Separation and Puriﬁcation Technology · 2018. 6. 19. · Seyed Mahdi Seyed Shahabadi, Jonathan A. Brant⁎ University of Wyoming, Department of Civil and Architectural Engineering,

Contents lists available at ScienceDirect

Separation and Purification Technology

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

Bio-inspired superhydrophobic and superoleophilic nanofibrous membranesfor non-aqueous solvent and oil separation from water

Seyed Mahdi Seyed Shahabadi, Jonathan A. Brant⁎

University of Wyoming, Department of Civil and Architectural Engineering, 1000 E. University Ave., Laramie, WY 82071, USA

A R T I C L E I N F O

Keywords:ElectrospinningElectrosprayingNanofibrous membraneOil/water separationSuperhydrophobicSuperoleophilic

A B S T R A C T

The separation of apolar, or non-aqueous, solvents from polar ones is an important challenge in water andwastewater treatment applications. Superhydrophobic and superoleophilic membranes present unique oppor-tunities for such separations. Inspired by the lotus leaf effect, superhydrophobic membranes were preparedhaving hierarchical surface roughness made of carbon black nanoparticles. A hydrophobic nanofibrous supportwas produced through electrospinning using polyvinylidene fluoride-co-hexafluoropropylene (PVDF-HFP).Hydrophobic carbon black nanoparticles were then coated onto the support via an electrospraying techniqueunder varying conditions to generate surfaces having unique micro- and nano-scale roughness features.Membranes made using a polymer concentration of 8% w/w and nanoparticle to polymer ratio of 1 showed thesmallest bead size (4.9 ± 2.0 µm2) and highest bead area density (74.3%), with corresponding average and rootmean square roughness values of 3.69 ± 0.07 µm and 4.63 ± 0.05 µm, respectively. When tested for surfacewettability, the prepared membrane showed water contact angle, sliding angle and contact angle hysteresisvalues of 160.8°, 7.0° and 5.3°, respectively; however, liquids with surface tensions ≤36.6mN/m had zerocontact angle on the membrane surface (superoleophilicity). The optimized membrane showed significantlyhigher gravity-driven flux (1275–2163 LMH) than the nanofibrous support membrane (933–1424 LMH) for thetested non-aqueous solvents.

1. Introduction

In the past decade the development and study of superhydrophobicsurfaces has grown substantially. Previous efforts have sought to usethese surfaces in diverse applications, such as in the formation of self-cleaning, anti-icing, and anti-fouling surfaces as well as for makingmembranes for oil/water separation [1–6]. Superhydrophobic surfaces,commonly defined as surfaces with a water contact angle (CA)≥ 150°and sliding angle (SA)≤ 10°, received increased attention after thereport of the “lotus effect” mechanism by Barthlott and Neinhuis [7].This effect is attributed to the combination of two characteristics: a lowsurface energy waxy layer and hierarchical surface roughness withmicro- and nano-scale structures [8]. Accordingly, artificial super-hydrophobic surfaces are usually manufactured in two stages: (1) fab-rication of hierarchical micro/nano-structures to improve roughnessand (2) modification of surface chemistry to reduce surface free energy[9,10]. In general, the effects of surface roughness on wettability differaccording to the liquid surface tension and solid surface free energy.Based on the Wenzel model, if the liquid spreads on the surface withCA < 90°, like low surface tension organic liquids on a solid surface

with high surface energy, roughening the surface increases the affinityof the solid surface toward the liquid. This increase in affinity is ob-served as a decrease in the CA. On the other hand, if the liquid CAis> 90°, like water on a low surface energy (hydrophobic) surface,surface roughness reduces surface wettability (low affinity). In thiscase, roughness can result in air being trapped between the liquid andsolid phases, and lead to a heterogeneous surface with low solid-liquidadhesion and high CA as illustrated by the Cassie-Baxter model [11].

Apart from water, the wettability of organic liquids, such as oil, tomembrane surfaces is of interest. The difference between the surfacetension of water (72.8mNm−1) and oil (< 35mNm−1) is the reasonwhy most superhydrophobic surfaces are at the same time oleophilic orsuperoleophilic (oil CA < 5°). As previously mentioned, the low sur-face energy of the solid is a key factor to achieving super-hydrophobicity; however, as these surfaces are normally characterizedby surface energy values> 35mNm−1, they tend to be oleophilic.Superhydrophobic/superoleophilic membranes are great candidates foroil/water separation either by adsorption, filtration, or a combinationthereof. Using superhydrophobic/superoleophilic membrane separationpresents a variety of advantages for oil/water separation compared to

https://doi.org/10.1016/j.seppur.2018.08.038Received 19 June 2018; Received in revised form 21 August 2018; Accepted 21 August 2018

⁎ Corresponding author.E-mail address: [email protected] (J.A. Brant).

Separation and Purification Technology 210 (2019) 587–599

Available online 23 August 20181383-5866/ © 2018 Elsevier B.V. All rights reserved.

T

http://www.sciencedirect.com/science/journal/13835866https://www.elsevier.com/locate/seppurhttps://doi.org/10.1016/j.seppur.2018.08.038https://doi.org/10.1016/j.seppur.2018.08.038mailto:[email protected]://doi.org/10.1016/j.seppur.2018.08.038http://crossmark.crossref.org/dialog/?doi=10.1016/j.seppur.2018.08.038&domain=pdf

more conventional techniques like gravity separation, skimming, coa-gulation, magnetic separation, flotation, and membrane filtration [12].For example, conventional techniques are limited by the droplet size ofthe non-aqueous liquid, surface fouling resulting in reduced waterfluxes, and the need for chemical addition resulting in the production oflow purity residual streams [13,14].

Various methods such as layer-by-layer assembly, low temperaturehydrothermal, dip coating and phase inversion have been used toproduce superhydrophobic surfaces [15–18]. Despite all of the previousstudies, the commercial production of superhydrophobic/super-oleophilic membranes has been hindered by complex production stepsand poor mechanical stability and flexibility in practical environments[19]. Following the two step process of introducing surface roughnessand reducing surface energy, all these methods consisted of a multi-stepfabrication process using various chemicals. For instance, Meng et. al.used the dip coating technique to produce superhydrophobic mem-branes containing TiO2 nanoparticles [18]. A precursor sol, a mixture ofAnhydrous ethanol, perchloric acid (HClO4), acetylacetone (AcAc), ti-tanium (IV) isopropoxide (TTIP) and Milli-Q water, was mixed with asolution of a templating agent in anhydrous ethanol to prepare a sol-gel.Afterwards, membranes were dip coated in the sol-gel for 8 s. Then,membranes dried at 120 °C for 16 h and rinsed and placed in a waterbath at 90 °C for 24 h. Finally, membranes were rinsed again and UVirradiated in water for 24 h. The whole process was repeated 4 times foreach sample to achieve surface roughness. The secondary step of re-ducing surface energy was applied to enhance surface hydrophobicity.For this reason, a solution of toluene and 1H,1H,2H,2H-per-fluorododecyltrichlorosilane (FTCS) was filtered through modifiedmembranes at 0 °C under low vacuum. The filtered membrane was keptat 120 °C for 2 h, then, backwashed with ethanol for 5min at 100 kPa.

One alternative synthesis technique that overcomes many of theaforementioned challenges to making superhydrophobic membranes iselectrospinning. Electrospinning is a simple and versatile technique tofabricate nonwoven, highly porous fibrous mats with inter-connectedpore structures [20,21]. Over the past decade, the progress in electro-spinning technology, such as moving collectors and multi-needle sys-tems, has offered new possibilities for mass production of nanofibers[22]. Furthermore, the emergence of needle-less electrospinning tech-niques has opened a new commercial outlook for nanofiber production[23]. Considering the promising outlook of nanofibrous membranes,investigating their potential applications is highly relevant.

Furthermore, while this technique has shown great promise in makingmacroporous membrane structures their performance in non-aqueousphase separations has not been extensively documented.

Recently, electrospraying of polymer and nanoparticles mixtureshas been emerged as a promising method for introducing hierarchicalsurface roughness, consisting of mico- and nano-scale roughness fea-tures. When a solution is under a relatively high electrostatic force, thespinning jet destabilizes and solution moves toward the collector as finedroplets, which is electrospraying [24,25]. During the droplets flighttime, solvent evaporates and beaded structures form after their de-position onto the collector. While the polymeric beads provide micro-scale roughness features, nanoparticles can protrude from the surface ofindividual beads and add roughness features in nanometer scale [24].To achieve superhydrophobicity, however, requires that the surfacechemistry of the nanoparticles be made to be hydrophobic, if it is al-ready not so. A summary of some of the different groups that have beenused to functionalize nanoparticles for making superhydrophobicmembranes is given in Table 1. Silane and fluorinated groups are themain functional groups used in previous studies to date. The toxicity offluorinated-based chemicals poses environmental consequences thatmay hinder their practical application [26,27]. Additionally, post-sur-face modification can reduce the stability of the membrane surfacecoating leading to a gradual loss of superhydrophobicity [27]. Hydro-philic nanomaterials have been generally used during membranesynthesis to facilitate their adhesion to the membrane support structure.

Nomenclature

3D three dimensionalAl2O3 aluminum oxideCA contact angleDMF N,N-dimethylformamideFESEM field emission scanning electron microscopyLSM laser scanning microscopyPVDF-HFP poly(vinylidene fluoride-co-hexafluoropropylene)SA sliding contact angleSDS sodium dodecyl sulfateSiO2 silicon dioxideTiO2 titanium dioxide

Symbols

A membrane areaDe density of ethanolDp density of dry membranedp maximum pore sizeflv liquid-air interfacefsl fractional contact area between the droplet and solid

J filtration fluxM0 mass of oil before separationMm mass of oil after separationMs mass of oil adsorbed in the membraneR separation efficiencyr Wenzel roughness factorSa average roughnessSq root mean square roughnessV the volume of the permeated solventWw wet weightWd dry weightε porosity

PΔ water entry pressureγ surface tensionθ contact angleθH2O water contact angleθY Young’s contact angleθW Wenzel contact angleθCB Cassie-Baxter contact angle

tΔ permeation time

Table 1Summary of functional group types used to reduce the surface energy of dif-ferent nanomaterials for fabricating superhydrophobic membranes.

Nanoparticle Functional group Watercontactangle (°)

Refs.

TiO2 (Heptadecafluoro-1,1,2,2-tetrahydrodecyl)triethoxysilane

162.7 [9]

(Tridecafluro-1,1,2,2-tetrahydrooctyl)triethoxysilane

163.2 [9]

(3,3,3-Trifluoropropyl) trimethoxysilane 165.3 [9]SiO2 Hexamethyldisilazane 160 [28]SiO2 Octamethylcyclotetrasiloxane 157 [29]TiO2 1H,1H,2H,2H perfluorooctyltriethoxysilane 153.4 [30]SiO2 1H,1H,H,2H-perfluorodecyltriethoxysilane NA [31]Al2O3 γ-Methacryloxypropyltrimethoxysilane 144 [32]

S.M. Seyed Shahabadi, J.A. Brant Separation and Purification Technology 210 (2019) 587–599

588

Electrospraying techniques open up the possibility of using intrinsicallyhydrophobic materials to create the superhydrophobic surface, therebyavoiding the drawbacks previously noted for the post-surface mod-ification approaches.

In this study, we present the application of naturally hydrophobiccarbon black nanoparticles to produce superhydrophobic/super-oleophilic membranes for separating non-aqueous phases from water.Hydrophobic poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP) was used as the polymer to produce a nanofibrous supportmembrane by electrospinning. An electrospraying technique was ap-plied to prepare the superhydrophobic coating layer containing PVDF-HFP and carbon black nanoparticles. The performance of the resultingmembranes was evaluated in terms of selectivity and flux for differentapolar phases from water. The nano-scale roughness caused by thenanoparticles on individual beads formed during the electrosprayingprocess provided hierarchical surface roughness and enhanced surfacehydrophobicity and oleophilicity. Synthesis conditions were optimizedto maximum superhydrophobicity, which was correlated to beadstructure and density on the membrane surface.

2. Materials and methods

2.1. Materials

Poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP,Mw=455,000 and average Mn=110,000) was provided by Sigma-Aldrich, USA. Carbon black nanoparticles (average diameter= 150 nm,purity > 65%) was purchased from US Research Nanomaterials, Inc.USA. N,N-dimethylformamide (DMF) (certified ACS grade,> 99.9%),acetone (certified ACS grade,> 99.7%), mineral oil (FCC/NF grade),hexane (certified grade) and sodium dodecyl sulfate (SDS) were sup-plied by Fisher Scientific, USA. Xylene (purity > 98.5%), toluene(purity > 99%) and ethanol (> 99.5% absolute) were obtained fromJ.T.Baker® (USA), Matheson, Coleman & Bell (USA) and Decon Labs,Inc. (USA), respectively. Moreover, methylene blue and Oil red O werepurchased from Chem-Impex Int'l. Inc. (USA) All test solutions weremade using ultrapure water having a resistivity of 18MΩ cm and anunmodified pH of 6.8.

2.2. Membrane fabrication

2.2.1. Nanofibrous membrane preparationA homogeneous solution of PVDF-HFP (20% w/w) in a mixture of

DMF and acetone (DMF to acetone ration of 4:1) was prepared by

stirring overnight at 60 °C. Afterwards, 8 mL of the solution was loadedinto a glass syringe for electrospinning using a Super-ES-2 machine (E-Spin Nanotech, India). As illustrated in Fig. 1, the solution was ejectedwith a fixed flow rate using a syringe pump. As the solution exited theneedle, which was connected to a high voltage supply, it formed aconical shape known as the Taylor cone due to the electrostatic fieldbetween the needle and the grounded collector. At high voltages, wherethe electrostatic force overcomes the surface tension of the solution, athin jet ejected from the cone and flew toward the collector. Mean-while, the solvent evaporated and the polymer fiber was stretched,elongated, whipped and finally deposited on the collector as randomfibers [20]. The solution and process parameters used for support layerfabrication, and the coded name for the prepared membranes, are listedin Table 2.

2.2.2. Superhydrophobic layer coating by electrosprayingThe prepared nanofibrous membrane was used as the support layer

onto which the superhydrophobic surface coatings were deposited.Solutions with different concentrations of polymer and carbon blacknanoparticles were used for superhydrophobic layer fabrication ap-plying the electrospraying process (Table 2). For making these solu-tions, homogeneous PVDF-HFP/DMF solutions were first prepared bystirring overnight at the temperature of 60 °C. Then, carbon black na-noparticles (using a nanoparticle to polymer ration of 1:1) were addedto the solution and it was mixed for 4 h. Finally, the solution was bathsonicated for 90mins prior to the electrospraying process. Compared tothe electrospinning process, the solution concentrations and the dis-tance between the needle tip and the collector, were set to be lower andthe applied voltage higher. Under these conditions, a dilute solutionwith weak viscoelastic force is affected by a strong electrostatic force(high applied voltage and short spinning distance). As a result, thespinning jet breaks-up and solution flies toward the collector as finedroplets containing polymer and nanoparticles (Fig. 1). As the dropletsmoved toward the collector, the solvent evaporated and a beadedstructure formed on the support membrane. Unlike the electrospinningprocess, acetone which is more volatile than DMF was not added intothe solution to have a degree of wetness and enhanced top layer ad-hesion to the support layer.

A Vibro Viscometer (A&D Company Limited, Japan) was used formeasuring the viscosities of the different liquids tested. Two sensorplates were vibrated at the same frequency driven by electromagneticforce. Then, the viscosity was determined based on the proportionalrelationship between the viscous resistance of the sample fluid and theamount of electric current required to drive and maintain the sensor

Fig. 1. Illustrations of the electrospinning and electrospraying processes used for synthesizing the nano-fibrous membrane supports and superhydrophobic surfacecoatings.


589

plates at a constant vibration amplitude. The measurements were donein triplicate at an ambient temperature of 22 °C.

2.3. Membrane characterization

Membrane surface morphology was characterized using field emis-sion scanning electron microscopy (FESEM, FEI Quanta FEG 450,Thermo Fisher Scientific, USA) with an accelerating voltage of 20 kV.For FESEM imaging, the membrane samples were prepared by coatingthem with carbon using a vacuum sputter-coating system. Fiber dia-meter, bead size and bead area density were determined through ana-lysis of FESEM images using Image J (National Institute of Health,USA). Membrane surface roughness was characterized using laserscanning microscopy (LSM, VK-X Series, Keyence, USA). As a non-contact surface roughness analysis tool, the LSM was used to char-acterize membrane surface roughness in terms of the following statis-tics: average roughness (Sa) and root mean square roughness (Sq).Samples were analyzed without modification and were analyzed in theair. A minimum of three areas per sample, and three samples wereanalyzed, to ascertain representative results.

Membrane hydrophobicity was characterized using an EASYDROPgoniometer (Krüss Scientific, USA) and with ultrapure water as theprobe liquid. A droplet volume of 5 µL was used for all measurements,which were done at an ambient temperature of 22 °C. Contact angle wasmeasured using the sessile drop technique. A minimum of three dro-plets were analyzed per sample, with a minimum of three samplesanalyzed to calculate average values. A similar technique was also usedto measure the contact angles formed between select membrane sam-ples and non-aqueous solvents to assess the affinities between the twophases. The goniometer was also used to measure the surface tension ofdifferent liquids using pendant drop tensiometry [33]. Sufficient liquidvolume was slowly ejected from the needle to reach just before thepoint of break-off from the needle tip. The droplet and needle tip wereset to occupy as much of the screen as possible to increase the availablepixels for analysis. Moreover, the brightness of the background illu-mination was optimized to increase the profile recognition by thesoftware. If the light intensity is too dark, then the contrast between thebackground and the drop will also be too weak. On the other hand, toobright background illumination can lead to over-illumination of thedrop, which then appears narrower than it actually is. Finally, thebaseline was set to be as close as possible to the needle tip, but far awayto exclude the part of the droplet whose shape is influenced by thecontact with needle. The correctness of the baseline position wasevaluated by checking the whether the fit line generated by the soft-ware exactly corresponds to the droplet profile.

Contact angle hysteresis was quantified by measuring both the ad-vancing and receding contact angles formed between a droplet of waterand the membrane surface [34]. The advancing contact angle was de-termined by increasing the volume of a water droplet that was de-posited onto the membrane surface. The point at which the three-phasecontact point between the membrane surface and water droplet startedto advance was determined as the advancing contact angle. After theadvancing contact angle was determined, the droplet volume was re-duced by pumping water out of it via the attached syringe pump. Thereceding contact angle was determined at the point where the three-phase contact point started to recede. The difference between the

advancing and receding contact angles was reported as the contactangle hysteresis. All hysteresis measurements were done using ultra-pure water.

Membrane porosity was determined using the gravimetric method[35]. Samples having a width and length of 2 cm×2 cm were soaked ina low surface tension wetting liquid, pure ethanol solution, for 30min.Then, the wet weight (Ww) was measured after removing the excessiveethanol on the surface with a tissue. Afterwards, the samples were driedin an oven for 2 h at 60 °C prior to the dry weight (Wd) measurement.Membrane porosity, ε, was then calculated according to Eq. (1) [35]

=⎡⎣ ⎤⎦

+

−

−ε

W WD

W WD

WD

w de

w de

dp (1)

where De and Dp are the density of ethanol and dry membranes. Allmeasurements were done in triplicate at ambient temperature (=22 °C)to calculate an average porosity for a given membrane sample. Mem-brane pore size was characterized using the Washburn equation (Eq.(2)) and pore-liquid entry pressure measurements. Liquid entry pres-sure was measured using a dead-end filtration cell (Sterlitech Cor-poration, USA). The feed solution was ultrapure water, which wasmaintained at a temperature of 22 °C. Pressure was applied to the feedsolution using compressed nitrogen gas. Pressure was gradually in-creased until permeation was observed. The pressure at which per-meation was observed was determined to be the pore liquid entrypressure, PΔ . A minimum of three different membrane samples werecharacterized to calculate an average PΔ value.

=Pβγcosθ

dΔ

4p (2)

where γ is the liquid surface tension, β is the pore shape factor, θ is themeasured liquid-solid contact angle, and dp is the maximum pore size ofthe sample.

The mechanical durability of the membrane samples was assessedby measuring the contact angle with water prior to, and after, ultra-sonic treatment (model CPX5800, Branson Ultrasonic Corporation,USA). Samples were placed in the bottom of a beaker containing water.Then the beaker was placed in a bath sonicator set at a power setting of180W power for various time periods up to 40mins. The chemicalstability of the membranes was assessed by measuring the contact anglewith water after immersing the sample in acidic and basic solutions.Solutions with pH values from 2 to 12 were prepared at 22 °C by addinghydrochloric acid and sodium hydroxide to water. The samples werefixed on the bottom of a petri dish and solutions were poured to com-pletely cover the membrane surface for 1 h. For each measurement,three samples were tested with at least three contact angles measuredper sample to report an average value.

2.4. Oil/Water separation

2.4.1. Oil adsorption to membraneThe affinity of five different solvents and oils, representing a variety

of viscosities and densities, to the superhydrophobic/oleophilic mem-branes was assessed. Select characteristics of the different solvents andoils are summarized in Table 3. Samples were dried in an oven for 1 h at60 °C prior to the dry weight (Wd) measurement. Then, they were

Table 2Solution and process parameters used for synthesizing the nano-fibrous membrane support and surface coatings.

Membranes Code Polymer Conc. (Wt. %) CB:PVDF-HFP ratio Solvent Voltage (kV) Spinning distance (cm) Feed rate (ml/h)

Support layer P20C0 20 0 DMF:Acetone (4:1) 11 12 3

Coated membranes P6C1 6 1 DMF 15 9 1.5P8C1 8 1 DMF 15 9 1.5P10C1 10 1 DMF 15 9 1.5


590

immersed in the solvent/oil of interest (30mL stored in a beaker cov-ered with Parafilm) at 22 °C for 60mins. Of note, it has been reportedthat superhydrophobic/superoleophilic surfaces are characterized byfast adsorption kinetics, where equilibrium is generally achieved in<10mins [36]. Samples were then removed and the wet weight (Ww)was measured. The sorption capacity was calculated according to Eq.(3). All tests were done in triplicate.

= −Sorption Capacity W WW

w d

d (3)

2.4.2. Solvent/water separationThe solvent/water separation performance of the membranes was

assessed using a custom-made filtration apparatus (see Video S1).Briefly, a solvent/water mixture was poured onto the membrane sur-face with an effective area of 13.85 cm2. Solvent/water mixtures weremade by mixing ultrapure water and different solvents (hexane,chloroform, toluene and xylene) at a water to solvent ration of 2:1. Fluxwas calculated by measuring the volume of the permeating phase with agraduated cylinder and using Eq. (4).

=J VA tΔ (4)

where J is the flux, V is the volume of the permeant phase(s), A is theactive membrane area and tΔ is the permeation time. It should be notedthat for the non-aqueous media having a density less than water, theseparation apparatus was tilted to a 45° angle to bring the non-aqueousphase in contact with the membrane surface. To better distinguish be-tween the aqueous and solvent/oil phases, the aqueous phase was dyedwith methylene blue and the solvent/oil phase with Oil Red O, re-spectively. All tests were done at a temperature of 22 °C and in triplicateto calculate an average flux. Separation efficiency R was quantifiedaccording to Eq. (5).

= +R M MM

m s

0 (5)

where Mm and Ms were the mass of oil after separation and the mass ofoil adsorbed in the membrane structure which was determined by thesorption capacity (g/g) of the membranes for a particular solvent, M0was the solvent/oil mass before separation. Eq. (5) is based on the as-sumption that no water passes through the membrane. The water depthin all separation experiments was less than 20 cm having a pressuremuch less than the LEP values of the membranes reported in Section3.3.

The separation of kerosene emulsions was also investigated to de-termine the permeation of non-aqueous solvents through the mem-brane, where kerosene was selected as representative of the broad classof apolar solvents. Water/kerosene emulsions were prepared by adding3mL of ultrapure water in 300mL kerosene followed by magneticstirring for 2 h and ultrasonication (model 5800, Branson UltrasonicCorporation, USA) for 1 h. The emulsion was poured onto the mem-brane surface using the previously described test apparatus. The sizedistributions and droplet concentrations in the feed and filtrate werecharacterized using an in-line camera (INFLOW™ Particle SizingCamera, J.M. Canty Inc., Buffalo, NY) capable of imaging and quanti-fying the size distribution of oil droplets in real time. The camera iscapable of detecting particles having a minimum diameter of 0.7 µm.Acquired images and videos were analyzed using CantyVisionClient™(J.M. Canty Inc., Buffalo, NY) software. In between measurements thesample cell was cleaned with ethanol. Three replicates were done at22 °C and average values were reported.

3. Results and discussion

3.1. Membrane surface morphology and roughness

FESEM images of the PVDF-HFP support membrane are shown inFig. 2. The support structure was comprised of fibers having an averagediameter of 437 ± 87 nm. The surfaces of the fibers were smooth andfree of beads, or structures, that may form as a result of polymer ag-gregation during the electrospinning/spraying process. The lack ofbeads on the fibers and the narrow distribution in their diameters isattributed to the synthesis conditions. As the polymer concentration inthe spinning solution increased so too did its viscosity and the asso-ciated viscoelastic forces. At the polymer concentration used here (20%w/w) the measured viscosity of the spinning solution was 1.0 Pa s. Thestrong viscoelastic force prevented break-up of the injected mixture (thespinning jet) by the electrostatic force leading to uniform fiber forma-tion [37–39]. Moreover, the addition of acetone facilitated solvent

Table 3Relevant characteristics of the tested solvents and oils. All values are reportedat 22 °C.

Solvent Viscosity (mPa·s) Density (kg/m3) Surface tension (mN/m)

Hexane 0.30 0.66 18.43Toluene 0.57 0.867 28.52Chloroform 0.76 1.49 26.67Xylene 0.57 0.86 30.10Kerosene 1.64 0.81 25.00Mineral oil 40.60 0.83 33.42

Fig. 2. FESEM images of the PVDF-HFP nanofibrous support layer prepared by electrospinning. The scale bar for (a) is 40 µm and that for (b) is 10 µm.


591

evaporation during the spinning jet flight toward the collector resultingin dry, bead-free fiber formation [40]. The absence of any structures onthe fibers was desired to prevent formation of surface defects in thesubsequently applied surface coatings. Although bead formation wasundesirable on the support structure surface, it is an attractive optionfor manipulating the hydrophobicity and roughness of the separatinglayer. Optimizing both of these characteristics required matching thenanoparticle properties to the selected synthesis conditions. Carbonblack nanoparticles are a carbonaceous material that is intrinsicallyhydrophobic [41]. The hydrophobic character and relative small size ofthe nanoparticles (150 nm) made them a good candidate to produce asuperhydrophobic surface.

The nano-carbon surface coatings on the PVDF-HFP support re-sulted in the formation of different surface morphologies, all char-acterized by micro-scale surface roughness (Figs. 3 and 4). Theroughness of the membrane surfaces was measured using laser scanningmicroscopy, which generates a 3D image of the membrane surface.Representative images and roughness statistics of the membranes aregiven in Fig. 5 and Table 4, respectively. These values represent in-creases of 204% and 193% in Sa and Sq, respectively, compared to thePVDF-HDP support membrane. In addition to the micro-scale roughnessimparted by the bead structures, the membrane surfaces were alsocharacterized by nano-scale roughness features on the beads themselves(inset images in Fig. 3a–c). The multi-scale structure created sites for airpocket formation between the membrane surface and the surroundingliquid phase. Thereby, water can only interact with the peaks of thesurface structures instead of wetting the entire surface [42]. Recentstudies indicate that hierarchical structure improves the stability of theCassie wetting mode and enhances superhydrophobic stability [43].

The viscosities of the spinning solutions used to make the threedifferent membranes were 29.8, 75.8 and 148.6 mPa s, respectively forthe P6C1, P8C1, and P10C1 membrane samples. These values were

significantly lower than that of the solution used for support layerfabrication (1.0 Pa s). When a solution with a low viscoelastic force isexposed to a high electrostatic force, caused in this case by the higherapplied voltage (15 kV) and lower spinning distance (9 cm) comparedto those in the electrospinning conditions for the support synthesis, thespinning jet breaks up and solution flies toward the collector as smalldroplets forming a beaded structure (Fig. 4). Bead area densities for theP6C1, P8C1 and P10C1 membranes were 44.9, 74.3 and 50.8%, re-spectively. The P8C1 membrane, which had the highest bead areadensity, also had the smallest average bead size (4.9 ± 2.0 µm2)compared to the other membranes, whose average bead sizes were96.7 ± 60.8 µm2 (P6C1) and 102.4 ± 156.8 µm2 (P10C1). The largerbead sizes and lower bead area densities of the P6C1 and P10C1membranes were attributed to the solvent and nanoparticle con-centrations used in the electrospraying solutions. The electrosprayingsolution used for making the P6C1 and P10C1 membranes had polymerconcentrations of 6% and 10%w/w (nanoparticles to polymer ratio of1:1), respectively. So the P6C1 membrane was characterized by a lowpolymer concentration relative to the solvent. Conversely, the P10C1membrane was made using a solution having a higher nanoparticleconcentration. It has been reported that bead size is inversely propor-tional to polymer concentration in the solution; decreasing polymerconcentration leads to larger bead sizes with lower area densities [37].In this case, a greater amount of solvent did not evaporate during thesolution flight toward the collector and remained in the membranestructure forming larger beads. The electrospraying solution used tomake the P10C1 membrane contained the highest amount of bothpolymer and carbon black nanoparticles. In this case, the larger beadsize and lower bead area density compared to the P8C1 membrane wasattributed to nanoparticle aggregation [24].

Fig. 3. FESEM images of the prepared dual layer membranes. Images (a), (b) and (c) are of the P6C1, P8C1 and P10C1 membranes, respectively, and (d) is a cross-sectional image of P6C1. The inset images are of the respective surface features on the different membranes. The scale bars for the insets images of figure (a), (b) and(c) are 2, 2 and 10 μm, respectively.


592

3.2. Membrane surface chemistry

Contact angle with water, θH2O data for each of the differentmembranes is summarized in Fig. 6a. The membrane support had anaverage θH2O of 139 ± 0.9° indicating that it was highly hydrophobic.This was due to a combination of the hydrophobic nature of PVDF-HFPpolymer and the micro-scale surface roughness formed by the randomly

arranged and stacked nanofibers. The HFP incorporated into the PVDFblocks increased the fluorine content of the PVDF-HFP homopolymerleading to enhanced hydrophobicity compared to pure PVDF [44].Moreover, it is reported that for the same polymer, non-woven nano-fibrous membranes exhibit higher hydrophobicity compared to mem-branes prepared by conventional methods due to the hierarchicalstructure of the randomly deposited nanofibers [45].

The carbon black nanoparticle coatings increased the hydro-phobicity of the PVDF-HFP membrane support irrespective of thesynthesis conditions used (Fig. 6a). The most hydrophobic membranewas the P8C1 membrane, followed by the P6C1 and P10C1 membranes,respectively. All the θH2O values were found to be statistically differentfrom one another at 95% confidence level. Because each surface wascomposed of the same material, the carbon black nanoparticles andPVDF-HFP with a ratio of 1:1, the differences in hydrophobicity wereattributed to differences in surface roughness features (Fig. 4). Thesedifferences, namely the density and distribution of surface features,determine together with the solid-liquid interfacial free energy, the

Fig. 4. Processed images of the (a) P6C1, (b) P8C1, and (c) P10C1 membranes. The inset figures are the bead size distributions (top right) and representativeillustrations of hierarchical roughness features (bottom).

Fig. 5. Representative laser scanning microscopy images of the (a) P20C0, (b) P6C1, (c) P8C1, and (d) P10C1 membranes.

Table 4Roughness statistics for the prepared membranes.

Membrane/Roughness statistic

P20C0 P6C1 P8C1 P10C1

Average roughness(Sa, µm)

1.21 ± 0.02 3.57 ± 0.05 4.82 ± 0.07 2.82 ± 0.03

Root mean squareroughness (Sq,µm)

1.58 ± 0.03 4.47 ± 0.04 6.02 ± 0.05 3.61 ± 0.04


593

ability of the liquid to penetrate into the surface depressions. This inturn affected the contact area between the liquid and the membranesurface.

Depending on whether the liquid impregnated the roughness fea-tures or not, the liquid contact angle on a rough surface is defined byWenzel and Cassie-Baxter models, respectively. Based on the Wenzeltheory for a homogenous wetting regime, the relationship betweenapparent contact angle (θW ) and the Young’s contact angle (θY ) is asfollows [46]:

=θ r θcos cosW Y (6)

where r is the roughness factor defined as the ration of the actual sur-face area to the theoretical area based on a perfectly flat surface.Conversely, when air is trapped within the surface features and betweenthe liquid droplet forming a heterogeneous wetting regime, the ap-parent contact angle (θCB) is defined by the Cassie-Baxter theory [47]:

= + −θ f θcos (cos 1) 1CB sl Y (7)

where fsl is the fractional contact area between the liquid and solid.Based on Eq. (7), for the apparent contact angle to approach 180° thesurface topography should be tailored such that the solid-liquid inter-face approaches zero.

As a water droplet sits on the surface, the trapped air between themicro-scale beads and nano-scale protrusions and valleys formed on thebead structures resulted in a heterogeneous solid-air interface in con-tact with the water droplet (Fig. 6b). Based on the Cassie-Baxter modelfor wettability of rough surfaces, the enhanced liquid-air interface ( flv)resulted in a greater contact angle with water as it decreases the valueof fsl ( = −f f1sl lv) [47]. Using Eq. (7), fsl was determined by measuringthe water contact angle on smooth (θY) and rough sample surfaces(θCB). The fsl values for P6C1, P8C1 and P10C1 membranes were 11.2,7.2 and 13.1%, respectively. The P8C1 membrane surface was char-acterized by the highest bead area density and the smallest bead sizeamongst the coated membranes. The higher bead area density, andsmaller roughness features (average bead size= 4.9 ± 2.0 µm2) re-sulted in the minimum fsl value and maximum θH2O for the P8C1

Fig. 6. (a) Contact angle with water for the nano-fibrous support and three different coated membranes. (b) Illustration of Cassie-Baxter wetting regime.

Fig. 7. KAO diagrams for the (a) P20C0 and (b) P8C1 membranes. The diagrams were generated using hexane, chloroform, toluene and SDS/ultrapure watermixtures (4, 3, 2, 1, 0.5, 0.1 and 0% w/w) at 22 °C.


594

sample amongst the coated membranes. The wettability and surfaceroughness of the P8C1 membrane was compared to that for the P20C0membrane using a KAO diagram (Fig. 7) [48]. To generate the diagram,the measured contact angles for different wetting liquids with varyingsurface tensions were measured on smooth and rough samples to re-spectively obtain θY and θRough values. The first region on top rightcorresponds to the soaking state where the soaking low surface tensionliquids into the material can be observed. The middle region refers tothe homogeneous wetting state described by the Wenzel theory (Eq.(6)). The third region at the bottom left is the Cassie-Baxter wettingstate where the apparent contact angle changes significantly andreaches values> 150°. From Fig. 7 both membranes showed soakingbehavior toward liquids with low surface tension (i.e., toluene,chloroform and hexane). The KAO diagram of P8C1 (Fig. 7b) was dif-ferent from that of the P20C0 (Fig. 7a) in two ways. First, the slope ofthe Wenzel region, which corresponds to the r value in Eq. (6), is higherfor P8C1. Based on the linear fitting, the r values for P20C0 and P8C1membranes were found to be 2.00 ± 0.29 and 3.03 ± 0.16. This in-dicates that the actual surface area has enhanced by 51% after elec-trospraying the top layer. Moreover, the data point for water (bottomleft), which has the highest surface tension among the tested liquids,shifted to the Cassie-Baxter region for the P8C1 membrane. It should bementioned that it is impossible to reach cos(θY) < −0.5, which cor-responds to the maximum contact angle of 120° on a smooth surfacewith the lowest possible surface energy [48].

Also, from Fig. 8 the P8C1 was characterized by a low sliding angleof 7.0°, which is the characteristic of a superhydrophobic surface. Aswater cannot wet, or contact, much of the available membrane surfacearea leaving a large volume of air trapped beneath the droplet, theadhesive force between the low surface energy solid and water is ex-tremely low resulting in the low sliding contact angle. This phenom-enon is known as the “lotus effect” [42]. The low sliding contact angleof the P8C1 membrane is reflective of its low contact angle hysteresis,which was 5.3° (Fig. S1). The higher sliding contact angles measured forthe P6C1 and P10C1 membranes could be due to the larger beads andlower bead area densities on their respective surfaces. For these mem-branes, water can more easily enter into the comparatively larger valleyareas, resulting in a higher solid-liquid adhesive force. This phenom-enon is termed the partial “petal effect” [49]. Compared to the lotusleaf, the micro- and nano-structure of rose petals are larger. As a result,water can enter into the micro-scale grooves, while it cannot enter thenanostructured ones, which is known as the Cassie impregnating wet-ting regime [50]. Subsequently, the higher solid-liquid adhesion pre-vents water droplets from rolling off of petal surfaces. It should bementioned that, P20C0 did not show any sliding angle even when themembrane was upside down (Fig. S2). The fibrous structure of P20C0did not possess any nanostructure roughness features on individual fi-bers (Fig. 2) and the micro-scale grooves produced by deposited fibersare highly wetted in the absent of air.

The aforementioned self-cleaning character of the P8C1 membraneindicates that polar contaminants would be removed from the mem-brane surface during operation [51]. Note that non-polar solutes/sol-vents would adhere to, or permeate through, the membrane. The self-cleaning character of the P8C1 membrane was further evaluated andcontrasted with that of the membrane support (the P20C0 membrane).This was done by pouring water droplets on the membrane surfaceswhen tilted at an angle of 7.0°. During this test, both surfaces werecoated with a montmorillonite clay. The clay was used to simulatefouling of the membrane by a polar moiety. As water slid over the P8C1membrane surface the clay partitioned into the aqueous phase and wassubsequently removed (Fig. 9a). Conversely, for the P20C0 membranethe clay was not removed due to the fact that water droplets morestrongly adhered to the membrane surface (Fig. 9b). In actual operationa portion of the clay would partition into the bulk solution and removedfrom the membrane surface in the tangential flow; however, a largerportion of the polar foulant would be expected to remain on the P20C0

membrane surface relative to the P8C1 membrane due to its strongerinteraction with the aqueous media.

3.3. Porosity, pore size and liquid entry pressure

The porosity of the P8C1 membrane (0.86 ± 0.02) was greaterthan that of the P6C1 (0.83 ± 0.01) and P10C1 (0.81 ± 0.02) mem-branes at a 95% confidence level. All the coated membranes had lowerporosities than the membrane support (0.88 ± 0.02). The reduction inporosity was attributed to the carbon black surface coating and thesubsequent constriction of the pore throats; however, the loss of por-osity was relatively low (≤8%) and the macroporous structure wasseemingly maintained. The LEP of the P20C0, P6C1, P8C1 and P10C1membranes was found to be 43.89 ± 0.39, 48.72 ± 0.40,51.48 ± 0.40 and 49.87 ± 1.05 kPa. Based on Eq. (2) and LEP andcontact angle results (Fig. 8), when assuming cylindrical pores (β=1)for all samples, the support had a maximum pore size of5.06 ± 0.04 µm, which was increased to 5.43 ± 0.05, 5.28 ± 0.05and 5.22 ± 0.04 µm for the P6C1, P8C1 and P10C1 membranes, re-spectively. These results are consistent with Liao et al. who found thatthe strong repulsive force between beads which have high accumulatedcharge during the electrospraying process increases the pore size of thecoated membranes [52].

3.4. Wettability as a function of liquids surface tension

Based on Eqs. (6) and (7), liquid contact angle is inversely propor-tional to its surface tension. To evaluate this relationship for the su-perhydrophobic membrane surfaces the surface tension of water wasmanipulated using an anionic surfactant, sodium dodecyl sulfate (SDS).From Fig. 10, the θH2O on the P8C1 membrane decreased as the SDSconcentration in the water increased. Recall that an increase in SDSconcentration resulted in a decrease in the surface tension of the water.When the surface tension of the water reached 36.6 mN/m θH2Odropped to zero. Therefore, the minimum liquid surface tension re-quired to prevent penetration into the pore throats of the membraneswas 36.6mN/m. Also shown in Fig. 10 are the contact angle, and sur-face tension, values for toluene, chloroform and hexane. As the surfacetensions of each of these three solvents is below the minimum identifiedvalue of 36.6mN/m they did not form a contact angle on the membranesurface and subsequently permeated through it. The importance of thisminimum surface tension lies in its use to distinguish between thosesolvents that will pass through the membrane and those that will berepelled by it. In other words, using such values and tailoring of the

P6C1 P8C1 P10C10

2

4

6

8

10

12

14

16

Slid

ing

Ang

le (°

)

Fig. 8. Sliding contact angles for water on the three different coated mem-branes.


595

membrane surface properties it is possible to achieve selective solventseparation from mixtures.

The affinity of the P8C1 membrane surface to an apolar solvent, inthis case hexane, and the low affinity for a polar solvent, water, is il-lustrated in Videos S2 and S3. The observed superhydrophobic andsuperoleophilic behaviors are due to the combined effects of membranesurface chemistry and topography. As previously noted, these char-acteristics combine to prevent high surface tension liquids(γTot≥ 36.6 mN/m), while permeating liquids with lower surface ten-sions. The surface free energy of the membrane was much lower thanwater surface tension and higher than that of the solvents/oils. For anon-wetting surface, the surface energy and solid-liquid interfacialtension is much lower than the liquid surface tension. As a result, theliquid droplet keeps its spherical shape. On the other hand, when solidsurface energy was higher than the liquid surface tension, the strongattractive force pulls the liquid and makes it spread out. Moreover,based on the Wenzel model (Eq. (6)), which is valid for low surfacetension liquids as they are expected to fill the surface grooves, enhancedsurface roughness (r) reduces wetting in a low wetting state ( > °θ 90Â )and improves wetting in a high wetting state ( < °θ 90Â ). In otherwords, membrane surface roughness not only has improved watercontact angle, it has also enhanced the affinity of the membrane surfacefor low surface tension oils, oleophilicity.

3.5. Mechanical and chemical durability

The robustness of the superhydrophobic coating was assessed byexposing P8C1 to ultrasonic vibration at various time. No significantchanged was observed in the water contact angle indicating the strongstructural integrity of the coated superhydrophobic layer (Fig. 11).Also, the contact angle of a single water droplet was measured over

time. As time passed and the droplet evaporated, it almost kept itsspherical shape and its contact angle decreased slightly indicating thestable superhydrophobicity.

The presence of dissolved acids or alkaline in practical oil/waterseparation may also negatively affect the membrane and result in theloss of superhydrophobicity [27]. The chemical resistance of themembrane was evaluated by measuring the water contact angle afterimmersing the membrane into acidic and alkaline solutions with dif-ferent PHs for 1 h (Fig. 12). Once again, the results demonstrated that

Fig. 9. Illustrations of the self-cleaning properties, based on water repellencies, of the (a) P8C1 and (b) P20C0 membranes.

Fig. 10. Contact angle of liquids with different surface tension on the P6C1 membrane, the horizontal at the surface tension of 36.6 mN/m indicates the surfacetension at which the water contact angle drops to zero.

Fig. 11. Changes in water contact angle after ultrasonic treatment in varioustime and changes in the contact angle of a single droplet over time.


596

the membrane preserved its superhydrophobicity, as no significantchange was observed in its water contact angle. The water/chloroformseparation was also performed after soaking the membrane for 1 h inacidic and basic solutions with a pH of 2 and 12, respectively, and nosignificant change was observed. The membrane showed a flux of1667 ± 14 and 1742 ± 11 LMH after being soaked in acidic and basicsolutions, respectively, further confirming the chemical stability of itsstructure.

3.6. Adsorption and permeation of apolar solvents

The superhydrophobic/superoleophilic character of the preparedmembranes, coupled with their high porosity, make them good candi-dates for removing non-polar compounds (e.g., BTEX compounds) fromaqueous mixtures. When the membrane is brought in contact with xy-lene on the water surface or under water chloroform, it selectivelyabsorbs them (Fig. 13a). The adsorption capacities of the differentmembranes for five compounds was assessed and the results are sum-marized in Fig. 13b. The t-test results for the reported values are listedin Table S1. To be consistent with previous studies, the reported sorp-tion capacities are the weight of the adsorbed oil based on the weight of

the membrane samples [6,19,36,53]. Therefore, the density and visc-osity of the solvents can effect the reported values. Of the five com-pounds, all membranes had the highest specific adsorption capacity(Fig. 13b) for the mineral oil. That is due to the significantly higherviscosity of mineral oil which can result in the formation of a thick layeron the membrane surface leading to an obvious enhancement in sorp-tion capacity. In addition, the slightly higher sorption capacity ofP20C0 compared to the coated membranes is because of its slightlyhigher porosity providing more voids for the solvent to be absorbed. Asit was mentioned, the coated membranes have slightly lower porositythan the nanofibrous support (P20C0) due to their relatively densecoated layer.

3.7. Membrane performance: Separation of apolar solvents from aqueousmixtures

Membrane performance was evaluated using aqueous mixtures ofchloroform, hexane, toluene, and kerosene. a representative video of anexperiment is given in Video S1. The steady-state flux and separationefficiencies for each of these non-aqueous solvents are summarized in

Fig. 12. Changes in water contact angle after immersing P8C1 in acidic andalkaline solutions.

P20C0 P6C1 P8C1 P10C10

2

4

6

8

10

12

14

HexaneChloroformXylenekeroseneMineral Oil

Ads

orpt

ion,

g S

orba

te/g

Sor

bent

(a) (b)Fig. 13. Selective xylene and chloroform in water sorption (a) and (b) adsorption capacity of the membranes for different types of oils.

Table 5Summary performance statistics for the membrane support and the super-hydrophobic/oleophilic nano carbon black membranes.

Membrane Solvent Flux LMH Separation efficiency (%)

P20C0 Hexane 1424 ± 10 99.95 ± 0.02Chloroform 1377 ± 13 99.95 ± 0.02Toluene 1022 ± 11 99.97 ± 0.01Xylene 933 ± 7 99.98 ± 0.01





597

Table 5. These values are comparable to, or even higher than, othersuperhydrophobic/oleophilic surfaces prepared with different methodssuch as 1215 LMH for chloroform separation from water by a poly-tetrafluoroethylene nanofibrous membrane [27], 1800 LMH and2648 LMH for hexane separation from water by coated stainless steelmesh [54] and carbon-silica nanofibrous membrane [14], respectively.For the optimized membrane in this study (P8C1), chloroform andhexane fluxes were found to be higher than 2000 LMH. Such a high fluxis due to the low mass transfer resistance because of the interconnectedpore structure and high porosity of the nanofibrous membrane. In ad-dition, the strong affinity of the membrane toward oils, super-oleophilicity, is another reason contributing to such a high flux. Thesurface roughness prepared by the electrospraying technique has alsoenhanced the contact area between the oil and the membrane resultingin an enhanced mass transfer. The variations in flux of different oils ismainly due to the differences in their viscosity. According to the Darcyexpression, flux is inversely proportional to liquid viscosity [14]. Othercoated membranes (P6C1 and P10C1) showed quite similar perfor-mance when applied for hexane/water separation performance. How-ever, the P20C0 membrane had significantly lower flux relative to thatfor the coated membranes. The improved flux of the coated membranesis due to their enhanced roughness, providing larger contact area, andlarger pore size compared to P20C0.

Moreover, the separation performance of water in oil emulsion wascarried out for the prepared membrane (P8C1). Water in keroseneemulsions with average droplet diameter of 19.0 μm (minimum dia-meter= 6.8 µm and maximum=37.3 µm) was used as the feed. Fig. S3shows the overall separation experiment. As it can be seen, no dropletwas observed in the separated kerosene, indicating the excellent se-paration efficiency of the membrane. The permeate flux was found to be1377 ± 18 LMH. The superhydrophobicity of the membrane is en-hanced when the membrane is soaked in kerosene. That is due to theextreme surface roughness which provides a heterogeneous surfaceconsisting of the membrane and oil in contact with the water droplet.The under oil water contact angle was found to be around 165°. Thisstrong hydrophobicity along with superoleophilicity of the membrane isthe main reason for such a high separation efficiency.

4. Conclusions

A facile approach for synthesizing superhydrophobic/super-oleophilic membranes for non-aqueous phase separation from waterwas developed. Optimal membrane performance was found when usinga carbon black nanoparticle concentration of 8% w/w. These conditionsproduced a hierarchical surface roughness characterized by the highestand lowest static and sliding water contact angles, respectively, in-dicating a superhydrophobic surface. Superhydrophobicity translatedinto high fluxes for non-aqueous phases and low to no permeation ofwater. Permeation was a function of both the affinity of the phase to themembrane surface and the pore-liquid entry pressure of the membrane,which was enhanced for water by the superhydrophobic character ofthe membrane surface.

Acknowledgements

We gratefully acknowledge the financial support provided for thisresearch by the Center of Excellence in Produced Water Management(CEPWM) at the University of Wyoming.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, in theonline version, at https://doi.org/10.1016/j.seppur.2018.08.038.

References

[1] S. Afzal, W.A. Daoud, S.J. Langford, Superhydrophobic and photocatalytic self-cleaning cotton, J. Mater. Chem. A 2 (2014) 18005–18011.

[2] Y.W. Wu, T. Hang, Z.Y. Yu, L. Xu, M. Li, Lotus leaf-like dual-scale silver film appliedas a superhydrophobic and self-cleaning substrate, Chem. Commun. 50 (2014)8405–8407.

[3] F. Xiao, S.J. Yuan, B. Liang, G.Q. Li, S.O. Pehkonen, T.J. Zhang, SuperhydrophobicCuO nanoneedle-covered copper surfaces for anticorrosion, J. Mater. Chem. A 3(2015) 4374–4388.

[4] Z.Y. Hu, X.M. Zhang, Z.Y. Liu, K.F. Huo, P.K. Chu, J. Zhai, L. Jiang, Regulatingwater adhesion on superhydrophobic TiO2 nanotube arrays, Adv. Funct. Mater. 24(2014) 6381–6388.

[5] S. Song, H. Yang, C.P. Su, Z.B. Jiang, Z. Lu, Ultrasonic-microwave assisted synthesisof stable reduced graphene oxide modified melamine foam with super-hydrophobicity and high oil adsorption capacities, Chem. Eng. J. 306 (2016)504–511.

[6] C.L. Zhou, Z.D. Chen, H. Yang, K. Hou, X.J. Zeng, Y.F. Zheng, J. Cheng, Nature-inspired strategy toward superhydrophobic fabrics for versatile oil/water separa-tion, ACS Appl. Mater. Interf. 9 (2017) 9184–9194.

[7] W. Barthlott, C. Neinhuis, Purity of the sacred lotus, or escape from contaminationin biological surfaces, Planta 202 (1997) 1–8.

[8] S.H. Li, J.Y. Huang, Z. Chen, G.Q. Chen, Y.K. Lai, A review on special wettabilitytextiles: theoretical models, fabrication technologies and multifunctional applica-tions, J. Mater. Chem. A 5 (2017) 31–55.

[9] N. Hamzah, C.P. Leo, Membrane distillation of saline with phenolic compoundusing superhydrophobic PVDF membrane incorporated with TiO2 nanoparticles:separation, fouling and self-cleaning evaluation, Desalination 418 (2017) 79–88.

[10] Q. Wang, Z. Dong, X.X. Yan, Y.J. Chang, L.L. Ren, J. Zhou, Biomimetic hydrophobicsurfaces with low or high adhesion based on poly(vinyl alcohol) and SiO2 nano-particles, J. Bionic. Eng. 14 (2017) 476–485.

[11] Y. Yuan, S.O. Choi, J. Kim, Analysis of contact area between water and irregularfibrous surface for prediction of wettability, RSC Adv. 6 (2016) 73313–73322.

[12] T.Y. Zhao, D.M. Zhang, C.M. Yu, L. Jiang, Facile fabrication of a polyethylene meshfor oil/water separation in a complex environment, ACS Appl. Mater. Interf. 8(2016) 24186–24191.

[13] C. Lee, S. Baik, Vertically-aligned carbon nano-tube membrane filters with super-hydrophobicity and superoleophilicity, Carbon 48 (2010) 2192–2197.

[14] M.H. Tai, P. Gao, B.Y.L. Tan, D.D. Sun, J.O. Leckie, Highly efficient and flexibleelectrospun carbon-silica nanofibrous membrane for ultrafast gravity-driven oil-water separation, ACS Appl. Mater. Interf. 6 (2014) 9393–9401.

[15] J. Pan, C.F. Xiao, Q.L. Huang, H.L. Liu, T. Zhang, ECTFE hybrid porous membranewith hierarchical micro/nano-structural surface for efficient oil/water separation,J. Membr. Sci. 524 (2017) 623–630.

[16] T. Ogawa, B. Ding, Y. Sone, S. Shiratori, Super-hydrophobic surfaces of layer-by-layer structured film-coated electrospun nanofibrous membranes, Nanotechnology18 (2007).

[17] A. Razmjou, E. Arifin, G.X. Dong, J. Mansouri, V. Chen, Superhydrophobic mod-ification of TiO2 nanocomposite PVDF membranes for applications in membranedistillation, J. Membr. Sci. 415 (2012) 850–863.

[18] S.W. Meng, J. Mansouri, Y. Ye, V. Chen, Effect of templating agents on the prop-erties and membrane distillation performance of TiO2-coated PVDF membranes, J.Membr. Sci. 450 (2014) 48–59.

[19] S.S. Ouyang, T. Wang, X.Y. Jia, Y. Chen, J.M. Yao, S. Wang, Self-indicating andrecyclable superhydrophobic membranes for effective oil/water separation in harshconditions, Mater. Design 96 (2016) 357–363.

[20] S. Ramakrishna, K. Fujihara, W.E. Teo, T. Yong, Z.W. Ma, R. Ramaseshan,Electrospun nanofibers: solving global issues, Mater. Today 9 (2006) 40–50.

[21] J.J. Xue, J.W. Xie, W.Y. Liu, Y.N. Xia, Electrospun nanofibers: new concepts ma-terials, and applications, Acc. Chem. Res. 50 (2017) 1976–1987.

[22] D. Pisignano, Polymer nanofibers: building blocks for nanotechnology, RSCNanosci. Nanotechnol. (2013) 1–427.

[23] N.N. Bui, J.R. McCutcheon, Nanoparticle-embedded nanofibers in highly permse-lective thin-film nanocomposite membranes for forward osmosis, J. Membr. Sci.518 (2016) 338–346.

[24] S.M.S. Shahabadi, H. Rabiee, S.M. Seyedi, A. Mokhtare, J.A. Brant,Superhydrophobic dual layer functionalized titanium dioxide/polyvinylidenefluoride-co-hexafluoropropylene (TiO2/PH) nanofibrous membrane for high fluxmembrane distillation, J. Membr. Sci. 537 (2017) 140–150.

[25] Y.D. Guo, D.Y. Tang, Z.L. Gong, Superhydrophobic films fabricated by electro-spraying poly(methyl methacrylate)-b-poly(dodecafluoroheptyl methacrylate) di-block copolymers, J. Phys. Chem. C 116 (2012) 26284–26294.

[26] H. Attia, S. Alexander, C.J. Wright, N. Hilal, Superhydrophobic electrospun mem-brane for heavy metals removal by air gap membrane distillation (AGMD),Desalination 420 (2017) 318–329.

[27] W.H. Qing, X.N. Shi, Y.J. Deng, W.D. Zhang, J.Q. Wang, C.Y.Y. Tang, Robust su-perhydrophobic-superoleophilic polytetrafluoroethylene nanofibrous membrane foroil/water separation, J. Membr. Sci. 540 (2017) 354–361.

[28] X.N. Wu, B. Zhao, L. Wang, Z.H. Zhang, J.J. Li, X.J. He, H.W. Zhang, X.H. Zhao,H.T. Wang, Superhydrophobic PVDF membrane induced by hydrophobic SiO2 na-noparticles and its use for CO2 absorption, Sep. Purif. Technol. 190 (2018)108–116.

[29] C.L. Su, J.J. Chang, K.X. Tang, F. Gao, Y.P. Li, H.B. Cao, Novel three-dimensionalsuperhydrophobic and strength-enhanced electrospun membranes for long-termmembrane distillation, Sep. Purif. Technol. 178 (2017) 279–287.


598

https://doi.org/10.1016/j.seppur.2018.08.038http://refhub.elsevier.com/S1383-5866(18)32136-1/h0005http://refhub.elsevier.com/S1383-5866(18)32136-1/h0005http://refhub.elsevier.com/S1383-5866(18)32136-1/h0010http://refhub.elsevier.com/S1383-5866(18)32136-1/h0010http://refhub.elsevier.com/S1383-5866(18)32136-1/h0010http://refhub.elsevier.com/S1383-5866(18)32136-1/h0015http://refhub.elsevier.com/S1383-5866(18)32136-1/h0015http://refhub.elsevier.com/S1383-5866(18)32136-1/h0015http://refhub.elsevier.com/S1383-5866(18)32136-1/h0020http://refhub.elsevier.com/S1383-5866(18)32136-1/h0020http://refhub.elsevier.com/S1383-5866(18)32136-1/h0020http://refhub.elsevier.com/S1383-5866(18)32136-1/h0025http://refhub.elsevier.com/S1383-5866(18)32136-1/h0025http://refhub.elsevier.com/S1383-5866(18)32136-1/h0025http://refhub.elsevier.com/S1383-5866(18)32136-1/h0025http://refhub.elsevier.com/S1383-5866(18)32136-1/h0030http://refhub.elsevier.com/S1383-5866(18)32136-1/h0030http://refhub.elsevier.com/S1383-5866(18)32136-1/h0030http://refhub.elsevier.com/S1383-5866(18)32136-1/h0035http://refhub.elsevier.com/S1383-5866(18)32136-1/h0035http://refhub.elsevier.com/S1383-5866(18)32136-1/h0040http://refhub.elsevier.com/S1383-5866(18)32136-1/h0040http://refhub.elsevier.com/S1383-5866(18)32136-1/h0040http://refhub.elsevier.com/S1383-5866(18)32136-1/h0045http://refhub.elsevier.com/S1383-5866(18)32136-1/h0045http://refhub.elsevier.com/S1383-5866(18)32136-1/h0045http://refhub.elsevier.com/S1383-5866(18)32136-1/h0050http://refhub.elsevier.com/S1383-5866(18)32136-1/h0050http://refhub.elsevier.com/S1383-5866(18)32136-1/h0050http://refhub.elsevier.com/S1383-5866(18)32136-1/h0055http://refhub.elsevier.com/S1383-5866(18)32136-1/h0055http://refhub.elsevier.com/S1383-5866(18)32136-1/h0060http://refhub.elsevier.com/S1383-5866(18)32136-1/h0060http://refhub.elsevier.com/S1383-5866(18)32136-1/h0060http://refhub.elsevier.com/S1383-5866(18)32136-1/h0065http://refhub.elsevier.com/S1383-5866(18)32136-1/h0065http://refhub.elsevier.com/S1383-5866(18)32136-1/h0070http://refhub.elsevier.com/S1383-5866(18)32136-1/h0070http://refhub.elsevier.com/S1383-5866(18)32136-1/h0070http://refhub.elsevier.com/S1383-5866(18)32136-1/h0075http://refhub.elsevier.com/S1383-5866(18)32136-1/h0075http://refhub.elsevier.com/S1383-5866(18)32136-1/h0075http://refhub.elsevier.com/S1383-5866(18)32136-1/h0080http://refhub.elsevier.com/S1383-5866(18)32136-1/h0080http://refhub.elsevier.com/S1383-5866(18)32136-1/h0080http://refhub.elsevier.com/S1383-5866(18)32136-1/h0085http://refhub.elsevier.com/S1383-5866(18)32136-1/h0085http://refhub.elsevier.com/S1383-5866(18)32136-1/h0085http://refhub.elsevier.com/S1383-5866(18)32136-1/h0090http://refhub.elsevier.com/S1383-5866(18)32136-1/h0090http://refhub.elsevier.com/S1383-5866(18)32136-1/h0090http://refhub.elsevier.com/S1383-5866(18)32136-1/h0095http://refhub.elsevier.com/S1383-5866(18)32136-1/h0095http://refhub.elsevier.com/S1383-5866(18)32136-1/h0095http://refhub.elsevier.com/S1383-5866(18)32136-1/h0100http://refhub.elsevier.com/S1383-5866(18)32136-1/h0100http://refhub.elsevier.com/S1383-5866(18)32136-1/h0105http://refhub.elsevier.com/S1383-5866(18)32136-1/h0105http://refhub.elsevier.com/S1383-5866(18)32136-1/h0110http://refhub.elsevier.com/S1383-5866(18)32136-1/h0110http://refhub.elsevier.com/S1383-5866(18)32136-1/h0115http://refhub.elsevier.com/S1383-5866(18)32136-1/h0115http://refhub.elsevier.com/S1383-5866(18)32136-1/h0115http://refhub.elsevier.com/S1383-5866(18)32136-1/h0120http://refhub.elsevier.com/S1383-5866(18)32136-1/h0120http://refhub.elsevier.com/S1383-5866(18)32136-1/h0120http://refhub.elsevier.com/S1383-5866(18)32136-1/h0120http://refhub.elsevier.com/S1383-5866(18)32136-1/h0125http://refhub.elsevier.com/S1383-5866(18)32136-1/h0125http://refhub.elsevier.com/S1383-5866(18)32136-1/h0125http://refhub.elsevier.com/S1383-5866(18)32136-1/h0130http://refhub.elsevier.com/S1383-5866(18)32136-1/h0130http://refhub.elsevier.com/S1383-5866(18)32136-1/h0130http://refhub.elsevier.com/S1383-5866(18)32136-1/h0135http://refhub.elsevier.com/S1383-5866(18)32136-1/h0135http://refhub.elsevier.com/S1383-5866(18)32136-1/h0135http://refhub.elsevier.com/S1383-5866(18)32136-1/h0140http://refhub.elsevier.com/S1383-5866(18)32136-1/h0140http://refhub.elsevier.com/S1383-5866(18)32136-1/h0140http://refhub.elsevier.com/S1383-5866(18)32136-1/h0140http://refhub.elsevier.com/S1383-5866(18)32136-1/h0145http://refhub.elsevier.com/S1383-5866(18)32136-1/h0145http://refhub.elsevier.com/S1383-5866(18)32136-1/h0145

[30] E.J. Lee, A.K. An, P. Hadi, S. Lee, Y.C. Woo, H.K. Shon, Advanced multi-nozzleelectrospun functionalized titanium dioxide/polyvinylidene fluoride-co-hexa-fluoropropylene (TiO2/PVDF-HFP) composite membranes for direct contact mem-brane distillation, J. Membr. Sci. 524 (2017) 712–720.

[31] W.W. Zhong, J.W. Hou, H.C. Yang, V. Chen, Superhydrophobic membranes viafacile bio-inspired mineralization for vacuum membrane distillation, J. Membr. Sci.540 (2017) 98–107.

[32] L. Kong, Y. Li, F. Qiu, T. Zhang, Q. Guo, X. Zhang, D. Yang, J. Xu, M. Xue,Fabrication of hydrophobic and oleophilic polyurethane foam sponge modified withhydrophobic Al2O3 for oil/water separation, J. Ind. Eng. Chem. (2017).

[33] J.D. Berry, M.J. Neeson, R.R. Dagastine, D.Y.C. Chan, R.F. Tabor, Measurement ofsurface and interfacial tension using pendant drop tensiometry, J. Colloid Interf.Sci. 454 (2015) 226–237.

[34] J.T. Korhonen, T. Huhtamaki, O. Ikkala, R.H.A. Ras, Reliable measurement of thereceding contact angle, Langmuir 29 (2013) 3858–3863.

[35] H. Ji, R. Zhao, Y.M. Li, B.L. Sun, Y.Z. Li, N. Zhang, J. Qiu, X. Li, C. Wang, Robust anddurable superhydrophobic electrospun nanofibrous mats via a simple Cu na-nocluster immobilization for oil-water contamination, Colloid Surf. A 538 (2018)173–183.

[36] C.R. Reshmi, S.P. Sundaran, A. Juraij, S. Athiyanathil, Fabrication of super-hydrophobic polycaprolactone/beeswax electrospun membranes for high-efficiencyoil/water separation, RSC Adv. 7 (2017) 2092–2102.

[37] V. Pillay, C. Dott, Y.E. Choonara, C. Tyagi, L. Tomar, P. Kumar, L.C. du Toit,V.M.K. Ndesendo, a review of the effect of processing variables on the fabrication ofelectrospun nanofibers for drug delivery applications, J. Nanomater. (2013).

[38] S.M.S. Shahabadi, S.A. Mousavi, D. Bastani, High flux electrospun nanofiberousmembrane: preparation by statistical approach, characterization, and microfiltra-tion assessment, J. Taiwan Inst. Chem. E 59 (2016) 474–483.

[39] S.M.S. Shahabadi, A. Kheradmand, V. Montazeri, H. Ziaee, Effects of process andambient parameters on diameter and morphology of electrospun polyacrylonitrilenanofibers, Polym. Sci. Ser. a+ 57 (2015) 155–167.

[40] E.J. Lee, A.K. An, T. He, Y.C. Woo, H.K. Shon, Electrospun nanofiber membranesincorporating fluorosilane-coated TiO2 nanocomposite for direct contact membranedistillation, J. Membr. Sci. 520 (2016) 145–154.

[41] Z.D. Shengpeng SHI, Hong YE, Chen ZHANG, Hangquan LI, A novel carbon black/polydimethylsiloxane composite membrane with high flux for the separation of

ethanol from water by pervaporation, Polym. J. 38 (2006) 949–955.[42] M. Zhang, S. Feng, L. Wang, Y. Zheng, Lotus effect in wetting and self-cleaning,

Biotribology 5 (2016) 31–43.[43] M.S. Bell, A. Shahraz, K.A. Fichthorn, A. Borhan, Effects of hierarchical surface

roughness on droplet contact angle, Langmuir 31 (2015) 6752–6762.[44] L. Shi, R. Wang, Y. Cao, C. Feng, D.T. Liang, J.H. Tay, Fabrication of poly(vinyli-

dene fluoride-co-hexafluropropylene) (PVDF-HFP) asymmetric microporous hollowfiber membranes, J. Membr. Sci. 305 (2007) 215–225.

[45] M. Kang, R. Jung, H.S. Kim, H.J. Jin, Preparation of superhydrophobic polystyrenemembranes by electrospinning, Colloid Surf. A 313 (2008) 411–414.

[46] R.N. Wenzel, Resistance of solid surfaces to wetting by water, Ind. Eng. Chem. 28(1936) 988–994.

[47] A.B.D. Cassie, S. Baxter, Wettability of porous surfaces, Trans. Faraday Soc. 40(1944) 546–551.

[48] J. Kujawa, W. Kujawski, Functionalization of ceramic metal oxide powders andceramic membranes by perfluoroalkylsilanes and alkylsilanes possessing differentreactive groups: physicochemical and tribological properties, ACS Appl. Mater.Interfaces 8 (2016) 7509–7521.

[49] M. Nosonovsky, B. Bhushan, Lotus versus rose: biomimetic surface, Effects GreenTribol. (2012) 25–40.

[50] L. Feng, Y. Zhang, J. Xi, Y. Zhu, N. Wang, F. Xia, L. Jiang, Petal effect: a super-hydrophobic state with high adhesive force, Langmuir 24 (2008) 4114–4119.

[51] X. Li, M.C. Garcia-Payo, M. Khayet, M. Wang, X. Wang, Superhydrophobic poly-sulfone/polydimethylsiloxane electrospun nanofibrous membranes for water desa-lination by direct contact membrane distillation, J. Membr. Sci. 542 (2017)308–319.

[52] Y. Liao, C.H. Loh, R. Wang, A.G. Fane, Electrospun superhydrophobic membraneswith unique structures for membrane distillation, ACS Appl. Mater. Interf. 6 (2014)16035–16048.

[53] J.Y. Wu, A.K. An, J.X. Guo, E.J. Lee, M.U. Farid, S. Jeong, CNTs reinforced super-hydrophobic-oleophilic electrospun polystyrene oil sorbent for enhanced sorptioncapacity and reusability, Chem. Eng. J. 314 (2017) 526–536.

[54] M. Xiang, M. Jiang, Y. Zhang, Y. Liu, F. Shen, G. Yang, Y. He, L. Wang, X. Zhang,S. Deng, Fabrication of a novel superhydrophobic and superoleophilic surface byone-step electrodeposition method for continuous oil/water separation, Appl. Surf.Sci. 434 (2018) 1015–1020.


599

http://refhub.elsevier.com/S1383-5866(18)32136-1/h0150http://refhub.elsevier.com/S1383-5866(18)32136-1/h0150http://refhub.elsevier.com/S1383-5866(18)32136-1/h0150http://refhub.elsevier.com/S1383-5866(18)32136-1/h0150http://refhub.elsevier.com/S1383-5866(18)32136-1/h0155http://refhub.elsevier.com/S1383-5866(18)32136-1/h0155http://refhub.elsevier.com/S1383-5866(18)32136-1/h0155http://refhub.elsevier.com/S1383-5866(18)32136-1/h0160http://refhub.elsevier.com/S1383-5866(18)32136-1/h0160http://refhub.elsevier.com/S1383-5866(18)32136-1/h0160http://refhub.elsevier.com/S1383-5866(18)32136-1/h0165http://refhub.elsevier.com/S1383-5866(18)32136-1/h0165http://refhub.elsevier.com/S1383-5866(18)32136-1/h0165http://refhub.elsevier.com/S1383-5866(18)32136-1/h0170http://refhub.elsevier.com/S1383-5866(18)32136-1/h0170http://refhub.elsevier.com/S1383-5866(18)32136-1/h0175http://refhub.elsevier.com/S1383-5866(18)32136-1/h0175http://refhub.elsevier.com/S1383-5866(18)32136-1/h0175http://refhub.elsevier.com/S1383-5866(18)32136-1/h0175http://refhub.elsevier.com/S1383-5866(18)32136-1/h0180http://refhub.elsevier.com/S1383-5866(18)32136-1/h0180http://refhub.elsevier.com/S1383-5866(18)32136-1/h0180http://refhub.elsevier.com/S1383-5866(18)32136-1/h0185http://refhub.elsevier.com/S1383-5866(18)32136-1/h0185http://refhub.elsevier.com/S1383-5866(18)32136-1/h0185http://refhub.elsevier.com/S1383-5866(18)32136-1/h0190http://refhub.elsevier.com/S1383-5866(18)32136-1/h0190http://refhub.elsevier.com/S1383-5866(18)32136-1/h0190http://refhub.elsevier.com/S1383-5866(18)32136-1/h0195http://refhub.elsevier.com/S1383-5866(18)32136-1/h0195http://refhub.elsevier.com/S1383-5866(18)32136-1/h0195http://refhub.elsevier.com/S1383-5866(18)32136-1/h0200http://refhub.elsevier.com/S1383-5866(18)32136-1/h0200http://refhub.elsevier.com/S1383-5866(18)32136-1/h0200http://refhub.elsevier.com/S1383-5866(18)32136-1/h0205http://refhub.elsevier.com/S1383-5866(18)32136-1/h0205http://refhub.elsevier.com/S1383-5866(18)32136-1/h0205http://refhub.elsevier.com/S1383-5866(18)32136-1/h0210http://refhub.elsevier.com/S1383-5866(18)32136-1/h0210http://refhub.elsevier.com/S1383-5866(18)32136-1/h0215http://refhub.elsevier.com/S1383-5866(18)32136-1/h0215http://refhub.elsevier.com/S1383-5866(18)32136-1/h0220http://refhub.elsevier.com/S1383-5866(18)32136-1/h0220http://refhub.elsevier.com/S1383-5866(18)32136-1/h0220http://refhub.elsevier.com/S1383-5866(18)32136-1/h0225http://refhub.elsevier.com/S1383-5866(18)32136-1/h0225http://refhub.elsevier.com/S1383-5866(18)32136-1/h0230http://refhub.elsevier.com/S1383-5866(18)32136-1/h0230http://refhub.elsevier.com/S1383-5866(18)32136-1/h0235http://refhub.elsevier.com/S1383-5866(18)32136-1/h0235http://refhub.elsevier.com/S1383-5866(18)32136-1/h0240http://refhub.elsevier.com/S1383-5866(18)32136-1/h0240http://refhub.elsevier.com/S1383-5866(18)32136-1/h0240http://refhub.elsevier.com/S1383-5866(18)32136-1/h0240http://refhub.elsevier.com/S1383-5866(18)32136-1/h0245http://refhub.elsevier.com/S1383-5866(18)32136-1/h0245http://refhub.elsevier.com/S1383-5866(18)32136-1/h0250http://refhub.elsevier.com/S1383-5866(18)32136-1/h0250http://refhub.elsevier.com/S1383-5866(18)32136-1/h0255http://refhub.elsevier.com/S1383-5866(18)32136-1/h0255http://refhub.elsevier.com/S1383-5866(18)32136-1/h0255http://refhub.elsevier.com/S1383-5866(18)32136-1/h0255http://refhub.elsevier.com/S1383-5866(18)32136-1/h0260http://refhub.elsevier.com/S1383-5866(18)32136-1/h0260http://refhub.elsevier.com/S1383-5866(18)32136-1/h0260http://refhub.elsevier.com/S1383-5866(18)32136-1/h0265http://refhub.elsevier.com/S1383-5866(18)32136-1/h0265http://refhub.elsevier.com/S1383-5866(18)32136-1/h0265http://refhub.elsevier.com/S1383-5866(18)32136-1/h0270http://refhub.elsevier.com/S1383-5866(18)32136-1/h0270http://refhub.elsevier.com/S1383-5866(18)32136-1/h0270http://refhub.elsevier.com/S1383-5866(18)32136-1/h0270

Bio-inspired superhydrophobic and superoleophilic nanofibrous membranes for non-aqueous solvent and oil separation from waterIntroductionMaterials and methodsMaterialsMembrane fabricationNanofibrous membrane preparationSuperhydrophobic layer coating by electrospraying

Membrane characterizationOil/Water separationOil adsorption to membraneSolvent/water separation

Results and discussionMembrane surface morphology and roughnessMembrane surface chemistryPorosity, pore size and liquid entry pressureWettability as a function of liquids surface tensionMechanical and chemical durabilityAdsorption and permeation of apolar solventsMembrane performance: Separation of apolar solvents from aqueous mixtures

ConclusionsAcknowledgementsSupplementary materialReferences

Documents

Separation and Puriﬁcation Technology · 2018. 6. 19. · Seyed Mahdi Seyed Shahabadi, Jonathan A. Brant⁎ University of Wyoming, Department of Civil and Architectural Engineering,