Contents lists available at ScienceDirect

Journal of Membrane Science

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

Functionalization of ultrafiltration membrane with polyampholyte hydrogeland graphene oxide to achieve dual antifouling and antibacterial properties

Wei Zhanga, Wei Chengb,c, Eric Ziemanna, Avraham Be’era, Xinglin Lub, Menachem Elimelechb,Roy Bernsteina,⁎

a Department of Desalination and Water Treatment, Zuckerberg Institute for Water Research, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University ofthe Negev, Sede-Boqer Campus 84990, IsraelbDepartment of Chemical and Environmental Engineering, Yale University, New Haven, CT 06520-8286, United Statesc State Key Laboratory of Urban Water Resource and Environment, School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin 150090,China

A R T I C L E I N F O

Keywords:FoulingGraphene oxideZwitterion hydrogelPES membraneAntifoulingAntibacterial

A B S T R A C T

Membrane modification using zwitterion polymers is an excellent strategy to reduce organic fouling and initialbacterial deposition, and functionalization of the membrane surface with antibacterial material can reducebiofilm formation. In this study, we combined these two approaches to develop an ultrafiltration poly-ethersulfone (PES) membrane with dual antifouling and antibacterial properties. At the beginning, a zwitterionpolyampholyte hydrogel was UV grafted onto PES membrane surface (p-PES membrane). Then, the hydrogel wasloaded with graphene oxide (GO) nanosheets using vacuum filtration strategy (GO-p-PES membrane). Ramanspectroscopy and scanning electron microscopy (SEM) confirmed the successful incorporation of the GO na-nosheets into the zwitterion hydrogel, and contact angle measurements indicated that the membrane hydro-philicity was enhanced. Static adsorption and dynamic filtration experiments demonstrated that the p-PES andGO-p-PES membranes exhibited similar organic fouling propensity. Moreover, the loading of GO induced an-tibacterial property for the membrane as evidenced by contact killing and antibiofouling filtration experiments.Leaching of GO was very low, with over 98% of the GO remaining on the membrane surface after 7 days. Ourfindings highlight the potential of this GO-functionalized polyampholyte hydrogel for long-term wastewatertreatment.

1. Introduction

Wastewater purification and reuse are becoming an attractive al-ternative water resource, mainly for agricultural production, and its useis expected to grow in the near future owing to water scarcity world-wide [1–4]. Membrane filtration, particularly low-pressure ultrafiltra-tion (UF) and microfiltration (MF), is one of the most advanced, cost-effective, and sustainable wastewater technologies for direct treatmentof wastewater [5], tertiary treatment of secondary effluents [6], use inmembrane bioreactors (MBRs) [7], and use as a pretreatment step be-fore wastewater desalination [8]. However, one of the main limitationsof the most commonly used polymeric UF membranes, such as poly-ethersulfone (PES) and polyvinyl difluoride (PVDF), is the high foulingpropensity of the membrane resulting from its intrinsic hydro-phobicity [9–11]. Membrane fouling is caused by the adsorption anddeposition of colloidal particles, suspended solids, macromolecules, and

microorganisms on the membrane surface [12,13]. Fouling results in asevere flux decline, increase in the energy consumption, and shortermembrane lifetime because of frequent chemical cleaning. These ne-gative effects enhance the operational costs and hinder the wide-scaleapplication of membrane-based technologies for wastewater treatment.

Increasing the surface hydrophilicity is an efficient approach toreducing membrane organic fouling and mitigating bacteria adhesion.These beneficial properties are typically attributed to the formation of ahydration layer, which imposes a thermodynamic and steric barrieragainst the adsorption of foulants [14–16]. Numerous studies have re-ported the antifouling modification of hydrophobic UF membranes toenhance hydrophilicity through direct surface coating or chemicalgrafting of hydrophilic polymers, such as polydopamine (PDA) [17],poly(ethylene glycol) (PEG) [18], zwitterionic polymers (e.g., sulfobe-taine) [19], or hydrophilic nanomaterials such as TiO2 [20], ZnO [21],and silica nanoparticles [22]. Although surface antifouling modification

https://doi.org/10.1016/j.memsci.2018.08.017Received 13 July 2018; Received in revised form 15 August 2018; Accepted 16 August 2018

⁎ Corresponding author.E-mail address: royber@bgu.ac.il (R. Bernstein).

Journal of Membrane Science 565 (2018) 293–302

Available online 18 August 20180376-7388/ © 2018 Elsevier B.V. All rights reserved.

T

using hydrophilic materials successfully delay or even prevent organicfouling and limit initial bacterial deposition, these modifications stillcannot completely inactivate bacteria that inevitably attach to themembrane surface, which eventually leads to formation of a biofilm[23]. Thus, developing novel membranes by introducing bactericidalcapabilities to the antifouling coating to achieve so-called dual func-tionality is a promising strategy in reducing both organic fouling andbiofouling [24,25].

Xu et al. entrapped polysulfone membrane with Ag/Cu2O hybridnanowires to obtain fouling resistance and antimicrobial properties[26]. In another study, Zou et al. synthesized a bifunctional membraneby integrating a zwitterion with antimicrobial peptides [27]. More re-cently, Wu et al. incorporated a silver-polydopamine nanohybrid in thepolysulfone membrane matrix, which resulted in excellent antibacterialand antibiofouling properties [28]. These results demonstrate the greatadvantage of dual functionality. However, the antibacterial strategies inthese studies relied on either dissolution-dependent (i.e., depleting)nanoparticles or environmentally sensitive peptides, which easily losetheir antimicrobial activity over time in a complex solution such aswastewater [29]. In addition, the release of biocides is also associatedwith potential impact to the environment [30].

Graphene oxide (GO), a two-dimensional carbon material, consistsof a high density of covalently attached oxygen-containing functionalgroups on its edges and basal planes, which make it highly hydrophilic[31–33]. Various studies have reported that introducing GO to mem-branes, either by blending in the bulk or grafting GO to the surface,improves the membrane antifouling property by reducing the surfaceroughness and hydrophobicity [34,35]. GO has also exhibited toxicitytoward bacterial cells, thereby effectively inhibiting bacterial growth.Unlike many antibacterial nanoparticles, the antibacterial activity ofGO nanosheets is not driven by dissolution but instead by contact be-tween the GO and bacterial cells [36,37]. The cytotoxicity of GO againsta wide variety of microorganisms has been attributed to the disruptedbacteria cell integrity resulting from the atomically sharp edges of GO,which penetrate the bacterial cell [38–40], and to oxidative reactionwith the cell membrane mediated by GO nanosheets [41–43]. Also, GOis a low-cost, easy to scale, and dispersible and stable in aqueous so-lution, making it an attractive antibacterial material [44], as demon-strated in many applications, including membranes [14,30].

Herein, a two-step surface modification strategy was developed forthe preparation of a dual antifouling and antibacterial coating for a UFPES membrane. In this approach, unlike the common blending method,only the membrane surface is modified, thereby allowing independentoptimization of the highly antifouling zwitterion hydrogel and the an-tibacterial GO. First, a zwitterionic polyampholyte hydrogel wasgrafted on the surface of a PES membrane by UV-photoinitiation. Then,GO nanosheets were loaded into the swollen polyampholyte hydrogelvia vacuum filtration. The successful incorporation of the GO into thegrafted hydrogel was shown by a series of characterization methods(i.e., Raman, SEM, ATR-FTIR, and contact angle measurements) as well

as visually by incorporating GO nanosheets into a bulk hydrogel. Theantifouling of the dual coating was demonstrated by static adsorptionand filtration experiments using protein solution as a model organicfoulant, and soluble microbial products extracted from an MBR and asolution containing bacteria intracellular products as real foulants. Theantibacterial property of the modified membrane was demonstratedusing the contact killing method and filtration experiments. The resultsclearly confirmed that the dual membrane functionalization with GOand zwitterion hydrogel is a promising strategy to prevent both organicfouling and biofouling in UF applications.

2. Materials and methods

2.1. Chemicals and materials

PES polymer was obtained from BASF, Germany, and N-methyl-2-pyrrolidone (NMP) was purchased from J.T. Baker, USA. Vinyl sulfonicacid (VSA) was purchased from Tokyo Chemical Industry, Japan, and[2-(methacryloyloxy) ethyl] trimethylammonium chloride (METMAC),N, N′-Methylenebis (acrylamide) (Bis), sulfuric acid (H2SO4), graphitepowder (200 mesh), and bovine serum albumin (BSA) were purchasedfrom Sigma. Na2HPO4 and KCl were purchased from Merck, USA, andHCl (37%), CaCl2, and NaCl were supplied by Bio-Lab, Israel. NaOHwas purchased from Dae-Jung Chemical and Metals, South Korea, andKH2PO4 was bought from Hopkin & Williams, USA. Soluble microbialproducts (SMP) extracted from a membrane bioreactor system wereprovided by Prof. Moshe Herzberg (Ben-Gurion University of the Negev,Israel).

2.2. Synthesis and characterization of GO nanosheets

GO nanosheets were synthesized using a modified Hummers’method that was previously described [38]. The purified GO suspensionwas then probe sonicated (VCX 130, Sonics, USA) for 120min at 130W(45 s on and 15 s off) in an ice bath. The size distribution of GO na-nosheets was measured using dynamic light scattering (DLS; MalvernZetasizer Nano-ZS, Germany).

2.3. Preparation of GO-p-PES membrane

2.3.1. Casting of PES membraneThe PES membrane was cast using the non-solvent phase inversion

method as described in our previous research [45]. The preparedmembranes were stored in Milli-Q water until use.

2.3.2. Surface modification of PES membraneA two-step protocol (Scheme 1) was used to modify the membrane

using the GO nanosheets functionalized polyampholyte hydrogel. Step1, grafting of the hydrogel was conducted as described in our previousresearch but with 1% Bis as the crosslinker [46]. After hydrogel

Scheme 1. Schematic illustration of the two-step procedure used to synthesize the GO-functionalized zwitterionic polyampholyte hydrogel PES membrane.

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

294

grafting, the modified PES membrane (p-PES membrane) was washed in50% ethanol/water solution for 12 h to remove the adsorbed polymersand then stored in Milli-Q water.

In Step 2, the GO nanosheets were loaded into the grafted zwitter-ionic polyampholyte hydrogel matrix using direct vacuum filtration. Amodified zwitterionic polyampholyte PES membrane (13.85 cm2) wasimmersed in 1M NaCl solution (17mL) for 2 h at 30 °C to achieve highswelling [47]. After that, 3 mL of GO suspension (10mgmL−1) wasadded to the NaCl solution, and eventually 20mL solution was vacuumfiltrated through the highly swollen membrane surface. As the mem-brane surface was rapidly covered with GO nanosheets, the filtrationprocess was conducted in four intervals (5 mL each), and between eachinterval the membrane surface was thoroughly washed with Milli-Qwater to remove the unloaded GO. At the end of the loading step, theGO-p-PES membranes were thoroughly washed with Milli-Q and storedin Milli-Q water until use.

The effect of the GO concentration on the GO loading in the hy-drogel and subsequently on the antibacterial property was also studied.The 3mL GO suspension was prepared using lower GO concentrations(1–5mgmL−1) and the loading was performed as described above. Theloading of GO (M, µgm−2) was calculated from the difference in theamounts of GO in the solution and in the water collected during thewashing step:

= −M m mA

2 1(1)

where m2 (µg) and m1 (µg) are the amounts of GO in the filtrate solutionand the collected washing solution, respectively, and A (m2) is themembrane area (13.85 cm2). The amount of GO was calculated using aUV spectrophotometer as described in Section 2.8.

2.4. Membrane characterization

The presence of GO on the membrane surface was confirmed byRaman spectroscopy (Jobin-Yvon LabRam HR–800 spectrometer using514 nm laser excitation) and scanning electron microscopy (SEM; JSM-7400 F, Japan). The grafting of the polyampholyte hydrogel was stu-died using attenuated total reflectance–Fourier transform infraredspectroscopy (ATR–FTIR; Vertex 70, Bruker). The instrument wasequipped with a single-reflection diamond-covered KRS–5 crystal (PIKETechnologies). Thirty scans were performed for each spectrum at aresolution of 4 cm−1. The contact angle of the membrane was measuredusing the sessile dropmethod: a 3–μL droplet of deionized water wasplaced on the membrane surface, and a digital image (OCA 15 Plus,Germany) was used to determine the contact angle.

2.5. Membrane filtration performance

The water permeability and molecular weight cut-off (MWCO) ofPES membrane, p-PES membrane, and GO-p-PES membrane were in-vestigated using a stirred (500 rpm) dead–end filtration cell (Amicon8050, Millipore) connected to a pressure controller and a peristalticpump (Masterflex, USA) solution reservoir (1000mL), respectively. Thetransmembrane pressure was controlled by manually adjusting thepressure controller. The permeate weight was measured using anelectronic balance (Ohaus, USA), which was connected to a computerand data collected automatically every minute.

The water permeability of the three membranes was evaluated aftera constant flux was recorded (at least 30min of filtration), and thepermeability (LP) was calculated using the following equation:

=× ∆ ×

L VA t PP (2)

where V (L) is the volume of the liquid, A (m2) is the effective filtrationarea (13.85 cm2), Δt (h) is the operation time, and P is the appliedpressure (~ 0.7 bar).

The MWCO of each membrane was estimated by measuring therejection of polyethylene glycol (PEG) with specific molecular weights(35, 100, and 200 kDa), where the MWCO was the value at which therejection was 90%. Following compaction with Milli-Q water, a0.1 g L−1 PEG solution was filtered for 2min at 0.7 bar, and then 5mLof the permeate was collected and measured using a total organiccarbon (TOC) analyzer (Analytik Jena, Germany). The rejection (R, %)was calculated using the following equation:

⎜ ⎟= ⎛⎝

− ⎞⎠

×RCC

1 100%p

f (3)

where Cp (g L−1) and Cf (g L−1) are the PEG concentration in thepermeate and feed solutions, respectively.

2.6. Assessing antifouling propensity

2.6.1. Static adsorption of proteinStatic adsorption experiments were performed to evaluate the or-

ganic fouling propensity of the membranes with isothiocyanate-con-jugated bovine serum albumin (FITC-BSA) as a model organic foulant.FITC-BSA was dissolved in a 10mM PBS solution (pH ~ 7.4) at aconcentration of 50mg L−1 in an aluminum foil sealed plastic tube.Subsequently, 8 mL of the as-prepared FITC-BSA solution was placed onthe membrane coupons (0.28 cm2), which were glued to a glass slide, indark conditions for 12 h on a rocking platform at 60 rpm. Thereafter,the FITC-BSA solution was discarded, and each membrane was gentlyrinsed with PBS solution four times. A fluorescence image of eachmembrane was then obtained using epifluorescence microscopy (ZeissAxio, Imager Z2), and the fluorescence intensity on each membranesurface was recorded. The BSA adsorption was calculated using afluorescence calibration curve between the BSA concentration andfluorescence intensity (Fig. S1) that was constructed for each type ofmembrane (PES, p-PES, and GO-p-PES).

2.6.2. Dynamic fouling experimentsThe membrane antifouling property was evaluated using an ultra-

filtration experiment at similar initial flux using the previously de-scribed dead-end filtration system. Three foulants were tested: BSA(1 g L−1 solution in 10mM PBS), SMP extracted from an MBR(0.02 g L−1 in 8.5 mM NaCl and 1.5mM CaCl2), and bacteria in-tracellular products (BIP, 0.25 g L−1 as protein concentration in 8.5 mMNaCl and 1.5mM CaCl2; for details see Supporting information). In-itially, each membrane was filtrated with Milli-Q water for at least30min to obtain a stable and similar initial flux (Jw0, L m−2 h−1). Then,the foulant solution was filtered for 2 h, during which the flux wasrecorded for each minute. After the foulant filtration step, the mem-brane was washed using the background solution for 10min (thewashing time was not counted in the filtration cycle), and then theMilli-Q water flux of the cleaned membrane was measured again (Jw, i).These steps were repeated three times for each membrane. The flux (Jw)of the membrane was determined using the following equation:

=× ∆

J VA tw (4)

The effect of the membrane modification on the fouling reversibilitywas estimated based on the flux recovery ratio (FRR) for each cycle (i):

=FRRJJi

W i

W

,

,0 (5)

2.7. Analysis of antibacterial and antibiofouling performance

2.7.1. Contact killing evaluationThe antibacterial activity of each membrane was studied using the

plate-counting method [48]. A single colony of E. coli was inoculated

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

295

into 2mL of sterile Luria Bertani (LB) broth and incubated at 30 °Covernight in an incubator at 200 rpm. Subsequently, the bacteria sus-pension (0.1 mL) was transferred to sterile LB broth (10mL) and re-grown for ~ 4 h to the exponential phase (OD=1 at 600 nm, corre-sponding to 4× 108 colony-forming units per milliliter (CFUmL−1)).Next, the culture was centrifuged, and the bacteria were washed threetimes with sterile 10mM PBS solution. After washing, the bacteria wereresuspended in the PBS solution and diluted to a concentration of 105

CFUmL−1. Then, 100 μL of the diluted bacterial suspension was placedon the membrane surface (1.54 cm2) for 3 h at room temperature in abiological hood. After 3 h, the bacterial suspension and membrane werecarefully transferred into Eppendorf tubes containing 0.9 mL of PBS andshaken vigorously for 30 s. Finally, 30 μL of the obtained bacteria so-lution was spread on an LB agar plate, and the colonies were countedafter incubation at 30 °C overnight.

The antibacterial activities of the three membranes for E. coli werefurther examined using fluorescent microscopy. After the contact-killing assay, the membrane and bacteria solution were immediatelystained with dyes (SYTO 9 for all cells and PI for dead cells) and left inthe dark for 15min. Subsequently, the membrane and bacteria solutionwere transferred into an Eppendorf tube containing 0.5mL PBS solutionand shaken vigorously. Then, the solution was centrifuged at 3000 rpmfor 5min, and 2 drops of the bottom concentrated bacteria were addedto a glass slide and examined using confocal laser scanning microscopy(CLSM; Zeiss LSM 510, Carl Zeiss, Thornwood, NY). The microscopewas equipped with a Plan-Apochromat 20×0.8 numerical apertureobjective.

2.7.2. Antibiofouling behavior in a cross-flow systemThe dynamic biofouling experiment was performed in a lab-scale

cross-flow system to evaluate the resistance of the GO-p-PES membraneto biofilm formation by E. coli, as previously reported [15]. Before theexperiment, the system was disinfected and washed with bleach solu-tion (10%, v/v), EDTA solution (5mM, pH 7), ethanol (70%, v/v), andMilli-Q water. Similar to the procedure described in Section 2.7.1, theE. coli pellets were first prepared and resuspended in sterile syntheticwastewater to reach a concentration of 3×106 CFUmL−1. The syn-thetic wastewater was prepared based on our previous research [49]and the composition of the wastewater is given in Table S1. The foulingexperiment for all the membranes (2.5 cm×2.5 cm) was conducted for12 h under a similar initial flux (~ 50 Lm−2 h−1). The decline of themembrane flux, which reflects the biofouling resistance propensity, wasmonitored and recorded each hour. Furthermore, the biofilms thatformed on the membrane surface were analyzed after each biofoulingexperiment. The membrane was gently rinsed with the backgroundsolution and subsequently stained with SYTO 9, PI, and concavalin A(Con A) for live bacteria, dead bacteria, and EPS, respectively, for45min in a dark environment after being removed from the system.After that, surface CLSM images were captured using a Zeiss LSM 510(Carl Zeiss, Thornwood, NY) equipped with a Plan-Apochromat20×0.8 numerical aperture objective. Finally, the confocal imageswere further analyzed using Auto PHLIP-mL, Image-J, and MATLABsoftware to calculate the formed biofilm parameters (e.g., dead/livebiovolume, EPS biovolume, and average thickness).

2.8. Stability of the GO-p-PES membrane

The long-term stability of the synthesized GO-p-PES membrane wasinvestigated by measuring the leached GO using UV–visible spectro-scopy. Modified membrane coupons (1.54 cm2) were submerged in a50mL plastic tube containing 3mL of salt solution (8.5 mM NaCl and1.5 mM CaCl2) and placed in a water bath shaker (25 °C) at 100 rpm fora few days. The leached GO present in the solution was measured usingUV–visible spectroscopy at a wavelength of 232 nm. The leached per-centage of GO on different days was calculated relative to the amount ofGO incorporated on the membrane after rinsing and using a standard

curve. The amount of GO loaded (mg cm−2) was estimated by sub-tracting the amount of GO in the washing solution from the amount inthe loading solution after the vacuum filtration. Similar to the mod-ification step of GO loaded into the polyampholyte hydrogel matrix(step two in Section 2.3.2), the GO suspension was filtrated through themembrane several times, and each time the washed solution and waterthat passed through the membrane during the vacuum filtration werecollected. Eventually, the amount of GO in the collected solution wascalculated based on the standard curve.

2.9. Statistical analyses

Differences between samples were determined using a one-wayanalysis of variance (ANOVA) followed by a post hoc test (Tukey HSD)using Origin software (Northampton, MA01060, USA). Significant dif-ferences (p=0.05) between sample groups are denoted in figures withan asterisk.

3. Results and discussion

3.1. Membrane surface characteristics

Fig. 1a presents the surface Raman spectra of the three membranes.A new peak at 1360 cm−1, attributed to the specific D band of the GO,appeared on the GO-p-PES membrane surface, confirming the suc-cessful loading of GO into the zwitterionic polyampholyte hydrogelmatrix [38,41]. The changes in the membrane surface morphologybefore and after immobilization of the GO nanosheets (~ 30 nm, asmeasured by DLS, see Fig. S2) into the hydrogel were further in-vestigated using SEM. From Figs. 1b and 1c, it seems that the porosity ofthe pristine PES membrane decreased following the grafting. However,this decrease in porosity, as will be shown later, does not change themembrane selectivity. We surmise that the collapse of the dry hydrogelon the membrane surface may cover the membrane surface, therebymaking the pores look smaller. In addition, as shown in Fig. 1d, thepores became indiscernible after the incorporation of GO. This findingconfirms the successful loading of the GO nanosheets into the zwitter-ionic polyampholyte hydrogel using the vacuum filtration method. Inorder to further verify that the GO nanosheets were incorporated intothe hydrogel, the GO suspension was also vacuum-filtered through abulk polyampholyte hydrogel (see Supporting information for details)and the hydrogel was sliced in the middle. The GO was clearly visible inthe matrix of the hydrogel, indicating that the GO nanosheets suc-cessfully loaded into the hydrogel using vacuum filtration method (Fig.S3).

The ATR-FTIR spectrum after each step was also measured. As ob-served in Fig. 2, two new peaks appeared for the p-PES membrane at1040 cm−1 and 954 cm−1, which are attributed to the sulfonic acidgroup of VSA and the quaternary amine of METMAC [23,50], respec-tively. These peaks indicate the surface grafting of the zwitterionicpolyampholyte hydrogel to the PES membrane. As expected, these twocharacteristic peaks are still visible in the spectrum of the GO-p-PESmembrane, suggesting that the loading of the GO did not affect thegrafted polyampholyte hydrogel property.

The surface hydrophilicity was investigated using the sessile dropcontact angle measurements. Fig. 3 shows that the membrane contactangle decreased from 78°± 4° to 63°± 3° after surface grafting of thepristine PES membrane with the polyampholyte hydrogel. The decreasein contact angle is probably attributed to the strong hydration capacityof the grafted hydrophilic hydrogel [51]. Moreover, compared with thep-PES membrane, modification with GO further enhanced the surfacehydrophilicity, indicating that the hydrophilic GO nanosheets wereintroduced into the hydrogel [52]. Nonetheless, hydrogels are usuallyextremely hydrophilic in water [53], and the addition GO is not likelyto further enhance the hydrophilicity of the membrane.

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

296

3.2. Water permeability and MWCO

As shown in Fig. 4a, the water permeability for the pristine PESmembrane was 133 ± 4 Lm−2 h−1 bar−1. The permeability decreasedto 93 ± 6 Lm−2 h−1 bar−1 and 70 ± 5 Lm−2 h−1 bar−1 followingthe zwitterionic polyampholyte hydrogel grafting and loading with the

GO nanosheets, respectively. The decline in water permeability wasmost likely due to the coating of the membrane surface with the graftedhydrogel while the GO nanosheets partially blocked the membranepores, as observed in the SEM images in Fig. 1b–d.

A decline in membrane permeability is typically accompanied by an

Fig. 1. Surface Raman spectra (a) and SEM images (b–d) of the pristine PES, p-PES, and GO-p-PES membranes. The scale bar in the SEM images is 1 µm.

Fig. 2. Surface ATR-FTIR spectra of the pristine PES, p-PES, and GO-p-PESmembranes.

Fig. 3. Water contact angles of the pristine PES, p-PES, and GO-p-PES mem-branes. The values are expressed as means± SD, n= 5. The asterisk indicatesthat the contact angle of the modified membrane with GO show a statisticallysignificant difference compared with that of the pristine PES membrane(p < 0.05).

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

297

increase in selectivity [54]. However, similar to the previous research[55], the MWCO of the p-PES membrane increased to ~210 kDa com-pared with an MWCO of 134 ± 16 kDa for the pristine PES membrane(Fig. 4b), probably owing to the degradation of the PES membraneduring the UV grafting process (which cannot be directly observed fromthe SEM images in Fig. 1). In contrast, the MWCO of the GO-p-PESmembrane decreased to ~ 120 kDa, most likely because of blockage ofthe membrane pores by the incorporated GO nanosheets.

3.3. Antifouling property

3.3.1. Static adsorption of proteinThe fouling propensity of the three membranes was evaluated using

static adsorption measurements with fluorescence BSA. Representativefluorescence images of the PES membrane, p-PES membrane and GO-p-PES membrane, as well as the relative BSA adsorption amounts, arepresented in Fig. 5. The pristine PES membrane had a bright greenfluorescence intensity (Fig. 5a), indicating significant adsorption ofprotein, possibly ascribing to the hydrophobic interaction between theprotein molecules and the PES membrane surface [56,57]. Comparedwith that of the pristine PES membrane, the fluorescence intensity ofthe p-PES membrane (Fig. 5b) was noticeably reduced, and almost nofluorescence was observed on the GO-p-PES membrane (Fig. 5c). Fur-thermore, the relative adsorption amount (Fig. 5d) also shown that theadsorption of the GO-p-PES membrane was reduced by more than 70%compared with that of the pristine PES membrane, probably owing tothe synergistic effects of the grafted zwitterionic polyampholyte hy-drogel and loaded hydrophilic GO nanosheets.

3.3.2. Dynamic fouling resistanceThe antifouling properties of the three membranes were evaluated

using filtration tests with two foulants under a similar initial flux (~50 Lm−2 h−1). As observed in Fig. 6, the flux of the three membranesdecreased rapidly during the initial stage of filtration with the BSAsolution, due to the combined effects of concentration polarization,protein adsorption, and pore blocking [58]. However, the flux re-coveries of the p-PES and GO-p-PES membranes following the washingstep for each filtration cycle were much higher than that of the pristinePES membrane. In addition, as shown in Table 1, the FRR values of thep-PES membrane and GO-p-PES membrane were higher (~ 80% forboth modified membranes) than that of the pristine PES membrane (~46%) after three cycles. The higher FRR value following the simplephysical washing indicates that the GO-loaded polyampholyte hydrogelcoating effectively weakened the interactions between the membranesurface and foulants; thus, the fouling layer was much more reversible

than that of the pristine PES membrane. The fouling propensity of thethree membranes was also tested with SMP extracted from an MBRsystem as a realistic foulant (Fig. S4a and S4b). As observed with theBSA solution, the two modified membranes still displayed a lower fluxdecline when the SMP solution was filtered and much higher FRR va-lues after three filtration cycles compared with the corresponding re-sults for the pristine membrane.

3.4. Antibacterial and antibiofouling properties

3.4.1. Surface antibacterial activityThe surface antibacterial activities of the three membranes were

evaluated using the contact killing approach. As shown in Fig. 7a, thepristine PES membrane and p-PES membrane had no toxicity toward E.coli, whereas a decline of nearly 80% of the E. coli colonies was ob-served for the GO-p-PES membrane. This finding indicates that theimmobilized GO nanosheets in the zwitterionic hydrogel matrix im-parted antimicrobial activity to the membrane. In addition, the anti-microbial activity of the GO-p-PES membrane was affected by the GOnanosheets content in the hydrogel (see Fig. S5), in agreement withprevious research [37]. Moreover, GO-p-PES membrane that was storedin water for over a month still showed a comparable antibacterialproperty to the membrane that was stored for only a few days, implyingthe stability of GO-enabled antimicrobial property.

The antibacterial activities of the membranes to E. coli were alsostudied using CLSM fluorescence analysis. Following three hours ofsurface contact with E. coli, the bacteria was stained with SYTO 9 (livecells, but can also stain dead cells [59–61]) and PI (dead cells). As seenfrom Fig. 7b, the number of dead bacteria was very low for the PES(Fig. 7b) and p-PES (Fig. 7f) membranes, whereas most of the cells onthe GO-p-PES membrane appeared to be dead (Fig. 7f). This resultfurther confirmed that the GO nanosheets modified zwitterionic poly-ampholyte PES membrane indeed effectively suppressed the bacteriaviability.

3.4.2. Antibiofouling propertyThe biofouling mitigation potential of the three membranes was

evaluated using dynamic biofouling assays in a cross-flow system withsynthetic wastewater. As observed in Fig. 8, the pristine PES membranedisplayed a flux decline of nearly 65% after 12 h of the fouling ex-periment. However, a flux decline of less than 25% was recorded for theGO-p-PES membrane, suggesting that the GO-functionalized poly-ampholyte hydrogel membrane exhibited good antibiofouling proper-ties.

The biofilm on the membrane surfaces at the end of the filtration

Fig. 4. (a) Water permeability and (b) PEG rejection of the pristine PES, p-PES, and GO-p-PES membranes. The values are expressed as means± SD, n=3.

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

298

experiments was analyzed using CLSM. As observed in Fig. 9a–c, thesurface biofilm layers of the p-PES and GO-p-PES membranes werethinner (~ 8 and 9 µm, respectively) than that of the pristine PESmembrane (~ 19 µm). In addition, the calculated biovolume constitu-tion (Fig. 9d) on the GO-p-PES membrane surface also indicated fewerbacteria (live and dead) and a lower amount of EPS than on the pristinePES membrane. These results may be attributed to the grafted zwit-terionic polyampholyte hydrogel conferring antifouling properties andthe loading of GO nanosheets providing antibacterial properties. Con-sequently, for the GO-p-PES membrane, in addition to less initial bac-teria being attached on the membrane surface because of the improvedsurface hydrophilicity and hydration layer, some of the adhesive pro-liferating bacteria were killed by the incorporated biocide GO na-nosheets.

It might be speculated that even though the bacteria were effec-tively killed by the modified membrane, the intracellular products ofdead bacteria would contaminate and foul the membrane [62]. Hence,the antifouling propensity of the GO-p-PES membrane toward BIP wasalso studied. The results (Fig. S6) clearly show that the GO-p-PESmembrane possessed a higher flux recovery than the pristine PESmembrane, implying the GO-p-PES membrane also possess a higherfouling resistance for BIP.

3.5. Stability of the GO-p-PES membrane

Leaching experiments were conducted to assess the stability of thesynthesized GO-p-PES membrane. The incorporated GO concentrationon the membrane surface was 0.57 ± 0.13mg cm−2 upon initialloading with 10mgmL−1 GO suspension. After 7 days, a very smallamount of the GO nanosheets leached, leaving 98% of the GO na-nosheets in the hydrogel (Fig. 10). The small amount of the leached GOmost likely consisted of GO nanosheets that were physically adsorbed tothe hydrogel surface and were not removed during the washing steps.

Fig. 5. (a–c) Fluorescence images and (d) re-lative adsorption percentage of the pristinePES, p-PES, and GO-p-PES membranes afterexposure to a 50mg L−1 FITC-BSA solution for12 h at room temperature. The values are ex-pressed as means± SD, n=7. The asteriskindicates that the relative adsorbed amounts ofthe modified membranes are significantly dif-ferent from the pristine PES membrane(p < 0.05). The scale bar in the images is500 µm.

Fig. 6. Time-dependent recycling fluxes of the pristine PES, p-PES, and GO-p-PES membranes during the BSA (1 g L−1 in 10mM PBS) ultrafiltration experi-ment. The tests were conducted under similar initial flux (50 Lm−2 h−1) byadjusting the TMP (0.2–0.7 bar).

Table 1Flux recovery ratio (FRR) values for the pristine PES, p-PES, and GO-p-PESmembranes after each BSA fouling experiment cycle. The values are expressedas means± SD, n= 3.

Membranes First-cycle FRR (%) Second-cycle FRR (%) Third-cycle FRR (%)

PES 69.3 ± 1.4 59.2 ± 1.3 45.6 ± 3.1p-PES 96.8 ± 2.2 87.1 ± 4.9 79.9 ± 3.3GO-p-PES 93.9 ± 3.2 84.1 ± 2.2 80.1 ± 4.5

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

299

This inference is supported by a subsequent leaching experiment, inwhich the hydrogel was soaked in Milli-Q water for 1 day, removed,and then soaked for 1 day in clean Milli-Q water. The GO concentrationfollowing the second soaking step was negligible. This finding, togetherwith the results of filtration experiments, suggests that the dual anti-fouling and antibacterial properties of the GO-p-PES membrane can bestable during long-term filtration.

4. Conclusion

This study demonstrated that grafting of a zwitterionic poly-ampholyte hydrogel PES membrane followed by loading of the hy-drogel with GO nanosheets can be employed as an effective method toproduce a GO-functionalized zwitterionic polyampholyte UF mem-brane. The successful preparation of the GO-p-PES membrane wasconfirmed by Raman spectroscopy, SEM, and ATR-FTIR analyses.Contact angle measurements demonstrated that the hydrophilicity ofthe GO-p-PES membrane increased due to the loaded GO into thegrafted hydrogel matrix. The static protein adsorption results revealedthat the GO-p-PES membrane had much lower adsorption propensitythan the pristine membrane. Three types of foulants were tested usingdynamic filtration experiments. The GO-p-PES membrane had muchhigher FRR than the pristine PES membrane for all three foulants. Inaddition, contact killing and antibiofouling experiments revealed theantibacterial property of the GO-p-PES membrane, and leaching ex-periments proved that the loaded GO was stable in the grafted hydrogelmatrix. Taken together, these findings suggest that the synthesized dualantifouling and antibacterial GO-p-PES membrane shows great poten-tial in treating water and wastewater.

Acknowledgements

We acknowledge financial support from the United States-IsraelBinational Agricultural Research and Development Fund (BARD), GrantIS-4977-16. AB is thankful for partial support from The Israel ScienceFoundation's Grant No. 373/16. W.Z. is supported by the KreitmanNegev Scholarship for Distinguished Ph.D. Students from BGU. W.C. issupported by the China Scholarship Council for providing a graduatefellowship.

Fig. 7. (a) Antibacterial activity of PES membrane, p-PES, and GO-p-PES membranes after surface contact with E. coli (~ 4×104 CFU/1.54 cm2) for 3 h. The valuesare expressed as means± SD, n=8. Confocal images displaying the dead and live E. coli cells (~4×107 CFU) on the (b–c) PES membrane, (d–e) p-PES membraneand (f–g) GO-p-PES membrane after 3 h of contact killing. The scale bar in the images is 5 µm.

Fig. 8. Flux decline of the pristine PES, p-PES, and GO-p-PES membranes in across-flow system. The flux was normalized by an initial water flux. The bio-fouling was conducted for 12 h with E. coli (3×106 CFUmL−1). The cross-flowvelocity rate was 11.1 cm s−1, and the temperature was maintained at25 ± 0.5 °C.

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

300

Appendix A. Supporting information

Supplementary data associated with this article can be found in theonline version at doi:10.1016/j.memsci.2018.08.017.

References

[1] I. Michael-Kordatou, C. Michael, X. Duan, X. He, D. Dionysiou, M. Mills, D. Fatta-Kassinos, Dissolved effluent organic matter: characteristics and potential implica-tions in wastewater treatment and reuse applications, Water Res. 77 (2015)213–248.

[2] W. Luo, H.V. Phan, G. Li, F.I. Hai, W.E. Price, M. Elimelech, L.D. Nghiem, An os-motic membrane bioreactor–membrane distillation system for simultaneous was-tewater reuse and seawater desalination: performance and implications, Environ.Sci. Technol. 51 (2017) 14311–14320.

[3] Y. Liu, X. Wang, H. Wu, High-performance wastewater treatment based on reusablefunctional photo-absorbers, Chem. Eng. J. 309 (2017) 787–794.

[4] H. Luo, J. Aboki, Y. Ji, R. Guo, G.M. Geise, Water and salt transport properties oftriptycene-containing sulfonated polysulfone materials for desalination membraneapplications, ACS Appl. Mater. Interfaces 10 (2018) 4102–4112.

[5] J. Lin, C.Y. Tang, W. Ye, S.-P. Sun, S.H. Hamdan, A. Volodin, C. Van Haesendonck,A. Sotto, P. Luis, B. Van der Bruggen, Unraveling flux behavior of superhydrophilicloose nanofiltration membranes during textile wastewater treatment, J. Membr. Sci.493 (2015) 690–702.

[6] J.L. Acero, F.J. Benitez, F.J. Real, F. Teva, Micropollutants removal from retentatesgenerated in ultrafiltration and nanofiltration treatments of municipal secondaryeffluents by means of coagulation, oxidation, and adsorption processes, Chem. Eng.J. 289 (2016) 48–58.

[7] R.W. Holloway, A.S. Wait, A.F. da Silva, J. Herron, M.D. Schutter, K. Lampi,T.Y. Cath, Long-term pilot scale investigation of novel hybrid ultrafiltration-osmotic

Fig. 9. (a–c) Representative CLSM orthogonal images of E. coli biofilms developed on the pristine PES, p-PES, and GO-p-PES membranes after 12 h of fouling (3× 106

CFUmL−1 E. coli), and the biofilm thickness. (d) Biovolume of the live cells, EPS, and dead cells for the three membranes (the number were averaged, with standarddeviation (SD) calculated from twelve random positions in duplication experiments).

Fig. 10. Percentage of GO nanosheets on the GO-p-PES membrane surfaceleached into solution during 7 days of shaking. The values are expressed asmeans± SD, n= 3. The inset shows the corresponding stand curve for the GOsuspension.

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

301

membrane bioreactors, Desalination 363 (2015) 64–74.[8] J. Löwenberg, J.A. Baum, Y.-S. Zimmermann, C. Groot, W. van den Broek,

T. Wintgens, Comparison of pre-treatment technologies towards improving reverseosmosis desalination of cooling tower blow down, Desalination 357 (2015)140–149.

[9] Z. Xu, J. Zhang, M. Shan, Y. Li, B. Li, J. Niu, B. Zhou, X. Qian, Organosilane-functionalized graphene oxide for enhanced antifouling and mechanical propertiesof polyvinylidene fluoride ultrafiltration membranes, J. Membr. Sci. 458 (2014)1–13.

[10] J.R. Du, S. Peldszus, P.M. Huck, X. Feng, Modification of membrane surfaces viamicroswelling for fouling control in drinking water treatment, J. Membr. Sci. 475(2015) 488–495.

[11] F. Zamani, A. Ullah, E. Akhondi, H.J. Tanudjaja, E.R. Cornelissen, A. Honciuc,A.G. Fane, J.W. Chew, Impact of the surface energy of particulate foulants onmembrane fouling, J. Membr. Sci. 510 (2016) 101–111.

[12] J.H. Jhaveri, Z. Murthy, A comprehensive review on anti-fouling nanocompositemembranes for pressure driven membrane separation processes, Desalination 379(2016) 137–154.

[13] V. Kochkodan, N. Hilal, A comprehensive review on surface modified polymermembranes for biofouling mitigation, Desalination 356 (2015) 187–207.

[14] P.K.S. Mural, A. Banerjee, M.S. Rana, A. Shukla, B. Padmanabhan, S. Bhadra,G. Madras, S. Bose, Polyolefin based antibacterial membranes derived from PE/PEOblends compatibilized with amine terminated graphene oxide and maleated PE, J.Mater. Chem. A 2 (2014) 17635–17648.

[15] C. Liu, A.F. Faria, J. Ma, M. Elimelech, Mitigation of biofilm development on thin-film composite membranes functionalized with zwitterionic polymers and silvernanoparticles, Environ. Sci. Technol. 51 (2016) 182–191.

[16] D.M. Davenport, J. Lee, M. Elimelech, Efficacy of antifouling modification of ul-trafiltration membranes by grafting zwitterionic polymer brushes, Sep. Purif.Technol. 189 (2017) 389–398.

[17] R. Shevate, M. Kumar, M. Karunakaran, M.N. Hedhili, K.-V. Peinemann,Polydopamine/cysteine surface modified isoporous membranes with self-cleaningproperties, J. Membr. Sci. 529 (2017) 185–194.

[18] S.R.-V. Castrillón, X. Lu, D.L. Shaffer, M. Elimelech, Amine enrichment and poly(ethylene glycol)(PEG) surface modification of thin-film composite forward osmosismembranes for organic fouling control, J. Membr. Sci. 450 (2014) 331–339.

[19] J. Zhao, Q. Shi, S. Luan, L. Song, H. Yang, H. Shi, J. Jin, X. Li, J. Yin, P. Stagnaro,Improved biocompatibility and antifouling property of polypropylene non-wovenfabric membrane by surface grafting zwitterionic polymer, J. Membr. Sci. 369(2011) 5–12.

[20] X. Li, J. Li, X. Fang, K. Bakzhan, L. Wang, B. Van der Bruggen, A synergetic analysismethod for antifouling behavior investigation on PES ultrafiltration membrane withself-assembled TiO2 nanoparticles, J. Colloid Interface Sci. 469 (2016) 164–176.

[21] X. Li, J. Li, B. Van der Bruggen, X. Sun, J. Shen, W. Han, L. Wang, Fouling behaviorof polyethersulfone ultrafiltration membranes functionalized with sol–gel formedZnO nanoparticles, RSC Adv. 5 (2015) 50711–50719.

[22] J. Lin, W. Ye, K. Zhong, J. Shen, N. Jullok, A. Sotto, B. Van der Bruggen,Enhancement of polyethersulfone (PES) membrane doped by monodisperse Stöbersilica for water treatment, Chem. Eng. Process. Process Intensif. 107 (2016)194–205.

[23] R. Bernstein, V. Freger, J.-H. Lee, Y.-G. Kim, J. Lee, M. Herzberg, ‘Should I stay orshould I go?’ Bacterial attachment vs biofilm formation on surface-modifiedmembranes, Biofouling 30 (2014) 367–376.

[24] Q. Yu, Z. Wu, H. Chen, Dual-function antibacterial surfaces for biomedical appli-cations, Acta Biomater. 16 (2015) 1–13.

[25] R. Hu, G. Li, Y. Jiang, Y. Zhang, J.-J. Zou, L. Wang, X. Zhang, Silver–zwitterionorganic–inorganic nanocomposite with antimicrobial and antiadhesive capabilities,Langmuir 29 (2013) 3773–3779.

[26] Z. Xu, S. Ye, G. Zhang, W. Li, C. Gao, C. Shen, Q. Meng, Antimicrobial polysulfoneblended ultrafiltration membranes prepared with Ag/Cu2O hybrid nanowires, J.Membr. Sci. 509 (2016) 83–93.

[27] P. Zou, W. Hartleb, K. Lienkamp, It takes walls and knights to defend a castle–-synthesis of surface coatings from antimicrobial and antibiofouling polymers, J.Mater. Chem. 22 (2012) 19579–19589.

[28] H. Wu, Y. Liu, J. Huang, L. Mao, J. Chen, M. Li, Preparation and characterization ofantifouling and antibacterial polysulfone ultrafiltration membranes incorporatedwith a silver–polydopamine nanohybrid, J. Appl. Polym. Sci. 135 (2018) 46430.

[29] W. Aoki, M. Ueda, Characterization of antimicrobial peptides toward the develop-ment of novel antibiotics, Pharmaceuticals 6 (2013) 1055–1081.

[30] F. Perreault, M.E. Tousley, M. Elimelech, Thin-film composite polyamide mem-branes functionalized with biocidal graphene oxide nanosheets, Environ. Sci.Technol. Lett. 1 (2013) 71–76.

[31] L.L. Zhang, R. Zhou, X. Zhao, Graphene-based materials as supercapacitor elec-trodes, J. Mater. Chem. 20 (2010) 5983–5992.

[32] L. Yu, Y. Zhang, B. Zhang, J. Liu, H. Zhang, C. Song, Preparation and character-ization of HPEI-GO/PES ultrafiltration membrane with antifouling and antibacterialproperties, J. Membr. Sci. 447 (2013) 452–462.

[33] X. Huang, K.L. Marsh, B.T. McVerry, E.M. Hoek, R.B. Kaner, Low-fouling anti-bacterial reverse osmosis membranes via surface grafting of graphene oxide, ACSAppl. Mater. Interfaces 8 (2016) 14334–14338.

[34] S. Zinadini, A.A. Zinatizadeh, M. Rahimi, V. Vatanpour, H. Zangeneh, Preparationof a novel antifouling mixed matrix PES membrane by embedding graphene oxidenanoplates, J. Membr. Sci. 453 (2014) 292–301.

[35] Y. Huang, H. Li, L. Wang, Y. Qiao, C. Tang, C. Jung, Y. Yoon, S. Li, M. Yu,Ultrafiltration membranes with structure‐optimized graphene‐oxide coatings forantifouling oil/water separation, Adv. Mater. Interfaces 2 (2015) 1400433.

[36] Y. Tu, M. Lv, P. Xiu, T. Huynh, M. Zhang, M. Castelli, Z. Liu, Q. Huang, C. Fan,H. Fang, Destructive extraction of phospholipids from Escherichia coli membranesby graphene nanosheets, Nat. Nanotechnol. 8 (2013) 594.

[37] S. Liu, T.H. Zeng, M. Hofmann, E. Burcombe, J. Wei, R. Jiang, J. Kong, Y. Chen,Antibacterial activity of graphite, graphite oxide, graphene oxide, and reducedgraphene oxide: membrane and oxidative stress, ACS Nano 5 (2011) 6971–6980.

[38] F. Perreault, A.F. De Faria, S. Nejati, M. Elimelech, Antimicrobial properties ofgraphene oxide nanosheets: why size matters, ACS Nano 9 (2015) 7226–7236.

[39] O. Akhavan, E. Ghaderi, Toxicity of graphene and graphene oxide nanowalls againstbacteria, ACS Nano 4 (2010) 5731–5736.

[40] X. Lu, X. Feng, J.R. Werber, C. Chu, I. Zucker, J.-H. Kim, C.O. Osuji, M. Elimelech,Enhanced antibacterial activity through the controlled alignment of graphene oxidenanosheets, Proc. Natl. Acad. Sci. USA 114 (2017) E9793–E9801.

[41] S.S. Nanda, D.K. Yi, K. Kim, Study of antibacterial mechanism of graphene oxideusing Raman spectroscopy, Sci. Rep. 6 (2016) 28443.

[42] K. Krishnamoorthy, M. Veerapandian, L.-H. Zhang, K. Yun, S.J. Kim, Antibacterialefficiency of graphene nanosheets against pathogenic bacteria via lipid peroxida-tion, J. Phys. Chem. C 116 (2012) 17280–17287.

[43] W. Hu, C. Peng, W. Luo, M. Lv, X. Li, D. Li, Q. Huang, C. Fan, Graphene-basedantibacterial paper, ACS Nano 4 (2010) 4317–4323.

[44] C. Zhao, X. Xu, J. Chen, G. Wang, F. Yang, Highly effective antifouling performanceof PVDF/graphene oxide composite membrane in membrane bioreactor (MBR)system, Desalination 340 (2014) 59–66.

[45] J. Liu, R. Bernstein, High-flux thin-film composite polyelectrolyte hydrogel mem-branes for ethanol dehydration by pervaporation, J. Membr. Sci. 534 (2017) 83–91.

[46] W. Zhang, Z. Yang, Y. Kaufman, R. Bernstein, Surface and anti-fouling properties ofa polyampholyte hydrogel grafted onto a polyethersulfone membrane, J. ColloidInterface Sci. 517 (2018) 155–165.

[47] V.A. Vasantha, S. Jana, A. Parthiban, J.G. Vancso, Water swelling, brine solubleimidazole based zwitterionic polymers–synthesis and study of reversible UCST be-haviour and gel–sol transitions, Chem. Commun. 50 (2014) 46–48.

[48] K. Gao, Y. Su, L. Zhou, M. He, R. Zhang, Y. Liu, Z. Jiang, Creation of active-passiveintegrated mechanisms on membrane surfaces for superior antifouling and anti-bacterial properties, J. Membr. Sci. 584 (2018) 621–631.

[49] S.E. Kwan, E. Bar-Zeev, M. Elimelech, Biofouling in forward osmosis and reverseosmosis: measurements and mechanisms, J. Membr. Sci. 493 (2015) 703–708.

[50] S. Ichimura, Y. Sota, J. Ishikawa, Y. Imanishi, K. Kitamura, S. Tsujii, T. Yamaguchi,Poly (p-phenylene sulfonic acid-ran-2, 5-benzophenone) pore-filling membraneswith highly packed acid structure and their polymer electrolyte fuel cell perfor-mances, Int. J. Hydrog. Energy 41 (2016) 21461–21469.

[51] C.-C. Yeh, A. Venault, Y. Chang, Structural effect of poly (ethylene glycol) seg-mental length on biofouling and hemocompatibility, Polym. J. 48 (2016) 551.

[52] F. Perreault, H. Jaramillo, M. Xie, M. Ude, L.D. Nghiem, M. Elimelech, Biofoulingmitigation in forward osmosis using graphene oxide functionalized thin-film com-posite membranes, Environ. Sci. Technol. 50 (2016) 5840–5848.

[53] D. Qin, Z. Liu, Z. Liu, H. Bai, D.D. Sun, Superior antifouling capability of hydrogelforward osmosis membrane for treating wastewaters with high concentration oforganic foulants, Environ. Sci. Technol. 52 (2018) 1421–1428.

[54] P.H. Castilho, T.R. Correia, M.T. Pessoa de Amorim, I.C. Escobar, J.A. Queiroz,I.J. Correia, A.M. Morão, Modification of microfiltration membranes by hydrogelimpregnation for pDNA purification, J. Appl. Polym. Sci. 132 (2015) 21.

[55] J. Pieracci, J.V. Crivello, G. Belfort, Photochemical modification of 10 kDa poly-ethersulfone ultrafiltration membranes for reduction of biofouling, J. Membr. Sci.156 (1999) 223–240.

[56] T. Cai, X. Li, C. Wan, T.-S. Chung, Zwitterionic polymers grafted poly (ether sulfone)hollow fiber membranes and their antifouling behaviors for osmotic power gen-eration, J. Membr. Sci. 497 (2016) 142–152.

[57] N.L. Le, M. Quilitzsch, H. Cheng, P.-Y. Hong, M. Ulbricht, S.P. Nunes, T.-S. Chung,Hollow fiber membrane lumen modified by polyzwitterionic grafting, J. Membr.Sci. 522 (2017) 1–11.

[58] L. Huang, S. Zhao, Z. Wang, J. Wu, J. Wang, S. Wang, In situ immobilization ofsilver nanoparticles for improving permeability, antifouling and anti-bacterialproperties of ultrafiltration membrane, J. Membr. Sci. 499 (2016) 269–281.

[59] P. Biswas, R. Bandyopadhyaya, Biofouling prevention using silver nanoparticleimpregnated polyethersulfone (PES) membrane: e. coli cell-killing in a continuouscross-flow membrane module, J. Colloid Interface Sci. 491 (2017) 13–26.

[60] J. Park, J. Kim, K. Singha, D.K. Han, H. Park, W.J. Kim, Nitric oxide integratedpolyethylenimine-based tri-block copolymer for efficient antibacterial activity,Biomaterials 34 (2013) 8766–8775.

[61] A. áde Leon, On the antibacterial mechanism of graphene oxide (GO)Langmuir–Blodgett films, Chem. Commun. 51 (2015) 2886–2889.

[62] M. Okshevsky, R.L. Meyer, The role of extracellular DNA in the establishment,maintenance and perpetuation of bacterial biofilms, Crit. Rev. Microbiol. 41 (2015)341–352.

W. Zhang et al. Journal of Membrane Science 565 (2018) 293–302

302