cost-effective and environmentally friendly [33].

19 Herein, inspired by the lotus leaf, we aim to fabricate a novel superhydrophobic

20 membrane for MD by conducting self-assembly on a commercial PVDF membrane.

21 The hierarchical micro-nanoscale rough surface structures are to be constructed by

22 synthetic silica composite particles comprising modified micro-sized silica particles

5
 1 (SiPs) with spherical polyvinylsilsesquioxane (PVSQ) nanoparticles grafted onto the

2 SiPs surfaces. The surface energy of the prepared membrane can be simultaneously

3 lowered by the hydrophobic groups, including vinyls and methoxyls, created from the

4 condensation reaction of the fluoride-free SiPs modifier vinyltrimethoxysilane

5 (VTMOS). Additionally, two suites of MD desalination tests were carried out, one

6 included a surfactant (sodium dodecyl sulfate, SDS), while the other was in the

7 presence of a type of natural organic matter (humic acid, HA). As a proof-of-concept,

8 the resultant superhydrophobic membrane presented robust MD performance

9 compared with the neat PVDF membrane. The facile and environmentally-friendly

10 approach developed in this study provides a new perspective for fabrication of

11 superhydrophobic membranes for robust MD operation. Due to the superior

12 anti-wetting and anti-fouling properties, the fabricated membrane is promising and

13 feasible for desalination with MD, especially for challenging saline wastewaters with

14 strong fouling or pore wetting potential.

15 2. Material and Methods

16 2.1. Materials

17 The commercial PVDF hydrophobic flat sheet membrane with a mean pore size of

18 about 0.75 µm was obtained from Millipore Co., Ltd. Silica particles (SiPs, 99.99%)

19 with a size of about 10 µm, ammonium hydroxide solution (AR grade, 25%),

20 vinyltrimethoxysilane (VTMOS, 98%), humic acid (HA, ≥90%) and calcium chloride

21 (Anhydrous, 96%) were supplied from Macklin Co, Ltd. Absolute ethanol (EtOH,

22 99%), sodium chloride (Anhydrous, >99.5%) and sodium hydroxide (AR grade, 96%)

6
 1 were purchased from Sinopharm Chemical Reagent Co., Ltd. Sodium dodecyl sulfate

2 (SDS, ≥99%) was acquired from Sigma-Aldrich. The deionized water was from a

3 Milli-Q ultrapure water system (Millipore, USA).

4 2.2. Membrane preparation

5 2.2.1. Preparation of the hydrophobic SiPs-PVSQ composite sol

6 First, SiPs were slowly dispersed in 200 mL EtOH under magnetic stirring for 0.5 h at

7 room temperature. NH3·H2O was then added into the SiPs suspension until the pH

8 reached 10. VTMOS of four times SiPs mass was then added into the suspension

9 dropwise. The mixture was sonicated for 2 h at 40 oC and then stirred for another 10 h

10 at the same temperature to promote complete modification of the SiPs. During the

11 SiPs hydrophobic modification by VTMOS, the spherical PVSQ nanoparticles were

12 introduced and grafted onto the SiPs surface as intended by the condensation reaction

13 of the VTMOS in the presence of NH3·H2O. The resulting dispersion is referred to as

14 the hydrophobic SiPs-PVSQ composite sol.

15 2.2.2. Fabrication of the superhydrophobic membrane

16 The commercial PVDF membrane was modified by the SiPs-PVSQ particles

17 following a procedure illustrated in Fig. 1. The detailed composition of different

18 SiPs-PVSQ composite sol used for membrane modification and the corresponding

19 resultant composite membrane are list in Table S1. The commercial PVDF membrane

20 with a size of 128×115 mm2 was first treated with a 12.0 mol/L NaOH solution at 60

o
21 C for 2 h to introduce hydroxyl groups to the membrane surface. The alkali-treated

22 PVDF (PVDF-OH) membrane was immersed into the SiPs-PVSQ composite sol

7
 1 solution for 0.5 h at 40 oC to absorb the SiPs-PVSQ composite particles onto the

2 PVDF-OH membrane. Finally, the modified membrane was air-dried at room

3 temperature.

4
5 Fig. 1. Graphical illustration of the self-assembly technique for the surface modification of a
6 commercial PVDF membrane.

7 2.3. Membrane characterization

8 The morphology, porosity, wettability and chemical composition were characterized

9 using the techniques as described below. The morphology of the pristine commercial

10 PVDF membrane and the modified membranes were investigated using scanning

11 electron microscopy (SEM) (Hitachi SU-70, Japan). The membrane surface

12 topography was characterized using atomic force microscopy (AFM) (NanoScope Ⅲ

13 a, Digital Instruments, USA). For each membrane sample, an area of 5×5 µm2 was

14 scanned using the tapping mode.

15 The porosity of the membranes was measured using the gravimetric method [34],

16 the membrane sample with a certain area should be fully wetted first and then the wet

17 and dry weights of the membrane sample were measured by using a Halogen

18 Moisture Meter (HR83, Mettler-Toledo, Switzerland). The mean pore size and pore

8
 1 size distribution of the membranes were determined by using a Capillary Flow

2 Porometer (CFP) (Porolux 1000, IB-FT GmbH, Germany) with a low-surface-tension

3 (16 mN/m) wetting agent. The liquid entry pressure of water (LEPW) of the membrane

4 was determined following the method described by Smolders and Franken [35], by

5 connecting an ultrafiltration cup to the CFP.

6 The water contact angle (WCA) of each membrane was measured using an optical

7 goniometer (OCA20, Data Physics Instruments Ltd., Germany). The water sliding angle

8 (WSA) of each membrane was determined by measuring the minimum tilting angle

9 (from a horizontal position) for a 10 µL water droplet to roll off. Both WCA and

10 WSA measurements were taken using five randomly selected spots and the average

11 values are reported.

12 Fourier transform infrared spectroscopy (FTIR) analyses were conducted to

13 analyze the distribution of functional groups on the membrane surface, using a FTIR

14 Spectrometer (Perkin Elmer, USA) with attenuated total reflectance (ATR). The IR

15 spectra of the pressed membrane samples were scanned at a resolution of 4 cm-1 from

16 650-4000 cm-1 with 64 scans for each sample.

17 2.4 Membrane performance in membrane distillation

18 Both the commercial PVDF membrane and the modified membranes were tested in

19 direct contact MD (DCMD) experiments using a setup illustrated in Fig. 2. The

20 membrane module has a plate-and-frame configuration with effective membrane area

21 of 7.47×10-3 m2. The feed and distillate solutions flowed co-currently through the

22 module with the same flow velocity of 0.35 m/s. The feed and distillate temperatures

9
 1 were maintained at 53 °C and 20 °C, respectively. The distillate weight and

2 conductivity were recorded automatically at set intervals, and the salinity of the

3 distillate was calculated from the measured conductivity based on a pre-established

4 calibration curve. The permeate flux and the salt rejection were calculated from the

5 rates of change in the distillate mass and salinity. The permeate flux was calculated by

6 the following equation:

∆W
7 J= (1)
A ⋅ ∆t

8 where J is the permeate flux (kg/m2h), ∆W is the mass of the permeate (kg), A is the
9 effective area of the used membrane (m2) and ∆t is the time interval (h). The salt
10 rejection R of the membrane was calculated according to the following equation:

C f −C p
11 R=
Cf

12 (2)
13 where Cf and Cp are the salt concentration of the feed and the permeate, respectively.

14 To compare the wetting resistance of the modified superhydrophobic membrane

15 and the commercial PVDF membrane, SDS was added to the feed solution (3.5 wt%

16 sodium chloride solution) to reduce its surface tension. The SDS concentration in the

17 feed after sequential additions was 0.1, 0.2, and 0.3 mM, and the surface tension of

18 the corresponding saline feed was about 42.4, 34.8, and 30.5 mN/m. The fouling

19 resistance of the modified membrane was evaluated using DCMD experiments with a

20 feed solution comprising NaCl (100 g/L), HA (50 mg/L) and CaCl2 (20 mM).

10
 1

2 Fig. 2. Schematic of the experimental setup for DCMD.

3 3. Results and discussion

4 3.1 Membrane surface morphology

5 The SEM images featuring the surface morphology of the commercial PVDF

6 membrane and the modified membranes are shown in Fig. 3. The pristine commercial

7 PVDF membrane has a sponge-like porous structure resulting from thermal induced

8 phase separation. In contrast, the surface of the modified membrane is fully covered

9 with micron-sized SiPs (9.56±4.82 µm) that are decorated with nano-sized PVSQ

10 particles, which resemble the papillae on natural lotus leaf surfaces [36]. The

11 micron-sized SiPs and nano-sized PVSQ roughness together create a two-scale

12 hierarchical textured surface, which is critical for imparting superhydrophobicity [37].

13 With increasing concentration of the SiPs-PVSQ composite sol, more and more

14 of the membrane surface is covered by the SiPs-PVSQ composite particles, which is

15 demonstrated by the energy dispersive spectroscopy (EDS) mapping images of the

16 relevant elements distribution on the surfaces of the commercial PVDF membrane

17 and the modified membranes (Fig. S1). The entire surfaces of the PVDF-M3 and

18 PVDF-M4 modified membranes are ultimately bestrewn with silica composite

11
 1 spheres. The hierarchical micro-nanoscale rough surface structure of the modified

2 membrane might contribute to sustaining a metastable Cassie-Baxter state for the

3 liquid-solid-vapor interface, which can reduce the solid-liquid contact area and

4 improve membrane surface hydrophobicity to some extent [38].

8 Fig. 3. SEM images of the surfaces of the commercial PVDF membrane and the modified
9 membranes, and the synthetic SiPs-PVSQ composite particles.

10 3.2 Membrane surface FTIR analysis

11 The FTIR spectrum of the commercial PVDF membrane presented main absorption

12 peaks at 765, 840, 875, 1070, 1183, 1404 and 3025 cm-1 wavenumbers (Fig. 4), which

12
 1 indicates that the polymer PVDF in the commercial membrane primarily consisted of

2 α-phase (peaks at 765, 1070 and 1404 cm-1) and β-phase (peaks at 840 and 1183 cm-1)

3 crystals. The peaks at 3025 cm-1 and 1183 cm-1 correspond to CF2 stretching vibration

4 and bending vibration, respectively. There is a strong absorption peak observed at 875

5 cm-1 characteristic of the C-C-C skeleton vibration, while the peak at 1404 cm-1

6 corresponds to CH2 deformation vibration.

7 In the FTIR spectrum of the alkali-treated PVDF-OH membrane, the CF2

8 stretching vibration (at 3025 cm-1) is weakened significantly and a broad adsorption

9 band around 3100-3500 cm-1 (-OH stretching vibration) can be observed, which

10 suggests that the alkali treatment led to the defluorination of PVDF polymer and

11 formation of surface hydroxyl groups. For the modified membranes, a strong and

12 broad absorption band is observed around 1100 cm-1 which is characteristic for

13 Si-O-Si anti-symmetric stretching vibration. In addition, the peaks at 1183 cm-1 and

14 1070 cm-1 overlaps with the absorption peak of Si-O-Si. The absorption peak around

15 800 cm-1 corresponds to the Si-C stretching vibration. In addition, with the composite

16 sol content increasing, both the adsorption peaks at 1602 cm-1 (-CH=CH2 stretching

17 vibration) and 2958 cm-1 (-CH3 anti-symmetric stretching vibration) become stronger,

18 which indicates the successful surface grafting of hydrophobic vinyl and methoxyl

19 groups.

13
 1

2 Fig. 4. ATR-FTIR spectra of the commercial PVDF membrane and the modified PVDF membranes.

3 3.3 Membrane surface wettability

4 The surface wettability is very different for the pristine PVDF membrane and the

5 modified membranes. As shown in Fig. 5, the commercial PVDF membrane has a

6 WCA of 136.3°, while all the WCAs of the modified membranes are higher than 150°.

7 Increasing the concentration of the SiPs-PVSQ composite sol in modification

8 increases the WCA and reduces the WSA. Specifically, the WCAs of the PVDF-M3

9 and PVDF-M4 membranes are both above 160°, whereas their WSAs are below 20°.

10 The very high WCA and low WSA suggest that wetting of the modified PVDF

11 membranes, especially PVDF-M3 and PVDF-M4, is in an excellent Cassie-Baxter

12 state [39]. The water droplet on the surface of the modified membranes was actually

13 supported by a composite interface containing the solid hierarchical texture and the

14 vapor/air entrapped among the composite particles, which meant an increase of the

15 liquid-vapor interface and a reduce of the solid-liquid contact area. Meanwhile, the air

16 packets trapped underneath the liquid droplet also increased liquid mobility on
14
 1 membrane surface. In contrast, the WSA of the pristine PVDF membrane was

2 unmeasurable (i.e. the water droplet does not roll off even on a vertical membrane

3 surface), which indicates that wetting of water on a pristine PVDF membrane is in a

4 Wenzel state [40].

80

160
60
Contact angle (o)

Sliding angle (o)

150
40

140
20

130 0
PVDF PVDF-M1 PVDF-M2 PVDF-M3 PVDF-M4
5 Membrane
6 Fig. 5. Water contact angles and sliding angles of the commercial PVDF membrane and the
7 modified membranes.

8 AFM topography measurements suggest that the surface modification increases

9 the surface roughness of the membranes (Fig. 6). Specifically, the average surface

10 roughness (Ra) of the membranes were 132 nm, 234 nm, 275 nm, 359 nm and 402 nm

11 for the pristine PVDF membrane, PVDF-M1, PVDF-M2, PVDF-M3 and PVDF-M4

12 membranes, respectively. The corresponding root mean square surface roughness (Rq)

13 of these membranes were 176 nm, 298 nm, 378 nm, 466 nm and 516 nm, respectively.

14 The presence of the SiPs-PVSQ composite particles significantly enhances the surface

15 roughness, which is consistent with the SEM morphology shown in Fig. 3. The

16 rougher surface morphology is capable of enhancing hydrophobicity and reducing

15
 1 liquid pinning based on the Cassie-Baxter theory [39].

3 Fig. 6. AFM images of the surfaces of the commercial PVDF membrane and the modified
4 membranes.

5 3.4 Membrane structural properties

6 After surface modification, the membrane LEPW value was improved, which

7 contributed to membrane performance stability during MD application. Compared

8 with the commercial PVDF membrane, the pores of the modified membranes became

9 smaller and pores distribution became narrower (Fig. S2 and Table 1). At the same

10 time, the porosity of the modified membranes also decreased. In general, the higher

11 thickness (due to the addition of the surface modification layer) and lower porosity

12 would increase the vapor transfer resistance during MD process [41], which may

13 compromise the distillate flux of the modified membranes. Because PVDF-M4 was

14 thicker and less porous than PVDF-M3 (Table 1) but the WCA and WSA of the two

15 membranes were similar (Fig. 5), we chose PVDF-M3 for more comprehensive

16 investigation of MD performance.
17 Table 1 The structural properties of the commercial PVDF membrane and the modified
18 membranes.
Porosity Thickness Mean pore diameter Maximum pore size
Membrane LEPW (bar)
(%) (µm) (µm) (µm)
PVDF 77.6±2.2 102.5±1.7 0.75±0.05 1.13±0.08 3.15±0.18
PVDF-M1 75.1±1.7 115.8±2.1 0.72±0.03 1.12±0.11 3.27±0.23
16
 PVDF-M2 73.3±2.6 118.6±3.4 0.70±0.07 1.10±0.09 3.34±0.17
PVDF-M3 70.8±1.5 120.3±2.2 0.65±0.06 1.05±0.07 3.51±0.13
PVDF-M4 66.2±3.1 134.5±2.8 0.58±0.02 1.07±0.10 3.53±0.15

1 3.5 Mechanism of membrane surface modification

2 Fig. 3 and Fig. 6 suggest that a hierarchical surface morphology was created by the

3 grafted SiPs-PVSQ composite particles via self-assembly. The proposed mechanism

4 of the surface modification is described here with illustration shown in Fig. 7. In the

5 presence of NH3·H2O, the SiPs dispersed in EtOH undergoes hydrolysis and acquires

6 surface hydroxyls (-OH). Therefore, the VTMOS can be grafted onto the SiPs surface

7 through the condensation reaction between the methoxy group of the VTMOS and the

8 hydroxyls coated on SiPs surface (Fig. 7a). Due to the abundant hydroxyl groups on

9 the SiPs surface, a large number of VTMOS molecules were grafted onto the SiPs

10 surface. The self-hydrolysis and condensation reactions of the VTMOS molecules

11 produced spherical nanoscale PVSQ particles as illustrated in Fig. 7b, leading to the

12 hierarchical surface structure with two scales of roughness. Meanwhile, the

13 hydrophobic groups such as vinyls and methoxyls resulting from the condensation

14 reaction of VTMOS lowered the surface energy of the SiPs-PVSQ composite particles.

15 The low surface energy and hierarchical surface roughness resulted in

16 superhydrophobicity of the modified surface.

17 As a result of the high alkalinity of the aqueous solution and high temperature,

18 the PVDF polymer in the commercial membrane surface was attacked by the strong

19 base. The deprotonation of CH2 group and occurred readily, and the polymer chain

20 fragment rearranged to yield C=C double bonds, which is known as an elimination

17
 1 reaction. In an alkaline environment, the C-F cleavage reaction replaced the fluoride

2 with a hydroxyl group due to the high activity of hydroxide [42]. Through

3 alkali-treatment, the hydroxyl groups were introduced into the commercial PVDF

4 membrane surface and the commercial membrane could be used for self-assembly

5 modification. The alkali-treated PVDF-OH membrane offered active sites for

6 self-assembly, and the membrane surface could be successfully grafted with the

7 hydrophobic SiPs-PVSQ composite particles via the condensation reaction between

8 the methoxyl groups of the SiPs-PVSQ composite particles and the hydroxyl groups

9 in the PVDF membrane surface (Fig. 7d).

10
11 (a) Hydrophobic modification of silica particles.

12
13 (b) Self-hydrolysis and condensation of the VTMOS.

18
 1
2 (c) Alkaline treatment of the PVDF membrane surface.

3
4 (d) SiPs-PVSQ composite particles grafted on PVDF membrane surface.

5 Fig. 7. Proposed mechanism for the surface modification of the commercial PVDF membrane.

6 3.6 Membrane performance in MD

7 3.6.1 Wetting Resistance

8 To compare the wetting resistance of the commercial PVDF membrane and the

9 modified membrane (PVDF-M3 as the selected example), DCMD experiments were

10 performed in the presence of the surfactant SDS. To accelerate the wetting process,

11 SDS was added to the feed solution progressively every two hours to increase the

12 SDS concentration (up to 0.3 mM). To keep the feed concentrations of salt and

13 surfactant in a reasonable range, the distillate was recirculated back to the feed

14 reservoir every half hour.

15 With the commercial PVDF membrane, wetting occurs even at the lowest SDS

16 concentration (0.1 mM), as indicated by the increased trans-membrane flux and

17 dramatically reduced salt rejection (Fig. 8). With increasing SDS concentration in the

18 feed solution, the membrane wetting became progressively more severe. These results

19 indicated that the commercial PVDF membrane was prone to wetting induced by the
19
 1 surfactant due to its weak surface hydrophobicity. When the SDS concentration

2 increased to 0.3 mM, the trans-membrane flux of the PVDF membrane increased

3 more than 500% and the salt rejection dropped below 30%. In contrast, the

4 performance of the PVDF-M3 membrane was very robust, even with a feed solution

5 of 0.3 mM SDS. Additionally, after wetting resistance test, we rinsed the PVDF-M3

6 membrane with DI water to clean the residual surfactant on membrane surface and

7 then dried it in air at room temperature. It was verified by membrane surface

8 wettability measurement that the PVDF-M3 membrane still maintained

9 superhydrophobic. The robust wetting resistance of the modified membrane may be

10 attributable to the re-entrant structure created by the assembly of micron-sized

11 SiPs-PVSQ spheres with low surface energy (the membrane surface tension is listed

12 in Supporting Information Table S2).

13

14 Fig. 8. Anti-wetting performance of the commercial PVDF membrane and the modified membrane
20
 1 at different SDS concentrations in DCMD. A 3.5 wt% sodium chloride solution with varying SDS
2 concentration was used as the feed. The flow rate of feed and permeate streams were both 90 L/h.
3 The feed and distillate temperatures were 53°C and 20°C, respectively. The initial permeate fluxes
4 for the commercial PVDF membrane and the modified membrane were 10.5 kg/m2h and 9.8
5 kg/m2h, respectively.

6 3.6.2 Fouling Resistance

7 To compare the fouling resistance of the commercial PVDF membrane and the

8 modified membrane, DCMD experiments were conducted for 150 h using a feed

9 solution comprising NaCl (100 g/L), CaCl2 (20 mM) and HA (50 mg/L). To maintain

10 constant feed concentration, the distillate was fed back to the feed tank every half

11 hour. The initial flux of the PVDF-M3 membrane was slightly lower compared with

12 the pristine PVDF membrane (Fig. 9), most likely due to the reduced membrane pore

13 size and porosity and the increased membrane thickness as summarized in Table 1.

14 However, during the 150 h DCMD experiment, the permeate flux of the pristine

15 PVDF membrane declined significantly, while the PVDF-M3 membrane exhibited a

16 relatively stable flux. At the end of the experiment, the permeate flux of the pristine

17 PVDF membrane declined by about 30%. In comparison, the flux decline of the

18 PVDF-M3 membrane was less than 10%.

21
 12 50

11 PVDF

Permeate conductivity (µS/cm)

PVDF-M3 40
Permeate flux (kg/m2⋅h)

10
30
9
20
8

10
7

6 0
0 30 60 90 120 150
1 Time (hr)
2 Fig. 9. Anti-fouling performance of the commercial PVDF membrane and the modified membrane
3 in DCMD. The mixed solution including NaCl (100g/L), CaCl2 (20 mM) and HA (50 mg/L) was
4 used as the feed. The flow rate of feed and permeate streams were both 90 L/h. The feed and
5 distillate temperatures were 53°C and 20°C, respectively.

6 Based on HSAB theory, as a hard acid, the divalent cation Ca2+ can easily

7 interact with a hard base such as -COO-. Because HA contains negatively charged

8 carboxyl groups, Ca2+ can act as a binding agent of two carboxyl functional groups to

9 form complexes, which may accelerate HA aggregation. The commercial PVDF

10 membrane is easily fouled as the accumulation of aggregated HA blocked the

11 membrane pores. Meanwhile, the foulants absorbing into the pores can also induce

12 pore wetting as indicated by the gradual increase in distillate conductivity [15]. In

13 contrast, superhydrophobic PVDF-M3 membrane has stronger fouling resistance

14 because the hierarchical micro-nanoscale rough structure minimizes the contact

15 between feed solution and the solid membrane matrix, i.e., there is significantly less

16 solid/water interface for the foulants to adhere to. Foulants that temporarily adhere to

22
 1 the water/air interface can be readily removed by the cross-flow. This also explains

2 the lack of pore wetting with the PVDF-M3 membrane as indicated by a measured

3 distillate conductivity that was practically unchanged during the entire experiment.

4 The SEM images of the commercial PVDF membrane and the PVDF-M3

5 membrane after the DCMD experiments (Fig. 10) also provide direct evidence that

6 the modified membrane was more resistant to fouling than the commercial PVDF

7 membrane. It is observed that the pristine PVDF membrane surface was completely

8 covered by clumpy foulants. The EDS analysis reveals that the clumpy deposits on the

9 membrane surface mainly comprised HA, CaCl2 and NaCl. In contrast, the PVDF-M3

10 membrane surface is visually free of such deposit. Besides, the EDS mapping images

11 of the relevant typical elements distribution on cross-sections of the used membranes

12 (Fig. S3) also indicated that the modified membrane had robust fouling resistance.

13 The different wetting and fouling resistance between the pristine PVDF

14 membrane and the PVDF-M3 membrane suggests that surface wettability of a

15 membrane has a critical impact on the membrane performance in long-term MD

16 operation with feed solution of strong fouling or wetting propensity. The robust

17 anti-wetting and anti-fouling performance of the PVDF-M3 membrane may enable

18 MD for desalinating “more challenging” hypersaline feed water, such as shale gas

19 produced water, electroplating wastewater, and wastewater from dyeing and other

20 industries.

23
 1

3 Fig. 10. SEM-EDS analysis of the membrane surface after DCMD experiments.

4 4. Conclusions

5 In this study, we developed a biomimetic superhydrophobic MD membrane using a

6 novel method of constructing multi-scale hierarchical re-entrant texture on the surface

7 of a PVDF membrane. The hierarchical micro-nanoscale structure was constructed by

8 grafting micron-sized silica particles (SiPs) decorated with polyvinylsilsesquioxane

9 (PVSQ) nanoparticles onto the membrane surface. The hydrophobic vinyl and

10 methoxyl groups generated from the condensation reaction of the

11 vinyltrimethoxysilane (VTMOS) impart low surface energy on the modified

12 membrane. The modified membrane demonstrates excellent water repellence as

13 characterized by high water contact angles and very low sliding angles. It also

14 delivers far better fouling and wetting resistance than the pristine PVDF membrane.

15 Our results suggest that the biomimetic superhydrophobic membrane is promising for
24
 1 enhancing the operational robustness of MD in the presence of fouling and wetting

2 agents, which may potentially enable the use of MD for desalinating hypersaline

3 brines of complex composition.

4 Acknowledgements

5 The authors wish to acknowledge the valuable feedbacks and suggestions in writing

6 the manuscript form Dr. Shihong Lin at Department of Civil and Environmental

7 Engineering, Vanderbilt University. Financial support provided by National Natural

8 Science Foundation of China (Grant No. 51678555) and the National Key R&D

9 Program of China (Grant No. 2016YFC0400500) are gratefully acknowledged.

10 References

11 [1] T. Tong, M. Elimelech, The global rise of zero liquid discharge for wastewater

12 management: Drivers, technologies, and future directions, Environ. Sci. Technol. 50

13 (2016) 6846-6855.

14 [2] G.U. Semblante, J.Z. Lee, L.Y. Lee, S.L. Ong, H.Y. Ng, Brine pre-treatment

15 technologies for zero liquid discharge systems, Desalination 441 (2018) 96-111.

16 [3] R. Schwantes, K. Chavan, D. Winter, C. Felsmann, J. Pfafferott, Techno-economic

17 comparison of membrane distillation and MVC in a zero liquid discharge application,

18 Desalination 428 (2018) 50-68.

19 [4] H.C. Duong, P. Cooper, B. Nelemans, T.Y. Cath, L.D. Nghiem, Evaluating energy

20 consumption of air gap membrane distillation for seawater desalination at pilot scale

21 level, Sep. Purif. Technol. 166 (2016) 55-62.

22 [5] E. Drioli, A. Ali, F. Macedonio, Membrane distillation: Recent developments and

25
 1 perspectives, Desalination 356 (2015) 56-84.

2 [6] A.M. Alklaibi, N. Lior, Membrane-distillation desalination: Status and potential,

3 Desalination 171 (2005) 111-131.

4 [7] G. Naidu, S. Jeong, Y. Choi, S. Vigneswaran, Membrane distillation for

5 wastewater reverse osmosis concentrate treatment with water reuse potential, J.

6 Membr. Sci. 524 (2017) 565-575.

7 [8] M.S. El-Bourawi, Z. Ding, R. Ma, M. Khayet, A framework for better

8 understanding membrane distillation separation process, J. Membr. Sci. 285 (2006)

9 4-29.

10 [9] B.B. Ashoor, S. Mansour, A. Giwa, V. Dufour, S.W. Hasan, Principles and

11 applications of direct contact membrane distillation (DCMD): A comprehensive

12 review, Desalination 398 (2016) 222-246.

13 [10] L. Eykens, K.D. Sitter, C. Dotremont, L. Pinoy, B.V. Bruggen, Membrane

14 synthesis for membrane distillation: A review, Sep. Purif. Technol. 182 (2017) 36-51.

15 [11] J. Zhang, N. Dow, M. Duke, E. Ostarcevic, J.D. Li, S. Gray, Identification of

16 material and physical features of membrane distillation membranes for high

17 performance desalination, J. Membr. Sci. 349 (2010) 295-303.

18 [12] J.P. Mericq, S. Laborie, C. Cabassud, Vacuum membrane distillation of seawater

19 reverse osmosis brines, Water Res. 44 (2010) 5260-5273.

20 [13] S. Yarlagadda, V.G. Gude, L.M. Camacho, S. Pinappu, S. Deng, Potable water

21 recovery from As, U, and F contaminated ground waters by direct contact membrane

22 distillation process, J. Hazard. Mater. 192 (2011) 1388-1394.

26
 1 [14] E.E. Meyer, K.J. Rosenberg, J. Israelachvili, Recent progress in understanding

2 hydrophobic interactions, Proc. Natl. Acad. Sci. U.S.A. 103 (2006) 15739-15746.

3 [15] G. Naidu, S. Jeong, S.J. Kim, I.S. Kim, S. Vigneswaran, Organic fouling

4 behavior in direct contact membrane distillation, Desalination 347 (2014) 230-239.

5 [16] Z. Wang, Y. Chen, F. Zhang, S. Lin, Significance of surface excess concentration

6 in the kinetics of surfactant induced pore wetting in membrane distillation,

7 Desalination 450 (2019) 46-53.

8 [17] M. Rezaei, D.M. Warsinger, J.H. Lienhard V, M.C. Duke, T. Matsuura, W.M.

9 Samhaber, Wetting phenomena in membrane distillation: Mechanisms, reversal, and

10 prevention, Water Res. 139 (2018) 329-352.

11 [18] Z. Wang, D. Hou, S. Lin, Composite membrane with underwater-oleophobic

12 surface for anti-oil-fouling membrane distillation, Environ. Sci. Technol. 50 (2016)

13 3866-3874.

14 [19] A. Zoungrana, M. Çakmakci, İ.H. Zengin, Ö. İnoğlu, H. Elcik, Treatment of

15 metal-plating waste water by modified direct contact membrane distillation, Chem.

16 Pap. 70 (2016) 1185-1195.

17 [20] B. Bhushan, Y.C. Jung, Natural and biomimetic artificial surfaces for

18 superhydrophobicity, self-cleaning, low adhesion, and drag reduction, Prog. Mater.

19 Sci. 56 (2011) 1-108

20 [21] P.F. Rios, H. Dodiuk, S. Kenig, S. Mccarthy, A. Dotan, Durable

21 ultra-hydrophobic surfaces for self-cleaning applications, Polym. Advan. Technol. 19

22 (2008) 1684-1691.

27
 1 [22] M. Qu, G. Zhao, X. Cao, J. Zhang, Biomimetic fabrication of lotus-leaf-like

2 structured polyaniline film with stable superhydrophobic and conductive properties,

3 Langmuir 24 (2008) 4185-4189.

4 [23] H.Y. Erbril, A.L. Demirel, Y. Avci, O. Mert, Transformation of a simple plastic

5 into a superhydrophobic surface, Science 299 (2003) 1377-1380.

6 [24] Y. Si, Z. Guo, Superhydrophobic nanocoatings: from materials to fabrications

7 and to applications, Nanoscale 7 (2015) 5922-5946.

8 [25] E. Ueda, P.A. Levkin, Emerging applications of

9 superhydrophilic-superhydrophobic micropatterns, Adv. Mater. 25 (2013) 1234-1247.

10 [26] Z. Ma, Y. Hong, L. Ma, M. Su, Superhydrophobic membranes with ordered

11 arrays of nanospiked microchannels for water desalination, Langmuir 25 (2009)

12 5446-5450.

13 [27] Y. Liao, R. Wang, A.G. Fane, Fabrication of bioinspired composite nanofiber

14 membranes with robust superhydrophobicity for direct contact membrane distillation,

15 Environ. Sci. Technol. 48 (2014) 6335-6341.

16 [28] L.D. Tijing, Y.C. Woo, W.G. Shim, T., He, J.S. Choi, S.H. Kim, H.K. Shon,

17 Superhydrophobic nanofiber membrane containing carbon nanotubes for

18 high-performance direct contact membrane distillation, J. Membr. Sci. 502 (2016)

19 158-170.

20 [29] Zhang, J., Song, Z., Li, B., Wang, Q., Wang, S., 2013. Fabrication and

21 characterization of superhydrophobic poly (vinylidene fluoride) membrane for direct

22 contact membrane distillation. Desalination 324, 1-9.

28
 1 [30] Yang, C., Li, X., Gilron, J., Kong, D., Yin, Y., Oren, Y., Linder, C., He, T., 2014.

2 CF4 plasma-modified superhydrophobic PVDF membranes for direct contact

3 membrane distillation. J. Membr. Sci. 456, 155-161.

4 [31] C. Boo, J. Lee, M. Elimelech, Engineering surface energy and nanostructure of

5 microporous films for expanded membrane distillation applications, Environ. Sci.

6 Technol. 50 (2016) 8112-8119.

7 [32] W. Zhang, Y. Li, J. Liu, B. Li, S. Wang, Fabrication of hierarchical poly

8 (vinylidene fluoride) micro/nano-composite membrane with anti-fouling property for

9 membrane distillation, J. Membr. Sci. 535 (2017) 258-267.

10 [33] Y.C. Woo, Y. Kim, M. Yao, L.D. Tijing, J.S. Choi, S. Lee, S.H. Kim, H.K. Shon,

11 Hierarchical composite membranes with robust omniphobic surface using

12 layer-by-layer assembly technique, Environ. Sci. Technol. 52 (2018) 2186-2196.

13 [34] J.L. Wang, L. Wang, W.X Ruan, C. Zhang, J.B. Ji, Rheology behavior of

14 high-density Polyethylene/Diluent blends and fabrication of hollow-fiber membranes

15 via thermally induced phase separation, J. Appl. Polym. Sci. 118 (2010) 2186-2194.

16 [35] K. Smolders, A.C.M. Franken, Terminology for Membrane Distillation,

17 Desalination 72 (1989) 249-262.

18 [36] Y. Yu, Z.H. Zhao, Q.S. Zheng, Mechanical and superhydrophobic stabilities of

19 two-scale surficial structure of lotus leaves, Langmuir 23 (2007) 8212-8216.

20 [37] M. Liu, S. Wang, L. Jiang, Nature-inspired superwettability systems, Nat. Rev.

21 Mater. 2 (2017) 17036.

22 [38] U. Cengiz, C.E. Cansoy, Applicability of Cassie-Baxter equation for

29
 1 superhydrophobic fluoropolymer-silica composite films, Appl. Surf. Sci. 335 (2015)

2 99-106.

3 [39] H.Y. Erbil, C.E. Cansoy, Range of applicability of the Wenzel and Cassie-Baxter

4 equations for superhydrophobic surfaces, Langmuir 25 (2009) 14135-14145.

5 [40] X. Dai, B.B. Stogin, S. Yang, T.S. Wong, Slippery wenzel state, ACS Nano 9

6 (2015) 9260-9267.

7 [41] J. Phattaranawik, R. Jiraratananon, A.G. Fane, Effect of pore size distribution and

8 air flux on mass transport in direct contact membrane distillation, J. Membr. Sci. 215

9 (2003) 75-85.

10 [42] G.J. Rossa, J.F. Watts, M.P. Hill, P. Morrissey, Surface modification of

11 poly(vinylidene fluoride) by alkaline treatment 1. The degradation mechanism,

12 Polymer 41 (2000) 1685-1696.

30
Highlights

A novel superhydrophobic membrane for robust membrane distillation was developed.

The biomimetic membrane surface was constructed via self-assembly method.

The modified membrane surface had hierarchical micro-nanoscale rough structures.

The membrane surface energy can be lowered by using a fluoride-free modifier.

The modified membrane showed superior anti-wetting and anti-fouling performance.

 Conflict of Interest Statement

Journal Title: Journal of Membrane Science

Paper Reference Number: MEMSCI_2019_1834_R1

Article Title: Biomimetic superhydrophobic membrane for membrane distillation with

robust wetting and fouling resistance

Authors: Deyin Hou, Kofi S.S. Christie, Kai Wang, Min Tang, Dewu Wang, Jun Wang
Corresponding Author: Deyin Hou (E-mail: dyhou@rcees.ac.cn)

The authors declared that they have no conflicts of interest to this work.

We confirm that we have acknowledged all organizations that funded our research, and
provided appropriate grant numbers.

All authors of this manuscript have directly participated in planning, execution, and
analysis of this study. The manuscript has been read and approved by all named authors
and that there are no other persons who satisfied the criteria for authorship but are not
listed. We further confirm that the order of authors listed in the manuscript has been
approved by all of us.

The contents of this manuscript have not been copyrighted or published previously. The
contents of this manuscript are not now under consideration for publication elsewhere.
The contents of this manuscript will not be copyrighted, submitted, or published
elsewhere while acceptance by Journal of Membrane Science.

Signature: , , , , ,

Date: