Composite membrane with electrospun multiscale-textured surface for robust oil-fouling resistance in membrane distillation

Author’s Accepted Manuscript Composite Membrane with Electrospun MultiscaleTextured Surface for Robust Oil-Fouling Resistance in Membrane Distillation Deyin Hou, Zhangxin Wang, Kunpeng Wang, Jun Wang, Shihong Lin www.elsevier.com/locate/memsci

PII: DOI: Reference:

S0376-7388(17)30974-2 https://doi.org/10.1016/j.memsci.2017.10.017 MEMSCI15644

To appear in: Journal of Membrane Science Received date: 5 April 2017 Revised date: 2 September 2017 Accepted date: 8 October 2017 Cite this article as: Deyin Hou, Zhangxin Wang, Kunpeng Wang, Jun Wang and Shihong Lin, Composite Membrane with Electrospun Multiscale-Textured Surface for Robust Oil-Fouling Resistance in Membrane Distillation, Journal of Membrane Science, https://doi.org/10.1016/j.memsci.2017.10.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. 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.

Composite Membrane with Electrospun Multiscale-Textured Surface for Robust Oil-Fouling Resistance in Membrane Distillation

Revised Manuscript Submitted to Journal of Membrane Science Sep. 2nd. 2017 Deyin Houa,b*, Zhangxin Wangc, Kunpeng Wanga,b, Jun Wangb, Shihong Linc*

a

Key Laboratory of Drinking Water Science and Technology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, PR China b

State Key Laboratory of Environmental Aquatic Chemistry, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, PR China c

Department of Civil and Environmental Engineering, Vanderbilt University, Nashville, Tennessee 37235-1831, United States

*Corresponding author: Deyin Hou; Shihong Lin E-mail: D. Hou, [email protected]; S. Lin, [email protected]. Tel.: D. Hou, +86 10 62917207; S. Lin, +1 615 3227226.

Abstract: In this study, we developed composite membranes with a hydrophobic substrate and a hydrophilic top surface using electrospinning to mitigate oil fouling in membrane distillation (MD). The electrospinning approach can be universally applied to any hydrophobic membrane substrate and to ensure the non-wetting condition of the substrate due to the electrospun fibrous structure. Using this approach, polytetrafluoroethylene (PTFE) hydrophobic substrate was coated with two different hydrophilic fibrous networks, including a cellulose acetate (CA) fibrous network and a nanocomposite fibrous network comprising CA and silica nanoparticles (SiNPs). We characterized the pristine and the modified membranes using contact angle measurements and tensiometer-based oil probe force spectroscopy, and tested the anti-fouling performance of the different membranes in MD experiments using a saline crude-oil emulsion as the feed solution. While both coatings offered significant improvement in oil fouling resistance compared to the substrate PTFE membrane, the nanocomposite CA-SiNPs fibrous coating outperformed the CA coating in terms of hydrophilicity, oil adhesion resistance, and anti-oil-fouling performance in MD experiments. Keywords: Membrane distillation; Membrane fouling; Oil adhesion; Hydrophilicity; Underwater oleophobicity

1. Introduction Membrane distillation (MD) is an emerging thermal desalination process using a hydrophobic microporous membrane. In a MD process, the driving force is the partial vapor pressure difference induced by the temperature difference across the hydrophobic membrane [1-3]. Unlike conventional distillation, MD can operate at relatively low temperatures and can thus tap into the vast amount of low-grade waste heat for desalination [4-7]. MD also has advantages over conventional pressure-driven membrane processes, such as nanofiltration (NF) or reverse osmosis (RO). For example, its low operating pressure reduces the capital cost due to the absence of expensive components, such as high pressure pumps and vessels, as well as pressure exchangers [8-10]. Recently, there have been extensive studies on the application of MD for water desalination, wastewater treatment, recovery of valuable resources from wastewater, and treatment of radioactive wastes [10-15]. In MD process, the membrane acts as a physical barrier to direct liquid permeation but provides a porous medium for vapor mass transfer [16]. Therefore, most commercially available MD membranes are made of hydrophobic materials, such

as

polypropylene

(PP),

polyvinylidene

fluoride

(PVDF),

and

polytetrafluoroethylene (PTFE) [17-19]. Among these three materials, PTFE has lowest surface tension (about 20 mN/m) and strongest thermal stability (with a melting point of about 327°C) [20]. However, hydrophobic membranes are inherently prone to fouling by hydrophobic organics, such as hydrocarbons and natural organic matter due to the long-range hydrophobic-hydrophobic interaction [21,22]. The necessity of using hydrophobic membranes in MD and the abundance of hydrophobic contaminants in many saline waste streams limit the application of MD to desalinate relatively clean feed water such as seawater [23]. In recent years, material scientists have developed high-performance underwater anti-oil-adhesion surfaces inspired by natural surfaces of similar interfacial properties, such as those of fish skin, mussel shell, and sharkskin [24-28]. These biological interfaces highlight the importance of both surface morphology and interfacial tension

in imparting excellent anti-adhesion or anti-wetting performance [29-31]. Following those principles, a wide variety of underwater superoleophobic meshes and membranes were fabricated for effective oil-water separation [32-37]. Underwater superoleophobic surfaces are typically in-air hydrophilic. The strong interaction between the water and the hydrophilic materials via hydrogen bonding creates a hydration layer that serves as robust barrier for the attachment of hydrophobic constituents [38]. It is therefore possible to develop a composite membrane by applying a hydrophilic skin layer onto a hydrophobic MD membrane to impart robust underwater fouling resistance. Such a strategy requires that the hydrophobicity of the substrate remains unaffected by the hydrophilic modification, as the hydrophobicity is imperative to prevent direct liquid permeation through the pores [39,40]. Following this approach, anti-fouling composite membranes have been developed using either in-situ formation of a hydrophilic skin layer [41] or spraying of a

hydrophilic

nanoparticle-polymer

composite

[42].

Membranes

with

hydrophilic-hydrophobic composite structure have also been developed for the purpose of enhancing vapor flux, with the rationale that a very thin hydrophobic layer can reduce the pathway for vapor transfer and thus enhance the vapor permeability of the membrane [43,44]. In other cases, composite materials have also been employed to fabricate superhydrophobic and omniphobic membranes to enhance MD performance [45-47]. One challenge of applying a hydrophilic skin onto a hydrophobic substrate using in-situ reaction or pressurized spraying approaches, especially in large scale membrane fabrication, is the possibility of introducing wetting defects due to the possible coating of the inner pores. If the pores are rendered hydrophilic through the depth of the membrane, even over very small percentage of membrane area, the salt rejection can be compromised to an unacceptable level because direct liquid permeation is significantly faster than vapor permeation. A very effective approach to overcome this potential problem is to employ electrospinning to form a hydrophilic fibrous skin layer, as it is extremely unlikely for electronspun fibers with ultrahigh

aspect ratios to penetrate through the tortuous pores. Electrospinning is a simple and versatile approach for creating nanometer- or submicrometer-sized fibrous network [48-50]. Electrospun fibrous networks have high porosity and interconnected open structure [51], which would minimize the additional mass transport resistance to the underlying hydrophobic membrane. Previous studies have demonstrated successful fabrication of electrospun fibrous membranes for microfiltration (MF), ultrafiltration (UF) and NF [52-56]. Here, we explore, for the first time, the possibility and effectiveness of using electrospinning to fabricate composite membranes with asymmetric wettability for anti-oil-fouling MD. In this study, we employ electrospinning to fabricate a composite MD membrane with exceptional oil fouling resistance. The composite membrane has a PTFE hydrophobic substrate and an underwater oleophobic layer formed by electrospinning a fibrous network of either cellulose acetate (CA) or a polymer-nanoparticle composite comprising CA and silica nanoparticles (SiNPs). As a derivative of natural polymer, CA can be easily processed into fibers, films and membranes from either melts or solutions. CA is one of the most commonly used hydrophilic materials for preparation of semipermeable membranes applicable to MF, UF and RO, and CA membranes have very low absorption characteristics and high thermal stability [57]. To electrospin a hydrophilic composite coating, SiNPs were chosen as the inorganic phase of the electrospun fibers because the silica particles had mild reactivity and well-known chemical properties, as well as good compatibility with organic solvents used to prepare the polymer solution and high interface adhesive force with polymer matrix [58]. The morphology and wettability of the composite membranes and the pristine PTFE hydrophobic membrane were investigated. The interactions between oil droplet and membrane surface underwater were measured to compare the propensities of oil adhesion to different membrane surfaces. Finally, direct contact MD (DCMD) experiments with saline feed solution containing a relatively high concentration of crude oil (1000 mg/L) were carried out to show the drastically different anti-fouling behaviors of the pristine PTFE and the composite membranes.

2. Materials and Methods 2.1 Materials and Chemicals PTFE hydrophobic flat sheet membrane used polyester fabric as support material was supplied by Sano Membrane Technology Engineering Co., Ltd. Chemically pure CA with an acylation degree of 39.2% was purchased from Sinopharm Chemical Reagent Co., Ltd. Tetraethyl orthosilicate (TEOS), N-methyl-2-pyrrolidone (NMP), ethanol, acetone, phosphoric acid (≥85 wt.% in H2O) and NaCl were analytically pure and purchased from Sigma-Aldrich. The crude oil, as the fouling agent, was acquired from Daqing Oil Filed of China National Petroleum Corporation. Deionized water was obtained from a Milli-Q ultrapure water purification system (Millipore, Billerica, MA). 2.2 Preparation of Composite Membranes by Electrospinning CA (2.0 g) was dissolved in a mixture of 2.5 mL of acetone and 15.5 mL of NMP. Silica nanoparticles (SiNPs) were synthesized using the Stöber method [59]. Specifically, an aliquot of 15.5 mL TEOS was added into 9.5 mL ethanol at room temperature under vigorous stirring with a bar. The solution was then poured into a mixture of 2.5 mL DI water, 9.75 mL ethanol and 0.6 mL phosphoric acid. The hydrolysis reaction was allowed to proceed for more than 3 h to reach completion. The morphology of the resultant SiNPs (Fig. S1, Supporting information) was observed with a H-7500 transmission electron microscopy (TEM) (Hitachi Ltd., Japan) and the average diameter of the SiNPs was about 183 nm determined with the aid of the computer software coupled to TEM. Without any post-treatment, 10 g SiNPs emulsion was blended with 20 g CA polymer solution under magnetic stirring at room temperature as the electrospinning precursor solution, with an addition 1.2 mL phosphoric acid added to improve its electrical conductivity. To illustrate the impact of SiNPs, we also prepared a precursor solution by mixing 2.0 g CA with 20.5 ml of acetone and 1.95 mL of NMP without adding any SiNPs. The electrospinning method to fabricate composite membranes with asymmetric wettability is schematically illustrated in Fig. 1. A hydrophobic PTFE flat sheet membrane was first attached onto an electrically conductive rotating drum.

Approximately 35 mL CA-SiNPs or CA polymer solution was loaded into a syringe with a stainless steel 18-gauge needle which served as the spinneret. The precursor solution was driven by syringe pump (CSP-100C, Shenzhen Carewell Electronics Co., Ltd., China) to the spinneret at a feed flow rate of 1.5 mL/h. A high voltage was applied between the spinneret, via an alligator clip, and the conductive drum collector. The voltages applied for electrospinning the CA-SiNPs and CA fibers were 20 kV and 15 kV, respectively. Other electrospinning parameters were controlled to be the same for both CA-SiNPs and CA, with a tip-to-collector distance of 15 cm, a chamber humidity of 30%, a chamber temperature of 30 °C, a drum rotating speed of 150 rmp and an electrospinning duration of 20 h. After electrospinning, the fabricated composite membranes were dried in air at room temperature to remove the residual solvents. In the resultant PTFE/CA-SiNPs composite membrane, the mass ratio of SiNPs/CA was 3:5 in the electrospun CA-SiNPs fibers.

Fig. 1. Schematic illustration of the procedure for fabricating the composite membranes with electrospun CA or CA-SiNPs fibrous network.

2.3 Membrane Characterization Scanning electron microscopy (SEM) images were measured on a field emission

scanning electron microscopy (Hitachi SU-70, Japan). An atomic force microscope (AFM) (NanoScope Ⅲa, Digital Instruments, USA) was employed to analyze the membrane surface morphology, all the membrane samples were measured by using a same tip and the surface roughness was obtained by tapping mode. Membrane mechanical properties were tested with an Instron tensiometer (Instron 5565-5kN, Instron Corporation, USA) at room temperature. The effective average pore size and pore size distribution were analyzed using a Capillary Flow Porometer (Porolux 1000, IB-FT GmbH, Germany). The membrane porosity was measured by gravimetric method using a wetting agent (Porefil, IB-GT GmbH, Germany) with a low surface tension of 16 dyn/cm and a density of 1.87 g/mL. The wettability of membrane surface was measured using the standard contact angle method with a goniometer (OCA20, DataPhysics Instruments Ltd., Germany). Both the in-air sessile drop contact angle with water and underwater contact angle with crude oil were measured. Because the density of the crude oil is lower than that of water, the underwater oil contact angle was measured using a floating oil droplet that was in contact with the down-facing membrane surface, similar to the captive bubble method. To better understand the interaction between the foulant (i.e. oil) and the membrane surface, we employed oil-probe force spectroscopy to conduct direct measurement of the oil-membrane interaction. Specifically, we used a highly sensitive force tensiometer (DCAT11, DataPhysics Instruments Ltd., Germany) to measure the interaction force between an oil droplet and a membrane surface submerged in water. An oil droplet (about 10 μL) was first suspended on a platinum-iridium ring as the force probe. The membrane sample was immobilized, using double-sided tape, to the bottom of a beaker placed on a vertically movable platform. The force measurement comprised two different stages defined by the direction of membrane movement (the position of the oil droplet was fixed). In the advancing stage, the platform moved upward at a constant speed of 0.01 mm/s, driving the membrane surface toward the suspended oil droplet until they contacted each other. Upon contact, the oil droplet could be either attracted or repelled by the approaching membrane, depending on the wetting property of the

membrane. In the receding stage, the membrane moved downward, away from the oil droplet until the droplet fully detached from the membrane surface. In the presence of oil-membrane attraction, the oil droplet could split to two parts, with one retained by the membrane surface and the other remaining on the platinum-iridium ring. 2.4 Preparation of Oily Feed Solution The oily and saline feed solution was prepared by vigorously mixing 4 g of crude oil and 4 L of NaCl aqueous solution (600 mM) using a laboratory high-shear emulsifier (AE500S-H, Angni Instruments, Shanghai, China) at 15 000 rpm for 30 min. The crude oil concentration of the prepared feed solution was 1000 mg/L. The saline crude-oil emulsion was stable for at least 24 h without observable phase separation, and the size distribution of the oil droplets, measured using dynamic light scattering with a ZetaSizer (Malvern Nano ZS, UK), can be found in supporting information (Fig. S2, Supporting information). 2.5 Membrane Distillation Experiment DCMD experiments were carried out to evaluate the permeate performance of prepared membranes (the schematic of the DCMD set-up can be found in Fig. S3). The membrane module adopted a plate-and-frame configuration consisting of two chambers, one for the feed and the other for distillate. The porous membrane, tightly clamped between the two chambers, had an effective area of 7.47×10-3 m2. Each chamber was composed of ten flow channels of a depth of 1.0 mm and a width of 7.0 mm. The feed and distillate streams were circulated in the membrane module by gear pumps (MP-215R, Shanghai Seisun Bumps, China). The circulation feed and distillate flow rates were adjusted by two rotameters (LZS-15, Yuyao Yinhuan Flowmeter, China) to 70 L/h, which yielded a flow velocity of about 0.28 m/s. The feed and distillate streams flowed co-currently through the module with their temperatures controlled by water baths. The temperatures of both streams were monitored at the inlets and outlets of the membrane module using four Pt-100 thermocouples (Digit RTD, model XMT-808, Yuyao Changjiang Temperature Meter Instruments, China) with an accuracy of ± 0.1ºC. The mass and conductivity of the distillate in the

distillate reservoir were constantly measured and recorded. The salinity of the distillate was calculated from the measured conductivity based on a pre-established calibration curve. The permeate flux and the salt rejection were calculated from the rates of change in the distillate mass and salinity.

3. Results and Discussions 3.1 Membrane Surface Morphology The SEM images featuring the surface morphology of the pristine PTFE membrane and composite membranes are shown in Fig. 2. The PTFE membrane was fibrous with a reticular fiber-nodule structure formed in a membrane fabrication process called co-stretching. The modified PTFE membrane with an electrospun CA fibrous top layer has a drastically different surface morphology as compared to that of the PTFE membrane. The average diameter of the CA fibers was 217.4±65.2 nm. To enhance the surface roughness required for hydrophilicity, SiNPs were impregnated into the CA nanofibers to construct the PTFE/CA-SiNPs composite membrane. Compared with the CA fibers, the average diameter of the CA-SiNPs fibers widened to 273.2±86.5 nm, and Fig. 2C-1 shows that the CA-SiNPs fibers also have a very different morphology. In contrast to the well-defined fibrous structure of the CA top layer (Fig. 2B-1 and 2B-2), the CA-SiNPs formed a fibrous network with not only fibers but also nodules and webbed structure (Fig. 2C-2).

Fig. 2. SEM images of the surfaces of (A-1) and (A-2) the commercial PTFE membrane, (B-1) and (B-2) the PTFE/CA composite membrane, (C-1) and (C-2) the PTFE/CA-SiNPs composite membrane.

The AFM images presented in Fig. 3 can be used to analyze the surface roughness of the pristine PTFE membrane and composite membranes. The average surface roughness (Ra) of the membranes were 257 nm, 681 nm and 769 nm for PTFE membrane, PTFE/CA and PTFE/CA-SiNPs composite membranes, respectively. The root mean square surface roughness (Rq) of these three membranes were 329 nm, 807 nm and 923 nm, respectively. The AFM images showed that the membrane surface roughness had been remarkably improved after electrospinning modification. Due to the presence of the SiNPs, the local roughness of the CA-SiNPs composite fibers was significantly higher than that of the electrospun pure CA fibers.

Fig. 3. AFM images of the surfaces of (A) the commercial PTFE membrane, (B) the PTFE/CA composite membrane, (C) the PTFE/CA-SiNPs composite membrane.

3.2 Membrane Structural Properties and Wettability The pore size distributions of the PTFE membrane and composite membranes are shown in Fig. S4 (Supporting information). The pore size distribution of the PTFE membrane was very narrow with a mean diameter of 0.48 μm as listed in Table 1. The pore size distribution of the composite membranes was not significantly affected by the modification, with the mean pore size of the PTFE/CA composite membrane being 0.45 μm and that of the PTFE/CA-SiNPs composite membrane being 0.47 μm. The porosity of the PTFE/CA composite membrane was much higher than that of both the PTFE membrane and the PTFE/CA-SiNPs composite membrane. This can be explained by the fact that the PTFE/CA membrane is much thicker due to the thick CA fibrous coating layer and that the CA coating is significantly more porous than the CA-SiNPs coating layer. The SiNPs can increase the viscosity of the electrospinning precursor solution, which induced the size of the electrospun fibers increasing. Meanwhile, the nascent CA-SiNPs fibers became prone to bonding to each other and the fibrous coating became denser and thinner. With the same spinning time (20 h), the presence of SiNPs resulted in a significantly thinner and less porous network of CA-SiNPs. As shown in Table 1, the membrane stretching strain at break was not affected by electrospinning modification. After electrospinning on the top of the PTFE membrane, the composite membrane stress at break declined as shown in Fig. S5 (Supporting information). However, the loads at break of all these three membranes were more than 280 N. The mechanical strength test results indicated that the composite membranes actually had a strong stretching strength and the relative low stress can be attributed to the thicker membrane thickness compared with the PTFE substrate. In essence, the mechanical strength of the composite membranes was determined by the PTFE substrate. The peel strength test results of the composite membranes can be found in Fig. S6 (Supporting information). The peel load of the CA and CA-SiNPs fibrous coatings from the PTFE substrate were about 1.2 N and 1.4 N, respectively, which meant that the electrospun coating adhered well to the PTFE substrate surface for the composite membranes. In addition, the SEM images of the cross-section of

composite membranes (Fig. S7, Supporting information) also showed that the modified membranes were successfully fabricated via electrospinning without obvious delamination. Table 1 Membrane structural properties and surface wettability. Membrane

PTFE

PTFE/CA

PTFE/CA-SiNPs

Mean pore diameter (μm)

0.48

0.45

0.47

Thickness (μm)

185.3±3.7

380.5±1.2

303.0±2.6

Bulk porosity (%)

54.3±3.8

72.2±2.6

50.6±3.3

Fibrous coating porosity (%)

--

81.7±3.2

45.1±2.8

Water contact angle (in air)

134.8±0.2°

62.6±0.3°

39.9±0.4°

Oil contact angle (under water)

35.7±0.5°

139.3±0.3°

154.2±0.3°

Peel load of fibrous coating (N)

--

1.2

1.4

Load at break (N)

283.9

285.3

280.5

Strain at break (%)

54.7

54.9

55.6

The PTFE membrane and composite membranes showed drastically different surface wetting properties. The in-air sessile drop water contact angle was 134.8±0.2° for the PTFE membrane (Fig. 4A), which indicates that the surface of PTFE membrane was hydrophobic in air. In contrast, the sessile drop water contact angles were 62.6±0.3° and 39.9±0.4° for the coated surfaces of the PTFE/CA (Fig. 4B) and PTFE/CA-SiNPs (Fig. 4C) composite membranes, respectively, which suggests that both the surfaces of the composite membranes were in-air hydrophilic. Compared with the PTFE/CA membrane, the surface of the PTFE/CA-SiNPs composite membrane exhibited stronger hydrophilicity due to the presence of SiNPs [60]. To analyze membrane surface wettability directly relevant to oil fouling, the underwater contact angles with crude oil for the three membranes were also measured. The underwater oil contact angle of the PTFE membrane surface was 35.7±0.5° (Fig. 4D), suggesting that the PTFE membrane was underwater oleophilic. In comparison, the surfaces of the PTFE/CA and PTFE/CA-SiNPs membranes were underwater oleophobic (Fig. 4E) and superoleophobic (Fig. 4F), respectively. The underwater oleophobicity of composite membrane was attributable to the strong hydration of the

surface coatings CA and CA-SiNPs. The attachment of oil and the spreading of an oil droplets on a hydrated surface were significantly mitigated by the presence of the hydration force.

In-Air Water Contact Angle

A

B

C

D

E

F

Underwater Oil Contact Angle Fig. 4. In-air water contact angle of (A) the commercial PTFE membrane, (B) the PTFE/CA composite membrane, and (C) the PTFE/CA-SiNPs composite membrane. Underwater oil contact angle of (D) the commercial PTFE membrane, (E) the PTFE/CA composite membrane, and (F) the PTFE/CA-SiNPs composite membrane.

3.3 Underwater Interaction between An Oil Droplet and Membrane Surface The interaction between an oil droplet and a membrane surface determines the fouling propensity of a membrane in the presence of oily foulants. We therefore conducted oil-probe force spectroscopy to compare such an interaction for the three different membranes, with the results presented in Fig. 5. First, the oil-droplet probe was driven toward the membrane at a constant speed (i.e. the approach event). The second event is the contact between the oil droplet and the membrane (i.e. the contact event). The third event involves further driving the force probe toward the membrane after the oil-membrane contact (i.e. the compression event). These three events compose the advancing stage of the force spectroscopy, in which the force probe was driven toward the membrane. In the receding stage, the force probe retracted from the membrane surface until the oil droplet was split, and then continued to move away

from the membrane surface. The characteristic shapes of the force curves were different depending on the interaction between the oil droplet and the tested membranes.

Fig. 5. Force-position curves recorded for the underwater interaction between the oil droplet and the membrane surface for (A) the commercial PTFE membrane, (B) the PTFE/CA composite membrane, and (C) the PTFE/CA-SiNPs composite membrane. The x axis represents the relative positon of the membrane, with position zero corresponding to the first contact between the oil droplet and the membrane.

Several characteristics of the force curve can be used to evaluate the oil-membrane interaction. First, if the membrane was attractive to the oil droplet, a positive adhesive force was observed upon the contact between the oil droplet and the membrane. In contrast, a negative “adhesive force” was observed if the membrane-oil interaction was repulsive. Second, the maximum adhesive force also provides semi-quantitation of the attractive interaction between the oil droplet and membrane. Last but not least, the difference in the baseline forces between the advancing and the receding stages (i.e. the baseline shift), which positively correlates with the volume of the oil droplet retained by the membrane surface, also serves semi-quantitative comparisons of the oil-membrane interaction. Table 2 summarizes the characteristics of the force curves for the three tested membranes. The PTFE membrane was attractive to the oil droplet due to the hydrophobicity of the both membrane surface and the oil droplet. The hydrophobic-hydrophobic interaction in an aqueous solution has been known to be strong, long-ranged, and attractive [61]. In comparison, both the PTFE/CA and PTFE/CA-SiNPs membrane were repulsive to the approaching oil droplet upon their

first contacts, which were attributable to the hydration force experienced by the oil droplet when it contacted the strongly hydrated CA and CA-SiNPs layers. Comparing the maximum adhesive force and the baseline shift suggests that the CA-SiNPs coating layer was more effective than the CA in resisting oil attachment. The stronger hydrophilicity of the CA-SiNPs network led to a stronger repulsive hydration force that significantly mitigated oil attachment onto the membrane surface. Table 2 Summary of force curve characteristics for the tested membranes. Membrane

PTFE

PTFE/CA

PTFE/CA-SiNPs

Attractive

Repulsive

Repulsive

Max adhesive force (μN)

186.5

71.3

31.4

Baseline shift

110.8

59.6

31.4

Force upon contact

3.4 Membrane Permeability and MD Performance without Foulant

2

Gas permeability (L/cm ·min)

80 PTFE PTFE/CA PTFE/CA-SiNPs

60

40

20

0 0

2

4

6

8

10

Pressure (Bar) Fig. 6. Nitrogen gas permeate flux of (A) the commercial PTFE membrane, (B) the PTFE/CA composite membrane, and (C) the PTFE/CA-SiNPs composite membrane.

The fluxes of nitrogen gas permeation at different pressures for the PTFE membrane and the composite membranes are shown in Fig. 6. The permeability of each membrane can be quantified by the slope of the flux vs. pressure plot. The results on Fig. 6 suggest that the permeability of the three membranes correlate negatively with the thickness of the membranes. Being the thickest one, the PTFE/CA composite

membrane yielded the lowest permeability, even though the CA coating layer was more porous than the CA-SiNPs coating. The membrane permeability test results also demonstrated that the fibrous coating could enhance gas transfer resistance and reduced gas permeate flux. The membranes were tested in DCMD experiments using a solution with 600 mM NaCl at 53ºC as the feed solution, whereas the distillate temperature was maintained to be 20ºC. Interestingly, the salt rejection rates of all these three membrane were 100%, which indicated that the hydrophilic fibrous coatings did not introduce wetting defects to the PTFE substrate. The permeate fluxes for the three membranes are shown in Fig. 7. The permeate flux of the PTFE/CA-SiNPs composite membrane was nearly the same as that of the PTFE membrane, and the flux of PTFE/CA was slightly lower due to the much thicker coating layer. It has been suggested that the hydrophilic layer on a composite membrane reduces the permeate flux due to the enhanced temperature polarization within such a layer because of the absence of hydrodynamic mixing [62].

2

Permeate flux ( L/m h)

25

20

15 PTFE PTFE/CA PTFE/CA-SiNPs

10

5 0

5

10

15

20

25

30

Operation time ( hr) Fig. 7. Permeate fluxes from DCMD tests of (A) the commercial PTFE membrane, (B) the PTFE/CA composite membrane, and (C) the PTFE/CA-SiNPs composite membrane. The salt solution with 600 mM NaCl was used as the feed. The flow rate of feed and permeate streams were both 70 L/h. The feed and distillate temperatures were 53°C and 20°C, respectively.

Compared with CA fibrous coating, the CA-SiNPs fibrous coating was denser and thinner, furthermore, it was also harder due to the presence of SiNPs, and closely resembled a membrane spacer plastered to the surface of the PTFE substrate in some extent. The CA-SiNPs coating can improve the fluid-dynamic conditions of the hot feed and provide effective boundary layer surface renewal and more uniform flow distribution [63,64], which was favorable for reducing the temperature polarization and the concentration polarization. For this reason, the PTFE/CA-SiNPs composite membrane can maintain a higher permeate flux compared to the PTFE/CA composite membrane. 3.5 Anti-Oil-Fouling Performance in Direct Contact Membrane Distillation To evaluate the oil fouling resistance of the membranes, we conducted DCMD experiments with a saline feed stream with a relatively high oil concentration. The saline oil-in-water emulsion comprised 600 mM NaCl and 1000 mg/L crude oil. The mean diameter of oil droplets in the emulsion was about 6.10.8 μm. Because the oil droplets are significantly larger than the membrane pores, the attachment of oil droplets onto the membrane surface is expected to block the membrane pores and lead to significant fouling. The three tested membranes present drastically different anti-fouling behaviors in DCMD experiments with the oily feed solution as shown in Fig. 8. Severe membrane fouling was observed with the PTFE membrane, as indicated by a rapid decline of permeate flux over time. Since the PTFE membrane was in-air hydrophobic and underwater oleophilic, the oil droplets tended to attach onto the PTFE membrane surface via the hydrophobic-hydrophobic interaction. The attached oil droplets blocked the membrane pores and reduced the open pores space for water vapor transfer, leading to the reduced permeate flux that eventually became zero. The oil fouling on the PTFE was clearly visible from photographic image of the membrane surface contacting the feed solution (Fig. 9A). The oil fouling was also largely irreversible, as rinsing the PTFE membrane surface did not seem to remove the oil foulant (Fig. 9D).

Normalized permeat flux ( J / Jo)

1.2 1.0 0.8 PTFE PTFE/CA PTFE/CA-SiNPs

0.6 0.4 0.2 0.0 0

5

10

15

20

25

30

Operation time ( hr) Fig. 8. Normalized permeate flux for the commercial PTFE membrane and the fabricated composite membranes in the DCMD experiments. The saline oil-in-water emulsion with 600 mM NaCl and 1000 mg/L crude oil was used as the feed. The flow rate at feed and permeate side were 70 L/h. The feed and distillate temperatures were 53°C and 20°C, respectively. The initial permeate fluxes for PTFE membrane, PTFE/CA and PTFE/CA-SiNPs composite membranes were 19.56 L/m2h, 17.36 L/m2h and 19.92 L/m2h, respectively.

In comparison, the DCMD performance of the composite membranes with underwater oleophobic or superoleophobic coating layers was significantly more stable. Appreciably slower flux decline was observed for the PTFE/CA membrane and no flux decline was observed with the PTFE/CA-SiNPs membrane over 30 h of DCMD experiments with the oily feed solution. The anti-oil-fouling property of the composite membranes can be attributed to the underwater oleophobic CA coating layer and the underwater superoleophobic CA-SiNPs coating layer. The hydrophilic hydroxyl groups of the CA and SiNPs strongly interacted with water and providing a hydration layer that prevents the oil droplet from attaching. The photographic images of the PTFE/CA and PTFE/CA-SiNPs membranes after the DCMD experiments (Fig. 9B and C) also provide direct evidence that the composite membranes were more resistant to fouling than the PTFE membrane, with

the PTFE/CA-SiNPs showing superior fouling resistance. When rinsed with DI water, the oil stains on the PTFE/CA membrane surface were mostly removed, with a few spots being fouled irreversibly (Fig. 9E); whereas the oil foulant on the PTFE/CA-SiNPs surface was completely removed (Fig. 9F).

Fig. 9. The photographic images of the membranes after DCMD experiments, and the photographic images of the fouled membranes after rinsing 2 minutes with DI water (A) and (D) the commercial PTFE membranes, (B) and (E) the PTFE/CA composite membranes, (C) and (F) the PTFE/CA-SiNPs composite membrane.

The fouling propensities of the three membranes were consistent with the results from the underwater oil CA measurements as well as the oil probe force spectroscopy. This suggests that the wetting property of a membrane surface has a critical impact on the oil-membrane affinity and consequently on the oil fouling propensity of a membrane surface. We note that all three membranes yielded a perfect salt rejection rate, which indicates the absence of any pore wetting during the DCMD fouling experiments.

4. Conclusions Novel composite membranes with asymmetric wettability have been developed in this study to mitigate oil fouling in MD processes. These membranes are composed of a PTFE substrate and an electrospun, in-air hydrophilic and underwater oleophobic top surface. Such a coating layer significantly changed the wetting property of the

substrate PTFE membrane, rendered it drastically less attractive to oil droplets, and consequently enhanced the membrane resistance against oil fouling. Of the two types of surface coating we developed, we found that the CA-SiNPs coating was more effective than the CA coating in imparting the oil fouling resistance as the SiNPs augmented the in-air hydrophilicity and underwater oleophobicity of the coating layer. While a PTFE membrane was severely fouled in DCMD experiments with a feed stream containing 1000 mg/L crude oil as foulant, the PTFE/CA-SiNPs membrane was able to sustain a stable operation over 30 h. The findings in this study provide important insights to the development of anti-fouling membranes for MD, which may potentially enable MD to desalinate more challenging saline wastewater with a strong fouling potential.

Acknowledgements Financial support provided by the National Natural Science Foundation of China (51678555 and 51478454) and National Key R&D Program of China (No. 2016YFC0400500) are gratefully acknowledged.

References [1] K. W. Lawson, D. R. Lloyd, Membrane distillation. J. Membr. Sci. 124 (1997) 1-25. [2] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensive review, Desalination 287 (2012) 2-18. [3] M. Bhadra, S. Roy, S. A. Mitra, Bilayered structure comprised of functionalized carbon nanotubes for desalination by membrane distillation, ACS Appl. Mater. Interfaces 8 (2016) 19507-19513. [4] S. Adham, S. A. Hussain, J. M. Matar, R. Dores, A. Janson, Application of membrane distillation for desalting brines from thermal desalination plants, Desalination 314 (2013) 101-108. [5] A. Christ, K. Regenauer-Lieb, T. C. Hui, Boosted multi-effect distillation for sensible low-grade heat sources: a comparison with feed pre-heating multi-effect distillation, Desalination 366 (2015) 32-46. [6] N. Dow, S. Gray, J. Li, J. Zhang, E. Ostarcevic, A. Liubinas, P. Atherton, R. Gareth, A. Gibbs, M. Duke, Pilot trial of membrane distillation driven by low grade waste heat: membrane fouling and energy assessment, Desalination 391 (2016) 30-42.

[7] S. Al-Obaidani, E. Curcio, F. Macedonio, G. D. Profio, H. Al-Hinai, E. Drioli, Potential of membrane distillation in seawater desalination: thermal efficiency, sensitivity study and cost estimation, J. Membr. Sci. 323 (2008) 85-98. [8] D. Singh, K. K. Sirkar, Desalination of brine and produced water by direct contact membrane distillation at high temperatures and pressures, J. Membr. Sci. 389 (2012) 380-388. [9] A. R. Bartmana, A. Zhu, P. D. Christofides, Y. Cohen, Minimizing energy consumption in reverse osmosis membrane desalination using optimization-based control, J. Process Contr. 20 (2010) 1261-1269. [10] G. Migliorini, E. Luzzo, Seawater reverse osmosis plant using the pressure exchanger for energy recovery: a calculation model, Desalination 165 (2004) 289-298. [11] K. Gethard, O. Sae-Khow, S. Mitra, Water desalination using carbon-nanotube-enhanced membrane distillation, ACS Appl. Mater. Interfaces 3 (2011) 110-114. [12] M. Gryta, M. Tomaszewska, J. Grzechulska, A. W. Morawski, Membrane distillation of NaCl solution containing natural organic matter, J. Membr. Sci. 181 (2001) 279-287. [13] Q. Ge, P. Wang, C. Wan, T. S. Chung, Polyelectrolyte-promoted forward osmosis-membrane distillation (FO-MD) hybrid process for dye wastewater treatment. Environ, Sci. Technol. 46 (2012) 6236-6243. [14] G. Lewandowicz, W. Białas, B. Marczewski, D. Szymanowska, Application of membrane distillation for ethanol recovery during fuel ethanol production, J. Membr. Sci. 375 (2011) 212-219. [15] H. Liu, J. Wang, Treatment of radioactive wastewater using direct contact membrane distillation, J. Hazard. Mater. 261 (2013) 307-315. [16] C. Boo, J. Lee, M. Elimelech, Omniphobic polyvinylidene fluoride (PVDF) membrane for desalination of shale gas produced water by membrane distillation, Environ. Sci. Technol. 50 (2016) 12275-12282. [17] Y. Liao, C. H. Loh, R. Wang, A. G. Fane, Electrospun superhydrophobic membranes with unique structures for membrane distillation, ACS Appl. Mater. Interfaces 6 (2014) 16035-16048. [18] N. Tang, Q. Jia, H. Zhang, J. Li, S. Cao, Preparation and morphological characterization of narrow pore size distributed polypropylene hydrophobic membranes for vacuum membrane distillation via thermally induced phase separation, Desalination 256 (2010) 27-36. [19] L. Eykens, K. D. Sitter, C. Dotremont, L. Pinoy, B. V. Bruggen, How to optimize the membrane properties for membrane distillation: a review, Ind. Eng. Chem. Res. 55

(2016) 9333-9343. [20] M. M. A. Shirazi, A. Kargari, M. Tabatabaei, Evaluation of commercial PTFE membranes in desalination by direct contact membrane distillation, Chem. Eng. Process. 76 (2014) 16-25. [21] E. E. Meyer, K. J. Rosenberg, J. Israelachvili, Recent progress in understanding hydrophobic interactions, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 15739-15746. [22] G. Zuo, R. Wang, Novel membrane surface modification to enhance anti-oil fouling property for membrane distillation application, J. Membr. Sci. 447 (2013) 26-35. [23] Z. Wang, D. Hou, S. Lin, Composite membrane with underwater-oleophobic surface for anti-oil-fouling membrane distillation, Environ. Sci. Technol. 50 (2016) 3866-3874. [24] Y. Cai, L. Lin, Z. Xue, M. Liu, S. Wang, L. Jiang, Filefish-inspired surface design for anisotropic underwater oleophobicity, Adv. Funct. Mater. 24 (2014) 809-816. [25] Z. Wang, X. Jiang, X. Cheng, C. H. Lau, L. Shao, Mussel-inspired hybrid coatings that transform membrane hydrophobicity into high hydrophilicity and underwater superoleophobicity for oil-in-water emulsion separation, ACS Appl. Mater. Interfaces 7 (2015) 9534-9545. [26] Y. C. Jung, B. Bhushan, Wetting behavior of water and oil droplets in three-phase interfaces for hydrophobicity/philicity and oleophobicity/philicity, Langmuir 25 (2009) 14165-14173. [27] K. Liu, Y. Tian, L. Jiang, Bio-inspired superoleophobic and smart saterials: design, fabrication, and application, Prog. Mater. Sci. 58 (2013) 503-564. [28] A. Tuteja, W. Choi, M. Ma, J. M. Mabry, S. A. Mazzella, G. C. Rutledge, G. H. McKinley, R. E. Cohen, Designing superoleophobic surfaces, Science 318 (2007) 1618-1622. [29] X. Hou, Smart gating multi-scale pore/channel-based membranes, Adv. Mater. 28 (2016) 7049-7064. [30] Z. Xue, S. Wang, L. Lin, L. Chen, M. Liu, L. Feng, L. Jiang, A novel superhydrophilic and underwater superoleophobic hydrogel-coated mesh for oil/water separation, Adv. Mater. 23 (2011) 4270-4273. [31] B. He, N. A. Patankar, J. Lee, Multiple equilibrium droplet shapes and design criterion for rough hydrophobic surfaces, Langmuir 19 (2003) 4999-5003. [32] S. Zhang, F. Lu, L. Tao, N. Liu, C. Gao, L. Feng, Y. Wei, Bio-inspired anti-oil-fouling chitosan-coated mesh for oil/water separation suitable for broad pH range and hyper-saline environments, ACS Appl. Mater. Interfaces 5 (2013) 11971-11976.

[33] L. Li, Z. Liu, Q. Zhang, C. Meng, T. Zhang, J. Zhai, Underwater superoleophobic porous membrane based on hierarchical TiO2 nanotubes: multifunctional integration of oil-water separation, flow-through photocatalysis and self-cleaning, J. Mater. Chem. A 3 (2015) 1279-1286. [34] L. Zhang, Y. Zhong, D. Cha, P. Wang, A self-cleaning underwater superoleophobic mesh for oil-water separation, Sci. Rep. 3 (2013) 1-5. [35] T. Yuan, J. Meng, T. Hao, Z. Wang, Y. Zhang, A scalable method toward superhydrophilic and underwater superoleophobic PVDF membranes for effective oil/water emulsion separation, ACS Appl. Mater. Interfaces 7 (2015) 14896-14904. [36] J. Li, H. M. Cheng, C. Y. Chan, P. F. Ng, L. Chen, B. Fei, J. H. Xin, Superhydrophilic and underwater superoleophobic mesh coating for efficient oil-water separation, RSC Adv. 5 (2015) 51537-51541. [37] X. Hou, Y. Hu, A. Grinthal, M. Khan, J. Aizenberg, Liquid-based gating mechanism with tunable multiphase selectivity and antifouling behaviour, Nature 519 (2015) 70-73. [38] S. Chen, L. Li, C. Zhao, J. Zheng, Surface hydration: principles and applications toward low-fouling/nonfouling biomaterials, Polymer 51 (2010) 5283-5293. [39] M. Khayet, Membranes and theoretical modelling of membrane distillation: a review, Adv. Colloid Interface Sci. 164 (2011) 56-88. [40] M. S. El-Bourawi, Z. W. Ding, R. Y. Ma, M. Khayet, A framework for better understanding membrane distillation separation process, J. Membr. Sci. 285 (2006) 4-29. [41] Z. Wang, S. Lin, The impact of low-surface-energy functional groups on oil fouling resistance in membrane distillation, J. Membr. Sci. 527 (2017) 68-77. [42] Z. Wang, S. Lin, Membrane fouling and wetting in membrane distillation and their mitigation by novel membranes with special wettability, Water Res. 112 (2017) 38-47. [43] S. Bonyadi, T. S. Chung, Flux enhancement in membrane distillation by fabrication of dual layer hydrophilic-hydrophobic hollow fiber membranes, J. Membr. Sci. 306 (2007) 134-146. [44] M. Qtaishat, D. Rana, M. Khayet, T. Matsuura, Preparation and characterization of novel hydrophobic/hydrophilic polyetherimide composite membranes for desalination by direct contact membrane distillation, J. Membr. Sci. 327 (2009) 264-273. [45] S. Lin, S. Nejati, C. Boo, Y. Hu, C. O. Osuji, M. Elimelech, Omniphobic membrane for robust membrane distillation, Environ. Sci. Technol. Lett. 1 (2014) 443-447. [46] J. Lee, C. Boo, W. H. Ryu, A. D. Taylor, M. Elimelech, Development of

omniphobic desalination membranes using a charged electrospun nanofiber scaffold, ACS Appl. Mater. Interfaces 8 (2016) 11154-11168. [47] J. Zhang, Z. Song, B. Li, Q. Wang, S. Wang, Fabrication and characterization of superhydrophobic poly (vinylidene fluoride) membrane for direct contact membrane distillation, Desalination 324 (2013) 1-9. [48] H. G. Wang, S. Yuan, D. L. Ma, X. B. Zhang, J. M. Yan, Electrospun materials for lithium and sodium rechargeable batteries: from structure evolution to electrochemical performance, Energy Environ. Sci. 8 (2015) 1660-1681. [49] S. Kaur, S. Sundarrajan, D. Rana, R. Sridhar, R. Gopal, T. Matsuura, S. Ramakrishna, Review: the characterization of electrospun nanofibrous liquid filtration membranes, J. Mater. Sci. 49 (2014) 6143-6159. [50] X. Li, Y. Zhang, H. Li, H. Chen, Y. Ding, W. Yang, Effect of oriented fiber membrane fabricated via needleless melt electrospinning on water filtration efficiency, Desalination 344 (2014) 266-273. [51] M. Essalhi, M. Khayet, Self-sustained webs of polyvinylidene fluoride electrospun nano-fibers: effects of polymer concentration and desalination by direct contact membrane distillation, J. Membr. Sci. 454 (2014) 133-143. [52] Y. Liu, R. Wang, H. Ma, B. S. Hsiao, B. Chu, High-flux microfiltration filters based on electrospun polyvinylalcohol nanofibrous membranes, Polymer 54 (2013) 548-556. [53] T. He, W. Zhou, A. Bahi, H. Yang, F. Ko, High permeability of ultrafiltration membranes based on electrospun PVDF modified by nanosized zeolite hybrid membrane scaffolds under low pressure, Chem. Eng. J. 252 (2014) 327-336. [54] S. Kaur, S. Sundarrajan, D. Rana, T. Matsuura, S. Ramakrishna, Influence of electrospun fiber size on the separation efficiency of thin film nanofiltration composite membrane, J. Membr. Sci. 392-393 (2012) 101-111. [55] M. H. Tai, P. Gao, B. Y. L. Tan, D. D. Sun, J. O. Leckie, Highly efficient and flexible electrospun carbon-silica nanofibrous membrane for ultrafast gravity-driven oil-water separation, ACS Appl. Mater. Interfaces 6 (2014) 9393-9401. [56] N. Daels, S. D. Vrieze, B. Decostere, P. Dejans, A. Dumoulin, K. D. Clerck, P. Westbroek, S. W.H. Van Hulle, The use of electrospun flat sheet nanofibre membranes in MBR applications, Desalination 257 (2010) 170-176. [57] W. K. Son, J. H. Youk, T. S. Lee, W. H. Park, Electrospinning of ultrafine cellulose acetate fibers: studies of a new solvent system and deacetylation of ultrafine cellulose acetate fibers, J. Polym. Sci. B Polym. Phys. 42 (2004) 5-11. [58] N. A. Hashim, Y. Liu, K. Li, Preparation of PVDF hollow fiber membranes using SiO2 particles: the effect of acid and alkali treatment on the membrane performances, Ind. Eng. Chem. Res. 50 (2011) 3035-3040.

[59] W. Stöber, A. Fink, E. Bohn, Controlled growth of monodisperse silica spheres in the micron size range, J. Colloid Interface Sci. 26 (1968) 62-69. [60] Y. Gu, J. Yang, S. Zhou, A facile immersion-curing approach to surface-tailored poly(vinyl alcohol)/silica underwater superoleophobic coatings with improved transparency and robustness, J. Mater. Chem. A 5 (2017) 10866-10875. [61] J. Israelachvili, R. Pashley, The hydrophobic interaction is long range, decaying exponentially with distance, Nature 300 (1982) 341-342. [62] X. Yang, R. Wang, A. G. Fane, Novel designs for improving the performance of hollow fiber membrane distillation modules, J. Membr. Sci. 384 (2011) 52-62. [63] L. Martínez, J. M. Rodríguez-Maroto, Characterization of membrane distillation modules and analysis of mass flux enhancement by channel spacers, J. Membr. Sci. 274 (2006) 123-137. [64] J. Phattaranawik, R. Jiraratananon, A.G. Fane, C. Halim, Mass flux enhancement using spacer filled channels in direct contact membrane distillation, J. Membr. Sci. 187 (2001) 193-201.

Graphical Abstract

Video Caption. Membranes".

The legend of the video was "Videos of Oil-Adhesion Behavior of the

Research Highlights  Composite membrane with electrospun multiscale-textured surface was fabricated.  The membrane surface was hydrophilic in air and underwater superoleophobic.  The composite membrane showed robust oil-fouling resistance in DCMD.  Oil probe force spectroscopy was introduced to evaluate oil-membrane interaction.