Desalination 399 (2016) 178–184

Contents lists available at ScienceDirect

Desalination

j ourna l homepage: www.e lsev ie r .com/ locate /desa l

Nanostructured PVDF membrane for MD application by an O2 and CF4plasma treatment

Seongpil Jeong a, Bongsu Shin b,c, Wonjin Jo b, Ho-Young Kim c, Myoung-Woon Moon b, Seockheon Lee a,⁎a Water and Resource Recycle Center, Korea Institute of Science and Technology, Seoul 02792, Republic of Koreab Materials and Life Science Research Division, Korea Institute of Science and Technology, Seoul 02792, Republic of Koreac School of Mechanical and Aerospace Engineering, Seoul National University, Seoul 08826, Republic of Korea

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• The PVDF membranes were modifiedby plasma treatments with O2 and CF4gases.

• The modified membrane have an en-hanced flux and hydrophobicity afterthe plasma treatment.

• The pore-size and porosity at the mem-brane surface increased after the plas-ma treatment.

• The hairy structure was formed at themembrane surface by the HMDSO coat-ing.

• The O2 plasma treatment with HMDSOcoating increased the receding contactangle.

⁎ Corresponding author.E-mail address: seocklee@kist.re.kr (S. Lee).

http://dx.doi.org/10.1016/j.desal.2016.09.0010011-9164/© 2016 Published by Elsevier B.V.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 26 April 2016Received in revised form 2 August 2016Accepted 5 September 2016Available online xxxx

Recently, various nanotechnologies have been utilized with regard to membrane modification due to their highactivities and the low cost of the nanomaterials involved. In order to enhance the hydrophobicity of the mem-brane surface for membrane distillation applications by decreasing the surface energy, a radio-frequency plas-ma-enhanced chemical vapor deposition (RF-PECVD) process is suggested with surface nanostructuring and asubsequent hydrophobic coating step. In this research, a commercial PVDF membrane was modified by plasmatreatments with the two different gases of O2 and CF4. The water contact angles of the active layers increasedfrom 73 to 117 and 101° and the fluxes of the treated membranes increased to 63 and 27.9% as compared to avirgin PVDF membrane when the feed used was D.I. water by the O2 and CF4 plasmamodifications, respectively.Defluorination at the long exposure time (120 min) of the plasma treatment and increase of the overall hydro-phobicity (the decrease of the contact angle hysteresis) by the HMDSO coating were the reasons of the flux var-iations for the plasma modified membranes.

© 2016 Published by Elsevier B.V.

Keywords:Membrane distillationO2 plasma treatmentCF4 plasma treatmentNanostructured membranePECVDHMDSO

1. Introduction

Membrane distillation (MD) is a thermally driven membrane pro-cess. The main mechanism of membrane distillation is the vapor pres-sure difference between the feed and the cooling sides. Water vapor

179S. Jeong et al. / Desalination 399 (2016) 178–184

evaporated from the hot feed sidemoves through a porous hydrophobicmembrane and condenses as water drops on the cold cooling side. Be-cause only vapor is transported across the membrane, the salt rejectionrate for non-volatile ions under the operating conditions of theMD pro-cess is theoretically 100%. In the desalination field, the MD process isconsidered as a next-generation technology due to the small flux de-crease, though the feed solution is highly saline. Recently, pilot plantshave been built worldwide in order to commercialize the MD systemin the desalination field or for other possible applications.

One of the major hurdles preventing the commercialization of MDtechnology is the absence of a high-fluxmembrane. In order to develophigh-performanceMDmembranes, various approaches have been used.Different polymeric and ceramic materials (PP, PVDF, PTFE, PE and zir-conia) have been utilized as the active layer of MD membranes [1–3].Flat-sheet and hollow-fiber membranes with a single layer or with abi-layer structure have been tested under various MD configurations[4–6]. Novel methods also have been used to fabricate a nanofiber or amesh of carbon nanotubes [7–9].

Surface modification refers to a technique to enhance the requiredperformance of a membrane by adding a small amount of a functionalgroup onto the surface or by changing the surface characteristics bymeans of physical or chemical methods. Due to thewidespread applica-tion of surface modification methods, several practical technologies formembranes have been developed. Onemethod involves the attachmentof a functional group of ions (zwitterions and hydrophilic matter) toprovide anti-fouling characteristics on the membrane surface [10,11].Membrane characteristics such as the hydrophobicity, surface chargeand roughness have also been shown to be feasible targets in the effortto develop a high-performance membrane [12–14]. In order to changethe surface characteristics of the membrane, various techniques havebeen suggested, such as surface polymerization, layer-by-layermethods, chemical vapor deposition (CVD), or a plasma treatment[14–19].

In this study, a radio-frequency plasma-enhanced chemical vapordeposition (RF-PECVD) process was selected for use to improve the hy-drophobicity of the MD membrane. In dry etching technologies such asRF-PECVD, the membrane sample can be immersed in various reactivegases (O2, CF4, etc.). The membrane samples can be modified by twomajor reactions (physical bombardment by ions and a chemical reac-tion caused by radicals) in the RF-PECVD chamber. It is known that rad-ical reactions cause greater modifications than physical ionbombardment. The emitted ions are responsible for the physical (ener-getic) bombarding of the surface, which influences the chemical etchingprocess. The generated radicals react with the surface to produce vola-tile products (CO, CO2, etc.), which are responsible for the dry etchingand which become attached to the membrane surface by forming sitesfor polymerizationwith other additives [20]. In earlier work, the plasmapolymerization of hexamethyldisiloxane (HMDSO), which has low sur-face energy of 24.2 mN/m, was shown to increase the degree of surfacehydrophobicity [21,22].

Other researchers have also utilized a plasma treatment to modifythe surfaces of MD membranes. Yang et al. modified an original hollowfiber membrane using a P2i plasma enhancement machine [16]. InYang's research, a hydrophobic polymeric nano-coating with 1H, 1H,2H, and 2H-perfluorodecyl acrylate was produced on the membranesurface by a plasma treatment. The LEP (liquid entry pressure) was en-hanced due to the increased contact angle and the decreased pore sizeof themodifiedmembrane.Wei et al. modified a hydrophilic polyether-sulfone (PES) membrane with a CF4 plasma treatment [17]. The contactangle of the plasma-treated membrane increased from 60° to 120° andwas maintained with excessive plasma power and exposure times. Theflux of the PES membrane modified with CF4 plasma was 66.7 kg/m2 hwhen the feed and permeate temperatureswere 73.8 and 20 °C, respec-tively. This represents one of the highest flux levels relative to those re-ported from the references in Wei's research when variouscharacteristics of the membranes and operating conditions are

considered. Yang et al. recently suggested that the flux of plasma-mod-ified membranes is enhanced due to the increase in the effective evap-oration caused by the change of the surface hydrophobicity [19].However, researchwith respect to the relationship betweenmembranecharacteristics and themembrane performance is still limited. In this re-search, we focused on the relationship between the hydrophobicity andmembrane flux. First, the static and dynamic hydrophobicities weremeasured for themodifiedmembrane surfaces in order to check the hy-drophobicity uniformity. We also tried to compare the modified mem-brane surfaces and their fluxes according to the introducing gases ofthe plasma treatment.

To investigate the performance enhancement caused by a plasmatreatment of a MDmembrane, a commercial PVDFmembrane was cho-sen and modified 1) by an O2 plasma treatment with a subsequent hy-drophobic coating and 2) solely by CF4 plasma. The exposure time ofthe each plasma treatment and the number of modifiedmembrane sur-face (active layer only or active/support layer together) for each mem-brane was varied. Furthermore, to understand the relationshipbetween the modified membrane characteristics and the membraneflux, SEM-EDX and contact angle analyses were conducted.

2. Experimental method

2.1. Membrane

A commercial PVDFmembranemanufactured by a membrane com-pany (S80306, Pall, U.S.A.) was selected as the candidate membrane forthe plasma treatment. The single-layer flat-sheet type of membranechosen for use here had a pore size of 0.45 μm and a membrane thick-ness of 147 μm.

2.2. Hydrophobic coating by a plasma treatment

The PVDFmembrane was modified using O2 and CF4 plasma via RF-PECVD. The schematic diagram of the experimental set-up is shown inFig. 1. The voltage of the CVD machine was fixed at 400 V (62 W) andthe vacuum condition in the CVD chamber was 20 mTorr. The mem-branes were modified by O2 plasma and CF4 plasma in the followingthree ways. First, a hydrophobic coating was deposited onto the mem-brane facing the feed side by means of an O2 plasma treatment (anO2‐hydrophobic coating). The procedure of the O2 plasma coating hadtwo steps. In the first step, the target pristine PVDF membrane wasplaced into the CVD chamber and was exposed to the plasma with O2

gas from 30 to 120 min according to the experimental conditions. Sec-ondly, the plasma-treated membrane was coated with HMDSO(hexamethyldisiloxane, Sigma Aldrich, U.S.A.) for 30 s. Because the O2

plasma-treated surface becomes hydrophilic, the low-surface-energymaterial of HMDSO, with a surface energy level of 24.4 mJ/m2, i.e.,lower than the surface energy (31.6 mJ/m2) of the PVDF membrane,was used as a coating [23]. Second, with this method, a hydrophobicand hydrophilic coatingwas deposited onto themembrane surfaces fac-ing the feed and the permeate sides via theO2 plasma treatment (an O2-hydrophobic/hydrophilic coating). One side of the target membranefacing the feed solutionwas coated using the same procedure describedabove, but for the other side of the target membrane facing the perme-ate solution, the HMDSO coating was omitted, leading to asymmetricwettability. The third way, involving the CF4 plasma treatment, wasone in which the HMDSO coating was not conducted because the CF4-treated surface became hydrophobic (a CF4‐hydrophobic coating). Theexposure time of the CF4 coating ranged from 60 to 120 m. Detailed in-formation about the testing conditions is given in Table 1.

2.3. Module and DCMD operation

A plate and frame type of acryl module was used for the direct con-tact membrane distillation (DCMD) test conducted. The dimensions of

Fig. 1. Schematic diagram of rf-PECVD and DCMD systems.

180 S. Jeong et al. / Desalination 399 (2016) 178–184

the water channel for the feed and permeate sides are identical at0.026 m × 0.077 m (the effective area of the membrane) × 0.003 m(the height of the water channel). The inlet and outlet are connectedto themodule perpendicularly to thewater flow. The targetmembraneswere placed in a DCMD module and operated for 90 min under thegiven experimental conditions. The operating conditions (i.e., feed andpermeate temperature, flow rate) are controlled using a water bathwith a stainless steel coil and by a flow-controllable gear pump(75211-15, Cole-Parmer Instrument Company, U.S.A.). The feed andpermeate temperatures in all tests were 60 and 20 °C, respectively.The cross-flow velocities of the feed and permeate were 0.34 m/s and0.26 m/s, respectively. The experimental data (i.e., the increase in theweight of the permeate solution, the EC in the feed and permeate) arerecorded by a balance and with an electronic conductivity meter(HQ40d, Hach, U.S.A.) and are automatically stored on a laptop comput-er. The details and operation of the experiment unit were sourced from

Table 1Detailed experimental conditions (plasma conditions and exposure times for the pristineand plasma modified membranes).

Abbreviation Plasma condition Exposuretime(min)

O2 plasma treatment CF4 plasmatreatment(hydrophobiccoating)

Hydrophobiccoating (H)

Hydrophobic/hydrophiliccoating (HH)

O2-H-30 O 30a

O2-H-120 O 120a

O2-HH-30,30 O 30, 30b

O2-HH-80,60 O 80, 60b

CF4-H-60 O 60c

CF4-H-90 O 90c

CF4-H-120 O 120c

a (T) O2 plasma (exposure time, A) + HMDSO coating (30 s).b (TF, TP) feed side: O2 plasma (exposure time, TF) + HMDSO coating (30 s)/permeate

side: O2 plasma (exposure time, TP).c (T) CF4 plasma (exposure time, A).

previous work [24]. The feed and permeate solutions were D.I. water.For the integrity test, 0.6 M of a NaCl solution was used in the feedwith the O2-plasma-treated membrane (O2 plasma for 30 min andHMDSO for 30 s). The permeate conductivity levels were maintainedat 1–2 μS/cm for an operation time of 90 min when the feed was the0.6 M NaCl solution.

2.4. Instrumental analysis

In order to measure the variation of the hydrophobicity, surfacemorphology andelemental composition of themodifiedmembrane sur-faces, the following instrumental analyses were conducted. Contactangle (CA) observations of pristine and variously modified membraneswere conducted using the sessile drop method with D.I. water and amicro-syringe on a static contact angle apparatus (DSA100, KRÜSS, Ger-many). The device measured the static contact angle 18–20 times pereach water drop and at least 2 water drops were observed for eachmembrane sample. The contact angle hysteresis (CAH) was also mea-suredwith a dynamic contact angle analyzer (Sigma700, Biolin Scientif-ic, Finland). The surfaces of the pristine PVDF and plasma-modifiedmembranes were observed by FE-SEM (S-4200, Hitachi, Japan) and E-SEM with Energy-Dispersive Spectroscopy (EDS) (XL-30, FEI, USA).

3. Results and discussion

3.1. Flux of the modified membrane after the plasma treatment

3.1.1. O2 hydrophobic coating and hydrophobic/hydrophilic coatingFig. 2 shows the flux variations according to the exposure time to O2

plasma at 30 s for the HMDSO coating (the O2-hydrophobic coating).The flux significantly increased when the exposure time of the O2 plas-mawas 120min (shown as a black triangle in Fig. 2). The average fluxesof themodifiedmembranes (O2-H-30 (shown as an empty square in Fig.2) and O2-H-120) were 21.9 and 33.2 kg/m2 h, respectively, while theaverage flux of the virgin membrane (shown as a gray diamond in Fig.2)was 20.4 kg/m2 h. Themodifiedmembranes treatedwith theO2 plas-mawith the HMDSO coating showed higher flux levels by 7.6 and 63.0%compared to the pristinemembrane depending on the O2 plasma treat-ment exposure time.

The flux behaviors according to the exposure time of the O2 plasmatreatment (both sides of themembrane)with theHMDSO coating (onlythe feed side of themembrane) for the PVDFmembrane are described inFig. 3 (O2-hydrophobic/hydrophilic coating). The average flux increasedfrom 23.0 to 28.0 kg/m2 h with longer exposure times of the O2 plasma(O2-HH-30, 30 and O2-HH-80, 60). The flux (23.0 kg/m2 h) of the mod-ifiedmembrane (O2-HH-30, 30)with both sides of themembrane treat-ed was slightly higher than that (21.9 kg/m2 h) of the modifiedmembrane with a single side of the membrane treated (O2-H-30) atan identical O2 exposure time of 30 min. One possible reason for the in-creased flux of the modified membrane with both sides treated is thedecrease in the effective membrane thickness, as the membrane onthe permeate side becamehydrophilic due to the grafting of hydrophilicpolar oxygenated species [25,26]. Consequently, after the O2 plasmatreatment with the HMDSO coating on the PVDF membrane, the fluxof the modified membrane increased with the O2 plasma exposuretime under the given experimental conditions.

3.1.2. CF4 hydrophobic coatingThe flux behavior according to the exposure time of the CF4 plasma

for the target membrane is shown in Fig. 4. The average fluxes of themodified membranes (CF4-H-60, CF4-H-90 and CF4-H-120) accordingto the CF4 exposure times (60, 90 and 120 min) were 27.9, 24.6 and25.5 kg/m2 h, respectively. All of the modified membranes showed en-hanced flux compared to the virgin membrane. Unlike the O2 plasmatreatments results, the flux did not distinctly increase according to theplasma exposure time under the given experimental conditions (CF4

Fig. 2. Flux behavior according to the exposure time of the O2 plasma treatment (singleside of the membrane) with the HMDSO coating on the PVDF membrane (feed water:D.I. water, corresponding temperatures of the feed and permeate: 60 °C and 20 °C,corresponding cross-flow velocities of the feed and permeate: 0.34 m/s and 0.26 m/s).

Fig. 4. Flux behavior according to the exposure time of the CF4 plasma treatment (singleside of the membrane) for the PVDF membrane (feed water: D.I. water, correspondingtemperatures of the feed and permeate: 60 °C and 20 °C, corresponding cross-flowvelocities of the feed and permeate: 0.34 m/s and 0.26 m/s).

181S. Jeong et al. / Desalination 399 (2016) 178–184

plasma exposure times of 60–120min). Plateaus of themembrane char-acteristics (surface porosity and pore size) and the performance (flux)according to the exposure time of the plasma treatment were observedin earlier studies when the exposure time was excessive [17,18]. How-ever, in case of the O2 plasma treatment, the plateaus was not observedunder given experimental conditions (exposure time from 30 min to120min). Moreover, the flux (33.2 kg/m2 h) of themodifiedmembranewhich underwent the O2 plasma treatment was higher than that(25.5 kg/m2 h) of that of the CF4 plasma treatmentwith a long exposuretime (120min). The fluxes of the plasma treated membranes by O2 andCF4 plasma were enhanced up to 63 to 27.9%, respectively. With in-crease of themembrane flux, the required membrane area could be de-creased. Therefore, the capital cost including the membrane(purchasing and replacing), modulation, piping and land (foot print)costs could decrease. Though the PTFE membrane has promising char-acteristics on hydrophobicity and flux, the plasmamodified PVDFmem-brane alsomight have chance to be commercialized due to the difficultyin fabrication and the limitation in the modification of the PTFE mem-brane. Many researchers have been modified a PVDF material for theMD application (18–19, 27).

In order to understand the reasons for theflux variation according tothe exposure time of the plasma treatment and type of the plasma, themembrane characteristics (i.e., the pore size, porosity and membranethickness) and the hydrophobicity (CA and CAH) were analyzed.

Fig. 3. Flux behavior according to the exposure time of the O2 plasma treatment (bothsides of the membrane) with the HMDSO coating (feed side of the membrane only) onthe PVDF membrane (feed water: D.I. water, corresponding temperatures of the feedand permeate: 60 °C and 20 °C, corresponding cross-flow velocities of the feed andpermeate: 0.34 m/s and 0.26 m/s).

3.2. Effect of the plasma treatment on the physical and chemical character-istics of the membrane

3.2.1. Surface morphology and membrane thickness according to the typeand exposure time of the plasma

The surfaces of the virgin and modified membranes were observedby SEM. The surface morphologies after the plasma treatments showeddistinct increases in the pore size and porosity compared to those of thevirgin membranes (Figs. 5 and 6). As shown in Fig. 5(a) (the virginmembrane) and Fig. 5(b) and (c) (the O2-plasma-modified mem-branes), the pore size and porosity on the surface of the PVDF mem-brane increased significantly with an increase in the exposure time ofthe O2 plasma up to 120 min. After the O2 plasma treatment (O2-H-120), hairy structures were found to have formed on the modifiedmembranes (Fig. 5(c)). The hairy structures are related to the surfacewettability, as the surface had become more hydrophobic [28]. This re-sult corresponded to the increased flux of the PVDFmembranewith theincreased exposure time of the O2 plasma treatment (Fig. 2). For the CF4plasma treatment of the PVDF membrane (Fig. 6), all of the modifiedmembranes had similar surface morphologies under the given the plas-ma exposure times after the plasma treatment. The results of observa-tions of the SEM images correspond to the flux results, indicating thatthe plasma treatment enhances the flux of the modified membrane bychanging its surface morphology. In summary, the increases in thepore size and porosity may affect the increase in the flux of the PVDFmembrane when the membrane was modified by the plasma treat-ment. The plasmaetchedmembrane surfaceswere similarwhen the ex-posure time of the plasma treatment was higher than 30 min, however,in the case of O2-H-120 experiment, the hairy structure was observed.

Additionally, themembrane thicknesses of the pristine andmodifiedmembranes were investigated in cross-sectional images taken from theSEM analysis. The membrane thicknesses of the tested membranesranged from 100 to 120 μm, and no significant change in themembranethickness occurred after the plasma treatment.

3.2.2. Hydrophobic characteristics of the modified membraneThemeasured static contact angle (CA) and contact angle hysteresis

(CAH) for the virgin and modified membranes and the fluxes undergiven experimental condition are listed in Table 2. The advancing andreceding contact angles are themaximumandminimumcontact angles,respectively. Additionally, the contact angle hysteresis can be calculatedfrom the difference between the advancing and receding contact angles.

The static contact angle of the O2-plasma-modified membrane in-creased with the plasma exposure time (Table 2). Camacho et al.found that the overall flux of the membrane distillation was dependent

(a)

(b)

(c)

5 µm

5 µm

5 µm

Fig. 5. Surface images of the virgin and modified membranes according to the O2 plasmaexposure time ((a) virgin PVDF membrane, (b) O2 plasma (30 min) + HMDSO (30 s),(c) O2 plasma (120 min) + HMDSO (30 s)).

(a)

(b)

(c)

5 µm

5 µm

5 µm

Fig. 6. Surface images of the membranes according to the CF4 plasma exposure time ((a)CF4 plasma (60 min), (b) CF4 plasma (90 min), (c) CF4 plasma (120 min)).

182 S. Jeong et al. / Desalination 399 (2016) 178–184

with surface contact angle because hydrophobic structures has lowerthermal conductivity according to their review paper [29]. Fan et al. fab-ricated and tested the various hydrophobic PVDF membranes and theflux of the membrane which poses higher hydrophobicity had a higherflux [27]. However, the static contact angle of the modified membranesdid not increase with an increase in the plasma treatment time in thecase of the CH4 plasma treatment. These results correspond to the sur-face morphology variations (Fig. 6).

The CAH of the modified membrane decreased after the O2 and CF4plasma treatment (120 min) and the advancing and receding contactangles were increased. In detail, the O2-H-120 membrane showed a

higher receding contact angle (109.4°) and a lower CAH (36.4°) thanthe CF4-H-120 membrane even though the advancing contact anglesof both membranes are similar. This suggests that the O2-H-120 mem-brane has a more evenly distributed hydrophobic surface than theCF4-H-120 membrane. Moreover, when we compared the performanceresult (flux) between the O2-H-120 membrane and the CF4-H-120membrane, it could be concluded that the low adhesive hydrophobicsurface having low CAH results in high flux during themembrane distil-lation process. The suggested reason for the flux increase with an in-creased hydrophobicity of the active layer of the MD membrane wasthe increase of the effective evaporation area [19]. However, the flux in-crease according to the CAHdecrease has not been suggested. Thewatervapor pressure near the relatively adhesive surface (CF4 plasma modi-fied membrane than O2 plasma modified membrane) could be de-creased due to the decrease of the adhesion force between the watervapor and the membrane.

Table 2Static contact angle, contact angle hysteresis and flux of the pristine and plasma modified membranes.

Virgin membrane O2-H-30 O2-H-120 O2-HH-80,60 CF4-H-60 CF4-H-90 CF4-H-120

Static contact angle (°) 73.4 ± 4.9 99.5 ± 3.3 117.2 ± 14.5 103.8 ± 1.5 101.3 ± 3.6 101.4 ± 16.2 99.6 ± 3.4Dynamic contact angle (°) Advancing contact angle 111.5 – 145.8 – – – 145.6

Receding contact angle 28.5 – 109.4 – – – 81.7Contact angle hysteresis 83.0 – 36.4 – – – 63.9

Flux (kg/m2h) 20.4 21.9 33.2 28.0 27.9 24.6 25.5

183S. Jeong et al. / Desalination 399 (2016) 178–184

3.2.3. EDX analysis of the plasma-modified membranesThe chemical compositions of the pristine andmodifiedmembranes

were investigated in an EDX analysis, as shown in Table 3. Carbon (68%)and fluoride (32%) were the major components of the virgin PVDFmembrane. The measured fluorine-to-carbon (F/C) atom ratio of thevirgin membrane was 0.5, similar to the F/C atom ratio (0.48) of thePVDF membrane from the research of Masuelli et al. [30]. For the mod-ified membranes, the F/C atom ratios increased after the shorter O2 andCF4 plasma treatments (30min and 60min, respectively) and decreasedafter the longer plasma treatment (120 min).

The increase in the F/C ratio at an early stage of the O2 and CF4 plas-ma treatments may be due to the strong etching effect of the plasma inthese cases [31]. In case of the O2 plasma treatment, the produced oxy-gen radical in the plasma chamber liberates carbon and hydrogen fromthe PVDFmembrane by forming carbon dioxide (CO2) [32]. Additional-ly, after the O2 plasma treatment, oxygen attaches onto the modifiedmembrane surface and the HMDSO containing the oxygen and silicaalso attach onto the O2-plasma-treated area. The attached HMDSOmol-ecules can boost the surface hydrophobicity on the membrane surface[21,22]. As shown in Table 3, the oxygen and silica contents of the O2-plasma-treated membrane occupied b5.9 and 1.8 atomic percentage,respectively, due to the HMDSO coating. The increased silica contentswith the increased exposure time can be explained by the enhanced ad-sorption sites for the HMDSO coatings formed by the increased expo-sure time of the O2 plasma. The decomposed HMDSO and itshydrophobic functional group could be coated as SiOx–C:H film [21].Therefore, additional hydrophobicity could be achieved. For the CF4plasma treatment, the etching, atomic insertion, deposition and poly-merization could be happened [19]. The results (Table 3) showed thatthe fluorination happened at the modified membrane surface possiblyby etching and F atom insertion. Therefore, it can be expected thatthin F layer may be formed at the modified hydrophobic membranesurface.

The decrease in the F/C ratio (1.1–N0.6) after 120min of exposure tothe O2 and CF4 plasma can be explained as defluorination [33]. The O/Cratio also decrease at 120min of exposure time to the O2 plasma, whichrepresents that the decrease of the O content (deoxygenation). In caseof the O2 plasma modified membrane, the major contributor of the hy-drophobicity could be the HMDSO coating because the more exposuretime of the O2 plasma resulted in more Si content at themembrane sur-face (Table 3). The one of the possible candidates of the hydrophobicfunctional group is [-SiF2-O-](n). The formation of [-SiF2-O-](n) structurewas suggested as themajor factor for super-hydrophobic surface [34]. Inthis research, the occurred defluorination for O2-H-120 membrane

Table 3EDX analysis results for the pristine and plasma modified membranes.

Types of theplasma

Exposure time(min)

C(%)

F(%)

O(%)

Si(%)

F/C(%/%)

O/C(%/%)

Non-treated – 68 32 0 0 0.5 0O2 30 47.7 48.4 3.5 0.4 1.0 0.1

120 58.6 36.8 2.8 1.8 0.6 0.05CF4 60 50.2 49.8 0 0 1.0 0.0

90 50.3 49.7 0 0 1.0 0.0120 61.8 38.2 0 0 0.6 0.0

supported that the formation of the SiFxOy complex. The other candi-date functional groups formed at the membrane surface can be com-bined hydrophilic polar oxygenated species [25,26] with fluoride, O–CH2 or O–CH2–CF2 [35]. And the possible reason of the deoxygenationfor O2-H-120 membrane could be the decrease of those functionalgroups at the membrane surface (Table 3).

In case of the CF4 plasma treatment, the fluorinated layer is themajor contributor for the hydrophobicity of the modified membrane.The site for the F attachment can be limited due to the number of the hy-drogen at the membrane surface is limited. Therefore, when thedefluorination happened (CF4-H-120), the hydrophobicity and the fluxwere slightly decreased (Table 2).

4. Conclusion

A plasma treatment was utilized as a surface modification tool for aPVDF membrane. The flux of a hydrophilic PVDF membrane increased63% with the exposure time of an O2 plasma treatment. The contactangle of the treated membrane was also increased from 73 to 117°with the exposure time of the O2 plasma treatment with a subsequentHMDSO coating. The flux of themodifiedmembrane was also increased27.9% by a CF4 treatment; however, no significant enhancementwas ob-servedwhen the exposure time ranged from60 to 120min. Both plasma(O2 and CF4) treatments enhanced the hydrophobicity of the PVDFmembrane surface from 73 to 117 and 101°, respectively. The flux ofthe O2 plasma-modified membrane after a long plasma exposure time(120 min) was higher than that of the CF4 plasma-modified membranebecause the HMDSO coating enhanced the overall hydrophobicity bydecreasing CAH of the modified membrane and defluorinationhappened.

Acknowledgements

We greatly appreciate the analytical support from Mr. YounggilYoon (E-SEM) andMs. SoheeKim (FE-SEM) at theKorea Institute of Sci-ence and Technology. This research was supported by the Korean Min-istry of the Environment as an “Eco-Innovation Program”(Environmental Research Laboratory) (414-111-011) and a grant[MPSS-CG-2016-02] through theDisaster and SafetyManagement Insti-tute funded by Ministry of Public Safety and Security of Koreangovernment.

References

[1] K.W. Lawson, D.R. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997) 1–25.[2] M. Khayet, Membranes and theoretical modeling of membrane distillation: a re-

view, Adv. Colloid Interf. Sci. 164 (2011) 56–88.[3] S.R. Krajewski, W. Kujawski, M. Bukowska, C. Picard, A. Larbot, Application of

fluoroalkylsilanes (FAS) grafted ceramic membranes in membrane distillation pro-cess of NaCl solutions, J. Membr. Sci. 281 (2006) 253–259.

[4] S. Bonyadi, T.S. Chung, Flux enhancement in membrane distillation by fabrication ofdual layer hydrophilic–hydrophobic hollow fiber membranes, J. Membr. Sci. 306(2007) 134–146.

[5] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensive review,Desalination 287 (2012) 2–18.

[6] E.K. Summers, H.A. Arafat, J.H. Lienhard V, Energy efficiency comparison of single-stage membrane distillation (MD) desalination cycles in different configurations,Desalination 290 (2012) 54–66.

184 S. Jeong et al. / Desalination 399 (2016) 178–184

[7] L. Francis, N. Ghaffour, A.S. Alsaadi, S.P. Nunes, G.L. Amy, PVDF hollow fiber andnanofiber membranes for fresh water reclamation using membrane distillation, J.Membr. Sci. 49 (2014) 2045–2053.

[8] L.D. Tijing, J.-S. Choi, S. Lee, S.-H. Kim, H.K. Shon, Recent progress of membrane dis-tillation using electrospun nanofibrous membrane, J. Membr. Sci. 453 (2014)435–462.

[9] C.H. Ahn, Y. Baek, C. Lee, S.O. Kim, S. Kim, S. Lee, S.-H. Kim, S.S. Bae, J. Park, J. Yoon,Carbon nanotube-based membranes: fabrication and application to desalination, J.Ind. Eng. Chem. 18 (2012) 1551–1559.

[10] Y.-F. Yang, Y. Li, Q.-L. Li, L.-S. Wan, Z.-K. Xu, Surface hydrophilization of microporouspolypropylene membrane by grafting zwitterionic polymer for anti-biofouling, J.Membr. Sci. 362 (2010) 255–264.

[11] B.-H. Jeong, E.M.V. Hoek, Y. Yan, A. Subramani, X. Huang, G. Hurwitz, A.K. Ghosh, A.Jawor, Interfacial polymerization of thin film nanocomposites: a new concept for re-verse osmosis membranes, J. Membr. Sci. 294 (2007) 1–7.

[12] L. De Lorenzo, E. Tocci, A. Gugliuzza, E. Drioli, Assembly of nanocomposite PEBAXmembranes: a complementary study of affinity and clusterization phenomena, J.Membr. Sci. 421–422 (2012) 75–84.

[13] E.M. VanWagner, A.C. Sagle, M.M. Sharma, Y.-H. La, B.D. Freeman, Surface modifica-tion of commercial polyamide desalination membranes using poly(ethylene glycol)diglycidyl ether to enhance membrane fouling resistance, J. Membr. Sci. 367 (2011)273–287.

[14] N. Hilal, M. Khayet, C.J. Wright, Membrane Modification: Technology and Applica-tion, CRC press, 2012.

[15] N. Joseph, P. Ahmadiannamini, R. Hoogenboom, I.F.J. Vankelecom, Layer-by-layerpreparation of polyelectrolyte multilayer membranes for separation, Polym. Chem.5 (2014) 1817–1831.

[16] X. Yang, R.Wang, L. Shi, A.G. Fane, M. Debowski, Performance improvement of PVDFhollow fiber based membrane distillation process, J. Membr. Sci. 369 (2011)437–447.

[17] X. Wei, B. Zhao, X.-M. Li, Z. Wang, B.-Q. He, T. He, B. Jiang, CF4 plasma surface mod-ification of asymmetric hydrophilic polyethersulfone membranes for direct contactmembrane distillation, J. Membr. Sci. 407–408 (2012) 164–175.

[18] C. Yang, X.-M. Li, J. Gilron, D.-f. Kong, Y. Yin, Y. Oren, C. Linder, T. He, CF4 plasma-modified superhydrophobic PVDF membranes for direct contact membrane distilla-tion, J. Membr. Sci. 456 (2014) 155–161.

[19] C. Yang, M. Tian, Y. Xie, X.-M. Li, B. Zhao, T. He, J. Liu, Effective evaporation of CF4plasma modified PVDF membranes in direct contact membrane distillation, J.Membr. Sci. 482 (2015) 25–32.

[20] N. Hilal, M. Khayet, C.J. Wright, Membrane Modification: Technology and Applica-tion, CRC press, Taylor & Francis Group, 2012 8 p.

[21] S.-C. Cha, E.K. Her, T.-J. Ko, S.J. Kim, H. Roh, K.-R. Lee, K.H. Oh, M.-W. Moon, Thermalstability of superhydrophobic, nanostructured surfaces, J. Colloid Interface Sci. 391(2013) 152–157.

[22] B. Shin, K.-R. Lee, M.-W. Moon, H.-Y. Kim, Extreme water repellency on nanostruc-tured low-surface-energy nonwoven fabrics, Soft Matter 8 (2012) 1817–1823.

[23] N. Hilal, A.F. Ismail, C.J. Wright, Membrane Fabrication, CRC Press, 2015.[24] S.P. Jeong, S.B. Lee, H.-T. Chon, S.H. Lee, Structural analysis and modeling of the com-

mercial high performance composite flat sheet membranes for membrane distilla-tion application, Desalination 349 (2014) 115–125.

[25] N. Vandencasteele, D. Merche, F. Reniers, XPS and contact angle study of N2 and O2

plasma-modified PTFE, PVDF and PVF surfaces, Surf. Interface Anal. 38 (2006)526–530.

[26] E.-S. Kim, Y.J. Kim, Q. Yub, B. Deng, Preparation and characterization of polyamidethin-film composite (TFC) membranes on plasma-modified polyvinylidene fluoride(PVDF), J. Membr. Sci. 344 (2009) 71–81.

[27] H. Fan, Y. Peng, Z. Li, P. Chen, Q. Jiang, S. Wang, Preparation and characterization ofhydrophobic PVDF membranes by vapor-induced phase separation and applicationin vacuum membrane distillation, J. Polym. Res. 20 (2013) 134–148.

[28] H.-W. Pei, H. Liu, Z.-Y. Lu, Y.-L. Zhu, Tuning surface wettability by designing hairystructures, Phys. Rev. E 91 (020401) (2015) 1–5.

[29] L.M. Camacho, L. Dumée, J. Zhang, J.-d. Li, M. Duke, J. Gomez, S. Gray, Advances inmembrane distillation for water desalination and purification applications, Water5 (2013) 94–196.

[30] M.A. Masuelli, M. Grasselli, J. Marchese, N.A. Ochoa, Preparation, structural and func-tional characterization of modified porous PVDF membranes by γ-irradiation, J.Membr. Sci. 389 (2012) 91–98.

[31] K.L. Mittal, Polymer Surface Modification: Relevance to Adhesion, 2nd edition VSP,2000 15 p.

[32] J. Torres, C.C. Perry, S.J. Bransfield, D.H. Fairbrother, Radical reactions with organicthin films: chemical interaction of atomic oxygen with an X-ray modified self-as-sembled monolayer, J. Phys. Chem. B 106 (2002) 6265–6272.

[33] N. Akashi, S.-i. Kuroda, Protein immobilization onto poly (vinylidene fluoride) mi-croporous membranes activated by the atmospheric pressure low temperatureplasma, Polymer 55 (2014) 2780–2791.

[34] S.-H. Gao, L.-H. Gao, K.-S. Zhou, Super-hydrophobicity and oleophobicity of siliconerubber modified by CF4 radio frequency plasma, Appl. Surf. Sci. 257 (2011)4945–4950.

[35] A. Kaynak, T. Mehmood, X.J. Dai, K. Magniez, A. Kouzani, Study of radio frequencyplasma treatment of PVDF film using Ar, O2 and (Ar + O2) gases for improved poly-pyrrole adhesion, Materials 6 (2013) 3482–3493.