Abstract

The covalently functionalized cellulose nanocrystal (CNC) composites were synthesized by bonding common bactericides, such as dodecyl dimethyl benzyl ammonium chloride (DDBAC), ZnO and graphene oxide (GO) nanosheets, onto the CNC's surface. Then, the DDBAC/CNC, ZnO/CNC and GO/CNC nanocomposites modified polyvinylidene fluoride (PVDF) ultrafiltration membranes were fabricated by a simple one-step non-solvent induced phase separation (NIPS) process. The resultant hybrid membranes possessed porous and rough surfaces with more finger-like macropores that even extended through the entire cross-section. The hydrophilicity, permeability, antibacterial and antifouling performance and mechanism of the hybrid ultrafiltration membranes were evaluated and compared in detail, aiming at screening a superior hybrid membrane for practical application in micro-polluted source water purification. Among these newly-developed hybrid membranes, GO/CNC/PVDF exhibited an enhanced perm-selectivity with a water flux of 230 L/(m2 h bar) and humic acid rejection of 92%, the improved antibacterial activity (bacteriostasis rate of 93%) and antifouling performance (flux recovery rate (FRR) of >90%) being due to the optimized pore structure, higher surface roughness, incremental hydrophilicity and electronegativity. A lower biofouling level after three weeks' filtration of the actual micro-polluted source water further demonstrated that embedding the hydrophilic and antibacterial GO/CNC nanocomposite into the polymer matrix is an effective strategy to improve membrane anti-biofouling ability.

INTRODUCTION

Membrane filtration technology has been widely used in removing soluble/suspended particulates, microorganisms and some organic matters from contaminated water. However, membrane fouling is the major obstacle to the widespread application of membranes in water recycling and wastewater treatment. Membrane fouling has been generally recognized to happen in two ways: one is the direct accumulation/deposition and adsorption of foulants onto the membrane surface or within membrane pores. Another way is biofouling, derived from the irreversible bacterial adhesion, growth, and multiplication of sessile cells to the membrane surface. Previous reports (Meng et al. 2009, 2017) demonstrated that the enhanced hydrophilicity and electronegativity can efficiently improve membrane surface water affinity and inhibit membrane pore blocking or cake layer formation, leading to a higher permeate flux. In addition, endowing a membrane antibacterial function can effectively prevent biofouling (Huang et al. 2017). Accordingly, developing a membrane with high hydrophilicity, strong electronegativity, and effective biocidal property is essential for membrane biofouling mitigation.

Cellulose nanocrystal (CNC), the crystalline portions of cellulose that are typically obtained by acid hydrolysis of cellulosic materials, is a promising nanofiller for improving water filtration membrane performances in view of its environmental friendly advantages, large aspect ratio, high chemical resistance, good mechanical strength and strong hydrophilicity (Bai et al. 2017; Voisin et al. 2017). It was reported (Lv et al. 2017, 2018a) that only a small amount of CNC added into the poly(vinylidene fluoride) (PVDF) membrane was enough to give the membrane high water flux, strong mechanical strength, and thermostability. Even so, however, the intrinsic CNC does not have any antibacterial activity, thus the functionalized CNC should be developed to endow the CNC composite with effective antibacterial ability. In the past decade, graphene oxide (GO) blended membranes have exhibited more superior performance in improving the mechanical strength, electronegativity, hydrophilicity and antimicrobial activity of membranes, due to their uniform dispersion, accelerated charge transfer and increased cytotoxic effect of physical piercing and oxidative stress by means of the adequate exposure of the active edges of GO nanosheets (Liu et al. 2011; Feng et al. 2017). However, GO nanosheets will leach from the membrane surface during the long-term filtration process and thus cannot maintain high efficiency in inhibiting bacterial colonization. In order to fully exert GO nanosheets' functions, it is highly desirable to immobilize GO nanosheets into the polymer matrix membranes via a reinforcing agent. The hydroxyl groups of CNC and the abundant oxygen-containing groups of GO sheets allow them to be integrated facilely in one pot via hydrogen bonding interactions (Ouyang et al. 2013), and the presence of CNC provides more cross-linking points between the GO sheets and CNC, which facilitates the efficient dispersion of GO sheets and CNC (Lv et al. 2018b). In this context, using GO/CNC composites as nanofillers would efficiently improve the hydrophilicity, antibacterial and antifouling properties of the functionalized PVDF membranes. Besides, ZnO is also frequently used as an antibacterial agent (Zhao et al. 2018), due to its outstanding optoelectronic property, high hydrophilicity, and broad antibacterial spectrum. The hybridization of cellulose and its derivatives with ZnO nanoparticles not only overcomes ZnO nanoparticle aggregation but also endows the ZnO/cellulose composite with superior antibacterial activity compared to both ZnO and cellulose (Lefatshe et al. 2017). On the other hand, organic antibacterial agents such as quaternary ammonium compounds (Ye et al. 2016; Zhang et al. 2016a, 2016b) are also widely used in antibacterial and anti-biofouling membranes in terms of their efficient bacteria killing capability and inhibitory effect on bacterial growth. As a quaternary ammonium type antibacterial agent, dodecyl dimethyl benzyl ammonium chloride (DDBAC) has a broad spectrum, highly efficient disinfection and algae killing activity. To the best of our knowledge, there is no previous report about direct blending of DDBAC/CNC composites with PVDF to simultaneously improve hydrophilicity, antibacterial and antifouling properties.

In the present study, three common antibacterial (DDBAC, ZnO and GO) functionalized CNC nanocomposites were first synthesized, and their microstructure and morphology properties were characterized by scanning electron microscope (SEM), Fourier transform infrared spectroscopy (FTIR), and X-ray diffraction (XRD). Then, the polyvinylidene fluoride (PVDF) hybrid ultrafiltration membranes were fabricated by the non-solvent induced phase separation (NIPS) method using DDBAC/CNC, ZnO/CNC and GO/CNC composites as nanofillers. The morphology, hydrophilicity, permeability, antibacterial activity, antifouling properties and mechanism of the various PVDF hybrid membranes were investigated and compared in detail, with the aim of screening an optimal antibacterial agent that can concurrently improve the hydrophilicity, antibacterial and antifouling ability of the membrane, and finally obtain a superior hybrid membrane for practical membrane filtration application in the treatment of micro-polluted source water.

MATERIALS AND METHODS

Preparation of ZnO, DDBAC and GO functionalized CNC nanocomposites

The cellulose nanocrystals were prepared using filter paper as raw materials according to the previous method (Lv et al. 2017). The preparation processes of DDBAC/CNC, ZnO/CNC and GO/CNC composites are provided in supplementary material S1 (available with the online version of this paper).

Preparation of PVDF hybrid membranes

The preparation processes of DDBAC/CNC, ZnO/CNC and GO/CNC nanocomposite-modified PVDF membranes are provided in supplementary material S2; all the fabricated membranes (Figure S1) were preserved in deionized water prior to characterization and use. (Supplementary material S1 and Figure S1 are available online.)

Characterization and performance analysis of membranes

The characterization and analysis methods of membrane morphology, porous structure, hydrophilicity, permeability, antibacterial and antifouling performances as well as membrane filtration application are provided in supplementary material S3-S6 (available online).

RESULTS AND DISCUSSION

Characterization of the functionalized CNC nanocomposites

The morphologies of the pristine CNC and DDBAC, ZnO and GO-functionalized CNC nanocomposites are shown in Figure 1. It can be clearly observed that the pristine CNC exhibited dense and hierarchical sheet-like structures due to the aggregation of the nanoscale rod-like CNC. The SEM images of functionalized CNC composites displayed three-dimensional loose sheet-like nanostructures, indicating that the introduction of DDBAC, ZnO and GO nanofillers notably facilitated CNC dispersion. Especially, ZnO and GO nanosheets were assembled on the agminated CNC surface, which is consistent with the fact that ZnO nanosheets can be synthesized through low-temperature hydrothermal treatment (Lefatshe et al. 2017). These loose and disordered stacking structures of the functionalized CNC nanocomposites may exert a positive effect on membrane permeability. The similar characteristic bands corresponding to CNC are also observed in Figure 2(a)–2(c) for the functionalized CNC composites. In each FTIR spectrum, the peaks at 3,337–3,423 and 1,380 cm–1 can be assigned to stretching and bending vibrations of hydroxyl groups, respectively (Lv et al. 2017, 2018a). The C–O–C stretching of pyranose and glucose ring skeletal vibration of CNC was observed at 1,050 cm−1 (Lv et al. 2018a). The peaks at 2,924, 2,853, 1,625 and 1,462 cm−1 in Figure 2(a) correspond to –CH3, –CH2, –C = C– and N–C, and C–H stretching vibration of DDBAC, respectively (Ye et al. 2016). The band at 1,634 cm−1 in Figure 2(b) is related to hydroxyl of adsorbed water and another band at around 421 cm−1 belonged to the Zn–O stretching mode, implying the existence of nanocrystalline ZnO in ZnO/CNC nanocomposites (Lefatshe et al. 2017). The FTIR spectrum of GO in Figure 2(c) presents the C = C bonds (1,628 cm−1), C–O vibrations (1,051 cm−1), and C = O stretching vibration (1,729 cm−1) (Lv et al. 2018b). A wider peak at around 3,400–3,500 cm−1 was observed in GO/CNC samples, implying the formation of hydrogen bond interactions between the hydroxyl groups of CNC and the abundant oxygen-containing groups of GO sheets (Ouyang et al. 2013; Lv et al. 2018b).

Figure 1

SEM images of (a) CNC, (b) DDBAC/CNC, (c) ZnO/CNC, and (d) GO/CNC nanocomposites. The insets in (a) and (b)–(d) show photographs of freeze-dried CNC powders and the partially amplified SEM images.

Figure 1

SEM images of (a) CNC, (b) DDBAC/CNC, (c) ZnO/CNC, and (d) GO/CNC nanocomposites. The insets in (a) and (b)–(d) show photographs of freeze-dried CNC powders and the partially amplified SEM images.

Figure 2

FTIR spectra and XRD patterns: (A,a) DDBAC/CNC, (B,b) ZnO/CNC and (C,c) GO/CNC nanocomposites.

Figure 2

FTIR spectra and XRD patterns: (A,a) DDBAC/CNC, (B,b) ZnO/CNC and (C,c) GO/CNC nanocomposites.

The crystal structure of the pristine CNC and the functionalized CNC composites was investigated by XRD. As shown in Figure 2(a)–2(c), the pristine CNC exhibited four distinct peaks at 2θ = 16.1°, 20.0°, 22.2° and 34.5°, which were attributed to (1Ī0), (110), (200) and (040) planes, respectively, in accordance with the characteristic diffraction peaks of monoclinic cellulose I lattice (Lv et al. 2018b). Many new intensive peaks at 2θ = 31.8°, 34.6°, 36.2°, 47.6°, 56.5°, 62.8°, and 67.9° were observed in ZnO/CNC nanocomposite, which correspond respectively to (100), (002), (101), (102), (110), (103), and (112) planes of the hexagonal wurtzite structure of ZnO (JCPDS No. 36-1451) (Lefatshe et al. 2017). A slight diffraction peak at 2θ = 11.7° appeared in the XRD pattern of the GO/CNC composite, which is attributed to the interlayer spacing between GO nanosheets (d002 = 0.756 nm) (Lv et al. 2018b), while the (002) reflection peak of GO at 2θ = 26.3° disappeared, implying the good exfoliation and dispersion of GO in the GO/CNC. Of note, the introduction of DDBAC has not altered the crystal structure of CNC, because no new peaks nor any peak shift were observed in the XRD pattern of DDBAC/CNC.

Characterization of PVDF hybrid membranes

The SEM images of the top surfaces, the bottom surfaces, and the cross-sections of the pristine CNC and the functionalized CNC nanocomposite modified PVDF membranes are presented in Figure 3. All the membranes' top surfaces are smooth and no distinct deformations were observed, while the bottom surfaces appeared porous. This result may be related to the strong non-solvent effect during the NIPS process, which led to the rapid and delayed phase separation on the top and bottom surfaces, respectively, when the initial casting films were immersed into the coagulation water bath (Huang et al. 2017). The cross-section SEM images displayed a typical asymmetrical porous structure with a dense superficial layer, a finger-like intermediate layer, and a sponge-like bottom support layer. Compared to the unmodified membrane (Figure S2a) (Figure S2 is available with the online version of this paper), all the modified PVDF membranes possessed more finger-like macropores that even extended through the entire cross-section. This phenomenon can be attributed to the excellent hydrophilicity of the pristine and functionalized CNC nanocomposites, which significantly altered the thermodynamic instability of casting solutions during the gelation process and thus effectively accelerated the inter-diffusion and de-mixing rate of solvent and non-solvent in the phase inversion process (Zhang et al. 2016a; Lv et al. 2018b). These results indicated that the functionalized CNC nanocomposites optimized the pore micro-structure, which is beneficial for improving membrane permeability.

Figure 3

SEM images of top surface, cross-section and bottom surface of different PVDF hybrid membranes.

Figure 3

SEM images of top surface, cross-section and bottom surface of different PVDF hybrid membranes.

Compared to the CNC/PVDF membrane, the functionalized CNC composite-modified PVDF membranes possessed larger average pore size and superior porosity (Figure 4(a)), and their hydraulic contact angles decreased, except DDBAC/CNC/PVDF membrane (Figure 4(b)). The probable reasons may be that: (i) the more hydrophilic ZnO/CNC and GO/CNC nanocomposites would tend to migrate to the membrane surface and decrease interface energy during the phase separation process (Hong & He 2012; Lv et al. 2018b), (ii) the hydrophobic segment on the long alkyl chain (C12) of DDBAC/CNC nanocomposites led to slightly increased contact angle (Zhang et al. 2016a, 2016b). Apparently, the superior hydrophilic GO/CNC nanocomposite would more effectively promote the formation of large pore channels and the amorphous nature of GO/CNC/PVDF membrane is beneficial for water permeability and antifouling property.

Figure 4

Average pore size and porosity (a), contact angle and water content (b) of various PVDF hybrid membranes.

Figure 4

Average pore size and porosity (a), contact angle and water content (b) of various PVDF hybrid membranes.

Anti-adhesion and anti-bacterial properties of PVDF hybrid membranes

Biofilm formation through the irreversible bacterial adhesion on membrane surface is believed to be the major factor in membrane fouling (Meng et al. 2009, 2017). The anti-adhesion performance against Escherichia coli. was firstly assessed for different PVDF hybrid membranes through SEM observation to investigate the antibacterial ability. As seen in Figure S2c, large amounts of bacteria with intact morphology were attached on the PVDF membrane surface after 12 h incubation in the bacterial media, which could be ascribed to no antibacterial activity and inherent hydrophobicity of the original PVDF membrane. However, the bacterial adhesion on the modified PVDF membranes decreased significantly (Figure 5(a)–5(d)), and more notably, the attached E. coli on DDBAC/CNC, ZnO/CNC and GO/CNC nanocomposite-modified PVDF membranes presented poor cellular integrity, implying serious destruction of the bacterial structure. It is reported although both hydrophilic and electrostatic interactions between bacterial and membrane surface are involved in bacterial adhesion, membrane hydrophilicity plays a more important role in the bacterial attachment process (Hou et al. 2017a, 2017b). Compared to the original PVDF membrane, the substantially improved hydrophilicity of the modified PVDF membranes can bond with water molecules, which generate a hydration layer on the membrane surface and prevent E. coli attachment in spite of the occurrence of potential electrostatic attraction between the positively charged ZnO/CNC and DDBAC/CNC nanocomposites and the negatively charged E. coli. On the other hand, the attached E. coli on CNC/PVDF maintained cell integrity better when compared with the other three modified PVDF membranes, due to the non-bactericidal nature of the pristine CNC.

Figure 5

The SEM images of E.coli adhesion and the results of biocidal efficacy against E. coli on: (A,a) CNC/PVDF, (B,b) DDBAC/CNC/PVDF, (C,c) ZnO/CNC/PVDF, and (D,d) GO/CNC/PVDF hybrid membranes, respectively.

Figure 5

The SEM images of E.coli adhesion and the results of biocidal efficacy against E. coli on: (A,a) CNC/PVDF, (B,b) DDBAC/CNC/PVDF, (C,c) ZnO/CNC/PVDF, and (D,d) GO/CNC/PVDF hybrid membranes, respectively.

To gain more insight into the antibacterial activity, the original and the modified PVDF membranes were also evaluated in term of bacteria inactivation and biocidal efficacy after being cultivated with E. coli through ‘sandwich’ systems (Hou et al. 2017a, 2017b). Figure 5(a)–5(d) show that in comparison with the original PVDF membrane (Figure S2d), a significant decline in bacterial colonies was observed for E. coli eluted from the DDBAC/CNC, ZnO/CNC and GO/CNC nanocomposite-modified PVDF membranes, while the CNC/PVDF membrane did not exhibit a dramatic reduction in E. coli colonies within 30 min of contact, as evidenced by about 2.1%, 73.6%, 80.7% and 92.8% of bacteriostasis rate (BR) for CNC, DDBAC/CNC, ZnO/CNC and GO/CNC-modified PVDF membranes (Figure S2e), respectively, indicating the superior antibacterial abilities of DDBAC/CNC, ZnO/CNC and GO/CNC nanocomposite-modified PVDF membranes. Nevertheless, the bacteria-killing approach of DDBAC, ZnO and GO concerns different inactivation mechanisms:

  • (i)

    The general antibacterial mechanism of fat ammonium bactericides with the positively charged quaternary ammonium groups involves electrostatic binding to the negatively charged cell wall of E. coli. The long alkyl chains (C12) of DDBAC provided good compatibility with the bacterial cell wall, which increased cell membrane permeability and then disrupted the cytoplasmic membrane and led to bacteria death (Zhang et al. 2016b).

  • (ii)

    The general antibacterial mechanism of ZnO is mainly related to the photo-generated and/or non-photo-generated superoxides (O2−), hydroxyl radicals (OH) and hydrogen peroxide (H2O2) (Lefatshe et al. 2017; Zhao et al. 2018), which damage cell membranes/proteins and cause bacterial death. The CNC promotes the good dispersion of ZnO nanosheets and provides a larger surface area, thus leading to generation of more reactive oxygen species and higher antibacterial activity.

  • (iii)

    The antibacterial mechanism of GO nanosheets involves the combined effect of physical membrane stress including wrapping, insertion and perforation/ incision as well as induction of oxidative stress (Liu et al. 2011; Feng et al. 2017). The electrostatic repulsion between the negatively charged GO/CNC and E. coli can effectively prevent the initial attachment, growth and spreading of E. coli onto the membrane surface. Additionally, the wrinkled GO nanosheets with sharp edges and defect sites on the basal plane could cause membrane stress and oxidative stress, which can destroy the bacteria cell integrity and effectively oxidize the thiol groups (–SH) of glutathione in bacteria to disulfide bond (–S–S–), resulting in irreversible damage to the bacterial cells (Feng et al. 2017). Therefore, the excellent hydrophilicity, cytotoxicity, and electronegativity of GO/CNC nanocomposites endowed the GO/CNC/PVDF membrane with outstanding antibacterial ability, which is favorable for preventing biofouling formation.

Antifouling performance of membranes

Three-stage dynamic filtration was carried out to explore the antifouling performances. As seen in Figure 6(a), the initial pure water flux of CNC/PVDF was around 75.0 L M–2 h–1 bar–1, while the DDBAC/CNC, ZnO/CNC and GO/CNC modified PVDF membranes presented higher fluxes of about 130, 175 and 230 L M–2 h–1bar–1, which are around 1.73, 2.33 and 3.07 times that of the CNC/PVDF membrane. When pure water was substituted by HA solution, an obvious decline in permeation flux was observed for all membranes due to the formation of a fouling layer caused by HA molecules' adsorption/deposition on the membrane surface/pore wall, which decreased the effective filtration area (Liu et al. 2015; Lv et al. 2018b). The recuperative pure water flux of all membranes exhibited an apparent upswing, but did not recover to the initial levels owing to the entrapment of HA molecules within the pores. Clearly, the tendency of pure water permeability and HA rejection (Figure 6(b)) agreed well with SEM images; that is, blending the functionalized CNC nanocomposites' optimized pore micro-structure, porosity, hydrophilicity and electronegativity of the PVDF hybrid membrane greatly.

Figure 6

(a) Permeation flux change during three-step ultrafiltration of HA solution and (b) time-dependent rejection rate of HA solution for different PVDF hybrid membranes; AFM images of (c) CNC, (d) DDBAC/CNC, (e) ZnO/CNC and (f) GO/CNC modified PVDF hybrid membranes.

Figure 6

(a) Permeation flux change during three-step ultrafiltration of HA solution and (b) time-dependent rejection rate of HA solution for different PVDF hybrid membranes; AFM images of (c) CNC, (d) DDBAC/CNC, (e) ZnO/CNC and (f) GO/CNC modified PVDF hybrid membranes.

Generally, the antifouling ability of the membranes was associated with surface roughness. The three-dimensional AFM images in Figure 6(c)-(f) show that compared to the CNC/PVDF membrane, the other three hybrid membranes displayed a prominent texture structure with more ridges and valleys, and thus presented higher roughness parameters (Ra). Previous studies (Zhang et al. 2013a, 2013b; Wu et al. 2015) confirmed that a higher surface roughness will not impose a negative effect on the separation performance of a hydrophilic membrane. On the contrary, it can effectively resist contaminants' adhesion and improves the membrane's permeating flux and antifouling ability, because the higher roughness produces more flow obstacles, which would effectively increase the hydraulic shear stress and mass transport through micro-turbulence and more cavities generated at the fluid/membrane interface, leading to a positive effect on permeation flux and fouling mitigation (Izák et al. 2008; Pourbozorg et al. 2016).

Flux recovery ratio (FRR) and filtration resistances were also used to clarify the antifouling mechanism of the PVDF hybrid membranes. As shown in Figure 7(a), the GO/CNC/PVDF exhibited the highest FRR value (above 90%) and the lowest HA adsorption value (ca. 75.3 mg m–2) after the ultrafiltration-regeneration process, which demonstrated the excellent antifouling property owing to the enhanced hydrophilicity and the optimized pore micro-structure after the introduction of GO/CNC nanocomposites. As seen in Figure 7(b), the functionalized CNC composite-modified PVDF membranes displayed significantly reduced filtration resistance (Rt), membrane resistance (Rm) and fouling resistance (Rf) values, suggesting a low fouling tendency and high permeability compared to the CNC/PVDF membrane. Of note, the GO/CNC/PVDF possessed the lowest filtration resistance and superior antifouling property, which could be ascribed to the abundant hydroxyl groups of CNC providing more binding sites for oxygen-containing functional groups of GO, and thus promoting the good dispersion of GO/CNC nanocomposites in PVDF membrane matrices (Lv et al. 2017, 2018a, 2018b), resulting in the enhanced pore size, porosity, hydrophilicity and electronegativity of the GO/CNC/PVDF membrane compared to the CNC/PVDF membrane. These outstanding properties effectively decreased the hydrophobic interaction between GO/CNC/PVDF membranes and HA molecules, which restrained the accumulation of the cake layer during the membrane filtration process. More importantly, the reversible cake layer can be facilely removed by hydraulic cleaning, thus the permeation flux was recovered to a greater extent, as depicted in Figure 7(a).

Figure 7

Antifouling properties of the different PVDF hybrid membranes: (a) water flux recovery ratio and HA adsorption and (b) membrane filtration resistance.

Figure 7

Antifouling properties of the different PVDF hybrid membranes: (a) water flux recovery ratio and HA adsorption and (b) membrane filtration resistance.

Membrane filtration application for actual micro-polluted source water

During membrane purification of the micro-polluted source water, a biofilm generally formed via the deposition/attachment of humic substances and microorganisms onto the membrane surface, resulting in the decrease in membrane permeability flux. In this work, the antifouling performances of different PVDF hybrid membranes were evaluated in a laboratory-scale membrane filtration system using actual reservoir water (Table S1, available online). As seen in Figure 8, the GO/CNC/PVDF membrane displayed a much more slowly-increasing rate of trans-membrane pressure (TMP) than the other three PVDF hybrid membranes, which is attributed to the enhanced hydrophilicity, electronegativity, antibacterial activity and the optimized pore structure, which effectively inhibit bacterial attachment/growth, alleviating membrane fouling and prolonging membrane operating time. The result demonstrated that embedding the hydrophilic and antibacterial GO/CNC nanocomposites into a polymer matrix is an effective strategy to improve membrane anti-biofouling ability.

Figure 8

The time-dependent permeation flux and TMP of different PVDF hybrid membranes during the treatment of the actual reservoir water in a laboratory-scale membrane filtration system.

Figure 8

The time-dependent permeation flux and TMP of different PVDF hybrid membranes during the treatment of the actual reservoir water in a laboratory-scale membrane filtration system.

CONCLUSION

The DDBAC/CNC, ZnO/CNC and GO/CNC nanocomposite-blended PVDF hybrid ultrafiltration membranes were fabricated, characterized and compared in term of membrane morphology, pore structure, hydrophilicity, antibacterial activity, and antifouling performance. Among these newly-developed hybrid membranes, the GO/CNC/PVDF exhibited superior antibacterial activity and antifouling performance, due to the improved pore micro-structure, surface roughness, hydrophilicity and electronegativity, leading to the increased filtration area, the decreased adhesion of bacteria/contaminants, and effective membrane fouling alleviation. The outstanding antifouling performance of the GO/CNC/PVDF membrane during the practical micro-polluted resource water filtration demonstrated that the hydrophilic and antibacterial GO/CNC nanocomposite is an effective additive for the anti-biofouling membrane.

ACKNOWLEDGEMENTS

This work was supported by the National Natural Science Foundation of China (No. 21437001) and the Program of Introducing Talents of Discipline to Universities (No. B13012).

REFERENCES

REFERENCES
Bai
L.
,
Bossa
N.
,
Qu
F.
,
Winglee
J.
,
Li
G.
,
Sun
K.
,
Liang
H.
&
Wiesner
M. R.
2017
Comparison of hydrophilicity and mechanical properties of nanocomposite membranes with cellulose nanocrystals (CNCs) and carbon nanotubes (CNTs)
.
Environmental Science and Technology
51
(
1
),
253
262
.
Feng
Y.
,
Wang
Z.
,
Zhang
R.
,
Lu
Y.
,
Huang
Y.
,
Shen
H.
,
Lv
X.
&
Liu
J.
2017
Anti-fouling graphene oxide based nanocomposites membrane for oil-water emulsion separation
.
Water Science and Technology
77
(
5–6
),
1179
1185
.
Hou
S.
,
Dong
X.
,
Zhu
J.
,
Zheng
J.
,
Bi
W.
,
Li
S.
&
Zhang
S.
2017a
Preparation and characterization of an antibacterial ultrafiltration membrane with N-chloramine functional groups
.
Journal of Colloid Interface Science
496
,
391
400
.
Hou
S.
,
Xing
J.
,
Dong
X.
,
Zheng
J.
&
Li
S.
2017b
Integrated antimicrobial and antifouling ultrafiltration membrane by surface grafting PEO and N-chloramine functional groups
.
Journal of Colloid Interface Science
500
,
333
340
.
Izák
P.
,
Godinho
M. H.
,
Brogueira
P.
,
Figueirinhas
J. L.
&
Crespo
J. G.
2008
3D topography design of membranes for enhanced mass transport
.
Journal of Membrane Science
321
(
2
),
337
343
.
Liu
S.
,
Zeng
T. H.
,
Hofmann
M.
,
Burcombe
E.
,
Wei
J.
,
Jiang
R.
,
Kong
J.
&
Chen
Y.
2011
Antibacterial activity of graphite, graphite oxide, graphene oxide, and reduced graphene oxide: membrane and oxidative stress
.
ACS Nano
5
(
9
),
6971
6980
.
Meng
F.
,
Chae
S. R.
,
Drews
A.
,
Kraume
M.
,
Shin
H. S.
&
Yang
F.
2009
Recent advances in membrane bioreactors (MBRs): membrane fouling and membrane material
.
Water Research
43
(
6
),
1489
1512
.
Meng
F.
,
Zhang
S.
,
Oh
Y.
,
Zhou
Z.
,
Shin
H. S.
&
Chae
S. R.
2017
Fouling in membrane bioreactors: an updated review
.
Water Research
114
,
151
180
.
Voisin
H.
,
Bergström
L.
,
Liu
P.
&
Mathew
A. P.
2017
Nanocellulose-based materials for water purification
.
Nanomaterials
7
(
3
),
1
19
.
Ye
X.
,
Qin
X.
,
Yan
X.
,
Guo
J.
,
Huang
L.
,
Chen
D.
,
Wu
T.
,
Shi
Q.
,
Tan
S.
&
Cai
X.
2016
Π–π conjugations improve the long-term antibacterial properties of graphene oxide/quaternary ammonium salt nanocomposites
.
Chemical Engineering Journal
304
,
873
881
.
Zhang
J.
,
Xu
Z.
,
Mai
W.
,
Min
C.
,
Zhou
B.
,
Shan
M.
,
Li
Y.
,
Yang
C.
,
Wang
Z.
&
Qian
X.
2013a
Improved hydrophilicity, permeability, antifouling and mechanical performance of PVDF composite ultrafiltration membranes tailored by oxidized low-dimensional carbon nanomaterials
.
Journal of Materials Chemistry A
1
(
9
),
3101
3111
.
Zhang
X.
,
Ma
J.
,
Tang
C. Y.
,
Wang
Z.
,
Ng
H. Y.
&
Wu
Z.
2016b
Antibiofouling polyvinylidene fluoride membrane modified by quaternary ammonium compound: direct contact-killing versus induced indirect contact-killing
.
Environmental Science and Technology
50
(
10
),
5086
5093
.
Zhao
S.
,
Guo
C.
,
Hu
Y.
,
Guo
Y.
&
Pan
Q.
2018
The preparation and antibacterial activity of cellulose/ZnO composite: a review
.
Open Chemistry
16
(
1
),
9
20
.

Author notes

The first two authors contributed equally to this work.

Supplementary data